From oded at geek.co.il  Sat Sep  1 17:39:57 2012
From: oded at geek.co.il (Oded Arbel)
Date: Sat Sep  1 17:40:06 2012
Subject: [odb-users] Programmatically generating ODB queries
Message-ID: <CAOqhUvyOB8Z2WUN=RUNWd3cTGgDdGUqepFtxdM1UFkYx0F=8Kw@mail.gmail.com>

Hi.

I'm trying to solve an issue where I'm support an API where users can
submit arbitrary queries (through an API called WBEM where users are
allowed to send "WQL" queries). The logical issue here is that the
query assumes as specific object structure which may or may not map
directly to the database entities - so I need to run some code to
parse and validate the input, then generate the database query.

To implement this I need:
1) something that parses the WQL into an abstract syntax tree
(obviously only a limited set of query operations is supported)
2) a map between "WQL names" to "database names", something like (
"WQL_Employee" => model::Employee ) => ( "WQL_first_name" =>
model::Employee::first_name, "WQL_last_name" => model::last_name, ...)
3) code that travels the AST and create an odb::query<Employee> object
which represents the the WQL text that the user had input

I have (1) working fine, and I'm having a bit of trouble with 2 and 3
because I'm finding it hard to do something like reflection on the ODB
types.

Currently for (2) I have something like this (the actual code is a lot
more complicated and uses macros and templates to be generic and easy
to write):
typedef boost::variant<odb::query<Employee>::first_name,
odb::query::<Employee>::last_name, [...] > validColumnType
typedef std::map<std::wstring, validColumnType> maptype;
maptype validFields();
validFields.insert(maptype::value_type("first",odb::query<Employee>::first_name));
validFields.insert(maptype::value_type("last",odb::query<Employee>::last_name));

And for (3) O have something like this (again, judicious use of
templates makes this sane for more then one model class):
odb::query<Employee> getQuery(const Expression& expression) const {
  if (expression.complex)
    return createComplexExpression(getQuery(expression.left()),
expression.op(), expression.right());
  Field left = expression.left();
  maptype::const_iterator column = validFields.find(left.getName());
  if (column == _fields.end())
    throw std::exception(("Invalid WQL field name");
  validColumnType col = column->second;
  return col.column() + expression.op() +
odb::query<ModelType>::_val(expression.right());
}
where right() is a constant value, such as int, op() returns a string
with the requested SQL operator and col is expected to be an
odb::query<ModelClass>::field that represents the field to be tested
against. (actually, boost::variant doesn't work as the last line would
have you believe and there is some "static_visitor" magic needed to
make that work, but operationally it behaves like the last line would
have worked if C++ had run-time type information).

So - is this completely crazy and there is a simple way to go about this?
-- 
Oded

From boris at codesynthesis.com  Mon Sep  3 08:18:24 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Mon Sep  3 08:00:52 2012
Subject: [odb-users] Can odb be used without touching the domain object
	files?
In-Reply-To: <CADH8t_WeE2YDJEGw0f3a8mN-HcD1-OLDupHy--E4ZXtYid3v3g@mail.gmail.com>
References: <CADH8t_WeE2YDJEGw0f3a8mN-HcD1-OLDupHy--E4ZXtYid3v3g@mail.gmail.com>
Message-ID: <boris.20120903140423@codesynthesis.com>

Hi Candy,

Candy Chiu <candy.chiu.ad@gmail.com> writes:

> I'd like to add db capability to a project without modifying the existing
> .h's.  Is there another way to achieve mapping in this case?

Yes, especially with the upcoming 2.1.0 release which adds support for
accessor/modifier functions. Basically, such non-intrusive mapping
relies on the following ODB features:

1. Ability to specify mapping pragmas outside the class and even in
   a separate file (see Chapter 12, "ODB Pragma Language" for details).

2. Ability to use accessor/modifier functions (or, more generally,
   expressions) to access data members in a class (see Sections 12.4.5
   "get/set/access" in the upcoming ODB 2.1.0 release).

As an example, let's assume we have the following person class that
is saved in person.hxx and which we cannot modify:

// person.hxx
//
class person
{
public:
  person ();

  const std::string& email () const;
  void email (const std::string&);

  const std::string& get_name () const;
  std::string& set_name ();

  unsigned short getAge () const;
  void setAge (unsigned short);

private:
  std::string email_;
  std::string name_;
  unsigned short age_;
};

To convert this class to ODB persistent class, we can create a separate
file, person-mapping.hxx, with the following content:

// person-mapping.hxx
//
#pragma db object(person)
#pragma db member(person::email_) id

Next we compile person.hxx with the ODB compiler like this (assuming
ODB 2.1.0 or later):

odb ... --odb-epilogue "#include \"person-mapping.hxx\"" person.hxx

And now we have person-odb.?xx files that we can use in a normal way.

We plan to release ODB 2.1.0 in the next week or two, however, if you
want to give this a try, I can package a pre-release.

Boris

From boris at codesynthesis.com  Mon Sep  3 08:27:23 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Mon Sep  3 08:09:50 2012
Subject: [odb-users] odb and QT
In-Reply-To: <CAHpdtTOva3jKc6V_Sh76-qEogxoT3M5doE4_S_9v1RT6MmqqRw@mail.gmail.com>
References: <CAHpdtTOva3jKc6V_Sh76-qEogxoT3M5doE4_S_9v1RT6MmqqRw@mail.gmail.com>
Message-ID: <boris.20120903142605@codesynthesis.com>

Hi Yihya,

Yihya Hugirat <hugirat@gmail.com> writes:

> how to use odb with qt in qt creator.

That depends on which platform you are using (Windows, Linux, etc),
which toolchain (e.g., MinGW or MSVC on Windows), and which database.

Boris

From boris at codesynthesis.com  Mon Sep  3 09:05:23 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Mon Sep  3 08:47:52 2012
Subject: [odb-users] Programmatically generating ODB queries
In-Reply-To: <CAOqhUvyOB8Z2WUN=RUNWd3cTGgDdGUqepFtxdM1UFkYx0F=8Kw@mail.gmail.com>
References: <CAOqhUvyOB8Z2WUN=RUNWd3cTGgDdGUqepFtxdM1UFkYx0F=8Kw@mail.gmail.com>
Message-ID: <boris.20120903144235@codesynthesis.com>

Hi Oded,

Oded Arbel <oded@geek.co.il> writes:

> I have (1) working fine, and I'm having a bit of trouble with 2 and 3
> because I'm finding it hard to do something like reflection on the ODB
> types.

That's probably because they don't provide any support for reflection
(and we have no plans to add it).


> So - is this completely crazy and there is a simple way to go about this?

I am not sure about crazy, but it looks completely incomprehensible to me.
The way I would do it is along these lines:

template <typename O, typename T>
struct member_expr
{
  virtual 
  ~member_expr ();

  virtual odb::query<O>
  create (std::string const& op, const T& rhs) = 0;
};

template <typename O, typename T, typename M>
struct member_expr_impl: member_expr<O, T>
{
  member_expr_impl (const M& m): m_ (m) {}

  virtual odb::query<O> 
  create (std::string const& op, const T& rhs) = 0
  {
    return m_ + op + odb::query<O>::_val (rhs);
  }

  const M& m_;
};

template <typename O, typename T, typename M>
std::unique_ptr<member_expr> 
make_member_expr (const M& m)
{
  return std::unique_ptr<member_expr> (
    new member_expr_impl<O, T, M> (m));
}

std::map<std::string, std::unique_ptr<member_expr>> map;

map["name"] = make_member_expr<Employee, std::string> (odb::query<Employee>::name);
map["age"] = make_member_expr<Employee, unsigned short> (odb::query<Employee>::age);

member_expr& ae = *map["age"];
odb::query<Employee> q1 = ae.create ("<", 30);

member_expr& ne = *map["name"];
odb::query<Employee> q2 = ne.create ("=", "John");

odb::query<Employee> q (q1 && q2);

Note that I haven't tried this code myself so making sure that it works
is left as an exercise for the reader.

Boris

From boris at codesynthesis.com  Tue Sep  4 05:58:32 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Tue Sep  4 05:40:58 2012
Subject: [odb-users] Re: Self persisting objects?
In-Reply-To: <CAOqhUvwToRzosFQAVrz6hhh-4Scj0Z6qMLtC4+P=7RXAH1f0qA@mail.gmail.com>
References: <CAOqhUvz+qfHUc5UhqPTmaRNoFRyeKiifxFJAFb6vRW+deUmYYw@mail.gmail.com>
	<CAOqhUvxubG0GoGu-cyQjfTb5yMHNfaHJNbpwi9oo7Cxhxmr_jg@mail.gmail.com>
	<CAOqhUvyP68W_+H0y3mweUv-wQU+3wxfmN5pmAjnv5z6MSjc84A@mail.gmail.com>
	<boris.20120830152755@codesynthesis.com>
	<CAOqhUvzoLNSOJEjaKj2NRc6e+JsFd_GeJrnNKYF7V=S2DCd+4Q@mail.gmail.com>
	<boris.20120831130415@codesynthesis.com>
	<CAOqhUvwToRzosFQAVrz6hhh-4Scj0Z6qMLtC4+P=7RXAH1f0qA@mail.gmail.com>
Message-ID: <boris.20120904115325@codesynthesis.com>

Hi Oded,

In the future please keep your replies CC'ed to the odb-users mailing
list as discussed in the posting guidelines:

http://www.codesynthesis.com/support/posting-guidelines.xhtml

Oded Arbel <oded@geek.co.il> writes:

> On Fri, Aug 31, 2012 at 2:15 PM, Boris Kolpackov <boris@codesynthesis.com> wrote:
>
> > I mean your persistent class (Employee) uses a smart pointer as its
> > object pointer (normally specified with the 'pointer' pragma, the
> > --default-pointer ODB compiler option, or if you are using Boost
> > or Qt smart-pointer profiles). See Section 3.2, "Object and View
> > Pointers" for details.
> 
> Yes, I do use std::shared_ptr as the shared pointer for ODB objects.
> 
> > If your persistent class uses a smart pointer, say, shared_ptr, then
> > initializing it from a reference that was passed to persist() (which
> > is needed in order to enter this object into the session) is a bad
> > idea.
> [...]
> > That's why if you are using smart pointers, persist() will enter
> > the object into the session only if you pass the smart pointer:
> 
> I see. that is indeed important to note.
> 
> > If you need to convert 'this' to a shared_ptr, look into the
> > shared_from_this mechanism.
> 
> I'll try that. I'm a bit worried about the requirement for there being
> a shared_ptr to the object otherwise shared_from_this will fail. So
> how will that work with this use case?
>
> Employee e("John","Doe");
> e.persist();
>
> I'm assuming it will break?

Yes, if you are using shared_from_this then all objects should be owned
by shared_ptr. Note also that this is not really specific to ODB. If you
need more information on shared_from_this, I suggest that you consult
the documentation (Boost has a good overview) and/or general C++ mailing
lists/forums.

Boris

From boris at codesynthesis.com  Tue Sep  4 06:04:39 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Tue Sep  4 05:47:01 2012
Subject: [odb-users] Re: Bulk insert support
Message-ID: <boris.20120904115843@codesynthesis.com>

Hi Candy,

In the future please keep your replies CC'ed to the odb-users mailing list.
It is also a good idea to start a new thread with a more descriptive subject
when asking an unrelated question. All of this and more is discussed in the
posting guidelines:

http://www.codesynthesis.com/support/posting-guidelines.xhtml

Candy Chiu <candy.chiu.ad@gmail.com> writes:

> Another question - Does the obdc connector support bulk insert?  I took a
> brief look at the source code, and didn't find the bulk insert anywhere.

No, ODB doesn't support bulk inserts yet. Also note that ODB uses prepared
statements and caches a lot of things (connections, statements, even memory
buffers), so I am actually not convinced that one would get a significant
performance gain from bulk inserts. I guess before we decide to work on
this support, we will need to run some real tests that show it's worth
the extra effort and complexity as well as memory overhead on the client
side.

Boris

From boris at codesynthesis.com  Tue Sep  4 10:32:46 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Tue Sep  4 10:15:10 2012
Subject: [odb-users] odb and QT
In-Reply-To: <f46d041825985f271a04c8cb18a2@google.com>
References: <boris.20120903142605@codesynthesis.com>
	<f46d041825985f271a04c8cb18a2@google.com>
Message-ID: <boris.20120904163051@codesynthesis.com>

Hi Yihya,

In the future please keep your replies CC'ed to the odb-users mailing
list as discussed in the posting guidelines:

http://www.codesynthesis.com/support/posting-guidelines.xhtml

Hugirat@gmail.com <Hugirat@gmail.com> writes:

> i'm using linux, postgresql and qmake also the toolchain form Fedora

I've created a Wiki page that shows how to use ODB with Qt Creator on
Linux:

http://wiki.codesynthesis.com/Using_ODB_with_Qt_Creator_on_Linux

Let me know if something doesn't work for you.

Boris

From hugirat at gmail.com  Tue Sep  4 10:28:26 2012
From: hugirat at gmail.com (Yihya Hugirat)
Date: Tue Sep  4 10:28:34 2012
Subject: [odb-users] odb and QT
In-Reply-To: <boris.20120904163051@codesynthesis.com>
References: <boris.20120903142605@codesynthesis.com>
	<f46d041825985f271a04c8cb18a2@google.com>
	<boris.20120904163051@codesynthesis.com>
Message-ID: <CAHpdtTPXdtOfkwC_xMkm7ZuhbM1LqRg6dCMZrVVztP20FUNG+Q@mail.gmail.com>

ok, i will do.
thank you for your fast replies. i will try it ant let you know.

Best regards
Yihya

On Tue, Sep 4, 2012 at 4:32 PM, Boris Kolpackov <boris@codesynthesis.com>wrote:

> Hi Yihya,
>
> In the future please keep your replies CC'ed to the odb-users mailing
> list as discussed in the posting guidelines:
>
> http://www.codesynthesis.com/support/posting-guidelines.xhtml
>
> Hugirat@gmail.com <Hugirat@gmail.com> writes:
>
> > i'm using linux, postgresql and qmake also the toolchain form Fedora
>
> I've created a Wiki page that shows how to use ODB with Qt Creator on
> Linux:
>
> http://wiki.codesynthesis.com/Using_ODB_with_Qt_Creator_on_Linux
>
> Let me know if something doesn't work for you.
>
> Boris
>



-- 
????? ????????? ?????????? ???? ??????? ???????????? ??? ????????? ???????
?????? ??????? ??????? ??? ?????????? ??????????? ????????? ????????
?????????? ?? ???????????


My public key please use it when you contact me:
??? ?????? ????? ?????? ???????? ??? ??????? ??:

09BA B710 4B98 ECDA 8C0A 6CB1 2E1E CF88 6A12 F2C3
From jason.young at wyle.com  Wed Sep  5 15:22:59 2012
From: jason.young at wyle.com (Young, Jason)
Date: Wed Sep  5 15:23:08 2012
Subject: [odb-users] mingw - configure error libodb is not found 
Message-ID: <BADF1341F1929445A2964274C7B75B3B4904D8@ALHV03-EX03.ai-engsvcs.com>

Hello,



I am attempting to compile the libodb-mysql-2.0.1 on Windows 7 using msys to compile this library for mingw.  I was able to successfully compile libodb, and it installed the following into the /usr/local/lib directory:



libodb.a

libodb.dll.a

libodb.la



I was then able to download the MySQL client libraries, and renamed the libmysql.lib to libmysqlclient_r.a.  When running configure for libodb-mysql-2.0.1, the linker is able to find libmysqlclient_r with no problem.  However...I keep running into the following error:



checking for libmysqlclient_r... yes

checking for libodb... no

configure: error: libodb is not found; consider using --with-libodb=DIR



I am using the following command to attempt to run configure:



./configure CPPFLAGS=-I/c/tools/mysql/include LDFLAGS=-L/c/tools/mysql/lib



I additionally attempted to use the --with-libodb=/usr/local/lib or --with-libodb=/c/msys/1.0/local (since my msys local directory is installed in C:\msys\1.0\local).



I cannot seem to get the ./configure to find the libodb.  I'm fairly new to msys and mingw.  Is there a step that I am missing that would enable me to link to the libodb in order to build the libodb-mysql-2.0.1?



Thank you,



Jason Young

jason.young@wyle.com<mailto:jason.young@wyle.com>

Wyle CAS

256-922-6617 (Wyle Office)

256-313-9055 (Redstone Arsenal Building 5400 Office)
From boris at codesynthesis.com  Thu Sep  6 06:07:28 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Thu Sep  6 05:49:27 2012
Subject: [odb-users] mingw - configure error libodb is not found
In-Reply-To: <BADF1341F1929445A2964274C7B75B3B4904D8@ALHV03-EX03.ai-engsvcs.com>
References: <BADF1341F1929445A2964274C7B75B3B4904D8@ALHV03-EX03.ai-engsvcs.com>
Message-ID: <boris.20120906115826@codesynthesis.com>

Hi Jason,

Young, Jason <jason.young@wyle.com> writes:

> I am attempting to compile the libodb-mysql-2.0.1 on Windows 7 using 
> msys to compile this library for mingw. I was able to successfully 
> compile libodb, and it installed the following into the /usr/local/lib
> directory:
>
> [...]
> 
> I cannot seem to get the ./configure to find the libodb.

I can see two reasons why configure would be unable to find libodb:

1. MinGW for some reason does not search in /usr/local by default. To test
   this, simply add -I/usr/local/include to CPPFLAGS and -L/usr/local/lib
   to LDFLAGS, for example:

   ./configure CPPFLAGS="-I/c/tools/mysql/include -I/usr/local/include" \
LDFLAGS="-L/c/tools/mysql/lib -L/usr/local/lib"


2. The other possibility is that the libodb library that you have built
   is somehow broken (this happens on MinGW all the time, especially
   when DLLs are involved).

   To figure out if that's the case, the best method is to open the
   config.log file and look at the actual compiler/linker diagnostics
   (search for the 'checking for libodb' string to find the relevant
   part). Note that there could be multiple tests and it is ok for
   some of them to fail.

   Also, if you can post the relevant fragment from config.log (or
   send the whole file), then I can take a look.

Boris

From jason.young at wyle.com  Thu Sep  6 09:12:37 2012
From: jason.young at wyle.com (Young, Jason)
Date: Thu Sep  6 09:12:42 2012
Subject: [odb-users] mingw - configure error libodb is not found
In-Reply-To: <boris.20120906115826@codesynthesis.com>
References: <BADF1341F1929445A2964274C7B75B3B4904D8@ALHV03-EX03.ai-engsvcs.com>,
	<boris.20120906115826@codesynthesis.com>
Message-ID: <BADF1341F1929445A2964274C7B75B3B4904F6@ALHV03-EX03.ai-engsvcs.com>

Boris,

Thank you for your response.

Option number 1 worked for me, though I had tried it yesterday
before I sent in the help request.  I did reboot after going home
yesterday, so I'm thinking that the reboot may have somehow
made a difference.  Anyway, I appreciate your response and
I look forward to getting started with odb.

Thanks,
Jason Young

jason.young@wyle.com

________________________________________
From: Boris Kolpackov [boris@codesynthesis.com]
Sent: Thursday, September 06, 2012 5:07 AM
To: Young, Jason
Cc: odb-users@codesynthesis.com
Subject: Re: [odb-users] mingw - configure error libodb is not found

Hi Jason,

Young, Jason <jason.young@wyle.com> writes:

> I am attempting to compile the libodb-mysql-2.0.1 on Windows 7 using
> msys to compile this library for mingw. I was able to successfully
> compile libodb, and it installed the following into the /usr/local/lib
> directory:
>
> [...]
>
> I cannot seem to get the ./configure to find the libodb.

I can see two reasons why configure would be unable to find libodb:

1. MinGW for some reason does not search in /usr/local by default. To test
   this, simply add -I/usr/local/include to CPPFLAGS and -L/usr/local/lib
   to LDFLAGS, for example:

   ./configure CPPFLAGS="-I/c/tools/mysql/include -I/usr/local/include" \
LDFLAGS="-L/c/tools/mysql/lib -L/usr/local/lib"


2. The other possibility is that the libodb library that you have built
   is somehow broken (this happens on MinGW all the time, especially
   when DLLs are involved).

   To figure out if that's the case, the best method is to open the
   config.log file and look at the actual compiler/linker diagnostics
   (search for the 'checking for libodb' string to find the relevant
   part). Note that there could be multiple tests and it is ok for
   some of them to fail.

   Also, if you can post the relevant fragment from config.log (or
   send the whole file), then I can take a look.

Boris

From candy.chiu.ad at gmail.com  Sat Sep  8 21:38:38 2012
From: candy.chiu.ad at gmail.com (Candy Chiu)
Date: Sat Sep  8 21:38:47 2012
Subject: [odb-users] odbc bulk insert
Message-ID: <CADH8t_VO5L1MrTXBt42RTXsxNqn6FYe3Dq_A2NhCC7mroy6sEw@mail.gmail.com>

Does the obdc connector support bulk insert?
From ststrou at sandia.gov  Fri Sep  7 17:37:01 2012
From: ststrou at sandia.gov (Stroud, Sean T)
Date: Sun Sep  9 04:15:26 2012
Subject: [odb-users] How to compile "hello" example on RHEL
Message-ID: <8EDFC8F388A1E44BB67C9BAA57091F9E285BCED4@EXMB03.srn.sandia.gov>

Hi,

I?m a new user, evaluating odb for potential use on our project.  The documentation and examples look really great.  Only one problem ? how do you compile them?  For the ?hello? example I see there is a ?README? file.  I would expect the README file to say how to compile and link the example but it doesn?t (it only tells how to run odb to generate the code).

My environment:
?       OS:                                  Linux (RHEL)
?       Compiler:                 g++ 4.1.2
?       ODB version:               1.7.0
?       ODB installation dir:  /home/ststrou/ODB

As a guess, I tried the following command in the ?/home/ststrou/ODB/odb-examples-1.7.0/hello? directory:

% g++ -c driver.cxx -I /home/ststrou/ODB/libodb-1.7.0 -I /home/ststrou/ODB/libodb-mysql-1.7.0

But that fails with the following output:

      In file included from /home/ststrou/ODB/libodb-1.7.0/odb/details/export.hxx:11,
                       from /home/ststrou/ODB/libodb-1.7.0/odb/forward.hxx:11,
                       from /home/ststrou/ODB/libodb-1.7.0/odb/traits.hxx:11,
                       from /home/ststrou/ODB/libodb-1.7.0/odb/database.hxx:14,
                       from driver.cxx:8:
      /home/ststrou/ODB/libodb-1.7.0/odb/details/config.hxx:17:34: error: odb/details/config.h: No such file or directory
      In file included from /home/ststrou/ODB/libodb-mysql-1.7.0/odb/mysql/version.hxx:11,
                       from person-odb.hxx:37,
                       from driver.cxx:14:
      /home/ststrou/ODB/libodb-mysql-1.7.0/odb/mysql/details/config.hxx:18:40: error: odb/mysql/details/config.h: No such file or directory
      database.hxx: In function ???std::auto_ptr<odb::database> create_database(int&, char**)???:
      database.hxx:85: error: ???db??? was not declared in this scope

I tried looking at the Makefile.am and Makefile.im files, but they look ridiculously complicated and I have no idea how to invoke them.

Thanks,

Sean

From boris at codesynthesis.com  Sun Sep  9 05:52:07 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Sun Sep  9 05:33:39 2012
Subject: [odb-users] How to compile "hello" example on RHEL
In-Reply-To: <8EDFC8F388A1E44BB67C9BAA57091F9E285BCED4@EXMB03.srn.sandia.gov>
References: <8EDFC8F388A1E44BB67C9BAA57091F9E285BCED4@EXMB03.srn.sandia.gov>
Message-ID: <boris.20120909112826@codesynthesis.com>

Hi Sean,

Stroud, Sean T <ststrou@sandia.gov> writes:

> Only one problem ? how do you compile them? For the ?hello? example
> I see there is a ?README? file. I would expect the README file to
> say how to compile and link the example but it doesn?t (it only
> tells how to run odb to generate the code).

The problems with giving such instructions in README is that they are
very platform-specific (i.e., on Linux one could use configure & make
while on Windows with VC++ one would use solution files). And these
are described in the INSTALL file in the root of the odb-examples
package.

But I agree, perhaps we should also include some information on how
to build each example manually, from the command line. I've added
this item to the TODO list.

> % g++ -c driver.cxx -I /home/ststrou/ODB/libodb-1.7.0 -I \
> /home/ststrou/ODB/libodb-mysql-1.7.0
> 
> But that fails with the following output:
> 
> [...]
>
> /home/ststrou/ODB/libodb-1.7.0/odb/details/config.hxx:17:34: error: odb/details/config.h: No such file or directory

I don't think you have built libodb and libodb-mysql (i.e., by running
configure and then make). For a high-level overview of how to do this,
see the "Installing ODB on UNIX" page:

http://www.codesynthesis.com/products/odb/doc/install-unix.xhtml

For more detailed information refer to the INSTALL file in each package.

You can just build libodb and libodb-mysql without installing them and
the above command will work. However, things will get progressively
more complicated from there. Specifically, the linking step will require
you to find and specify .so files. Running the example will fail unless
you add the paths where these libraries are located to LD_LIBRARY_PATH.

Instead, it is much easier to install the headers and the libraries to
the location where g++ will search for them automatically (/usr/local
by default).

Normally, to build and install these libraries all you have to do is run
these commands in each directory:

./configure
make
sudo make install

If later you want to uninstall them, then just run:

sudo make uninstall

Ok, assuming that you have libodb and libodb-mysql installed, the command
line to compile and link the hello example will then look like this:

g++ -DDATABASE_MYSQL -c driver.cxx person-odb.cxx
g++ -o driver driver.o person-odb.o -lodb-mysql -lodb

Note the DATABASE_MYSQL define -- it selects which database the example
should target.

Boris

From boris at codesynthesis.com  Sun Sep  9 05:57:28 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Sun Sep  9 05:38:55 2012
Subject: [odb-users] odbc bulk insert
In-Reply-To: <CADH8t_VO5L1MrTXBt42RTXsxNqn6FYe3Dq_A2NhCC7mroy6sEw@mail.gmail.com>
References: <CADH8t_VO5L1MrTXBt42RTXsxNqn6FYe3Dq_A2NhCC7mroy6sEw@mail.gmail.com>
Message-ID: <boris.20120909115545@codesynthesis.com>

Hi Candy,

Candy Chiu <candy.chiu.ad@gmail.com> writes:

> Does the obdc connector support bulk insert?

You have already asked this question and I have already answered it.
Please see:

http://www.codesynthesis.com/pipermail/odb-users/2012-September/000747.html

Boris

From macromarship at gmail.com  Tue Sep 11 02:17:06 2012
From: macromarship at gmail.com (Scott Zhang)
Date: Tue Sep 11 02:17:14 2012
Subject: [odb-users] compiling odb on opensuse 12.1 x64 problem
Message-ID: <CABy3iKzA12mX1JO7s1hHv3oNU58bLH=1oQ1somxZBLOu7FV-Hg@mail.gmail.com>

hello. All.
     I have been looking for an orm software for c++. Get tried
Stactiverecord, which is easy to use but it is weak in query and missing
some features(like oracle support, order by) I need.
     After search around on google, I found odb is the only full featured
library support many db which my project targeting.
     I have downloaded the source and want to build it on my opensuse 12.1
x64 with gcc 4.6.1. The first problem is configure script complains "g++
support plugins=no". I don't know why odb need this feature. But by pass
--enable-static configure script passed.
    Then I type "make". But it stop first on missing "coretypes.h", I
downloaded then it stop on missing "bversion.h".
    I don't understand why build odb is so hard. I seldom get problem when
build package like this.
    Is there any reason a "bersion.h" or "coretypes.h" is a must? or just
because my opensue remove that files?


Regards.
Scott
From macromarship at gmail.com  Tue Sep 11 02:29:49 2012
From: macromarship at gmail.com (Scott Zhang)
Date: Tue Sep 11 02:29:57 2012
Subject: [odb-users] Re: compiling odb on opensuse 12.1 x64 problem
In-Reply-To: <CABy3iKzA12mX1JO7s1hHv3oNU58bLH=1oQ1somxZBLOu7FV-Hg@mail.gmail.com>
References: <CABy3iKzA12mX1JO7s1hHv3oNU58bLH=1oQ1somxZBLOu7FV-Hg@mail.gmail.com>
Message-ID: <CABy3iKyg2pLUTdSptcQqWw1djJV8qpsfjmWn5pfis836-6k6yg@mail.gmail.com>

I only means the "odb" compiler itself is hard to compile.

libodb
libodb-mysql
libodb-oracle  build smoothly.

Regards.
Scott

On Tue, Sep 11, 2012 at 2:17 PM, Scott Zhang <macromarship@gmail.com> wrote:

> hello. All.
>      I have been looking for an orm software for c++. Get tried
> Stactiverecord, which is easy to use but it is weak in query and missing
> some features(like oracle support, order by) I need.
>      After search around on google, I found odb is the only full featured
> library support many db which my project targeting.
>      I have downloaded the source and want to build it on my opensuse 12.1
> x64 with gcc 4.6.1. The first problem is configure script complains "g++
> support plugins=no". I don't know why odb need this feature. But by pass
> --enable-static configure script passed.
>     Then I type "make". But it stop first on missing "coretypes.h", I
> downloaded then it stop on missing "bversion.h".
>     I don't understand why build odb is so hard. I seldom get problem when
> build package like this.
>     Is there any reason a "bersion.h" or "coretypes.h" is a must? or just
> because my opensue remove that files?
>
>
> Regards.
> Scott
>
From macromarship at gmail.com  Tue Sep 11 02:54:37 2012
From: macromarship at gmail.com (Scott Zhang)
Date: Tue Sep 11 02:54:44 2012
Subject: [odb-users] Re: compiling odb on opensuse 12.1 x64 problem
In-Reply-To: <CABy3iKyg2pLUTdSptcQqWw1djJV8qpsfjmWn5pfis836-6k6yg@mail.gmail.com>
References: <CABy3iKzA12mX1JO7s1hHv3oNU58bLH=1oQ1somxZBLOu7FV-Hg@mail.gmail.com>
	<CABy3iKyg2pLUTdSptcQqWw1djJV8qpsfjmWn5pfis836-6k6yg@mail.gmail.com>
Message-ID: <CABy3iKxNwb3qiU_nn45yvqTnJfZJn1imZqYeyWiT_RQzDA72hQ@mail.gmail.com>

I just download the prebuild odb compiler and it runs ok on my system.
Now I am trying example now.

On Tue, Sep 11, 2012 at 2:29 PM, Scott Zhang <macromarship@gmail.com> wrote:

> I only means the "odb" compiler itself is hard to compile.
>
> libodb
> libodb-mysql
> libodb-oracle  build smoothly.
>
> Regards.
> Scott
>
>
> On Tue, Sep 11, 2012 at 2:17 PM, Scott Zhang <macromarship@gmail.com>wrote:
>
>> hello. All.
>>      I have been looking for an orm software for c++. Get tried
>> Stactiverecord, which is easy to use but it is weak in query and missing
>> some features(like oracle support, order by) I need.
>>      After search around on google, I found odb is the only full featured
>> library support many db which my project targeting.
>>      I have downloaded the source and want to build it on my opensuse
>> 12.1 x64 with gcc 4.6.1. The first problem is configure script complains
>> "g++ support plugins=no". I don't know why odb need this feature. But by
>> pass --enable-static configure script passed.
>>     Then I type "make". But it stop first on missing "coretypes.h", I
>> downloaded then it stop on missing "bversion.h".
>>     I don't understand why build odb is so hard. I seldom get problem
>> when build package like this.
>>     Is there any reason a "bersion.h" or "coretypes.h" is a must? or just
>> because my opensue remove that files?
>>
>>
>> Regards.
>> Scott
>>
>
>
From boris at codesynthesis.com  Tue Sep 11 04:34:20 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Tue Sep 11 04:15:35 2012
Subject: [odb-users] Re: compiling odb on opensuse 12.1 x64 problem
In-Reply-To: <CABy3iKxNwb3qiU_nn45yvqTnJfZJn1imZqYeyWiT_RQzDA72hQ@mail.gmail.com>
References: <CABy3iKzA12mX1JO7s1hHv3oNU58bLH=1oQ1somxZBLOu7FV-Hg@mail.gmail.com>
	<CABy3iKyg2pLUTdSptcQqWw1djJV8qpsfjmWn5pfis836-6k6yg@mail.gmail.com>
	<CABy3iKxNwb3qiU_nn45yvqTnJfZJn1imZqYeyWiT_RQzDA72hQ@mail.gmail.com>
Message-ID: <boris.20120911103123@codesynthesis.com>

Hi Scott,

I did some digging and it appears that OpenSUSE builds its GCC with
plugins disabled (unlike all other major distributions). And ODB needs
this support. I also see that other people ran into the same issue with
other tools. So I've filed a bug report:

https://bugzilla.novell.com/show_bug.cgi?id=779692

In the meantime, using the pre-built version of ODB seems like the
easiest option on OpenSUSE.

Boris

From macromarship at gmail.com  Tue Sep 11 09:52:56 2012
From: macromarship at gmail.com (Scott Zhang)
Date: Tue Sep 11 09:53:04 2012
Subject: [odb-users] Re: compiling odb on opensuse 12.1 x64 problem
In-Reply-To: <boris.20120911103123@codesynthesis.com>
References: <CABy3iKzA12mX1JO7s1hHv3oNU58bLH=1oQ1somxZBLOu7FV-Hg@mail.gmail.com>
	<CABy3iKyg2pLUTdSptcQqWw1djJV8qpsfjmWn5pfis836-6k6yg@mail.gmail.com>
	<CABy3iKxNwb3qiU_nn45yvqTnJfZJn1imZqYeyWiT_RQzDA72hQ@mail.gmail.com>
	<boris.20120911103123@codesynthesis.com>
Message-ID: <CABy3iKzPa9HnzDFi_o_d6FPLMJmK2pwxkxFMbKTqvjcPsq4vOQ@mail.gmail.com>

yes. Prebuit version works fine.
Thanks.

I am learning to use odb in my project now.


Thanks.
Regards
Scott

On Tue, Sep 11, 2012 at 4:34 PM, Boris Kolpackov <boris@codesynthesis.com>wrote:

> Hi Scott,
>
> I did some digging and it appears that OpenSUSE builds its GCC with
> plugins disabled (unlike all other major distributions). And ODB needs
> this support. I also see that other people ran into the same issue with
> other tools. So I've filed a bug report:
>
> https://bugzilla.novell.com/show_bug.cgi?id=779692
>
> In the meantime, using the pre-built version of ODB seems like the
> easiest option on OpenSUSE.
>
> Boris
>
From danielpeterjames at gmail.com  Tue Sep 11 10:08:13 2012
From: danielpeterjames at gmail.com (Daniel James)
Date: Tue Sep 11 10:08:41 2012
Subject: [odb-users] performance of updates in table with many rows for
	sqlite db
Message-ID: <CAAqNS9abv+KyR9V-880w6aJ=zuUGKaR-cnW_KjoLQ8y+Cth8Eg@mail.gmail.com>

Hello

I'm not sure if this is a reasonable question but here goes:

Suppose I have the following object:

#pragma db object pointer(std::shared_ptr)
class Intensity
{

public:
    //
    // Relationships
    //
    odb::lazy_weak_ptr<Probe> probe;


    //
    // Data members.
    //
    float signal, backgroundSignal, normalisedSignal, predictedSignal;
    bool flag, extra_flag;
    unsigned long id();
private:
    friend class odb::access;

#pragma db id auto
    unsigned long id_;
};

I store about 6 million of these in a table. Then I try and update a
subset of about 40 000.

What is the quickest way to do this? At the moment it is proceeding at
a rate of about one update per second with the following SQL being
generated at each update:

UPDATE "Intensity" SET
"probe"=?,"signal"=?,"backgroundSignal"=?,"normalisedSignal"=?,"predictedSignal"=?,"flag"=?,"extra_flag"=?
WHERE "id"=?


Do I need to:

rethink the database design?
somehow use an index on the Record table to speed things up?
other strategy?

Many thanks,

Daniel

From candy.chiu.ad at gmail.com  Tue Sep 11 12:18:48 2012
From: candy.chiu.ad at gmail.com (Candy Chiu)
Date: Tue Sep 11 12:18:58 2012
Subject: [odb-users] Re: odbc bulk insert
In-Reply-To: <boris.20120909115545@codesynthesis.com>
References: <CADH8t_VO5L1MrTXBt42RTXsxNqn6FYe3Dq_A2NhCC7mroy6sEw@mail.gmail.com>
	<boris.20120909115545@codesynthesis.com>
Message-ID: <CADH8t_WYi-0paKiZSEJToUJ9HeRi+7L2muS9b5bGTprT-GmteA@mail.gmail.com>

Boris,

I have ran a bulk insert experiment against the native library.  The bulk
insert performs about 40 times faster than the row-by-row insertion even
using a cached connection and prepared statement.  Memory allocation may be
the issue, but I doubt it caused such a big difference.

Candy

On Sunday, September 9, 2012, Boris Kolpackov wrote:

> Hi Candy,
>
> Candy Chiu <candy.chiu.ad@gmail.com <javascript:;>> writes:
>
> > Does the obdc connector support bulk insert?
>
> You have already asked this question and I have already answered it.
> Please see:
>
> http://www.codesynthesis.com/pipermail/odb-users/2012-September/000747.html
>
> Boris
>
From macromarship at gmail.com  Wed Sep 12 03:43:38 2012
From: macromarship at gmail.com (Scott Zhang)
Date: Wed Sep 12 03:43:47 2012
Subject: [odb-users] Pagination and "join","groupby" in odb
Message-ID: <CABy3iKzZYvp-RcK_2eX=GCotSqVwL+XY=89Shx3DvOtVJx3RiA@mail.gmail.com>

Hello. All.
    I have quickly went over the odb-manual.pdf. Basically the mapping of
class to table(activerecord), and Query on single table/class is what I
need. Another few features I need are "pagination" which limit the returned
number and starting number of query. And join/group by etc features. I
don't see where are them. Can anyone help on this? Thanks.


    And I have some options about odb.  Sorry I am a mostly
.net(subsonic)/php(codeignitor)/java(struts2) developer.  so I am fan of
easy use of orm/activerecord provided with these library. On C++
developing, I usually use mysqlcppconn library previously but that need to
write sql all the time. Then I want to change to use a orm where odb falls
in.
    In my view of odb, basically the activerecord and query are cool. The
view/container/relations/inheritance looks provided too much features and
included too much new rules different from other frameworks/library. I
found it is hard to remember them all and most of work can be done only use
the activerecord and query only.
    ODB's design works smiliar like django(python) and RubyOnRails, you
draft the class using your rules first, then generate sql to create the
database, then everything works cool. But if you create the database first,
especially coding with a php front end done, then code with c++ and odb to
fit your existing tables relation with your class, that would be a big
headache(for relations, container, inheritance ....). So I think only the
basic activerecord and query should be enough. Just my option because I
don't think I will use these complex container/relation .... .  I will
control the relations from c++ code level.

Thanks.
Regards.
Scott
From macromarship at gmail.com  Wed Sep 12 04:00:48 2012
From: macromarship at gmail.com (Scott Zhang)
Date: Wed Sep 12 04:00:56 2012
Subject: [odb-users] Re: Pagination and "join","groupby" in odb
In-Reply-To: <CABy3iKzZYvp-RcK_2eX=GCotSqVwL+XY=89Shx3DvOtVJx3RiA@mail.gmail.com>
References: <CABy3iKzZYvp-RcK_2eX=GCotSqVwL+XY=89Shx3DvOtVJx3RiA@mail.gmail.com>
Message-ID: <CABy3iKzbUuapN+35JMB36BKRk5=2YY2iewVyVHKt5yjQKGSVUg@mail.gmail.com>

Hello.
   I searched over net and found this discussion
http://www.codesynthesis.com/pipermail/odb-users/2012-January/000431.html

   Actually, I would suggest odb  borrow some implementation from
codeigniter because it support msql/mysql/oracle/sqlite/pgsql. Most of
table mapping are dynamic but that's ok. I just mean the pagination it
implements can be used on all databases. if it can, so should odb. Just
generate the correct sql then problem solved.

   And I see "view" can be a wrapper for any customized sql to map result
into a static class. That works.


Regards
Scott



On Wed, Sep 12, 2012 at 3:43 PM, Scott Zhang <macromarship@gmail.com> wrote:

> Hello. All.
>     I have quickly went over the odb-manual.pdf. Basically the mapping of
> class to table(activerecord), and Query on single table/class is what I
> need. Another few features I need are "pagination" which limit the returned
> number and starting number of query. And join/group by etc features. I
> don't see where are them. Can anyone help on this? Thanks.
>
>
>     And I have some options about odb.  Sorry I am a mostly
> .net(subsonic)/php(codeignitor)/java(struts2) developer.  so I am fan of
> easy use of orm/activerecord provided with these library. On C++
> developing, I usually use mysqlcppconn library previously but that need to
> write sql all the time. Then I want to change to use a orm where odb falls
> in.
>     In my view of odb, basically the activerecord and query are cool. The
> view/container/relations/inheritance looks provided too much features and
> included too much new rules different from other frameworks/library. I
> found it is hard to remember them all and most of work can be done only use
> the activerecord and query only.
>     ODB's design works smiliar like django(python) and RubyOnRails, you
> draft the class using your rules first, then generate sql to create the
> database, then everything works cool. But if you create the database first,
> especially coding with a php front end done, then code with c++ and odb to
> fit your existing tables relation with your class, that would be a big
> headache(for relations, container, inheritance ....). So I think only the
> basic activerecord and query should be enough. Just my option because I
> don't think I will use these complex container/relation .... .  I will
> control the relations from c++ code level.
>
> Thanks.
> Regards.
> Scott
>
>
>
From boris at codesynthesis.com  Wed Sep 12 07:26:34 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Wed Sep 12 07:07:34 2012
Subject: [odb-users] Re: odbc bulk insert
In-Reply-To: <CADH8t_WYi-0paKiZSEJToUJ9HeRi+7L2muS9b5bGTprT-GmteA@mail.gmail.com>
References: <CADH8t_VO5L1MrTXBt42RTXsxNqn6FYe3Dq_A2NhCC7mroy6sEw@mail.gmail.com>
	<boris.20120909115545@codesynthesis.com>
	<CADH8t_WYi-0paKiZSEJToUJ9HeRi+7L2muS9b5bGTprT-GmteA@mail.gmail.com>
Message-ID: <boris.20120912132446@codesynthesis.com>

Hi Candy,

Candy Chiu <candy.chiu.ad@gmail.com> writes:

> I have ran a bulk insert experiment against the native library.  The bulk
> insert performs about 40 times faster than the row-by-row insertion even
> using a cached connection and prepared statement.  Memory allocation may be
> the issue, but I doubt it caused such a big difference.

Interesting. Would you be able to send me your test so that I can compare
it to an equivalent code that uses ODB?

Boris

From boris at codesynthesis.com  Wed Sep 12 07:39:14 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Wed Sep 12 07:20:09 2012
Subject: [odb-users] performance of updates in table with many rows for
	sqlite db
In-Reply-To: <CAAqNS9abv+KyR9V-880w6aJ=zuUGKaR-cnW_KjoLQ8y+Cth8Eg@mail.gmail.com>
References: <CAAqNS9abv+KyR9V-880w6aJ=zuUGKaR-cnW_KjoLQ8y+Cth8Eg@mail.gmail.com>
Message-ID: <boris.20120912132725@codesynthesis.com>

Hi Daniel,

Daniel James <danielpeterjames@gmail.com> writes:

> I store about 6 million of these in a table. Then I try and update a
> subset of about 40 000.
> 
> What is the quickest way to do this? At the moment it is proceeding at
> a rate of about one update per second with the following SQL being
> generated at each update:
> 
> [...]
> 
> Do I need to:
> 
> rethink the database design?
> somehow use an index on the Record table to speed things up?
> other strategy?

There are two areas I would explore first:

1. How do you find records that you need to update? Do you know their
   object ids and just load them? Or do you do a query? If it is the
   latter, then it would be good to understand which part of that
   time is spent querying and which -- updating. Also, in this case,
   an index can definitely help.

2. How many updates do you perform per transaction? If you do each
   update in a separate transaction, then that would explain the
   times you are seeing. In this case, the database has to wait
   until (some part of) the data is physically written to the disk
   before it can return from the call to commit(). Similarly, if
   you do all 40,000 updates in a single transaction, things can
   get slow because of all the housekeeping data that the database
   has to keep around until the transaction is committed or rolled
   back.

   So, generally, there is an optimum number of updates that you need
   to perform per transaction to achieve the best performance. And the
   way to find this number is through some experimentation.

Boris

From boris at codesynthesis.com  Wed Sep 12 07:53:18 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Wed Sep 12 07:34:13 2012
Subject: [odb-users] Re: Pagination and "join","groupby" in odb
In-Reply-To: <CABy3iKzbUuapN+35JMB36BKRk5=2YY2iewVyVHKt5yjQKGSVUg@mail.gmail.com>
References: <CABy3iKzZYvp-RcK_2eX=GCotSqVwL+XY=89Shx3DvOtVJx3RiA@mail.gmail.com>
	<CABy3iKzbUuapN+35JMB36BKRk5=2YY2iewVyVHKt5yjQKGSVUg@mail.gmail.com>
Message-ID: <boris.20120912134136@codesynthesis.com>

Hi Scott,

Scott Zhang <macromarship@gmail.com> writes:

>    Actually, I would suggest odb  borrow some implementation from
> codeigniter because it support msql/mysql/oracle/sqlite/pgsql. Most of
> table mapping are dynamic but that's ok. I just mean the pagination it
> implements can be used on all databases. if it can, so should odb. Just
> generate the correct sql then problem solved.

ODB, being a C++ ORM, has very different goals compared to CodeIgniter,
which targets PHP. Specifically, ODB's goal is to deliver high 
performance and using dynamically-generated queries is a non-starter.
Pagination can be implemented quite easily (as shown on the page you
found) for databases that provide support for it. For other databases,
things are a bit more complex but also doable.


> And I see "view" can be a wrapper for any customized sql to map result
> into a static class. That works.

Yes, that's what you can use to implement custom joins, aggregate queries,
etc.


> But if you create the database first, especially coding with a php front
> end done, then code with c++ and odb to fit your existing tables relation
> with your class, that would be a big headache (for relations, container,
> inheritance ....).

I disagree. ODB uses what we call "canonical" relational techniques to
implement object relations, containers, and inheritance. In other words,
if you create a database schema manually and follow good relational
database design, then mapping that schema to C++ classes with ODB should
be a fairly straightforward matter.

Now if the schema is produced by another ORM that you are using and it
doesn't follow good relational database design, then there is nothing
ODB can do. But that's really the problem with another ORM, not ODB.

Boris

From macromarship at gmail.com  Wed Sep 12 12:23:51 2012
From: macromarship at gmail.com (Scott Zhang)
Date: Wed Sep 12 12:24:01 2012
Subject: [odb-users] Re: Pagination and "join","groupby" in odb
In-Reply-To: <boris.20120912134136@codesynthesis.com>
References: <CABy3iKzZYvp-RcK_2eX=GCotSqVwL+XY=89Shx3DvOtVJx3RiA@mail.gmail.com>
	<CABy3iKzbUuapN+35JMB36BKRk5=2YY2iewVyVHKt5yjQKGSVUg@mail.gmail.com>
	<boris.20120912134136@codesynthesis.com>
Message-ID: <CABy3iKzfUf26Tm5_wOvG8eQCowmyq=Og-+PkafKBUk9Ehq0v9Q@mail.gmail.com>

Hi. thanks for reply.

But I think it is better if odb integrate the "order by" and "pagination"
into its query object so we don't need to insert the "sql" into "query".
Especially for pagination, odb should hide the different implement of
database from end users.

Is there a plan for this(order by, pagination)?


Regards.
Scott

On Wed, Sep 12, 2012 at 7:53 PM, Boris Kolpackov <boris@codesynthesis.com>wrote:

> Hi Scott,
>
> Scott Zhang <macromarship@gmail.com> writes:
>
> >    Actually, I would suggest odb  borrow some implementation from
> > codeigniter because it support msql/mysql/oracle/sqlite/pgsql. Most of
> > table mapping are dynamic but that's ok. I just mean the pagination it
> > implements can be used on all databases. if it can, so should odb. Just
> > generate the correct sql then problem solved.
>
> ODB, being a C++ ORM, has very different goals compared to CodeIgniter,
> which targets PHP. Specifically, ODB's goal is to deliver high
> performance and using dynamically-generated queries is a non-starter.
> Pagination can be implemented quite easily (as shown on the page you
> found) for databases that provide support for it. For other databases,
> things are a bit more complex but also doable.
>
>
> > And I see "view" can be a wrapper for any customized sql to map result
> > into a static class. That works.
>
> Yes, that's what you can use to implement custom joins, aggregate queries,
> etc.
>
>
> > But if you create the database first, especially coding with a php front
> > end done, then code with c++ and odb to fit your existing tables relation
> > with your class, that would be a big headache (for relations, container,
> > inheritance ....).
>
> I disagree. ODB uses what we call "canonical" relational techniques to
> implement object relations, containers, and inheritance. In other words,
> if you create a database schema manually and follow good relational
> database design, then mapping that schema to C++ classes with ODB should
> be a fairly straightforward matter.
>
> Now if the schema is produced by another ORM that you are using and it
> doesn't follow good relational database design, then there is nothing
> ODB can do. But that's really the problem with another ORM, not ODB.
>
> Boris
>
From macromarship at gmail.com  Thu Sep 13 06:32:40 2012
From: macromarship at gmail.com (Scott Zhang)
Date: Thu Sep 13 06:32:47 2012
Subject: [odb-users] About using odb in same program to connect to
 mysql/sqlite/pgsql at same time
Message-ID: <CABy3iKx+mFwu1-PGKs6y7aP7oO=GW5a0Jcq0mpRdVXerZovvBQ@mail.gmail.com>

Hello.
    I am reading code to try to add pagination to odb. For orderby and
groupby etc, as they are trial in most case, so append sql string is
acceptable. Because I am going to replace my 2 projects DB access (one
mysql, another oracle) library with odb, so I think add pagination will
worth that.

    Well. odb in my view is a really complex and heavy used template
design. I am not sure the benefit of this. If the template was used to
improve the flexibly to support static compiling check etc, because of so
many different data types. I do think we have easier design. If you could
look at the C# subsonic's activerecord template, from a "ODB/Subsonice
compiler" level, the generate code can itself be static typed so could need
less or none template. Just subsonic can reverse generate class from
database. odb can generate persistent class from transient class.

   And from the code flow. Looks I only need to add a Take(int num_),
Offset(int offset_), (maybe OrderBy(string, enum orderby?)  and more ), in
the "query" class then we can support "pagination". Because
  for the query to run:
eg:
  result r (db->query<person> (query::age > 30));
although the first sight to convert "query::age>30" or even more complex
expression into a where clause looks mysterious. But after reading code,
query::age itself is a query_column, which can parse ">30" and return a
query. Then by the override operators, query can parse the left expressions
itself. Then the place for the Offset, Take(orderby, groupby ....) could be
after all where clause processed. We just need to call on  query object like

   result r(db->query<person>(
(query::aget>30).orderby(query::aget::column(), DESC).offset(10).take(10)) )

then we can get the 10 - 20 rows.

    And another question, from the code, looks odb::query linked with its
mysql implementation  so the query it returned will be a mysql query. But
if we link with both mysql and oracle at same time, eg for an application
need to use mysql and oracle at same time. Can odb::query work correctly?


Thanks.
Regards
Scott
From macromarship at gmail.com  Thu Sep 13 06:38:56 2012
From: macromarship at gmail.com (Scott Zhang)
Date: Thu Sep 13 06:39:04 2012
Subject: [odb-users] Re: About using odb in same program to connect to
 mysql/sqlite/pgsql at same time
In-Reply-To: <CABy3iKx+mFwu1-PGKs6y7aP7oO=GW5a0Jcq0mpRdVXerZovvBQ@mail.gmail.com>
References: <CABy3iKx+mFwu1-PGKs6y7aP7oO=GW5a0Jcq0mpRdVXerZovvBQ@mail.gmail.com>
Message-ID: <CABy3iKy8kHqpE+1yVC-xp+0iJLCQS_haEkhqKYGG2mgtr235pw@mail.gmail.com>

for last question, looks it should work. I notice the query_column has a
mysql namespace. And the generate person-odb.hxx, it explicitly use
mysql::query_column which when parse will generate mysql::query.

On Thu, Sep 13, 2012 at 6:32 PM, Scott Zhang <macromarship@gmail.com> wrote:

> Hello.
>     I am reading code to try to add pagination to odb. For orderby and
> groupby etc, as they are trial in most case, so append sql string is
> acceptable. Because I am going to replace my 2 projects DB access (one
> mysql, another oracle) library with odb, so I think add pagination will
> worth that.
>
>     Well. odb in my view is a really complex and heavy used template
> design. I am not sure the benefit of this. If the template was used to
> improve the flexibly to support static compiling check etc, because of so
> many different data types. I do think we have easier design. If you could
> look at the C# subsonic's activerecord template, from a "ODB/Subsonice
> compiler" level, the generate code can itself be static typed so could need
> less or none template. Just subsonic can reverse generate class from
> database. odb can generate persistent class from transient class.
>
>    And from the code flow. Looks I only need to add a Take(int num_),
> Offset(int offset_), (maybe OrderBy(string, enum orderby?)  and more ), in
> the "query" class then we can support "pagination". Because
>   for the query to run:
> eg:
>   result r (db->query<person> (query::age > 30));
> although the first sight to convert "query::age>30" or even more complex
> expression into a where clause looks mysterious. But after reading code,
> query::age itself is a query_column, which can parse ">30" and return a
> query. Then by the override operators, query can parse the left expressions
> itself. Then the place for the Offset, Take(orderby, groupby ....) could be
> after all where clause processed. We just need to call on  query object like
>
>    result r(db->query<person>(
> (query::aget>30).orderby(query::aget::column(), DESC).offset(10).take(10)) )
>
> then we can get the 10 - 20 rows.
>
>     And another question, from the code, looks odb::query linked with its
> mysql implementation  so the query it returned will be a mysql query. But
> if we link with both mysql and oracle at same time, eg for an application
> need to use mysql and oracle at same time. Can odb::query work correctly?
>
>
> Thanks.
> Regards
> Scott
>
>
>
From boris at codesynthesis.com  Thu Sep 13 08:11:22 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Thu Sep 13 07:52:10 2012
Subject: [odb-users] Re: Pagination and "join","groupby" in odb
In-Reply-To: <CABy3iKzfUf26Tm5_wOvG8eQCowmyq=Og-+PkafKBUk9Ehq0v9Q@mail.gmail.com>
References: <CABy3iKzZYvp-RcK_2eX=GCotSqVwL+XY=89Shx3DvOtVJx3RiA@mail.gmail.com>
	<CABy3iKzbUuapN+35JMB36BKRk5=2YY2iewVyVHKt5yjQKGSVUg@mail.gmail.com>
	<boris.20120912134136@codesynthesis.com>
	<CABy3iKzfUf26Tm5_wOvG8eQCowmyq=Og-+PkafKBUk9Ehq0v9Q@mail.gmail.com>
Message-ID: <boris.20120913140014@codesynthesis.com>

Hi Scott,

Scott Zhang <macromarship@gmail.com> writes:

> Especially for pagination, odb should hide the different implement of
> database from end users.

That would be a good idea if all the databases that ODB supports had
a way to implement pagination without incurring any extra overheads
(e.g., dynamic queries, etc). Unfortunately, in some databases (e.g.,
Oracle, SQL Server prior to 2012) pagination requires a non-trivial
effort to implement which would in turn also require changes to the
fairly optimal query model that ODB currently implements.


> Is there a plan for this (order by, pagination)?

ORDER BY, yes. Pagination, probably not at this stage.

ORDER BY is already quite easy to do:

query<person> ((query::age < 30) + "ORDER BY" + query::name);

But, I agree, we should wrap it in some syntactic sugar. Something
along these lines:

query<person> (query::age < 30, order_by (query::name));

Boris

From boris at codesynthesis.com  Thu Sep 13 08:40:00 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Thu Sep 13 08:20:56 2012
Subject: [odb-users] About using odb in same program to connect to
	mysql/sqlite/pgsql at same time
In-Reply-To: <CABy3iKx+mFwu1-PGKs6y7aP7oO=GW5a0Jcq0mpRdVXerZovvBQ@mail.gmail.com>
References: <CABy3iKx+mFwu1-PGKs6y7aP7oO=GW5a0Jcq0mpRdVXerZovvBQ@mail.gmail.com>
Message-ID: <boris.20120913142145@codesynthesis.com>

Hi Scott,

Scott Zhang <macromarship@gmail.com> writes:

>     Well. odb in my view is a really complex and heavy used template
> design.

Actually, while ODB does use templates, most of it is what I would call
"traditional" template usage. Things like traits, class templates, etc.
In particular, there is very little template metaprogramming, mostly to
keep things portable and compilation times sane.

In fact, some people suggest that we don't use templates enough! In
their optinion, we should get rid of the ODB compiler and re-implement
all this functionality using metaprogramming (and preprocessor). Go
figure.


> And from the code flow. Looks I only need to add a Take(int num_),
> Offset(int offset_), (maybe OrderBy(string, enum orderby?)  and more ), in
> the "query" class then we can support "pagination". Because for the
> query to run: eg:
>   result r (db->query<person> (query::age > 30));
> although the first sight to convert "query::age>30" or even more complex
> expression into a where clause looks mysterious. But after reading code,
> query::age itself is a query_column, which can parse ">30" and return a
> query.

Yes, you seem to have figured it out, even with all these pesky templates.


> Then by the override operators, query can parse the left expressions
> itself. Then the place for the Offset, Take(orderby, groupby ....) could be
> after all where clause processed. We just need to call on query object like
> 
> result r(db->query<person>(
> (query::aget>30).orderby(query::aget::column(), DESC).offset(10).take(10)))

Well, this would require (custom) modifications to ODB. You could also
achieve pretty much the same but with a slightly different syntax without
any ODB modifications. For example:

template <typename T, odb::mysql::database_type_id ID>
odb::mysql::query 
order_by (const odb::mysql::query_column<T, ID>& c)
{
  return "ORDER BY" + c;
}

odb::mysql::query 
limit (std::size_t n)
{
  return "LIMIT" + odb::mysql::query::_val (n);
}

odb::mysql::query 
offset (std::size_t n)
{
  return "OFFSET" + odb::mysql::query::_val (n);
}

And then:

db->query<person> ((query::age > 30) + 
                   order_by (query::name) + 
                   offset (10) + 
                   limit (20));

>     And another question, from the code, looks odb::query linked with its
> mysql implementation  so the query it returned will be a mysql query. But
> if we link with both mysql and oracle at same time, eg for an application
> need to use mysql and oracle at same time. Can odb::query work correctly?

No, this is not yet properly supported by ODB, though it is high up on our
TODO list. For more information, see this earlier post:

http://www.codesynthesis.com/pipermail/odb-users/2012-August/000720.html

Boris

From macromarship at gmail.com  Thu Sep 13 09:58:31 2012
From: macromarship at gmail.com (Scott Zhang)
Date: Thu Sep 13 09:58:40 2012
Subject: [odb-users] About using odb in same program to connect to
	mysql/sqlite/pgsql at same time
In-Reply-To: <boris.20120913142145@codesynthesis.com>
References: <CABy3iKx+mFwu1-PGKs6y7aP7oO=GW5a0Jcq0mpRdVXerZovvBQ@mail.gmail.com>
	<boris.20120913142145@codesynthesis.com>
Message-ID: <CABy3iKz7gR0JThgaTzEtcz0ULq3kn+2qOA6u1vPqa+1xVwK-GQ@mail.gmail.com>

Yes.

For pagination, I don't think the difference is so large.

I have been working with SqlServer 2005 - 2008 since 2005. use row_number()
function previous is the only way to do it, and it works fine for moderate
data size. For now, sql2linq support directly pagination which I don't know
how is that implemented.
for non linq app, I think row_num should work.
http://www.witheringtree.com/2012/02/overloading-codeigniter-db-driver/


Mysql/sqlite/pgsql should support limit.

Oracle have something like this
$newsql = "SELECT * FROM (select inner_query.*, rownum rnum FROM ($sql)
inner_query WHERE rownum < $limit)";

                if ($offset != 0)
                {
                        $newsql .= " WHERE rnum >= $offset";
                }




I am aware we can insert any sql into "query" to solve most of questions.
Just it make orm looks weird. That's why I suggest order by and pagination
integration. All orm like softwares provide these function instead of
inject sql.


On Thu, Sep 13, 2012 at 8:40 PM, Boris Kolpackov <boris@codesynthesis.com>wrote:

> Hi Scott,
>
> Scott Zhang <macromarship@gmail.com> writes:
>
> >     Well. odb in my view is a really complex and heavy used template
> > design.
>
> Actually, while ODB does use templates, most of it is what I would call
> "traditional" template usage. Things like traits, class templates, etc.
> In particular, there is very little template metaprogramming, mostly to
> keep things portable and compilation times sane.
>
> In fact, some people suggest that we don't use templates enough! In
> their optinion, we should get rid of the ODB compiler and re-implement
> all this functionality using metaprogramming (and preprocessor). Go
> figure.
>
>
> > And from the code flow. Looks I only need to add a Take(int num_),
> > Offset(int offset_), (maybe OrderBy(string, enum orderby?)  and more ),
> in
> > the "query" class then we can support "pagination". Because for the
> > query to run: eg:
> >   result r (db->query<person> (query::age > 30));
> > although the first sight to convert "query::age>30" or even more complex
> > expression into a where clause looks mysterious. But after reading code,
> > query::age itself is a query_column, which can parse ">30" and return a
> > query.
>
> Yes, you seem to have figured it out, even with all these pesky templates.
>
>
> > Then by the override operators, query can parse the left expressions
> > itself. Then the place for the Offset, Take(orderby, groupby ....) could
> be
> > after all where clause processed. We just need to call on query object
> like
> >
> > result r(db->query<person>(
> > (query::aget>30).orderby(query::aget::column(),
> DESC).offset(10).take(10)))
>
> Well, this would require (custom) modifications to ODB. You could also
> achieve pretty much the same but with a slightly different syntax without
> any ODB modifications. For example:
>
> template <typename T, odb::mysql::database_type_id ID>
> odb::mysql::query
> order_by (const odb::mysql::query_column<T, ID>& c)
> {
>   return "ORDER BY" + c;
> }
>
> odb::mysql::query
> limit (std::size_t n)
> {
>   return "LIMIT" + odb::mysql::query::_val (n);
> }
>
> odb::mysql::query
> offset (std::size_t n)
> {
>   return "OFFSET" + odb::mysql::query::_val (n);
> }
>
> And then:
>
> db->query<person> ((query::age > 30) +
>                    order_by (query::name) +
>                    offset (10) +
>                    limit (20));
>
> >     And another question, from the code, looks odb::query linked with its
> > mysql implementation  so the query it returned will be a mysql query. But
> > if we link with both mysql and oracle at same time, eg for an application
> > need to use mysql and oracle at same time. Can odb::query work correctly?
>
> No, this is not yet properly supported by ODB, though it is high up on our
> TODO list. For more information, see this earlier post:
>
> http://www.codesynthesis.com/pipermail/odb-users/2012-August/000720.html
>
> Boris
>
From chris.richards at yellowfeather.co.uk  Mon Sep 17 12:57:31 2012
From: chris.richards at yellowfeather.co.uk (Chris Richards)
Date: Mon Sep 17 12:57:49 2012
Subject: [odb-users] Which version of SQLite is required?
Message-ID: <5057567B.20900@yellowfeather.co.uk>

Hi,

I'm just trying to configure and make the libodb-sqlite files on CentOS 
5 and am getting an error stating that sqlite3_open_v2 and 
sqlite3_prepare_v2 are not declared. The version of SQLite on my system 
is 3.3.6 and on the ODB website 
<http://www.codesynthesis.com/products/odb/features.xhtml> it states 
that ODB supports SQLite 3.0.7 and later. It looks as though the 
sqlite3_open_v2 and sqlite3_prepare_v2 were introduced in SQLite v3.5.0 
http://www.sqlite.org/34to35.html.

Am I missing something or is SQlite v3.5 now required?

Regards,
Chris


From boris at codesynthesis.com  Mon Sep 17 13:39:49 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Mon Sep 17 13:19:54 2012
Subject: [odb-users] Which version of SQLite is required?
In-Reply-To: <5057567B.20900@yellowfeather.co.uk>
References: <5057567B.20900@yellowfeather.co.uk>
Message-ID: <boris.20120917193437@codesynthesis.com>

Hi Chris,

Chris Richards <chris.richards@yellowfeather.co.uk> writes:

> I'm just trying to configure and make the libodb-sqlite files on CentOS  
> 5 and am getting an error stating that sqlite3_open_v2 and  
> sqlite3_prepare_v2 are not declared. The version of SQLite on my system  
> is 3.3.6 and on the ODB website  
> <http://www.codesynthesis.com/products/odb/features.xhtml> it states  
> that ODB supports SQLite 3.0.7 and later. It looks as though the  
> sqlite3_open_v2 and sqlite3_prepare_v2 were introduced in SQLite v3.5.0  
> http://www.sqlite.org/34to35.html.

Yes, ODB uses both sqlite3_open_v2 and sqlite3_prepare_v2. And you
are right, that means ODB only supports v3.5.3 or later. I will update
the documentation.

If you are going to build SQLite from source and you plan to access
the database from multiple threads, then make sure that you enable
the unlock notification functionality in SQLite:

./configure CFLAGS="-O3 -DSQLITE_ENABLE_UNLOCK_NOTIFY=1"

Boris

From danielpeterjames at gmail.com  Mon Sep 17 13:43:56 2012
From: danielpeterjames at gmail.com (Daniel James)
Date: Mon Sep 17 13:44:24 2012
Subject: [odb-users] performance of updates in table with many rows for
	sqlite db
In-Reply-To: <boris.20120912132725@codesynthesis.com>
References: <CAAqNS9abv+KyR9V-880w6aJ=zuUGKaR-cnW_KjoLQ8y+Cth8Eg@mail.gmail.com>
	<boris.20120912132725@codesynthesis.com>
Message-ID: <CAAqNS9ZeZR3hLK9ADFipEdeZ4y-kgVNsQi_QGR=dTTPW0-HRZQ@mail.gmail.com>

On 12 September 2012 12:39, Boris Kolpackov <boris@codesynthesis.com> wrote:
> Hi Daniel,
>
> Daniel James <danielpeterjames@gmail.com> writes:
>
>> I store about 6 million of these in a table. Then I try and update a
>> subset of about 40 000.
>>
>> What is the quickest way to do this? At the moment it is proceeding at
>> a rate of about one update per second with the following SQL being
>> generated at each update:
>>
>> [...]
>>
[...]
>    So, generally, there is an optimum number of updates that you need
>    to perform per transaction to achieve the best performance. And the
>    way to find this number is through some experimentation.

Many thanks for the advice. I made a simpler model and found that I
was able to do 40 000 row updates to a table with 6 000 000 rows in
about 2-3 seconds. This seems quite fast actually! I found about 8 000
updates per transaction was effective.

Unfortunately, batching gave little improvement in the unsimplified
model. I think the reason things are going so slowly with the full
model, (about 1 update per second), is down to indices and/or foreign
keys pointing in to the large table. I will make further investigation
into this.

I can't remember reading about an ODB interface to drop and rebuild
indices before and after large updates. Would this be a useful or
doable functionality do you think? Otherwise I'll try the native SQL
functionality.


Daniel

From boris at codesynthesis.com  Tue Sep 18 05:38:17 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Tue Sep 18 05:18:12 2012
Subject: [odb-users] ODB 2.1.0 released
Message-ID: <boris.20120918113634@codesynthesis.com>

Hi,

We have released ODB 2.1.0. The NEWS file entries for this release are
as follows:

 * The ODB compiler is now capable of automatically discovering accessor and
   modifier functions for inaccessible data members in persistent classes,
   composite value types, and views. It will then use these accessors and
   modifiers in the generated code instead of trying to access such data
   members directly. The names of these functions are derived from the
   data member names and, by default, the ODB compiler will look for names
   in the form: get_foo/set_foo, getFoo/setFoo, getfoo/setfoo, and just
   foo. You can also add custom name derivations with the --accessor-regex
   and --modifier-regex ODB compiler options. For more information, refer
   to Section 3.2, "Declaring Persistent Objects and Values" in the ODB
   manual.

 * New pragmas, get, set, and access, allow the specification of custom
   accessor and modifier expressions for data members in persistent classes,
   composite value types, and views. For more information, refer to Section
   12.4.5, "get/set/access" in the ODB manual as well as the 'access' example
   in the odb-examples package.

 * New pragma, virtual, allows the declaration of virtual data members. A
   virtual data member is an imaginary data member that is only used for
   the purpose of database persistence. This mechanism can be useful to
   aggregate or dis-aggregate real data members, implement the pimpl idiom,
   and to handle third-party types for which names of real data members may
   not be known. For more information, refer to Section 12.4.13, "virtual"
   in the ODB manual as well as the 'access' and 'pimpl' examples in the
   odb-examples package.

 * Support for defining database indexes. Both simple and composite indexes
   can be defined with support for database-specific index types, methods,
   and options. For more information, refer to Section 12.6, "Index
   Definition Pragmas" as well as Sections [13-17].16, "<Database> Index
   Definition" in the ODB manual.

 * Support for mapping extended database types, such as geospatial types,
   user-defined types, and collections. This mechanism allows you to map
   any database type to one of the types for which ODB provides built-in
   support (normally string or binary). The text or binary representation
   of the data can then be extracted into a C++ data type of your choice.
   For more information, refer to Section 12.7, "Database Type Mapping
   Pragmas" in the ODB manual.

 * The Boost profile now provides persistence support for the Boost Multi-
   Index container (boost::multi_index_container). For more information,
   refer to Section 19.3, "Multi-Index Container Library" in the ODB manual.

 * The Boost profile now provides persistence support for the Boost uuid type
   (boost::uuids::uuid). For more information, refer to Section 19.6, "Uuid
   Library" in the ODB manual as well as the 'boost' example in the odb-
   examples package.

 * The Qt profile now provides persistence support for the QUuid type. For
   more information, refer to Section 20.1, "Basic Types" in the ODB manual
   as well as the 'qt' example in the odb-examples package.

 * SQLite improvements: Persistence support for std::wstring on Windows
   (Section 14.1, "SQLite Type Mapping"). Ability to pass the database
   name as std::wstring on Windows (Section 14.2, "SQLite Database Class").
   Ability to specify the virtual filesystem (vfs) module in the database
   constructors (Section 14.2, "SQLite Database Class").

 * Support for mapping C++11 std::array<char, N> and std::array<unsigned
   char, N> types to BLOB/BINARY database types. For more information,
   refer to Sections [13-17].1, "<Database> Type Mapping" in the ODB manual.

 * Support for mapping the char[16] array to PostgreSQL UUID and SQL Server
   UNIQUEIDENTIFIER types. For more information, refer to Sections 15.1,
   "PostgreSQL Type Mapping" and 17.1, "SQL Server Type Mapping" in the
   ODB manual.

 * New option, --output-name, specifies the alternative base name used to
   construct the names of the generated files. Refer to the ODB compiler
   command line interface documentation (man pages) for details.

 * New option, --generate-schema-only, instructs the ODB compiler to
   generate the database schema only. Refer to the ODB compiler command
   line interface documentation (man pages) for details.

 * New option, --at-once, triggers the generation of code for all the input
   files as well as for all the files that they include at once. Refer to
   the ODB compiler command line interface documentation (man pages) for
   details.

 * New options, --sql-interlude and --sql-interlude-file, allow the insertion
   of custom SQL between the DROP and CREATE statements in the generated
   database schema file.

 * New options, --omit-drop and --omit-create, trigger the omission of DROP
   and CREATE statements, respectively, from the generated database schema.

 * New ODB manual Section, 6.3 "Circular Relationships", explains how to
   handle persistent classes with circular dependencies that are defined
   in separate headers.

 * The id() pragma that was used to declare a persistent class without an
   object id has been renamed to no_id.

 * New pragma, definition, allows the specification of an alternative code
   generation point for persistent classes, views, and composite value
   types. This mechanism is primarily useful for converting third-party
   types to ODB composite value types. For more information, refer to
   Section 12.3.7, "Definition" in the ODB manual.

 * The session constructor now accepts an optional bool argument (true by
   default) which indicates whether to make this session current for this
   thread. For more information, refer to Chapter 10, "Session" in the ODB
   manual.

 * Simplified Oracle automatically-assigned object id implementation that
   does not rely on triggers.

 * Support for mapping boost::posix_time::ptime and QDateTime to the DATE
   Oracle type. For more information, refer to Sections 19.4.4 (Boost) and
   20.4.4 (Qt) in the ODB manual.

 * Default SQLite mapping for float and double now allows NULL since SQLite
   treats NaN FLOAT values as NULL. For more information, refer to Section
   14.1, "SQLite Type Mapping" in the ODB manual.

We have also added Visual Studio 2012 and Clang 3.1 to the list of compilers
that we use for testing each release. Specifically, all the runtime libraries,
examples, and tests now come with project/solution files for Visual Studio
2012 in addition to 2010 and 2008. For the complete list of compilers that
were used to test this release, refer to the ODB Platforms and Compilers
page:

http://www.codesynthesis.com/products/odb/platforms.xhtml

A more detailed discussion of the major new features can be found in the
following blog post:

http://www.codesynthesis.com/~boris/blog/2012/09/18/odb-2-1-0-released/

We would also like to thank everyone who reported bugs, suggested fixes
or new features, as well as tested early versions of this release.

Source code and pre-compiled binary packages for this release are available
from the ODB Download page:

http://www.codesynthesis.com/products/odb/download.xhtml

SHA1 checksums for the files in this release are as follows:

87d58649f5c0f405e8453e02d270da2f87a4ccbb  libodb-2.1.0.tar.bz2
338f800db1d7d1b4cd78851a6421dcbc9b3d0996  libodb-2.1.0.tar.gz
cc510d90c7333c3ee779ed21eff6bc8101fb4b9b  libodb-2.1.0.zip
5c3eb40b834b385e7ac4b3c20f6cd82651abb66d  libodb-boost-2.1.0.tar.bz2
bf0aaf000bf4ae5f29109fb213bec01b626c5331  libodb-boost-2.1.0.tar.gz
20cd82d6d7605c569f023e07a91ad666a9dac5d4  libodb-boost-2.1.0.zip
1bf4e89a16693948b31af2f250ab1fbd5b3615c8  libodb-mssql-2.1.0.tar.bz2
8f14c156d02db9669c15e6a21beb239c83586113  libodb-mssql-2.1.0.tar.gz
7beb4b7462a5a884095ca944c101265d2068f2d3  libodb-mssql-2.1.0.zip
e58eca3465a6c9b278113bc3faa26455a94d040c  libodb-mysql-2.1.0.tar.bz2
529a8a3141de27494107d0c9a210dd0698cbfb55  libodb-mysql-2.1.0.tar.gz
580cef317c20f199ae3098faa57c4d4c80564381  libodb-mysql-2.1.0.zip
f2c92818e89fa06f90d20c48e63c64775f75375c  libodb-oracle-2.1.0.tar.bz2
96022e1e8b7fbac4a72b8302019710b9607f760c  libodb-oracle-2.1.0.tar.gz
18198da3a867e2208affcc38370668cc6747e93c  libodb-oracle-2.1.0.zip
464a970ee1cfef88e7a53ab6f4cd0ae95320cce9  libodb-pgsql-2.1.0.tar.bz2
26c4ac2db75951b93f5a6b6c7aa917e3f7f4bef6  libodb-pgsql-2.1.0.tar.gz
8fd686ab87b52b144e6976a47c1db3745d483d62  libodb-pgsql-2.1.0.zip
ba04acd724ad7253ce7b6d534ee3354c3fcd0951  libodb-qt-2.1.0.tar.bz2
a89dbfc22b85f5efb2788b8c08c7280159caf032  libodb-qt-2.1.0.tar.gz
0dd1c1cb7672b6b1eec57f1e670d67dcc6ac1cbf  libodb-qt-2.1.0.zip
58085bdb5d890cc952594cc288d28db9c93d3349  libodb-sqlite-2.1.0.tar.bz2
f8c16947ec322d989d9526a0f250b132e1bfae64  libodb-sqlite-2.1.0.tar.gz
874e0bfdf740053e6e5bdc3cf134cb954dfdbd2c  libodb-sqlite-2.1.0.zip
4556ab5a60a19170eb0f7376756a07829d87418b  odb-2.1.0-i686-linux-gnu.tar.bz2
2bd23a560579b3ab5f9d66de223d8875f8356a2d  odb-2.1.0-i686-macosx.tar.bz2
9e5be40268ee0650b1f8cb61f92dff06c851ede0  odb-2.1.0-i686-solaris.tar.bz2
15b550ab80b48598aadd84e2dd4f1b4960854c38  odb-2.1.0-i686-windows.zip
82021b5c393d50206d0ba73e9d3a78258dbeb2c9  odb-2.1.0-sparc-solaris.tar.bz2
5e435d952e5262b5de49df0392e375e0d4440515  odb-2.1.0.tar.bz2
cd4cb4d039e184a3d221bb75bd3c4c400fdf2cd5  odb-2.1.0.tar.gz
f5f1c62b8eeaa03b0646d7bb1f43aeac3e6378bb  odb-2.1.0-x86_64-linux-gnu.tar.bz2
d8daf4a2be8c055a39f139f0986fa151fe8b3342  odb-2.1.0.zip
0994d39cc7ab3607b5c6e88f7c16edf32e0d6336  odb-examples-2.1.0.tar.bz2
e097d5771f9cb9e636c1a293ca9e3f6f407d22eb  odb-examples-2.1.0.tar.gz
8d0cf9419c445e92c5e5cc991dd314b4410d2c39  odb-examples-2.1.0.zip
3c5ecd517fc72764ae4ecfca28fba75c2a22454f  odb-tests-2.1.0.tar.bz2
06ebaea3383264ceec9ff9ff05668cf80f461014  odb-tests-2.1.0.tar.gz
3cd8ca94d885d81ffb92dd631b9ba2d2cd50aa25  odb-tests-2.1.0.zip

Enjoy,
	Boris

From danielpeterjames at gmail.com  Tue Sep 18 05:21:59 2012
From: danielpeterjames at gmail.com (Daniel James)
Date: Tue Sep 18 05:22:26 2012
Subject: [odb-users] performance of updates in table with many rows for
	sqlite db
In-Reply-To: <CAAqNS9ZeZR3hLK9ADFipEdeZ4y-kgVNsQi_QGR=dTTPW0-HRZQ@mail.gmail.com>
References: <CAAqNS9abv+KyR9V-880w6aJ=zuUGKaR-cnW_KjoLQ8y+Cth8Eg@mail.gmail.com>
	<boris.20120912132725@codesynthesis.com>
	<CAAqNS9ZeZR3hLK9ADFipEdeZ4y-kgVNsQi_QGR=dTTPW0-HRZQ@mail.gmail.com>
Message-ID: <CAAqNS9ZuwCTzvGugnSoUK2pMrT8ik_mppBxgZMfGZxWgNZ5R1w@mail.gmail.com>

On 17 September 2012 18:43, Daniel James <danielpeterjames@gmail.com> wrote:
> On 12 September 2012 12:39, Boris Kolpackov <boris@codesynthesis.com> wrote:
>> Hi Daniel,
>>
>> Daniel James <danielpeterjames@gmail.com> writes:
>>
>>> I store about 6 million of these in a table. Then I try and update a
>>> subset of about 40 000.
>>>
>>> What is the quickest way to do this? At the moment it is proceeding at
>>> a rate of about one update per second with the following SQL being
>>> generated at each update:
>>>
>>> [...]
>>>
> [...]
>>    So, generally, there is an optimum number of updates that you need
>>    to perform per transaction to achieve the best performance. And the
>>    way to find this number is through some experimentation.
>
> Many thanks for the advice. I made a simpler model and found that I
> was able to do 40 000 row updates to a table with 6 000 000 rows in
> about 2-3 seconds. This seems quite fast actually! I found about 8 000
> updates per transaction was effective.
>
> Unfortunately, batching gave little improvement in the unsimplified
> model. I think the reason things are going so slowly with the full
> model, (about 1 update per second), is down to indices and/or foreign
> keys pointing in to the large table. I will make further investigation
> into this.
>
> I can't remember reading about an ODB interface to drop and rebuild
> indices before and after large updates. Would this be a useful or
> doable functionality do you think? Otherwise I'll try the native SQL
> functionality.

Actually, I think an even better fix for my problem would be to be
able to update individual columns. I understand ODB includes all
columns in the update statement for any column. Is it correct that
index updates could be avoided if 'dumb' columns could be updated in
isolation?

Daniel

>
>
> Daniel

From boris at codesynthesis.com  Tue Sep 18 06:19:12 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Tue Sep 18 05:59:08 2012
Subject: [odb-users] performance of updates in table with many rows for
	sqlite db
In-Reply-To: <CAAqNS9ZuwCTzvGugnSoUK2pMrT8ik_mppBxgZMfGZxWgNZ5R1w@mail.gmail.com>
References: <CAAqNS9abv+KyR9V-880w6aJ=zuUGKaR-cnW_KjoLQ8y+Cth8Eg@mail.gmail.com>
	<boris.20120912132725@codesynthesis.com>
	<CAAqNS9ZeZR3hLK9ADFipEdeZ4y-kgVNsQi_QGR=dTTPW0-HRZQ@mail.gmail.com>
	<CAAqNS9ZuwCTzvGugnSoUK2pMrT8ik_mppBxgZMfGZxWgNZ5R1w@mail.gmail.com>
Message-ID: <boris.20120918120632@codesynthesis.com>

Hi Daniel,

Daniel James <danielpeterjames@gmail.com> writes:

> Actually, I think an even better fix for my problem would be to be
> able to update individual columns.

This is not yet supported directly by ODB.


> I understand ODB includes all columns in the update statement for
> any column. Is it correct that index updates could be avoided if
> 'dumb' columns could be updated in isolation?

I am pretty sure most databases are able to detect that the value
being updated has not changed and thus avoid any expensive operations
(e.g., write the new value to disk or rebuild indexes). However, in
the case of SQLite, anything is possible. I've been surprised by this
database multiple times already. So you may want to write to the SQLite
mailing list and confirm if this is the case.


> I can't remember reading about an ODB interface to drop and rebuild
> indices before and after large updates. Would this be a useful or
> doable functionality do you think? Otherwise I'll try the native SQL
> functionality.

The just-released ODB 2.1.0 provides support for defining database
indexes on data members. However, this is purely for schema generation.
If you need anything fancy like dropping the index before a bunch of
updates and then re-building them after, then you will need to do it
yourself using native SQL.

Boris

From pstath at axxcelera.com  Tue Sep 18 16:37:25 2012
From: pstath at axxcelera.com (Stath, Paul)
Date: Tue Sep 18 16:37:35 2012
Subject: [odb-users] odb-test-2.1.0 make check deadlocks in thread test.
Message-ID: <3233D27CC5658E4598557F8521F6B07E2143E65213@RIC-MS01.abw.int>

I downloaded ODB 2.1.0 and successfully compiled it from source.

When I run 'make check' using the 2.1.0 version of odb-test, the 'common/threads' test hangs until I terminate it with SIGINT.

~/src/odb-tests-2.1.0/common/threads$ make check
make  check-am
make[1]: Entering directory `/home/pstath/src/odb-tests-2.1.0/common/threads'
make  check-TESTS
make[2]: Entering directory `/home/pstath/src/odb-tests-2.1.0/common/threads'
./driver --options-file ../../db.options
^Cmake[2]: *** [check-TESTS] Interrupt
make[1]: *** [check-am] Interrupt
make: *** [check] Interrupt

My setup is as follows:

Ubuntu 12.04 LTS

~/src/odb-tests-2.1.0/common/threads$ g++ -v
Using built-in specs.
COLLECT_GCC=g++
COLLECT_LTO_WRAPPER=/usr/lib/gcc/i686-linux-gnu/4.6/lto-wrapper
Target: i686-linux-gnu
Configured with: ../src/configure -v --with-pkgversion='Ubuntu/Linaro 4.6.3-1ubuntu5' --with-bugurl=file:///usr/share/doc/gcc-4.6/README.Bugs --enable-languages=c,c++,fortran,objc,obj-c++ --prefix=/usr --program-suffix=-4.6 --enable-shared --enable-linker-build-id --with-system-zlib --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --with-gxx-include-dir=/usr/include/c++/4.6 --libdir=/usr/lib --enable-nls --with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-gnu-unique-object --enable-plugin --enable-objc-gc --enable-targets=all --disable-werror --with-arch-32=i686 --with-tune=generic --enable-checking=release --build=i686-linux-gnu --host=i686-linux-gnu --target=i686-linux-gnu
Thread model: posix
gcc version 4.6.3 (Ubuntu/Linaro 4.6.3-1ubuntu5)

The SQLite library is compiled with SQLITE_ENABLE_UNLOCK_NOTIFY defined.

~/src/sqlite$ nm /usr/local/lib/libsqlite3.so | grep unlock_notify
000472c0 T sqlite3_unlock_notify

It is probably something I'm doing wrong.
Let me know if you need more information.
From boris at codesynthesis.com  Wed Sep 19 02:07:43 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Wed Sep 19 01:47:34 2012
Subject: [odb-users] odb-test-2.1.0 make check deadlocks in thread test.
In-Reply-To: <3233D27CC5658E4598557F8521F6B07E2143E65213@RIC-MS01.abw.int>
References: <3233D27CC5658E4598557F8521F6B07E2143E65213@RIC-MS01.abw.int>
Message-ID: <boris.20120919080129@codesynthesis.com>

Hi Paul,

Stath, Paul <pstath@axxcelera.com> writes:

> When I run 'make check' using the 2.1.0 version of odb-test, the 
> 'common/threads' test hangs until I terminate it with SIGINT.

This test performs a large number of database operations from multiple
threads and using different connection pool configurations. So it takes
some time to complete. For example, on my box with 15K SAS drives it
take a bit over a minute. But if I run with the database on a slow-
ish USB drive, it can take several minutes. Can you wait maybe 5
minutes and see if it completes successfully?


> The SQLite library is compiled with SQLITE_ENABLE_UNLOCK_NOTIFY defined.
> 
> ~/src/sqlite$ nm /usr/local/lib/libsqlite3.so | grep unlock_notify
> 000472c0 T sqlite3_unlock_notify
> 
> It is probably something I'm doing wrong. Let me know if you need
> more information.

I don't see anything wrong with your setup. If waiting for the test
to finish does not help, can you let me know which version of SQLite
you are using so that I can try to reproduce this problem?

Boris

From danielpeterjames at gmail.com  Wed Sep 19 09:50:36 2012
From: danielpeterjames at gmail.com (Daniel James)
Date: Wed Sep 19 09:51:05 2012
Subject: [odb-users] performance of updates in table with many rows for
	sqlite db
In-Reply-To: <boris.20120918120632@codesynthesis.com>
References: <CAAqNS9abv+KyR9V-880w6aJ=zuUGKaR-cnW_KjoLQ8y+Cth8Eg@mail.gmail.com>
	<boris.20120912132725@codesynthesis.com>
	<CAAqNS9ZeZR3hLK9ADFipEdeZ4y-kgVNsQi_QGR=dTTPW0-HRZQ@mail.gmail.com>
	<CAAqNS9ZuwCTzvGugnSoUK2pMrT8ik_mppBxgZMfGZxWgNZ5R1w@mail.gmail.com>
	<boris.20120918120632@codesynthesis.com>
Message-ID: <CAAqNS9bKYyQwBBLX3S69vVU=MtuRUGng-FzY0V2h1crzvMmGAg@mail.gmail.com>

On 18 September 2012 11:19, Boris Kolpackov <boris@codesynthesis.com> wrote:
> Hi Daniel,
>
> Daniel James <danielpeterjames@gmail.com> writes:
>
>> Actually, I think an even better fix for my problem would be to be
>> able to update individual columns.
>
> This is not yet supported directly by ODB.
>
>
>> I understand ODB includes all columns in the update statement for
>> any column. Is it correct that index updates could be avoided if
>> 'dumb' columns could be updated in isolation?
>
> I am pretty sure most databases are able to detect that the value
> being updated has not changed and thus avoid any expensive operations
> (e.g., write the new value to disk or rebuild indexes). However, in
> the case of SQLite, anything is possible. I've been surprised by this
> database multiple times already. So you may want to write to the SQLite
> mailing list and confirm if this is the case.
>

It appears that updating a row with a foreign key reference is
activating a trigger, even if the child key is unchanged. I have been
able to work around this with:

PRAGMA foreign_keys=OFF;

I've only been able to make this pragma work if used directly after
constructing the db, like this:

AppDelHelper(std::string dbName) : dbp_(new odb::sqlite::database(dbName))
    {
        odb::connection_ptr c(dbp_->connection());
        c->execute("PRAGMA foreign_keys=OFF;");
    }

In other, later, parts of the program the pragma has no effect. I
think this is to do with the following from the sqlite docs:

"It is not possible to enable or disable foreign key constraints in
the middle of a multi-statement transaction (when SQLite is not in
autocommit mode). Attempting to do so does not return an error; it
simply has no effect."

How can I ensure the db _is_ in autocommit mode? Simply being outside
a transaction does not seem to be enough. I'm only accessing the db
with a single thread, is it possible that ODB is keeping the
transaction open beyond than the transaction::commit() call?

Thanks again,

Daniel

From boris at codesynthesis.com  Wed Sep 19 10:33:54 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Wed Sep 19 10:14:27 2012
Subject: [odb-users] performance of updates in table with many rows for
	sqlite db
In-Reply-To: <CAAqNS9bKYyQwBBLX3S69vVU=MtuRUGng-FzY0V2h1crzvMmGAg@mail.gmail.com>
References: <CAAqNS9abv+KyR9V-880w6aJ=zuUGKaR-cnW_KjoLQ8y+Cth8Eg@mail.gmail.com>
	<boris.20120912132725@codesynthesis.com>
	<CAAqNS9ZeZR3hLK9ADFipEdeZ4y-kgVNsQi_QGR=dTTPW0-HRZQ@mail.gmail.com>
	<CAAqNS9ZuwCTzvGugnSoUK2pMrT8ik_mppBxgZMfGZxWgNZ5R1w@mail.gmail.com>
	<boris.20120918120632@codesynthesis.com>
	<CAAqNS9bKYyQwBBLX3S69vVU=MtuRUGng-FzY0V2h1crzvMmGAg@mail.gmail.com>
Message-ID: <boris.20120919163001@codesynthesis.com>

Hi Daniel,

Daniel James <danielpeterjames@gmail.com> writes:

> How can I ensure the db _is_ in autocommit mode? Simply being outside
> a transaction does not seem to be enough. I'm only accessing the db
> with a single thread, is it possible that ODB is keeping the
> transaction open beyond than the transaction::commit() call?

You need to turn the foreign key checking off before the transaction
starts and turn it back on after it has finished:

odb::connection_ptr c (db->connection ());
c->execute ("PRAGMA foreign_keys=OFF;");

transaction t (c->begin ()); // Note: not db->begin ()!

...

t.commit ();

c->execute ("PRAGMA foreign_keys=ON;");

Boris

From simon at 2ndQuadrant.com  Wed Sep 19 10:48:47 2012
From: simon at 2ndQuadrant.com (Simon Riggs)
Date: Wed Sep 19 10:48:54 2012
Subject: [odb-users] ODB 2.1 on PostgreSQL
Message-ID: <CA+U5nMJZaB2295JCxgQOsO1g4Wttnd7TzHGEz+VbKSKr7HQgxw@mail.gmail.com>

Hi,

Very happy to see new release of ODB on PostgreSQL.

I wanted to pass on a few comments on the PostgreSQL limitations
section, listed here
http://www.codesynthesis.com/products/odb/doc/manual.xhtml#15.5
to allow some further understanding on those areas.

15.5.1 Query Result Caching

The PostgreSQL ODB runtime implementation will always return a cached
query result (Section 4.4, "Query Result") even when explicitly
requested not to. This is a limitation of the PostgreSQL client
library (libpq) which does not support uncached (streaming) query
results.

PostgreSQL 9.2 now supports streaming...
http://www.postgresql.org/docs/9.2/static/libpq-single-row-mode.html


15.5.3 Unique Constraint Violations

Due to the granularity of the PostgreSQL error codes, it is impossible
to distinguish between the duplicate primary key and other unique
constraint violations. As a result, when making an object persistent,
the PostgreSQL ODB runtime will translate all unique constraint
violation errors to the object_already_persistent exception (Section
3.14, "ODB Exceptions").

> create table test2 (id integer primary key, id2 integer unique);
CREATE TABLE
> insert into test2 values (1, 1);
INSERT 0 1
ERROR:  duplicate key value violates unique constraint "test2_pkey"
DETAIL:  Key (id)=(1) already exists.
> insert into test2 values (2, 1);
ERROR:  duplicate key value violates unique constraint "test2_id2_key"
DETAIL:  Key (id2)=(1) already exists.

The error message itself doesn't differentiate between primary keys
and unique constraints, but the object naming convention shows that
PKs have the "_pkey" suffix by default. Changing that isn't
recommended, so you could probably rely on it.

Hope that helps,

-- 
 Simon Riggs                   http://www.2ndQuadrant.com/
 PostgreSQL Development, 24x7 Support, Training & Services

From adisabat at libero.it  Wed Sep 19 13:37:31 2012
From: adisabat at libero.it (adisabat)
Date: Wed Sep 19 13:37:49 2012
Subject: [odb-users] Problem compiler example for ODB C++  2.1.0 
Message-ID: <MALYAJ$2198F54BF6B67D9BFF11B37E9F115567@libero.it>


Hello everybody,

i'm trying to use the ODB C++ but while compiling the examples for version 2.1.0 i have got the following error messages from odb compiler:


make[2]: Entering directory `/root/ODB-C++/ODB_C++/odb-examples-2.1.0/access'
odb     --database mysql --generate-schema --generate-query --table-prefix access_ person.hxx
Using built-in specs.
COLLECT_GCC=/usr/local/bin/g++
Target: i686-pc-linux-gnu
Configured with: ./configure --with-gmp-include=/usr/local/include --with-gmp-lib=/usr/local/lib --with-mpfr-lib=/usr/local/lib --with-mpfr-include=/usr/local/include --with-mpc-include=/usr/local/include --with-mpc-lib=/usr/local/lib
Thread model: posix
gcc version 4.7.1 (GCC)
COLLECT_GCC_OPTIONS='-v' '-E' '-P' '-std=gnu++98' '-shared-libgcc' '-mtune=generic' '-march=pentiumpro'
 /usr/local/libexec/gcc/i686-pc-linux-gnu/4.7.1/cc1plus -E -quiet -v -P -D_GNU_SOURCE - -mtune=generic -march=pentiumpro -std=gnu++98
ignoring duplicate directory "/usr/lib/qt/include"
ignoring nonexistent directory "/usr/local/lib/gcc/i686-pc-linux-gnu/4.7.1/../../../../i686-pc-linux-gnu/include"
#include "..." search starts here:
#include <...> search starts here:
 /usr/lib/qt/include
 /usr/local/lib/gcc/i686-pc-linux-gnu/4.7.1/../../../../include/c++/4.7.1
 /usr/local/lib/gcc/i686-pc-linux-gnu/4.7.1/../../../../include/c++/4.7.1/i686-pc-linux-gnu
 /usr/local/lib/gcc/i686-pc-linux-gnu/4.7.1/../../../../include/c++/4.7.1/backward
 /usr/local/lib/gcc/i686-pc-linux-gnu/4.7.1/include
 /usr/local/include
 /usr/local/lib/gcc/i686-pc-linux-gnu/4.7.1/include-fixed
 /usr/include
End of search list.
COMPILER_PATH=/usr/local/libexec/gcc/i686-pc-linux-gnu/4.7.1/:/usr/local/libexec/gcc/i686-pc-linux-gnu/4.7.1/:/usr/local/libexec/gcc/i686-pc-linux-gnu/:/usr/local/lib/gcc/i686-pc-linux-gnu/4.7.1/:/usr/local/lib/gcc/i686-pc-linux-gnu/
LIBRARY_PATH=/usr/local/lib/gcc/i686-pc-linux-gnu/4.7.1/:/usr/local/lib/gcc/i686-pc-linux-gnu/4.7.1/../../../:/lib/:/usr/lib/
COLLECT_GCC_OPTIONS='-v' '-E' '-P' '-std=gnu++98' '-shared-libgcc' '-mtune=generic' '-march=pentiumpro'
odb: error: unable to parse profile paths
make[2]: *** [person-odb.hxx] Error 1
make[2]: Leaving directory `/root/ODB-C++/ODB_C++/odb-examples-2.1.0/access'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/root/ODB-C++/ODB_C++/odb-examples-2.1.0'
make: *** [all] Error 2

as you can see i'm suing the GCC 4.7.1 compiler and my system is Linux 2.6.38.2 and i cannot upgrade it because i have other applications running on.

Any suggest is welcome, many thanks in advance.

Rgds
 Antonio 



From pstath at axxcelera.com  Wed Sep 19 14:52:15 2012
From: pstath at axxcelera.com (Stath, Paul)
Date: Wed Sep 19 14:52:27 2012
Subject: [odb-users] odb-test-2.1.0 make check deadlocks in thread test.
In-Reply-To: <boris.20120919080129@codesynthesis.com>
References: <3233D27CC5658E4598557F8521F6B07E2143E65213@RIC-MS01.abw.int>
	<boris.20120919080129@codesynthesis.com>
Message-ID: <3233D27CC5658E4598557F8521F6B07E2143E65277@RIC-MS01.abw.int>

Boris --

Thanks for the quick reply.

I edited the driver.cxx source, and turned on tracing on the db object.
I found that the test was running and I was impatient.

This test takes a very long time to run on my system, which is utilizing
a 7200 RPM SATA/600 drive.

-- Paul


> -----Original Message-----
> From: Boris Kolpackov [mailto:boris@codesynthesis.com]
> Sent: Wednesday, September 19, 2012 2:08 AM
> To: Stath, Paul
> Cc: odb-users@codesynthesis.com
> Subject: Re: [odb-users] odb-test-2.1.0 make check deadlocks in thread
> test.
> 
> Hi Paul,
> 
> Stath, Paul <pstath@axxcelera.com> writes:
> 
> > When I run 'make check' using the 2.1.0 version of odb-test, the
> > 'common/threads' test hangs until I terminate it with SIGINT.
> 
> This test performs a large number of database operations from multiple
> threads and using different connection pool configurations. So it takes
> some time to complete. For example, on my box with 15K SAS drives it
> take a bit over a minute. But if I run with the database on a slow-
> ish USB drive, it can take several minutes. Can you wait maybe 5
> minutes and see if it completes successfully?
> 
> 
> > The SQLite library is compiled with SQLITE_ENABLE_UNLOCK_NOTIFY
> defined.
> >
> > ~/src/sqlite$ nm /usr/local/lib/libsqlite3.so | grep unlock_notify
> > 000472c0 T sqlite3_unlock_notify
> >
> > It is probably something I'm doing wrong. Let me know if you need
> > more information.
> 
> I don't see anything wrong with your setup. If waiting for the test
> to finish does not help, can you let me know which version of SQLite
> you are using so that I can try to reproduce this problem?
> 
> Boris

From boris at codesynthesis.com  Wed Sep 19 15:37:18 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Wed Sep 19 15:17:11 2012
Subject: [odb-users] odb-test-2.1.0 make check deadlocks in thread test.
In-Reply-To: <3233D27CC5658E4598557F8521F6B07E2143E65277@RIC-MS01.abw.int>
References: <3233D27CC5658E4598557F8521F6B07E2143E65213@RIC-MS01.abw.int>
	<boris.20120919080129@codesynthesis.com>
	<3233D27CC5658E4598557F8521F6B07E2143E65277@RIC-MS01.abw.int>
Message-ID: <boris.20120919213408@codesynthesis.com>

Hi Paul,

Stath, Paul <pstath@axxcelera.com> writes:

> This test takes a very long time to run on my system, which is utilizing
> a 7200 RPM SATA/600 drive.

It is quite a pathological test in that it is very write-heavy. Its
main purpose is to stress the ODB/database combination as much as
possible.

Boris

From boris at codesynthesis.com  Wed Sep 19 15:45:46 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Wed Sep 19 15:25:32 2012
Subject: [odb-users] Problem compiler example for ODB C++  2.1.0
In-Reply-To: <MALYAJ$2198F54BF6B67D9BFF11B37E9F115567@libero.it>
References: <MALYAJ$2198F54BF6B67D9BFF11B37E9F115567@libero.it>
Message-ID: <boris.20120919213813@codesynthesis.com>

Hi Antonio,

adisabat <adisabat@libero.it> writes:

> i'm trying to use the ODB C++ but while compiling the examples for
> version 2.1.0 i have got the following error messages from odb
> compiler:
> 
> odb --database mysql --generate-schema --generate-query --table-prefix access_ person.hxx
> Using built-in specs.
> COLLECT_GCC=/usr/local/bin/g++
> Target: i686-pc-linux-gnu
>
> [...]
>
> odb: error: unable to parse profile paths

This is bizarre. It looks like GCC for some reason writes its 
diagnostics (the stuff that it prints when you specify the -v option)
to STDOUT instead of STDERR. Can you try the following command for me:

g++ -v -x c++ -E -P /dev/null 1>/tmp/stdout.log 2>/tmp/stderr.log

And then let me know whether stderr.log or stdout.log contains the
output?

Also, if it is stderr.log, then can you try to run the above ODB
command directly from the command line instead of by running make?

Boris

From adisabat at libero.it  Thu Sep 20 05:05:15 2012
From: adisabat at libero.it (adisabat)
Date: Thu Sep 20 05:05:28 2012
Subject: [odb-users] Problem compiler example for ODB C++  2.1.0
Message-ID: <MAN58R$2913F9126380EE562EC8F13DDFDBED71@libero.it>

Hi Boris, 

many thanks for your help,

In my system the g++ command i wrapped into perl script command to color the compilation output to detect wanrning,info,errors. 

This seems not good for the odb compiler.

Now the odb compiler works properly. 

Thanks again,
Antonio






---------- Initial Header -----------

>From      : "Boris Kolpackov" boris@codesynthesis.com
To          : "adisabat" adisabat@libero.it
Cc          : "odb-users" odb-users@codesynthesis.com
Date      : Wed, 19 Sep 2012 21:45:46 +0200
Subject : Re: [odb-users] Problem compiler example for ODB C++  2.1.0







> Hi Antonio,
> 
> adisabat <adisabat@libero.it> writes:
> 
> > i'm trying to use the ODB C++ but while compiling the examples for
> > version 2.1.0 i have got the following error messages from odb
> > compiler:
> > 
> > odb --database mysql --generate-schema --generate-query --table-prefix access_ person.hxx
> > Using built-in specs.
> > COLLECT_GCC=/usr/local/bin/g++
> > Target: i686-pc-linux-gnu
> >
> > [...]
> >
> > odb: error: unable to parse profile paths
> 
> This is bizarre. It looks like GCC for some reason writes its 
> diagnostics (the stuff that it prints when you specify the -v option)
> to STDOUT instead of STDERR. Can you try the following command for me:
> 
> g++ -v -x c++ -E -P /dev/null 1>/tmp/stdout.log 2>/tmp/stderr.log
> 
> And then let me know whether stderr.log or stdout.log contains the
> output?
> 
> Also, if it is stderr.log, then can you try to run the above ODB
> command directly from the command line instead of by running make?
> 
> Boris
> 



From boris at codesynthesis.com  Thu Sep 20 07:21:29 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Thu Sep 20 07:01:12 2012
Subject: [odb-users] Problem compiler example for ODB C++  2.1.0
In-Reply-To: <MAN58R$2913F9126380EE562EC8F13DDFDBED71@libero.it>
References: <MAN58R$2913F9126380EE562EC8F13DDFDBED71@libero.it>
Message-ID: <boris.20120920131813@codesynthesis.com>

Hi Antonio,

adisabat <adisabat@libero.it> writes:

> In my system the g++ command i wrapped into perl script command to
> color the compilation output to detect wanrning,info,errors. 
> 
> This seems not good for the odb compiler.

This will work fine with the ODB compiler provided this script sends
STDERR output from GCC to STDERR, not STDOUT, which is the right thing
to do regardless of whether you use ODB or not.

If this your own script or a part of some package?

Boris

From boris at codesynthesis.com  Thu Sep 20 07:50:44 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Thu Sep 20 07:30:20 2012
Subject: [odb-users] ODB 2.1 on PostgreSQL
In-Reply-To: <CA+U5nMJZaB2295JCxgQOsO1g4Wttnd7TzHGEz+VbKSKr7HQgxw@mail.gmail.com>
References: <CA+U5nMJZaB2295JCxgQOsO1g4Wttnd7TzHGEz+VbKSKr7HQgxw@mail.gmail.com>
Message-ID: <boris.20120920133840@codesynthesis.com>

Hi Simon,

Simon Riggs <simon@2ndquadrant.com> writes:

> I wanted to pass on a few comments on the PostgreSQL limitations
> section to allow some further understanding on those areas.

Thanks, I appreciate this. Please see my commens/questions below.


> PostgreSQL 9.2 now supports streaming...
> http://www.postgresql.org/docs/9.2/static/libpq-single-row-mode.html

Interesting. Do you know if this requires both 9.2 libqp and the
server or just libpq? I.e., will this work if I connect using
9.2 libpq to 8.4 server?

Also, I am wondering if, when we do need cached result, using
PQsendQueryPrepared() and then calling PQgetResult() multiple
times will be significantly slower than a single call to 
PQexecPrepared()?


> ERROR:  duplicate key value violates unique constraint "test2_pkey"
> DETAIL:  Key (id)=(1) already exists.
> ERROR:  duplicate key value violates unique constraint "test2_id2_key"
> DETAIL:  Key (id2)=(1) already exists.
> 
> The error message itself doesn't differentiate between primary keys
> and unique constraints, but the object naming convention shows that
> PKs have the "_pkey" suffix by default. Changing that isn't
> recommended, so you could probably rely on it.

Relying on a naming convention is probably a bad idea. There are,
however, column names in the error message so we can use that (we
know the name of the column which is the primary key). Things get
complicated though if we have a composite primary key...

It would have been much easier if Postgres just had a separate
error code for a primary key violation, like MySQL ;-).

Boris

From ststrou at sandia.gov  Thu Sep 20 14:43:50 2012
From: ststrou at sandia.gov (Stroud, Sean T)
Date: Thu Sep 20 15:37:11 2012
Subject: [odb-users] Problem installing ODB 2.1.0 somewhere other than
	/usr/local/
Message-ID: <8EDFC8F388A1E44BB67C9BAA57091F9E285BECB8@EXMB03.srn.sandia.gov>

Hi,

I am trying to install ODB 2.1.0 into my home directory instead of /usr/local/ (since I don't have root privilege), but am getting an error.  Here are the commands I am using

% pwd
/home/ststrou/ODB2.1
% ls
libodb-2.1.0
libodb-mysql-2.1.0
libodb-pgsql-2.1.0
odb-2.1.0-x86_64-linux-gnu
odb-examples-2.1.0
% mkdir local
% cd libodb-2.1.0
% ./configure --prefix=/home/ststrou/ODB2.1/local
% make
% make install
% cd ../libodb-mysql-2.1.0
% ./configure --prefix=/home/ststrou/ODB2.1/local --with-libodb=/home/ststrou/ODB2.1/libodb-2.1.0
% make
% make install

The problem occurs with the very last command ("make install").  The error is:

/usr/bin/ld: cannot find -lodb
collect2: ld returned 1 exit status
libtool: install: error: relink `libodb-mysql.la' with the above command before installing it
make[2]: *** [install-libLTLIBRARIES] Error 1
make[2]: Leaving directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
make[1]: *** [install-am] Error 2
make[1]: Leaving directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
make: *** [install-recursive] Error 1

Am I doing something wrong?

Thanks.

Sean

PS:  Here is the full output of all the commands above:

% pwd
/home/ststrou/ODB2.1

% ls
libodb-2.1.0
libodb-mysql-2.1.0
libodb-pgsql-2.1.0
odb-2.1.0-x86_64-linux-gnu
odb-examples-2.1.0

% mkdir local

% cd libodb-2.1.0

% ./configure --prefix=/home/ststrou/ODB2.1/local
checking for a BSD-compatible install... /usr/bin/install -c
checking whether build environment is sane... yes
checking for a thread-safe mkdir -p... /bin/mkdir -p
checking for gawk... gawk
checking whether make sets $(MAKE)... yes
checking how to create a ustar tar archive... gnutar
checking build system type... x86_64-unknown-linux-gnu
checking host system type... x86_64-unknown-linux-gnu
checking for style of include used by make... GNU
checking for gcc... gcc
checking whether the C compiler works... yes
checking for C compiler default output file name... a.out
checking for suffix of executables...
checking whether we are cross compiling... no
checking for suffix of object files... o
checking whether we are using the GNU C compiler... yes
checking whether gcc accepts -g... yes
checking for gcc option to accept ISO C89... none needed
checking dependency style of gcc... gcc3
checking for a sed that does not truncate output... /bin/sed
checking for grep that handles long lines and -e... /bin/grep
checking for egrep... /bin/grep -E
checking for fgrep... /bin/grep -F
checking for ld used by gcc... /usr/bin/ld
checking if the linker (/usr/bin/ld) is GNU ld... yes
checking for BSD- or MS-compatible name lister (nm)... /usr/bin/nm -B
checking the name lister (/usr/bin/nm -B) interface... BSD nm
checking whether ln -s works... yes
checking the maximum length of command line arguments... 98304
checking whether the shell understands some XSI constructs... yes
checking whether the shell understands "+="... yes
checking for /usr/bin/ld option to reload object files... -r
checking for objdump... objdump
checking how to recognize dependent libraries... pass_all
checking for ar... ar
checking for strip... strip
checking for ranlib... ranlib
checking command to parse /usr/bin/nm -B output from gcc object... ok
checking how to run the C preprocessor... gcc -E
checking for ANSI C header files... yes
checking for sys/types.h... yes
checking for sys/stat.h... yes
checking for stdlib.h... yes
checking for string.h... yes
checking for memory.h... yes
checking for strings.h... yes
checking for inttypes.h... yes
checking for stdint.h... yes
checking for unistd.h... yes
checking for dlfcn.h... yes
checking for objdir... .libs
checking if gcc supports -fno-rtti -fno-exceptions... no
checking for gcc option to produce PIC... -fPIC -DPIC
checking if gcc PIC flag -fPIC -DPIC works... yes
checking if gcc static flag -static works... yes
checking if gcc supports -c -o file.o... yes
checking if gcc supports -c -o file.o... (cached) yes
checking whether the gcc linker (/usr/bin/ld -m elf_x86_64) supports shared libraries... yes
checking whether -lc should be explicitly linked in... no
checking dynamic linker characteristics... GNU/Linux ld.so
checking how to hardcode library paths into programs... immediate
checking whether stripping libraries is possible... yes
checking if libtool supports shared libraries... yes
checking whether to build shared libraries... yes
checking whether to build static libraries... yes
checking for g++... g++
checking whether we are using the GNU C++ compiler... yes
checking whether g++ accepts -g... yes
checking dependency style of g++... gcc3
checking whether we are using the GNU C++ compiler... (cached) yes
checking whether g++ accepts -g... (cached) yes
checking dependency style of g++... (cached) gcc3
checking how to run the C++ preprocessor... g++ -E
checking for ld used by g++... /usr/bin/ld -m elf_x86_64
checking if the linker (/usr/bin/ld -m elf_x86_64) is GNU ld... yes
checking whether the g++ linker (/usr/bin/ld -m elf_x86_64) supports shared libraries... yes
checking for g++ option to produce PIC... -fPIC -DPIC
checking if g++ PIC flag -fPIC -DPIC works... yes
checking if g++ static flag -static works... yes
checking if g++ supports -c -o file.o... yes
checking if g++ supports -c -o file.o... (cached) yes
checking whether the g++ linker (/usr/bin/ld -m elf_x86_64) supports shared libraries... yes
checking dynamic linker characteristics... GNU/Linux ld.so
checking how to hardcode library paths into programs... immediate
configure: creating ./config.lt
config.lt: creating libtool
checking for the pthreads library -lpthreads... no
checking for the pthreads library -lpthread... yes
checking if more special flags are required for pthreads... -D_REENTRANT
checking for __thread keyword... yes
checking whether to use rpath... yes
configure: creating ./config.status
config.status: creating libodb.pc
config.status: creating Makefile
config.status: creating odb/Makefile
config.status: creating odb/config.h
config.status: odb/config.h is unchanged
config.status: creating odb/details/config.h
config.status: odb/details/config.h is unchanged
config.status: executing depfiles commands
config.status: executing libtool commands
config.status: executing libtool-rpath-patch commands

% make
Making all in odb
make[1]: Entering directory `/home/ststrou/ODB2.1/libodb-2.1.0/odb'
make  all-am
make[2]: Entering directory `/home/ststrou/ODB2.1/libodb-2.1.0/odb'
make[2]: Leaving directory `/home/ststrou/ODB2.1/libodb-2.1.0/odb'
make[1]: Leaving directory `/home/ststrou/ODB2.1/libodb-2.1.0/odb'
make[1]: Entering directory `/home/ststrou/ODB2.1/libodb-2.1.0'
make[1]: Nothing to be done for `all-am'.
make[1]: Leaving directory `/home/ststrou/ODB2.1/libodb-2.1.0'

% make install
Making install in odb
make[1]: Entering directory `/home/ststrou/ODB2.1/libodb-2.1.0/odb'
make[2]: Entering directory `/home/ststrou/ODB2.1/libodb-2.1.0/odb'
test -z "/home/ststrou/ODB2.1/local/lib" || /bin/mkdir -p "/home/ststrou/ODB2.1/local/lib"
 /bin/sh ../libtool   --mode=install /usr/bin/install -c   libodb.la '/home/ststrou/ODB2.1/local/lib'
libtool: install: /usr/bin/install -c .libs/libodb-2.1.so /home/ststrou/ODB2.1/local/lib/libodb-2.1.so
libtool: install: (cd /home/ststrou/ODB2.1/local/lib && { ln -s -f libodb-2.1.so libodb.so || { rm -f libodb.so && ln -s libodb-2.1.so libodb.so; }; })
libtool: install: /usr/bin/install -c .libs/libodb.lai /home/ststrou/ODB2.1/local/lib/libodb.la
libtool: install: /usr/bin/install -c .libs/libodb.a /home/ststrou/ODB2.1/local/lib/libodb.a
libtool: install: chmod 644 /home/ststrou/ODB2.1/local/lib/libodb.a
libtool: install: ranlib /home/ststrou/ODB2.1/local/lib/libodb.a
libtool: install: warning: remember to run `libtool --finish /usr/local/lib'
test -z "/home/ststrou/ODB2.1/local/include/odb" || /bin/mkdir -p "/home/ststrou/ODB2.1/local/include/odb"
/bin/mkdir -p '/home/ststrou/ODB2.1/local/include/odb/details'
 /usr/bin/install -c -m 644  details/config.h '/home/ststrou/ODB2.1/local/include/odb/details'
test -z "/home/ststrou/ODB2.1/local/include/odb" || /bin/mkdir -p "/home/ststrou/ODB2.1/local/include/odb"
 /usr/bin/install -c -m 644  std-map-traits.hxx std-set-traits.hxx no-id-object-result.txx lazy-ptr-impl.txx post.hxx session.txx no-op-cache-traits.hxx transaction.ixx view-result.hxx cache-traits.hxx std-array-traits.hxx std-unordered-set-traits.hxx lazy-ptr.txx database.txx traits.hxx tracer.hxx result.hxx session.hxx object-result.hxx lazy-ptr.hxx exception.hxx transaction.hxx connection.hxx polymorphic-map.hxx connection.ixx session.ixx core.hxx simple-object-result.hxx pointer-traits.hxx pre.hxx statement.hxx database.hxx database.ixx simple-object-result.txx view-result.txx lazy-ptr-impl.ixx lazy-ptr.ixx forward.hxx schema-catalog.hxx std-vector-traits.hxx '/home/ststrou/ODB2.1/local/include/odb/.'
/bin/mkdir -p '/home/ststrou/ODB2.1/local/include/odb/details/meta'
 /usr/bin/install -c -m 644  details/meta/remove-const.hxx details/meta/polymorphic-p.hxx details/meta/remove-volatile.hxx details/meta/answer.hxx details/meta/remove-pointer.hxx details/meta/remove-const-volatile.hxx details/meta/class-p.hxx '/home/ststrou/ODB2.1/local/include/odb/details/meta'
/bin/mkdir -p '/home/ststrou/ODB2.1/local/include/odb/details'
 /usr/bin/install -c -m 644  details/type-info.hxx details/shared-ptr.hxx details/exception.hxx details/mutex.hxx details/shared-ptr-fwd.hxx details/condition.hxx details/thread.hxx details/tls.hxx details/wrapper-p.hxx details/unused.hxx details/export.hxx details/lock.hxx details/transfer-ptr.hxx details/buffer.hxx details/unique-ptr.hxx details/config.hxx '/home/ststrou/ODB2.1/local/include/odb/details'
/bin/mkdir -p '/home/ststrou/ODB2.1/local/include/odb/tr1'
 /usr/bin/install -c -m 644  tr1/lazy-ptr.txx tr1/memory.hxx tr1/lazy-ptr.hxx tr1/pointer-traits.hxx tr1/lazy-ptr.ixx tr1/wrapper-traits.hxx tr1/lazy-pointer-traits.hxx '/home/ststrou/ODB2.1/local/include/odb/tr1'
/bin/mkdir -p '/home/ststrou/ODB2.1/local/include/odb/compilers/vc'
 /usr/bin/install -c -m 644  compilers/vc/post.hxx compilers/vc/pre.hxx '/home/ststrou/ODB2.1/local/include/odb/compilers/vc'
 /usr/bin/install -c -m 644  lazy-ptr-impl.hxx polymorphic-object-result.hxx exceptions.hxx wrapper-traits.hxx polymorphic-info.hxx polymorphic-map.ixx container-traits.hxx query.hxx no-id-object-result.hxx polymorphic-object-result.txx std-unordered-map-traits.hxx lazy-pointer-traits.hxx schema-catalog-impl.hxx callback.hxx nullable.hxx std-forward-list-traits.hxx version.hxx std-list-traits.hxx polymorphic-map.txx '/home/ststrou/ODB2.1/local/include/odb/.'
/bin/mkdir -p '/home/ststrou/ODB2.1/local/include/odb/details/posix'
 /usr/bin/install -c -m 644  details/posix/mutex.hxx details/posix/thread.ixx details/posix/condition.hxx details/posix/thread.hxx details/posix/tls.ixx details/posix/mutex.ixx details/posix/tls.txx details/posix/tls.hxx details/posix/exceptions.hxx details/posix/condition.ixx '/home/ststrou/ODB2.1/local/include/odb/details/posix'
/bin/mkdir -p '/home/ststrou/ODB2.1/local/include/odb/details/shared-ptr'
 /usr/bin/install -c -m 644  details/shared-ptr/base.txx details/shared-ptr/counter-type.hxx details/shared-ptr/base.ixx details/shared-ptr/base.hxx '/home/ststrou/ODB2.1/local/include/odb/details/shared-ptr'
make[2]: Leaving directory `/home/ststrou/ODB2.1/libodb-2.1.0/odb'
make[1]: Leaving directory `/home/ststrou/ODB2.1/libodb-2.1.0/odb'
make[1]: Entering directory `/home/ststrou/ODB2.1/libodb-2.1.0'
make[2]: Entering directory `/home/ststrou/ODB2.1/libodb-2.1.0'
make[2]: Nothing to be done for `install-exec-am'.
test -z "/home/ststrou/ODB2.1/local/share/doc/libodb" || /bin/mkdir -p "/home/ststrou/ODB2.1/local/share/doc/libodb"
 /usr/bin/install -c -m 644 GPLv2 LICENSE README NEWS version '/home/ststrou/ODB2.1/local/share/doc/libodb'
test -z "/home/ststrou/ODB2.1/local/lib/pkgconfig" || /bin/mkdir -p "/home/ststrou/ODB2.1/local/lib/pkgconfig"
 /usr/bin/install -c -m 644 libodb.pc '/home/ststrou/ODB2.1/local/lib/pkgconfig'
make[2]: Leaving directory `/home/ststrou/ODB2.1/libodb-2.1.0'
make[1]: Leaving directory `/home/ststrou/ODB2.1/libodb-2.1.0'

% cd ../libodb-mysql-2.1.0

% ./configure --prefix=/home/ststrou/ODB2.1/local --with-libodb=/home/ststrou/ODB2.1/libodb-2.1.0
checking for a BSD-compatible install... /usr/bin/install -c
checking whether build environment is sane... yes
checking for a thread-safe mkdir -p... /bin/mkdir -p
checking for gawk... gawk
checking whether make sets $(MAKE)... yes
checking how to create a ustar tar archive... gnutar
checking build system type... x86_64-unknown-linux-gnu
checking host system type... x86_64-unknown-linux-gnu
checking for style of include used by make... GNU
checking for gcc... gcc
checking whether the C compiler works... yes
checking for C compiler default output file name... a.out
checking for suffix of executables...
checking whether we are cross compiling... no
checking for suffix of object files... o
checking whether we are using the GNU C compiler... yes
checking whether gcc accepts -g... yes
checking for gcc option to accept ISO C89... none needed
checking dependency style of gcc... gcc3
checking for a sed that does not truncate output... /bin/sed
checking for grep that handles long lines and -e... /bin/grep
checking for egrep... /bin/grep -E
checking for fgrep... /bin/grep -F
checking for ld used by gcc... /usr/bin/ld
checking if the linker (/usr/bin/ld) is GNU ld... yes
checking for BSD- or MS-compatible name lister (nm)... /usr/bin/nm -B
checking the name lister (/usr/bin/nm -B) interface... BSD nm
checking whether ln -s works... yes
checking the maximum length of command line arguments... 98304
checking whether the shell understands some XSI constructs... yes
checking whether the shell understands "+="... yes
checking for /usr/bin/ld option to reload object files... -r
checking for objdump... objdump
checking how to recognize dependent libraries... pass_all
checking for ar... ar
checking for strip... strip
checking for ranlib... ranlib
checking command to parse /usr/bin/nm -B output from gcc object... ok
checking how to run the C preprocessor... gcc -E
checking for ANSI C header files... yes
checking for sys/types.h... yes
checking for sys/stat.h... yes
checking for stdlib.h... yes
checking for string.h... yes
checking for memory.h... yes
checking for strings.h... yes
checking for inttypes.h... yes
checking for stdint.h... yes
checking for unistd.h... yes
checking for dlfcn.h... yes
checking for objdir... .libs
checking if gcc supports -fno-rtti -fno-exceptions... no
checking for gcc option to produce PIC... -fPIC -DPIC
checking if gcc PIC flag -fPIC -DPIC works... yes
checking if gcc static flag -static works... yes
checking if gcc supports -c -o file.o... yes
checking if gcc supports -c -o file.o... (cached) yes
checking whether the gcc linker (/usr/bin/ld -m elf_x86_64) supports shared libraries... yes
checking whether -lc should be explicitly linked in... no
checking dynamic linker characteristics... GNU/Linux ld.so
checking how to hardcode library paths into programs... immediate
checking whether stripping libraries is possible... yes
checking if libtool supports shared libraries... yes
checking whether to build shared libraries... yes
checking whether to build static libraries... yes
checking for g++... g++
checking whether we are using the GNU C++ compiler... yes
checking whether g++ accepts -g... yes
checking dependency style of g++... gcc3
checking whether we are using the GNU C++ compiler... (cached) yes
checking whether g++ accepts -g... (cached) yes
checking dependency style of g++... (cached) gcc3
checking how to run the C++ preprocessor... g++ -E
checking for ld used by g++... /usr/bin/ld -m elf_x86_64
checking if the linker (/usr/bin/ld -m elf_x86_64) is GNU ld... yes
checking whether the g++ linker (/usr/bin/ld -m elf_x86_64) supports shared libraries... yes
checking for g++ option to produce PIC... -fPIC -DPIC
checking if g++ PIC flag -fPIC -DPIC works... yes
checking if g++ static flag -static works... yes
checking if g++ supports -c -o file.o... yes
checking if g++ supports -c -o file.o... (cached) yes
checking whether the g++ linker (/usr/bin/ld -m elf_x86_64) supports shared libraries... yes
checking dynamic linker characteristics... GNU/Linux ld.so
checking how to hardcode library paths into programs... immediate
configure: creating ./config.lt
config.lt: creating libtool
checking for the pthreads library -lpthreads... no
checking for the pthreads library -lpthread... yes
checking if more special flags are required for pthreads... -D_REENTRANT
checking for __thread keyword... yes
checking for libmysqlclient_r... yes
checking for libodb... yes
checking whether to use rpath... yes
configure: creating ./config.status
config.status: creating libodb-mysql.pc
config.status: creating Makefile
config.status: creating odb/mysql/Makefile
config.status: creating odb/mysql/config.h
config.status: odb/mysql/config.h is unchanged
config.status: creating odb/mysql/details/config.h
config.status: odb/mysql/details/config.h is unchanged
config.status: executing depfiles commands
config.status: executing libtool commands
config.status: executing libtool-rpath-patch commands

% make
Making all in odb/mysql
make[1]: Entering directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
make  all-am
make[2]: Entering directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
make[2]: Leaving directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
make[1]: Leaving directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
make[1]: Entering directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0'
make[1]: Nothing to be done for `all-am'.
make[1]: Leaving directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0'

% make install
Making install in odb/mysql
make[1]: Entering directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
make[2]: Entering directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
test -z "/home/ststrou/ODB2.1/local/lib" || /bin/mkdir -p "/home/ststrou/ODB2.1/local/lib"
 /bin/sh ../../libtool   --mode=install /usr/bin/install -c   libodb-mysql.la '/home/ststrou/ODB2.1/local/lib'
libtool: install: warning: relinking `libodb-mysql.la'
libtool: install: (cd /home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql; /bin/sh /home/ststrou/ODB2.1/libodb-mysql-2.1.0/libtool  --tag CXX --mode=relink g++ -g -O2 -D_REENTRANT -release 2.1 -no-undefined -L/usr/lib/mysql -L/home/ststrou/ODB2.1/libodb-2.1.0//odb -o libodb-mysql.la -rpath /home/ststrou/ODB2.1/local/lib connection.lo connection-factory.lo database.lo enum.lo error.lo exceptions.lo query.lo query-const-expr.lo simple-object-statements.lo statement.lo statements-base.lo tracer.lo traits.lo transaction.lo transaction-impl.lo details/options.lo -lodb -lmysqlclient_r -lpthread )
libtool: relink: g++ -shared -nostdlib /usr/lib/gcc/x86_64-redhat-linux/4.1.2/../../../../lib64/crti.o /usr/lib/gcc/x86_64-redhat-linux/4.1.2/crtbeginS.o  .libs/connection.o .libs/connection-factory.o .libs/database.o .libs/enum.o .libs/error.o .libs/exceptions.o .libs/query.o .libs/query-const-expr.o .libs/simple-object-statements.o .libs/statement.o .libs/statements-base.o .libs/tracer.o .libs/traits.o .libs/transaction.o .libs/transaction-impl.o details/.libs/options.o   -Wl,-rpath -Wl,/usr/local/lib -L/usr/lib/mysql -L/home/ststrou/ODB2.1/libodb-2.1.0//odb -L/usr/local/lib -lodb -lmysqlclient_r -lpthread -L/usr/lib/gcc/x86_64-redhat-linux/4.1.2 -L/usr/lib/gcc/x86_64-redhat-linux/4.1.2/../../../../lib64 -L/lib/../lib64 -L/usr/lib/../lib64 -lstdc++ -lm -lc -lgcc_s /usr/lib/gcc/x86_64-redhat-linux/4.1.2/crtendS.o /usr/lib/gcc/x86_64-redhat-linux/4.1.2/../../../../lib64/crtn.o    -Wl,-soname -Wl,libodb-mysql-2.1.so -o .libs/libodb-mysql-2.1.so
/usr/bin/ld: cannot find -lodb
collect2: ld returned 1 exit status
libtool: install: error: relink `libodb-mysql.la' with the above command before installing it
make[2]: *** [install-libLTLIBRARIES] Error 1
make[2]: Leaving directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
make[1]: *** [install-am] Error 2
make[1]: Leaving directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
make: *** [install-recursive] Error 1




From boris at codesynthesis.com  Thu Sep 20 16:08:31 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Thu Sep 20 15:48:08 2012
Subject: [odb-users] Problem installing ODB 2.1.0 somewhere other than
	/usr/local/
In-Reply-To: <8EDFC8F388A1E44BB67C9BAA57091F9E285BECB8@EXMB03.srn.sandia.gov>
References: <8EDFC8F388A1E44BB67C9BAA57091F9E285BECB8@EXMB03.srn.sandia.gov>
Message-ID: <boris.20120920215921@codesynthesis.com>

Hi Sean,

Stroud, Sean T <ststrou@sandia.gov> writes:

> I am trying to install ODB 2.1.0 into my home directory instead 
> of /usr/local/ (since I don't have root privilege), but am getting
> an error.  Here are the commands I am using
> 
> % cd libodb-2.1.0
> % ./configure --prefix=/home/ststrou/ODB2.1/local
> % make
> % make install
> % cd ../libodb-mysql-2.1.0
> % ./configure --prefix=/home/ststrou/ODB2.1/local --with-libodb=/home/ststrou/ODB2.1/libodb-2.1.0
> % make
> % make install
> 
> The problem occurs with the very last command ("make install").  The
> error is:
> 
> /usr/bin/ld: cannot find -lodb
> collect2: ld returned 1 exit status
> libtool: install: error: relink `libodb-mysql.la' with the above command before installing it

I think you get this error because you install libodb in one place
(~/ODB2.1/local) but tell configure (with the --with-libodb option)
that it is in another. The --with-libodb option should really only
be used when you want to use non-installed build of this library.

In your case, instead of using --with-libodb, you should add
~/ODB2.1/local/include to CPPFLAGS and ~/ODB2.1/local/lib to LDFLAGS.
Also adding ~/ODB2.1/local/lib to LD_LIBRARY_PATH is a good idea
(you may want to add it in your .bashrc). Here is how I would build
things in your situation:

% export LD_LIBRARY_PATH=/home/ststrou/ODB2.1/local/lib:$LD_LIBRARY_PATH
% cd libodb-2.1.0
% ./configure --prefix=/home/ststrou/ODB2.1/local
% make
% make install
% cd ../libodb-mysql-2.1.0
% ./configure --prefix=/home/ststrou/ODB2.1/local \
CPPFLAGS=-I/home/ststrou/ODB2.1/libodb-2.1.0/include \
LDFLAGS=-L/home/ststrou/ODB2.1/libodb-2.1.0/lib
% make
% make install

Boris

From ststrou at sandia.gov  Thu Sep 20 16:36:10 2012
From: ststrou at sandia.gov (Stroud, Sean T)
Date: Thu Sep 20 16:37:00 2012
Subject: [EXTERNAL] Re: [odb-users] Problem installing ODB 2.1.0
	somewhere other than /usr/local/
In-Reply-To: <boris.20120920215921@codesynthesis.com>
References: <8EDFC8F388A1E44BB67C9BAA57091F9E285BECB8@EXMB03.srn.sandia.gov>
	<boris.20120920215921@codesynthesis.com>
Message-ID: <8EDFC8F388A1E44BB67C9BAA57091F9E285BED35@EXMB03.srn.sandia.gov>

Hi Boris,

Seems like the --with-libodb option ought to work, since I am pointing at the directory where I build libodb (not the directory where I installed it).  Nevertheless, I tried the ./configure command you gave below, but got the following error:

    % ./configure --prefix=/home/ststrou/ODB2.1/local CPPFLAGS=-I/home/ststrou/ODB2.1/libodb-2.1.0/include LDFLAGS=-L/home/ststrou/ODB2.1/libodb-2.1.0/lib
    ...
    configure: error: libodb is not found; consider using --with-libodb=DIR

In fact, I had tried this before and that is the reason I began trying the --with-libodb option in the first place.  
Looking more closely at the "-I" and "-L" options in the above command, I see that they refer to directories that do not exist.  So perhaps you meant this instead?

    % ./configure --prefix=/home/ststrou/ODB2.1/local CPPFLAGS=-I/home/ststrou/ODB2.1/libodb-2.1.0 LDFLAGS=-L/home/ststrou/ODB2.1/libodb-2.1.0

Unfortunately that fails with the same error " libodb not found".
Hmm. I see that those directories exist, but the libodb.la file is actually located in the "odb" subdirectory.  So next I tried this:

    % ./configure --prefix=/home/ststrou/ODB2.1/local CPPFLAGS=-I/home/ststrou/ODB2.1/libodb-2.1.0/odb LDFLAGS=-L/home/ststrou/ODB2.1/libodb-2.1.0/odb

Dang, still fails with "libodb not found", even though the library is clearly in there, as shown below:

   -rw-rw-r-- 1 ststrou ststrou 940 Sep 20 10:14 /home/ststrou/ODB2.1/libodb-2.1.0/odb/libodb.la

Perhaps I should point it to the installed libodb include/libs dirs instead, i.e. like this?

   %./configure --prefix=/home/ststrou/ODB2.1/local CPPFLAGS=-I/home/ststrou/ODB2.1/local/include LDFLAGS=/home/ststrou/ODB2.1/local/lib

Nope, that fails as well, this time with the original error ("cannot find -lodb").

Well I'm almost out of ideas, but here are two thoughts:
- I notice in your examples you have a backslash in the command before the CPPFLAGS and before the LDFLAGS.  Is that important?  The examples in the INSTALL file don't have it.  I tried with and without it (but not very rigorously) and didn't seem to make a difference.
- Also, your suggestion about setting LD_LIBRARY_FLAGS env var should not be necessary for building these libs, correct?  My understanding is that is only to help when linking a client program against the odb libraries.  I tried with and without it (but not very rigorously) and didn't seem to make a difference.

Sean


-----Original Message-----
From: Boris Kolpackov [mailto:boris@codesynthesis.com] 
Sent: Thursday, September 20, 2012 2:09 PM
To: Stroud, Sean T
Cc: odb-users@codesynthesis.com
Subject: [EXTERNAL] Re: [odb-users] Problem installing ODB 2.1.0 somewhere other than /usr/local/

Hi Sean,

Stroud, Sean T <ststrou@sandia.gov> writes:

> I am trying to install ODB 2.1.0 into my home directory instead of 
> /usr/local/ (since I don't have root privilege), but am getting an 
> error.  Here are the commands I am using
> 
> % cd libodb-2.1.0
> % ./configure --prefix=/home/ststrou/ODB2.1/local
> % make
> % make install
> % cd ../libodb-mysql-2.1.0
> % ./configure --prefix=/home/ststrou/ODB2.1/local 
> --with-libodb=/home/ststrou/ODB2.1/libodb-2.1.0
> % make
> % make install
> 
> The problem occurs with the very last command ("make install").  The 
> error is:
> 
> /usr/bin/ld: cannot find -lodb
> collect2: ld returned 1 exit status
> libtool: install: error: relink `libodb-mysql.la' with the above 
> command before installing it

I think you get this error because you install libodb in one place
(~/ODB2.1/local) but tell configure (with the --with-libodb option) that it is in another. The --with-libodb option should really only be used when you want to use non-installed build of this library.

In your case, instead of using --with-libodb, you should add ~/ODB2.1/local/include to CPPFLAGS and ~/ODB2.1/local/lib to LDFLAGS.
Also adding ~/ODB2.1/local/lib to LD_LIBRARY_PATH is a good idea (you may want to add it in your .bashrc). Here is how I would build things in your situation:

% export LD_LIBRARY_PATH=/home/ststrou/ODB2.1/local/lib:$LD_LIBRARY_PATH
% cd libodb-2.1.0
% ./configure --prefix=/home/ststrou/ODB2.1/local
% make
% make install
% cd ../libodb-mysql-2.1.0
% ./configure --prefix=/home/ststrou/ODB2.1/local \ CPPFLAGS=-I/home/ststrou/ODB2.1/libodb-2.1.0/include \ LDFLAGS=-L/home/ststrou/ODB2.1/libodb-2.1.0/lib
% make
% make install

Boris



From simon at 2ndQuadrant.com  Thu Sep 20 16:46:49 2012
From: simon at 2ndQuadrant.com (Simon Riggs)
Date: Thu Sep 20 16:47:01 2012
Subject: [odb-users] ODB 2.1 on PostgreSQL
In-Reply-To: <boris.20120920133840@codesynthesis.com>
References: <CA+U5nMJZaB2295JCxgQOsO1g4Wttnd7TzHGEz+VbKSKr7HQgxw@mail.gmail.com>
	<boris.20120920133840@codesynthesis.com>
Message-ID: <CA+U5nMLr6Hx96EMnvCTzmEaMquX47Xz29JMCYL9HeVhKOk7PPQ@mail.gmail.com>

On 20 September 2012 12:50, Boris Kolpackov <boris@codesynthesis.com> wrote:

>> PostgreSQL 9.2 now supports streaming...
>> http://www.postgresql.org/docs/9.2/static/libpq-single-row-mode.html
>
> Interesting. Do you know if this requires both 9.2 libqp and the
> server or just libpq?

Yes, just libpq

> I.e., will this work if I connect using
> 9.2 libpq to 8.4 server?

Yes

> Also, I am wondering if, when we do need cached result, using
> PQsendQueryPrepared() and then calling PQgetResult() multiple
> times will be significantly slower than a single call to
> PQexecPrepared()?

Feedback on it is interesting. View is there is some overhead, but its
OK for general use.

>> ERROR:  duplicate key value violates unique constraint "test2_pkey"
>> DETAIL:  Key (id)=(1) already exists.
>> ERROR:  duplicate key value violates unique constraint "test2_id2_key"
>> DETAIL:  Key (id2)=(1) already exists.
>>
>> The error message itself doesn't differentiate between primary keys
>> and unique constraints, but the object naming convention shows that
>> PKs have the "_pkey" suffix by default. Changing that isn't
>> recommended, so you could probably rely on it.
>
> Relying on a naming convention is probably a bad idea. There are,
> however, column names in the error message so we can use that (we
> know the name of the column which is the primary key). Things get
> complicated though if we have a composite primary key...

OK

> It would have been much easier if Postgres just had a separate
> error code for a primary key violation, like MySQL ;-).

How important is that? The patch for that is simple, but the
management of the change would require us to have a config parameter
to provide backwards compatibility. So it wouldn't be an easy sell....

-- 
 Simon Riggs                   http://www.2ndQuadrant.com/
 PostgreSQL Development, 24x7 Support, Training & Services

From boris at codesynthesis.com  Fri Sep 21 03:47:45 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Fri Sep 21 03:27:19 2012
Subject: [odb-users] ODB 2.1 on PostgreSQL
In-Reply-To: <CA+U5nMLr6Hx96EMnvCTzmEaMquX47Xz29JMCYL9HeVhKOk7PPQ@mail.gmail.com>
References: <CA+U5nMJZaB2295JCxgQOsO1g4Wttnd7TzHGEz+VbKSKr7HQgxw@mail.gmail.com>
	<boris.20120920133840@codesynthesis.com>
	<CA+U5nMLr6Hx96EMnvCTzmEaMquX47Xz29JMCYL9HeVhKOk7PPQ@mail.gmail.com>
Message-ID: <boris.20120921094244@codesynthesis.com>

Hi Simon,

Simon Riggs <simon@2ndQuadrant.com> writes:

> > It would have been much easier if Postgres just had a separate
> > error code for a primary key violation, like MySQL ;-).
> 
> How important is that? The patch for that is simple, but the
> management of the change would require us to have a config parameter
> to provide backwards compatibility. So it wouldn't be an easy sell....

SQLite uses an interesting mechanism called extended error codes.
Basically, for some "primary" error codes one can also ask for an
extended code which further qualifies the error (e.g., primary key
violation vs foreign key vs unique, etc).

But, I agree, it would probably be easier to hack something up in
ODB than to make this change in PostgreSQL.

From boris at codesynthesis.com  Fri Sep 21 04:00:22 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Fri Sep 21 03:40:07 2012
Subject: [EXTERNAL] Re: [odb-users] Problem installing ODB 2.1.0
	somewhere other than /usr/local/
In-Reply-To: <8EDFC8F388A1E44BB67C9BAA57091F9E285BED35@EXMB03.srn.sandia.gov>
References: <8EDFC8F388A1E44BB67C9BAA57091F9E285BECB8@EXMB03.srn.sandia.gov>
	<boris.20120920215921@codesynthesis.com>
	<8EDFC8F388A1E44BB67C9BAA57091F9E285BED35@EXMB03.srn.sandia.gov>
Message-ID: <boris.20120921094753@codesynthesis.com>

Hi Sean,

Stroud, Sean T <ststrou@sandia.gov> writes:

> Seems like the --with-libodb option ought to work, since I am pointing
> at the directory where I build libodb (not the directory where I installed
> it).

It works for *building* libodb-mysql but not for *installing* it since
an installed library should not reference a non-installed one.


> Nevertheless, I tried the ./configure command you gave below, but got
> the following error:

Sorry, I made a mistake in the command. Instead of ~/ODB2.1/libodb-2.1.0
in CPPFLAGS and LDFLAGS it should have been ~/ODB2.1/local:

./configure --prefix=/home/ststrou/ODB2.1/local CPPFLAGS=-I/home/ststrou/ODB2.1/local/include LDFLAGS=-L/home/ststrou/ODB2.1/local/lib


> - I notice in your examples you have a backslash in the command
>   before the CPPFLAGS and before the LDFLAGS. Is that important?

Backslashes are used to break a long command into multiple lines,
for better readability.


> - Also, your suggestion about setting LD_LIBRARY_FLAGS env var 
> should not be necessary for building these libs, correct?  My
> understanding is that is only to help when linking a client
> program against the odb libraries.

Modern linkers are quite tricky in this regard and may require a
shared library to be loadable just to link it to another library.
See the -rpath-link ld option for details. So it is safer to just
add /home/ststrou/ODB2.1/local/lib to LD_LIBRARY_PATH before
building.

Note also that it is LD_LIBRARY_PATH, not LD_LIBRARY_FLAGS.

Boris

From ststrou at sandia.gov  Fri Sep 21 10:17:17 2012
From: ststrou at sandia.gov (Stroud, Sean T)
Date: Fri Sep 21 10:17:46 2012
Subject: [EXTERNAL] Re: [odb-users] Problem installing ODB 2.1.0
	somewhere other than /usr/local/
In-Reply-To: <boris.20120921094753@codesynthesis.com>
References: <8EDFC8F388A1E44BB67C9BAA57091F9E285BECB8@EXMB03.srn.sandia.gov>
	<boris.20120920215921@codesynthesis.com>
	<8EDFC8F388A1E44BB67C9BAA57091F9E285BED35@EXMB03.srn.sandia.gov>
	<boris.20120921094753@codesynthesis.com>
Message-ID: <8EDFC8F388A1E44BB67C9BAA57091F9E285C4174@EXMB03.srn.sandia.gov>

Hi Boris,

Thanks for all your patience with this.  Unfortunately I am still getting errors, even with the latest command you provided.  Here is the script I use to build/install all the ODB libs:

#!/bin/ksh -v
pwd
ls
rm -rf local
mkdir local
export LD_LIBRARY_PATH=/home/ststrou/ODB2.1/local/lib:$LD_LIBRARY_PATH
cd libodb-2.1.0
./configure --prefix=/home/ststrou/ODB2.1/local
make
make install
cd ../libodb-mysql-2.1.0
./configure --prefix=/home/ststrou/ODB2.1/local CPPFLAGS=-I/home/ststrou/ODB2.1/local/include LDFLAGS=-L/home/ststrou/ODB2.1/local/lib
make
make install

I invoke the above script from the /home/ststrou/ODB2.1/ directory.  The last "make install" command fails with " /usr/bin/ld: cannot find -lodb ".  Below is the full output from running the script:

#!/bin/ksh -v

pwd
/home/ststrou/ODB2.1
ls
buildit
download
libodb-2.1.0
libodb-mysql-2.1.0
libodb-pgsql-2.1.0
local
odb-2.1.0-x86_64-linux-gnu
odb-examples-2.1.0
out
rm -rf local
mkdir local
export LD_LIBRARY_PATH=/home/ststrou/ODB2.1/local/lib:$LD_LIBRARY_PATH
cd libodb-2.1.0
./configure --prefix=/home/ststrou/ODB2.1/local
checking for a BSD-compatible install... /usr/bin/install -c
checking whether build environment is sane... yes
checking for a thread-safe mkdir -p... /bin/mkdir -p
checking for gawk... gawk
checking whether make sets $(MAKE)... yes
checking how to create a ustar tar archive... gnutar
checking build system type... x86_64-unknown-linux-gnu
checking host system type... x86_64-unknown-linux-gnu
checking for style of include used by make... GNU
checking for gcc... gcc
checking whether the C compiler works... yes
checking for C compiler default output file name... a.out
checking for suffix of executables... 
checking whether we are cross compiling... no
checking for suffix of object files... o
checking whether we are using the GNU C compiler... yes
checking whether gcc accepts -g... yes
checking for gcc option to accept ISO C89... none needed
checking dependency style of gcc... gcc3
checking for a sed that does not truncate output... /bin/sed
checking for grep that handles long lines and -e... /bin/grep
checking for egrep... /bin/grep -E
checking for fgrep... /bin/grep -F
checking for ld used by gcc... /usr/bin/ld
checking if the linker (/usr/bin/ld) is GNU ld... yes
checking for BSD- or MS-compatible name lister (nm)... /usr/bin/nm -B
checking the name lister (/usr/bin/nm -B) interface... BSD nm
checking whether ln -s works... yes
checking the maximum length of command line arguments... 98304
checking whether the shell understands some XSI constructs... yes
checking whether the shell understands "+="... yes
checking for /usr/bin/ld option to reload object files... -r
checking for objdump... objdump
checking how to recognize dependent libraries... pass_all
checking for ar... ar
checking for strip... strip
checking for ranlib... ranlib
checking command to parse /usr/bin/nm -B output from gcc object... ok
checking how to run the C preprocessor... gcc -E
checking for ANSI C header files... yes
checking for sys/types.h... yes
checking for sys/stat.h... yes
checking for stdlib.h... yes
checking for string.h... yes
checking for memory.h... yes
checking for strings.h... yes
checking for inttypes.h... yes
checking for stdint.h... yes
checking for unistd.h... yes
checking for dlfcn.h... yes
checking for objdir... .libs
checking if gcc supports -fno-rtti -fno-exceptions... no
checking for gcc option to produce PIC... -fPIC -DPIC
checking if gcc PIC flag -fPIC -DPIC works... yes
checking if gcc static flag -static works... yes
checking if gcc supports -c -o file.o... yes
checking if gcc supports -c -o file.o... (cached) yes
checking whether the gcc linker (/usr/bin/ld -m elf_x86_64) supports shared libraries... yes
checking whether -lc should be explicitly linked in... no
checking dynamic linker characteristics... GNU/Linux ld.so
checking how to hardcode library paths into programs... immediate
checking whether stripping libraries is possible... yes
checking if libtool supports shared libraries... yes
checking whether to build shared libraries... yes
checking whether to build static libraries... yes
checking for g++... g++
checking whether we are using the GNU C++ compiler... yes
checking whether g++ accepts -g... yes
checking dependency style of g++... gcc3
checking whether we are using the GNU C++ compiler... (cached) yes
checking whether g++ accepts -g... (cached) yes
checking dependency style of g++... (cached) gcc3
checking how to run the C++ preprocessor... g++ -E
checking for ld used by g++... /usr/bin/ld -m elf_x86_64
checking if the linker (/usr/bin/ld -m elf_x86_64) is GNU ld... yes
checking whether the g++ linker (/usr/bin/ld -m elf_x86_64) supports shared libraries... yes
checking for g++ option to produce PIC... -fPIC -DPIC
checking if g++ PIC flag -fPIC -DPIC works... yes
checking if g++ static flag -static works... yes
checking if g++ supports -c -o file.o... yes
checking if g++ supports -c -o file.o... (cached) yes
checking whether the g++ linker (/usr/bin/ld -m elf_x86_64) supports shared libraries... yes
checking dynamic linker characteristics... GNU/Linux ld.so
checking how to hardcode library paths into programs... immediate
configure: creating ./config.lt
config.lt: creating libtool
checking for the pthreads library -lpthreads... no
checking for the pthreads library -lpthread... yes
checking if more special flags are required for pthreads... -D_REENTRANT
checking for __thread keyword... yes
checking whether to use rpath... yes
configure: creating ./config.status
config.status: creating libodb.pc
config.status: creating Makefile
config.status: creating odb/Makefile
config.status: creating odb/config.h
config.status: odb/config.h is unchanged
config.status: creating odb/details/config.h
config.status: odb/details/config.h is unchanged
config.status: executing depfiles commands
config.status: executing libtool commands
config.status: executing libtool-rpath-patch commands
make
Making all in odb
make[1]: Entering directory `/home/ststrou/ODB2.1/libodb-2.1.0/odb'
make  all-am
make[2]: Entering directory `/home/ststrou/ODB2.1/libodb-2.1.0/odb'
make[2]: Leaving directory `/home/ststrou/ODB2.1/libodb-2.1.0/odb'
make[1]: Leaving directory `/home/ststrou/ODB2.1/libodb-2.1.0/odb'
make[1]: Entering directory `/home/ststrou/ODB2.1/libodb-2.1.0'
make[1]: Nothing to be done for `all-am'.
make[1]: Leaving directory `/home/ststrou/ODB2.1/libodb-2.1.0'
make install
Making install in odb
make[1]: Entering directory `/home/ststrou/ODB2.1/libodb-2.1.0/odb'
make[2]: Entering directory `/home/ststrou/ODB2.1/libodb-2.1.0/odb'
test -z "/home/ststrou/ODB2.1/local/lib" || /bin/mkdir -p "/home/ststrou/ODB2.1/local/lib"
 /bin/sh ../libtool   --mode=install /usr/bin/install -c   libodb.la '/home/ststrou/ODB2.1/local/lib'
libtool: install: /usr/bin/install -c .libs/libodb-2.1.so /home/ststrou/ODB2.1/local/lib/libodb-2.1.so
libtool: install: (cd /home/ststrou/ODB2.1/local/lib && { ln -s -f libodb-2.1.so libodb.so || { rm -f libodb.so && ln -s libodb-2.1.so libodb.so; }; })
libtool: install: /usr/bin/install -c .libs/libodb.lai /home/ststrou/ODB2.1/local/lib/libodb.la
libtool: install: /usr/bin/install -c .libs/libodb.a /home/ststrou/ODB2.1/local/lib/libodb.a
libtool: install: chmod 644 /home/ststrou/ODB2.1/local/lib/libodb.a
libtool: install: ranlib /home/ststrou/ODB2.1/local/lib/libodb.a
libtool: install: warning: remember to run `libtool --finish /usr/local/lib'
test -z "/home/ststrou/ODB2.1/local/include/odb" || /bin/mkdir -p "/home/ststrou/ODB2.1/local/include/odb"
/bin/mkdir -p '/home/ststrou/ODB2.1/local/include/odb/details'
 /usr/bin/install -c -m 644  details/config.h '/home/ststrou/ODB2.1/local/include/odb/details'
test -z "/home/ststrou/ODB2.1/local/include/odb" || /bin/mkdir -p "/home/ststrou/ODB2.1/local/include/odb"
 /usr/bin/install -c -m 644  std-map-traits.hxx std-set-traits.hxx no-id-object-result.txx lazy-ptr-impl.txx post.hxx session.txx no-op-cache-traits.hxx transaction.ixx view-result.hxx cache-traits.hxx std-array-traits.hxx std-unordered-set-traits.hxx lazy-ptr.txx database.txx traits.hxx tracer.hxx result.hxx session.hxx object-result.hxx lazy-ptr.hxx exception.hxx transaction.hxx connection.hxx polymorphic-map.hxx connection.ixx session.ixx core.hxx simple-object-result.hxx pointer-traits.hxx pre.hxx statement.hxx database.hxx database.ixx simple-object-result.txx view-result.txx lazy-ptr-impl.ixx lazy-ptr.ixx forward.hxx schema-catalog.hxx std-vector-traits.hxx '/home/ststrou/ODB2.1/local/include/odb/.'
/bin/mkdir -p '/home/ststrou/ODB2.1/local/include/odb/details/meta'
 /usr/bin/install -c -m 644  details/meta/remove-const.hxx details/meta/polymorphic-p.hxx details/meta/remove-volatile.hxx details/meta/answer.hxx details/meta/remove-pointer.hxx details/meta/remove-const-volatile.hxx details/meta/class-p.hxx '/home/ststrou/ODB2.1/local/include/odb/details/meta'
/bin/mkdir -p '/home/ststrou/ODB2.1/local/include/odb/details'
 /usr/bin/install -c -m 644  details/type-info.hxx details/shared-ptr.hxx details/exception.hxx details/mutex.hxx details/shared-ptr-fwd.hxx details/condition.hxx details/thread.hxx details/tls.hxx details/wrapper-p.hxx details/unused.hxx details/export.hxx details/lock.hxx details/transfer-ptr.hxx details/buffer.hxx details/unique-ptr.hxx details/config.hxx '/home/ststrou/ODB2.1/local/include/odb/details'
/bin/mkdir -p '/home/ststrou/ODB2.1/local/include/odb/tr1'
 /usr/bin/install -c -m 644  tr1/lazy-ptr.txx tr1/memory.hxx tr1/lazy-ptr.hxx tr1/pointer-traits.hxx tr1/lazy-ptr.ixx tr1/wrapper-traits.hxx tr1/lazy-pointer-traits.hxx '/home/ststrou/ODB2.1/local/include/odb/tr1'
/bin/mkdir -p '/home/ststrou/ODB2.1/local/include/odb/compilers/vc'
 /usr/bin/install -c -m 644  compilers/vc/post.hxx compilers/vc/pre.hxx '/home/ststrou/ODB2.1/local/include/odb/compilers/vc'
 /usr/bin/install -c -m 644  lazy-ptr-impl.hxx polymorphic-object-result.hxx exceptions.hxx wrapper-traits.hxx polymorphic-info.hxx polymorphic-map.ixx container-traits.hxx query.hxx no-id-object-result.hxx polymorphic-object-result.txx std-unordered-map-traits.hxx lazy-pointer-traits.hxx schema-catalog-impl.hxx callback.hxx nullable.hxx std-forward-list-traits.hxx version.hxx std-list-traits.hxx polymorphic-map.txx '/home/ststrou/ODB2.1/local/include/odb/.'
/bin/mkdir -p '/home/ststrou/ODB2.1/local/include/odb/details/posix'
 /usr/bin/install -c -m 644  details/posix/mutex.hxx details/posix/thread.ixx details/posix/condition.hxx details/posix/thread.hxx details/posix/tls.ixx details/posix/mutex.ixx details/posix/tls.txx details/posix/tls.hxx details/posix/exceptions.hxx details/posix/condition.ixx '/home/ststrou/ODB2.1/local/include/odb/details/posix'
/bin/mkdir -p '/home/ststrou/ODB2.1/local/include/odb/details/shared-ptr'
 /usr/bin/install -c -m 644  details/shared-ptr/base.txx details/shared-ptr/counter-type.hxx details/shared-ptr/base.ixx details/shared-ptr/base.hxx '/home/ststrou/ODB2.1/local/include/odb/details/shared-ptr'
make[2]: Leaving directory `/home/ststrou/ODB2.1/libodb-2.1.0/odb'
make[1]: Leaving directory `/home/ststrou/ODB2.1/libodb-2.1.0/odb'
make[1]: Entering directory `/home/ststrou/ODB2.1/libodb-2.1.0'
make[2]: Entering directory `/home/ststrou/ODB2.1/libodb-2.1.0'
make[2]: Nothing to be done for `install-exec-am'.
test -z "/home/ststrou/ODB2.1/local/share/doc/libodb" || /bin/mkdir -p "/home/ststrou/ODB2.1/local/share/doc/libodb"
 /usr/bin/install -c -m 644 GPLv2 LICENSE README NEWS version '/home/ststrou/ODB2.1/local/share/doc/libodb'
test -z "/home/ststrou/ODB2.1/local/lib/pkgconfig" || /bin/mkdir -p "/home/ststrou/ODB2.1/local/lib/pkgconfig"
 /usr/bin/install -c -m 644 libodb.pc '/home/ststrou/ODB2.1/local/lib/pkgconfig'
make[2]: Leaving directory `/home/ststrou/ODB2.1/libodb-2.1.0'
make[1]: Leaving directory `/home/ststrou/ODB2.1/libodb-2.1.0'
cd ../libodb-mysql-2.1.0
./configure --prefix=/home/ststrou/ODB2.1/local CPPFLAGS=-I/home/ststrou/ODB2.1/local/include LDFLAGS=-L/home/ststrou/ODB2.1/local/lib
checking for a BSD-compatible install... /usr/bin/install -c
checking whether build environment is sane... yes
checking for a thread-safe mkdir -p... /bin/mkdir -p
checking for gawk... gawk
checking whether make sets $(MAKE)... yes
checking how to create a ustar tar archive... gnutar
checking build system type... x86_64-unknown-linux-gnu
checking host system type... x86_64-unknown-linux-gnu
checking for style of include used by make... GNU
checking for gcc... gcc
checking whether the C compiler works... yes
checking for C compiler default output file name... a.out
checking for suffix of executables... 
checking whether we are cross compiling... no
checking for suffix of object files... o
checking whether we are using the GNU C compiler... yes
checking whether gcc accepts -g... yes
checking for gcc option to accept ISO C89... none needed
checking dependency style of gcc... gcc3
checking for a sed that does not truncate output... /bin/sed
checking for grep that handles long lines and -e... /bin/grep
checking for egrep... /bin/grep -E
checking for fgrep... /bin/grep -F
checking for ld used by gcc... /usr/bin/ld
checking if the linker (/usr/bin/ld) is GNU ld... yes
checking for BSD- or MS-compatible name lister (nm)... /usr/bin/nm -B
checking the name lister (/usr/bin/nm -B) interface... BSD nm
checking whether ln -s works... yes
checking the maximum length of command line arguments... 98304
checking whether the shell understands some XSI constructs... yes
checking whether the shell understands "+="... yes
checking for /usr/bin/ld option to reload object files... -r
checking for objdump... objdump
checking how to recognize dependent libraries... pass_all
checking for ar... ar
checking for strip... strip
checking for ranlib... ranlib
checking command to parse /usr/bin/nm -B output from gcc object... ok
checking how to run the C preprocessor... gcc -E
checking for ANSI C header files... yes
checking for sys/types.h... yes
checking for sys/stat.h... yes
checking for stdlib.h... yes
checking for string.h... yes
checking for memory.h... yes
checking for strings.h... yes
checking for inttypes.h... yes
checking for stdint.h... yes
checking for unistd.h... yes
checking for dlfcn.h... yes
checking for objdir... .libs
checking if gcc supports -fno-rtti -fno-exceptions... no
checking for gcc option to produce PIC... -fPIC -DPIC
checking if gcc PIC flag -fPIC -DPIC works... yes
checking if gcc static flag -static works... yes
checking if gcc supports -c -o file.o... yes
checking if gcc supports -c -o file.o... (cached) yes
checking whether the gcc linker (/usr/bin/ld -m elf_x86_64) supports shared libraries... yes
checking whether -lc should be explicitly linked in... no
checking dynamic linker characteristics... GNU/Linux ld.so
checking how to hardcode library paths into programs... immediate
checking whether stripping libraries is possible... yes
checking if libtool supports shared libraries... yes
checking whether to build shared libraries... yes
checking whether to build static libraries... yes
checking for g++... g++
checking whether we are using the GNU C++ compiler... yes
checking whether g++ accepts -g... yes
checking dependency style of g++... gcc3
checking whether we are using the GNU C++ compiler... (cached) yes
checking whether g++ accepts -g... (cached) yes
checking dependency style of g++... (cached) gcc3
checking how to run the C++ preprocessor... g++ -E
checking for ld used by g++... /usr/bin/ld -m elf_x86_64
checking if the linker (/usr/bin/ld -m elf_x86_64) is GNU ld... yes
checking whether the g++ linker (/usr/bin/ld -m elf_x86_64) supports shared libraries... yes
checking for g++ option to produce PIC... -fPIC -DPIC
checking if g++ PIC flag -fPIC -DPIC works... yes
checking if g++ static flag -static works... yes
checking if g++ supports -c -o file.o... yes
checking if g++ supports -c -o file.o... (cached) yes
checking whether the g++ linker (/usr/bin/ld -m elf_x86_64) supports shared libraries... yes
checking dynamic linker characteristics... GNU/Linux ld.so
checking how to hardcode library paths into programs... immediate
configure: creating ./config.lt
config.lt: creating libtool
checking for the pthreads library -lpthreads... no
checking for the pthreads library -lpthread... yes
checking if more special flags are required for pthreads... -D_REENTRANT
checking for __thread keyword... yes
checking for libmysqlclient_r... yes
checking for libodb... yes
checking whether to use rpath... yes
configure: creating ./config.status
config.status: creating libodb-mysql.pc
config.status: creating Makefile
config.status: creating odb/mysql/Makefile
config.status: creating odb/mysql/config.h
config.status: odb/mysql/config.h is unchanged
config.status: creating odb/mysql/details/config.h
config.status: odb/mysql/details/config.h is unchanged
config.status: executing depfiles commands
config.status: executing libtool commands
config.status: executing libtool-rpath-patch commands
make
Making all in odb/mysql
make[1]: Entering directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
make  all-am
make[2]: Entering directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
make[2]: Leaving directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
make[1]: Leaving directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
make[1]: Entering directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0'
make[1]: Nothing to be done for `all-am'.
make[1]: Leaving directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0'
make install
Making install in odb/mysql
make[1]: Entering directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
make[2]: Entering directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
test -z "/home/ststrou/ODB2.1/local/lib" || /bin/mkdir -p "/home/ststrou/ODB2.1/local/lib"
 /bin/sh ../../libtool   --mode=install /usr/bin/install -c   libodb-mysql.la '/home/ststrou/ODB2.1/local/lib'
libtool: install: warning: relinking `libodb-mysql.la'
libtool: install: (cd /home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql; /bin/sh /home/ststrou/ODB2.1/libodb-mysql-2.1.0/libtool  --tag CXX --mode=relink g++ -g -O2 -D_REENTRANT -release 2.1 -no-undefined -L/usr/lib/mysql -L/home/ststrou/ODB2.1/libodb-2.1.0//odb -o libodb-mysql.la -rpath /home/ststrou/ODB2.1/local/lib connection.lo connection-factory.lo database.lo enum.lo error.lo exceptions.lo query.lo query-const-expr.lo simple-object-statements.lo statement.lo statements-base.lo tracer.lo traits.lo transaction.lo transaction-impl.lo details/options.lo -lodb -lmysqlclient_r -lpthread )
libtool: relink: g++ -shared -nostdlib /usr/lib/gcc/x86_64-redhat-linux/4.1.2/../../../../lib64/crti.o /usr/lib/gcc/x86_64-redhat-linux/4.1.2/crtbeginS.o  .libs/connection.o .libs/connection-factory.o .libs/database.o .libs/enum.o .libs/error.o .libs/exceptions.o .libs/query.o .libs/query-const-expr.o .libs/simple-object-statements.o .libs/statement.o .libs/statements-base.o .libs/tracer.o .libs/traits.o .libs/transaction.o .libs/transaction-impl.o details/.libs/options.o   -Wl,-rpath -Wl,/usr/local/lib -L/usr/lib/mysql -L/home/ststrou/ODB2.1/libodb-2.1.0//odb -L/usr/local/lib -lodb -lmysqlclient_r -lpthread -L/usr/lib/gcc/x86_64-redhat-linux/4.1.2 -L/usr/lib/gcc/x86_64-redhat-linux/4.1.2/../../../../lib64 -L/lib/../lib64 -L/usr/lib/../lib64 -lstdc++ -lm -lc -lgcc_s /usr/lib/gcc/x86_64-redhat-linux/4.1.2/crtendS.o /usr/lib/gcc/x86_64-redhat-linux/4.1.2/../../../../lib64/crtn.o    -Wl,-soname -Wl,libodb-mysql-2.1.so -o .libs/libodb-mysql-2.1.so
/usr/bin/ld: cannot find -lodb
collect2: ld returned 1 exit status
libtool: install: error: relink `libodb-mysql.la' with the above command before installing it
make[2]: *** [install-libLTLIBRARIES] Error 1
make[2]: Leaving directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
make[1]: *** [install-am] Error 2
make[1]: Leaving directory `/home/ststrou/ODB2.1/libodb-mysql-2.1.0/odb/mysql'
make: *** [install-recursive] Error 1


-----Original Message-----
From: Boris Kolpackov [mailto:boris@codesynthesis.com] 
Sent: Friday, September 21, 2012 2:00 AM
To: Stroud, Sean T
Cc: odb-users@codesynthesis.com
Subject: Re: [EXTERNAL] Re: [odb-users] Problem installing ODB 2.1.0 somewhere other than /usr/local/

Hi Sean,

Stroud, Sean T <ststrou@sandia.gov> writes:

> Seems like the --with-libodb option ought to work, since I am pointing 
> at the directory where I build libodb (not the directory where I 
> installed it).

It works for *building* libodb-mysql but not for *installing* it since an installed library should not reference a non-installed one.


> Nevertheless, I tried the ./configure command you gave below, but got 
> the following error:

Sorry, I made a mistake in the command. Instead of ~/ODB2.1/libodb-2.1.0 in CPPFLAGS and LDFLAGS it should have been ~/ODB2.1/local:

./configure --prefix=/home/ststrou/ODB2.1/local CPPFLAGS=-I/home/ststrou/ODB2.1/local/include LDFLAGS=-L/home/ststrou/ODB2.1/local/lib


> - I notice in your examples you have a backslash in the command
>   before the CPPFLAGS and before the LDFLAGS. Is that important?

Backslashes are used to break a long command into multiple lines, for better readability.


> - Also, your suggestion about setting LD_LIBRARY_FLAGS env var should 
> not be necessary for building these libs, correct?  My understanding 
> is that is only to help when linking a client program against the odb 
> libraries.

Modern linkers are quite tricky in this regard and may require a shared library to be loadable just to link it to another library.
See the -rpath-link ld option for details. So it is safer to just add /home/ststrou/ODB2.1/local/lib to LD_LIBRARY_PATH before building.

Note also that it is LD_LIBRARY_PATH, not LD_LIBRARY_FLAGS.

Boris



From boris at codesynthesis.com  Fri Sep 21 13:43:02 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Fri Sep 21 13:22:31 2012
Subject: [EXTERNAL] Re: [odb-users] Problem installing ODB 2.1.0
	somewhere other than /usr/local/
In-Reply-To: <8EDFC8F388A1E44BB67C9BAA57091F9E285C4174@EXMB03.srn.sandia.gov>
References: <8EDFC8F388A1E44BB67C9BAA57091F9E285BECB8@EXMB03.srn.sandia.gov>
	<boris.20120920215921@codesynthesis.com>
	<8EDFC8F388A1E44BB67C9BAA57091F9E285BED35@EXMB03.srn.sandia.gov>
	<boris.20120921094753@codesynthesis.com>
	<8EDFC8F388A1E44BB67C9BAA57091F9E285C4174@EXMB03.srn.sandia.gov>
Message-ID: <boris.20120921193246@codesynthesis.com>

Hi Sean,

I just tried this on my box and everything works fine. I also compared
my output to yours and I see some differences. Specifically, when you
run 'make install' in libodb-mysql-2.1.0, in your case libtool needs
to relink the library for some reason. Also the command that it runs
to accomplish this contains this option:

-L/home/ststrou/ODB2.1/libodb-2.1.0//odb

This should clearly not be there since we didn't specify it.

What I think happens is that you have some stale configuration/build
from your earlier attempts using --with-libodb. Generally, in this kind
of troubleshooting, it is always a good idea to start each attempt with
a clean source directories to exclude any possible interference from
the previous attempts.

So, to test this, can you try the following for me. Also can you use
bash instead of ksh; I don't think it matters but I just want to
eliminate all the differences between our configurations.

cd /home/ststrou/ODB2.1/
rm -r libodb-2.1.0 libodb-mysql-2.1.0 local
tar xfj libodb-2.1.0.tar.bz2
tar xfj libodb-mysql-2.1.0.tar.bz2
export LD_LIBRARY_PATH=/home/ststrou/ODB2.1/local/lib:$LD_LIBRARY_PATH
cd libodb-2.1.0
./configure --prefix=/home/ststrou/ODB2.1/local
make
make install
cd ../libodb-mysql-2.1.0
./configure --prefix=/home/ststrou/ODB2.1/local CPPFLAGS=-I/home/ststrou/ODB2.1/local/include LDFLAGS=-L/home/ststrou/ODB2.1/local/lib
make
make install

Boris

From ststrou at sandia.gov  Fri Sep 21 15:56:11 2012
From: ststrou at sandia.gov (Stroud, Sean T)
Date: Fri Sep 21 15:56:42 2012
Subject: [EXTERNAL] Re: [odb-users] Problem installing ODB 2.1.0
	somewhere other than /usr/local/
In-Reply-To: <boris.20120921193246@codesynthesis.com>
References: <8EDFC8F388A1E44BB67C9BAA57091F9E285BECB8@EXMB03.srn.sandia.gov>
	<boris.20120920215921@codesynthesis.com>
	<8EDFC8F388A1E44BB67C9BAA57091F9E285BED35@EXMB03.srn.sandia.gov>
	<boris.20120921094753@codesynthesis.com>
	<8EDFC8F388A1E44BB67C9BAA57091F9E285C4174@EXMB03.srn.sandia.gov>
	<boris.20120921193246@codesynthesis.com>
Message-ID: <8EDFC8F388A1E44BB67C9BAA57091F9E285C82B0@EXMB03.srn.sandia.gov>

Hi Boris,

Looks like your hunch was right - the install worked this time.  Thanks for all your help on this!

Sean

-----Original Message-----
From: Boris Kolpackov [mailto:boris@codesynthesis.com] 
Sent: Friday, September 21, 2012 11:43 AM
To: Stroud, Sean T
Cc: odb-users@codesynthesis.com
Subject: Re: [EXTERNAL] Re: [odb-users] Problem installing ODB 2.1.0 somewhere other than /usr/local/

Hi Sean,

I just tried this on my box and everything works fine. I also compared my output to yours and I see some differences. Specifically, when you run 'make install' in libodb-mysql-2.1.0, in your case libtool needs to relink the library for some reason. Also the command that it runs to accomplish this contains this option:

-L/home/ststrou/ODB2.1/libodb-2.1.0//odb

This should clearly not be there since we didn't specify it.

What I think happens is that you have some stale configuration/build from your earlier attempts using --with-libodb. Generally, in this kind of troubleshooting, it is always a good idea to start each attempt with a clean source directories to exclude any possible interference from the previous attempts.

So, to test this, can you try the following for me. Also can you use bash instead of ksh; I don't think it matters but I just want to eliminate all the differences between our configurations.

cd /home/ststrou/ODB2.1/
rm -r libodb-2.1.0 libodb-mysql-2.1.0 local tar xfj libodb-2.1.0.tar.bz2 tar xfj libodb-mysql-2.1.0.tar.bz2 export LD_LIBRARY_PATH=/home/ststrou/ODB2.1/local/lib:$LD_LIBRARY_PATH
cd libodb-2.1.0
./configure --prefix=/home/ststrou/ODB2.1/local
make
make install
cd ../libodb-mysql-2.1.0
./configure --prefix=/home/ststrou/ODB2.1/local CPPFLAGS=-I/home/ststrou/ODB2.1/local/include LDFLAGS=-L/home/ststrou/ODB2.1/local/lib
make
make install

Boris



From steve.xu.cd at gmail.com  Sun Sep 23 14:18:20 2012
From: steve.xu.cd at gmail.com (steve xu)
Date: Sun Sep 23 13:57:35 2012
Subject: [odb-users] memory leak detected
Message-ID: <CAMzAr7vg+qTBMnWCW54X6CNr5NrRV6ftbvQks_LXe9fHHnF=pg@mail.gmail.
	com>

Hi

I am a newbie for ODB library, I have a project use ODB with mysql support.
When I compile a simple application with MFC framework, the memory leak is
detected, but when i compile the application with win32 console, it has no
memory leak.
 follow is the logs from the Visual studio 2008 ouput window:
"mfctest.exe?: "E:\Temp\ODB\odb-study\hello\Debug\mfctest.exe??
"mfctest.exe?: "C:\Windows\System32\ntdll.dll?
"mfctest.exe?: "C:\Windows\System32\kernel32.dll?
"mfctest.exe?: "C:\Windows\System32\KernelBase.dll?
"mfctest.exe?:
"E:\Temp\ODB\odb-study\hello\Debug\odb-mysql-d-2.0-vc9.dll??
"mfctest.exe?: "E:\Temp\ODB\odb-study\hello\Debug\odb-d-2.0-vc9.dll??
"mfctest.exe?:
"C:\Windows\winsxs\x86_microsoft.vc90.debugcrt_1fc8b3b9a1e18e3b_9.0.30729.6161_none_2a4f639a55563668\msvcr90d.dll?
"mfctest.exe?:
"C:\Windows\winsxs\x86_microsoft.vc90.debugcrt_1fc8b3b9a1e18e3b_9.0.30729.6161_none_2a4f639a55563668\msvcp90d.dll??
"mfctest.exe?: "E:\Temp\ODB\odb-study\hello\Debug\libmysql.dll?
"mfctest.exe?: "C:\Windows\System32\user32.dll?
"mfctest.exe?: "C:\Windows\System32\gdi32.dll?
"mfctest.exe?: "C:\Windows\System32\lpk.dll?
"mfctest.exe?: "C:\Windows\System32\usp10.dll?
"mfctest.exe?: "C:\Windows\System32\msvcrt.dll?
"mfctest.exe?: "C:\Windows\System32\advapi32.dll?
"mfctest.exe?: "C:\Windows\System32\sechost.dll?
"mfctest.exe?: "C:\Windows\System32\rpcrt4.dll?
"mfctest.exe?: "C:\Windows\System32\ws2_32.dll?
"mfctest.exe?: "C:\Windows\System32\nsi.dll?
"mfctest.exe?: "E:\Temp\ODB\odb-study\hello\Debug\db.dll??
"mfctest.exe?:
"C:\Windows\winsxs\x86_microsoft.vc90.debugmfc_1fc8b3b9a1e18e3b_9.0.30729.6161_none_2f2f658c522a659b\mfc90ud.dll??
"mfctest.exe?: "C:\Windows\System32\shlwapi.dll?
"mfctest.exe?:
"C:\Windows\winsxs\x86_microsoft.windows.common-controls_6595b64144ccf1df_6.0.7601.17514_none_41e6975e2bd6f2b2\comctl32.dll?
"mfctest.exe?: "C:\Windows\System32\msimg32.dll?
"mfctest.exe?: "C:\Windows\System32\oleaut32.dll?
"mfctest.exe?: "C:\Windows\System32\ole32.dll?
"mfctest.exe?: "C:\Windows\System32\imm32.dll?
"mfctest.exe?: "C:\Windows\System32\msctf.dll?
"mfctest.exe?: "C:\Windows\System32\nlaapi.dll?
"mfctest.exe?: "C:\Windows\System32\NapiNSP.dll?
"mfctest.exe?: "C:\Windows\System32\pnrpnsp.dll?
"mfctest.exe?: "C:\Windows\System32\mswsock.dll?
"mfctest.exe?: "C:\Windows\System32\dnsapi.dll?
"mfctest.exe?: "C:\Windows\System32\winrnr.dll?
"mfctest.exe?: "C:\Windows\System32\rasadhlp.dll?
"mfctest.exe?: "C:\Windows\System32\uxtheme.dll?
"mfctest.exe?: "C:\Windows\System32\dwmapi.dll?
"mfctest.exe?:
"C:\Windows\winsxs\x86_microsoft.vc90.mfcloc_1fc8b3b9a1e18e3b_9.0.30729.6161_none_49768ef57548175e\MFC90CHS.DLL?????????????????
"mfctest.exe?: "C:\Windows\System32\shell32.dll?
"mfctest.exe?: "C:\Windows\System32\cryptbase.dll?
"mfctest.exe?: "C:\Windows\System32\clbcatq.dll?
Detected memory leaks!
Dumping objects ->
{143} normal block at 0x00645670, 40 bytes long.
 Data: <                > 03 00 00 00 00 00 00 00 CD CD CD CD 00 CD CD CD
{142} normal block at 0x00645600, 52 bytes long.
 Data: < Rd  Td  Rd     > F0 52 64 00 B0 54 64 00 F0 52 64 00 00 00 00 00
{141} normal block at 0x00645590, 52 bytes long.
 Data: < Ud  Sd `Sd     > 20 55 64 00 D0 53 64 00 60 53 64 00 00 00 00 00
{140} normal block at 0x00645520, 52 bytes long.
 Data: < Rd  Ud  Rd     > F0 52 64 00 90 55 64 00 F0 52 64 00 00 00 00 00
{139} normal block at 0x006454B0, 52 bytes long.
 Data: <@Td  Sd  Vd     > 40 54 64 00 D0 53 64 00 00 56 64 00 00 00 00 00
{138} normal block at 0x00645440, 52 bytes long.
 Data: < Rd  Td  Rd     > F0 52 64 00 B0 54 64 00 F0 52 64 00 00 00 00 00
{137} normal block at 0x006453D0, 52 bytes long.
 Data: < Td  Rd  Ud     > B0 54 64 00 F0 52 64 00 90 55 64 00 00 00 00 00
{136} normal block at 0x00645360, 52 bytes long.
 Data: < Rd  Ud  Rd     > F0 52 64 00 90 55 64 00 F0 52 64 00 00 00 00 00
{135} normal block at 0x006452F0, 52 bytes long.
 Data: <@Td  Sd `Sd     > 40 54 64 00 D0 53 64 00 60 53 64 00 CD CD CD CD
{134} normal block at 0x00645288, 40 bytes long.
 Data: <        HRd  wRQ> 01 00 00 00 04 00 00 00 48 52 64 00 E9 77 52 51
{133} normal block at 0x00645248, 1 bytes long.
 Data: < > 00
{131} normal block at 0x00645130, 24 bytes long.
 Data: <         wRQ    > 01 00 00 00 04 00 00 00 E9 77 52 51 00 00 00 00
{130} normal block at 0x006450B0, 68 bytes long.
 Data: < Pd  Pd  Pd     > B0 50 64 00 B0 50 64 00 B0 50 64 00 CD CD CD CD
{129} normal block at 0x00645058, 28 bytes long.
 Data: <                > 00 00 00 00 CD CD CD CD CD CD CD CD CD CD CD CD
Object dump complete.

Steve

From boris at codesynthesis.com  Mon Sep 24 07:11:27 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Mon Sep 24 06:50:25 2012
Subject: [odb-users] memory leak detected
In-Reply-To: <CAMzAr7vg+qTBMnWCW54X6CNr5NrRV6ftbvQks_LXe9fHHnF=pg@mail.gmail.com>
References: <CAMzAr7vg+qTBMnWCW54X6CNr5NrRV6ftbvQks_LXe9fHHnF=pg@mail.gmail.
	com>
Message-ID: <boris.20120924131012@codesynthesis.com>

Hi Steve,

steve xu <steve.xu.cd@gmail.com> writes:

> When I compile a simple application with MFC framework, the memory leak is
> detected, but when i compile the application with win32 console, it has no
> memory leak.

Can you send a small test case that reproduces this problem?

Boris

From rene at catatonic.dk  Tue Sep 25 03:06:11 2012
From: rene at catatonic.dk (Rene Jensen)
Date: Tue Sep 25 03:06:21 2012
Subject: [odb-users] Assign raw foreign key id without using a loaded object
Message-ID: <CA+AwhR2EjFx5Vc=SHTV0Rkg-DEdS4ytsZS_N_SiyLrhBWC4h0w@mail.gmail.com>

Hi Boris.

This is a scenario which comes up very often when I use a mix of SQL
queries for some parts of the UI and ODB for others. I may have missed the
answer in the docs and source code.

Say I have an object with a QLazyWeakPointer to some other class.
Can I assign a reference to a different id without having to load the
referred object into memory first, by simply using e.g. an integer key (if
my foreign key database type was INTEGER, of course):

QSharedPointer<Format> F = ... load_format(10) ... bla bla

P->format = 10;
P->format = F;

Best regards,
Rene Jensen
From boris at codesynthesis.com  Tue Sep 25 04:51:34 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Tue Sep 25 04:30:24 2012
Subject: [odb-users] Assign raw foreign key id without using a loaded
	object
In-Reply-To: <CA+AwhR2EjFx5Vc=SHTV0Rkg-DEdS4ytsZS_N_SiyLrhBWC4h0w@mail.gmail.com>
References: <CA+AwhR2EjFx5Vc=SHTV0Rkg-DEdS4ytsZS_N_SiyLrhBWC4h0w@mail.gmail.com>
Message-ID: <boris.20120925104654@codesynthesis.com>

Hi Rene,

Rene Jensen <rene@catatonic.dk> writes:

> Say I have an object with a QLazyWeakPointer to some other class.
> Can I assign a reference to a different id without having to load the
> referred object into memory first, by simply using e.g. an integer key (if
> my foreign key database type was INTEGER, of course):
> 
> QSharedPointer<Format> F = ... load_format(10) ... bla bla
> 
> P->format = 10;
> P->format = F;

Yes, you can initialize an unloaded lazy pointer with just an ID (and
a reference to the database):

P->format = QLazyWeakPointer<Format> (db, 10);

This is discussed at the end of Section 6.4, "Lazy Pointers" in the
ODB manual.

Boris

From rene at catatonic.dk  Tue Sep 25 05:10:30 2012
From: rene at catatonic.dk (Rene Jensen)
Date: Tue Sep 25 05:10:41 2012
Subject: [odb-users] Assign raw foreign key id without using a loaded
	object
In-Reply-To: <boris.20120925104654@codesynthesis.com>
References: <CA+AwhR2EjFx5Vc=SHTV0Rkg-DEdS4ytsZS_N_SiyLrhBWC4h0w@mail.gmail.com>
	<boris.20120925104654@codesynthesis.com>
Message-ID: <CA+AwhR2jDPJzrx0Z=N1p-Yx+p2r7eHfmvW9K2RJmkHxzA_LF6w@mail.gmail.com>

On Tue, Sep 25, 2012 at 10:51 AM, Boris Kolpackov
<boris@codesynthesis.com>wrote:

> Hi Rene,
>
> Rene Jensen <rene@catatonic.dk> writes:
>
> > Say I have an object with a QLazyWeakPointer to some other class.
> > Can I assign a reference to a different id without having to load the
> > referred object into memory first, by simply using e.g. an integer key
> (if
> > my foreign key database type was INTEGER, of course):
> >
> > QSharedPointer<Format> F = ... load_format(10) ... bla bla
> >
> > P->format = 10;
> > P->format = F;
>
> Yes, you can initialize an unloaded lazy pointer with just an ID (and
> a reference to the database):
>
> P->format = QLazyWeakPointer<Format> (db, 10);
>
> This is discussed at the end of Section 6.4, "Lazy Pointers" in the
> ODB manual.
>


Indeed, it both works and is discussed in the manual. I am using ODB 2.0.0.
Thank you very much. This is an undispensable feature.

Regards,
Rene Jensen
From boris at codesynthesis.com  Tue Sep 25 07:35:52 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Tue Sep 25 07:14:37 2012
Subject: [odb-users] memory leak detected
In-Reply-To: <boris.20120924131012@codesynthesis.com>
References: <CAMzAr7vg+qTBMnWCW54X6CNr5NrRV6ftbvQks_LXe9fHHnF=pg@mail.gmail.com>
	<boris.20120924131012@codesynthesis.com>
Message-ID: <boris.20120925133422@codesynthesis.com>

Hi,

Boris Kolpackov <boris@codesynthesis.com> writes:

> steve xu <steve.xu.cd@gmail.com> writes:
> 
> > When I compile a simple application with MFC framework, the memory leak is
> > detected, but when i compile the application with win32 console, it has no
> > memory leak.
> 
> Can you send a small test case that reproduces this problem?

This turned out to be an application issue (the memory was released after
the leak detection was performed).

Boris

From macromarship at gmail.com  Wed Sep 26 03:45:06 2012
From: macromarship at gmail.com (Scott Zhang)
Date: Wed Sep 26 03:45:27 2012
Subject: [odb-users] about use libodb-oracle
Message-ID: <CABy3iKyoAEJND7=_sDf+a7O_g5_ZBcW5aB3+8vCJ0UTYKDAxeA@mail.gmail.com>

hello.
   I am using odb-oracle in my c++ code.
   Basically I have a member table in my db, I added this table manually,
which has following fields. ID is autoincrement field.primary key
"
IDNUMBER(38,0)No - 1 USER_IDVARCHAR2(20)Yes - - FULLNAMEVARCHAR2(50)Yes - -
PASSWORDVARCHAR2(100)Yes - - EMAILVARCHAR2(50)Yes - - BIRTHDAYDATEYes - -
ADDRESSVARCHAR2(500)Yes - - GENDERVARCHAR2(1)Yes - - PHONEVARCHAR2(20)Yes - -
REG_DATETIMEDATEYes - - LAST_ACTIVITY_DATETIMEDATEYes - "
   Most of them are operated by php code. In my c++ code, I only need to
check USER_ID and PASSWORD, and may update the password field.  So I create
a member class
#pragma db object
class MEMBER
{
public:
        friend class odb::access;
        MEMBER() {}
#pragma db id auto
unsigned long ID_;
std::string USER_ID_;
std::string PASSWORD_;

};

I tried to save a record into database and it works. But the data inserted
doesn't looks correct.

1    marship    Scott Zhang    c81e728d9d4c2f636f067f89cc14862c
macromarship@gmail.com    09/18/2012    xi'an    m    13679178184
09/18/2012    09/18/2012
Edit    2    test    -     test    -     -     -     -     -     -     -
Edit    6349320    test    -     test    -     -     -     -     -     -
    -
Edit    6348160    test    -     test

First 2 rows are inserted into by PHP code, whose ID are correct.
Then 3rd and 4th rows are insert by odb code. I don't know why ID looks so
big and not continuous. And the insert code only works for first time, when
I run same program for second time, the insert failed with exception
"
terminate called after throwing an instance of
'odb::object_already_persistent'
  what():  object already persistent
Aborted
"

Here is the code of my main function:
 std::auto_ptr<odb::database> db(new
odb::oracle::database("****","****","localhost:1521/XE",0));
        MEMBER m;
        m.USER_ID_ = "test12";
        m.PASSWORD_ = "test12";
        MEMBER m1;
        m1.USER_ID_ = "test13";
        m1.PASSWORD_ = "test13";
        transaction t(db->begin());
        db->persist(m);
        db->persist(m1);
        t.commit();


Can anyone help explain this?


Thanks.
Regards.
Scott
From macromarship at gmail.com  Wed Sep 26 05:44:23 2012
From: macromarship at gmail.com (Scott Zhang)
Date: Wed Sep 26 05:44:36 2012
Subject: [odb-users] Re: about use libodb-oracle
In-Reply-To: <CABy3iKyoAEJND7=_sDf+a7O_g5_ZBcW5aB3+8vCJ0UTYKDAxeA@mail.gmail.com>
References: <CABy3iKyoAEJND7=_sDf+a7O_g5_ZBcW5aB3+8vCJ0UTYKDAxeA@mail.gmail.com>
Message-ID: <CABy3iKycfHi4ZWhT-x7LgTt72tYGrcjTxT1xHtif0styfiY2-w@mail.gmail.com>

Looks odb-oracle generated sql has some effect. When I import the sql and
run odb oracle with it. the id is generated successfully.


On Wed, Sep 26, 2012 at 3:45 PM, Scott Zhang <macromarship@gmail.com> wrote:

> hello.
>    I am using odb-oracle in my c++ code.
>    Basically I have a member table in my db, I added this table manually,
> which has following fields. ID is autoincrement field.primary key
> "
> IDNUMBER(38,0)No - 1 USER_IDVARCHAR2(20)Yes - - FULLNAMEVARCHAR2(50)Yes - -
> PASSWORDVARCHAR2(100)Yes - - EMAILVARCHAR2(50)Yes - - BIRTHDAYDATEYes - -
> ADDRESSVARCHAR2(500)Yes - - GENDERVARCHAR2(1)Yes - - PHONEVARCHAR2(20)Yes- -
> REG_DATETIMEDATEYes - - LAST_ACTIVITY_DATETIMEDATEYes - "
>    Most of them are operated by php code. In my c++ code, I only need to
> check USER_ID and PASSWORD, and may update the password field.  So I create
> a member class
> #pragma db object
> class MEMBER
> {
> public:
>         friend class odb::access;
>         MEMBER() {}
> #pragma db id auto
> unsigned long ID_;
> std::string USER_ID_;
> std::string PASSWORD_;
>
> };
>
> I tried to save a record into database and it works. But the data inserted
> doesn't looks correct.
>
> 1    marship    Scott Zhang    c81e728d9d4c2f636f067f89cc14862c
> macromarship@gmail.com    09/18/2012    xi'an    m    13679178184
> 09/18/2012    09/18/2012
> Edit    2    test    -     test    -     -     -     -     -     -     -
> Edit    6349320    test    -     test    -     -     -     -     -     -
>     -
> Edit    6348160    test    -     test
>
> First 2 rows are inserted into by PHP code, whose ID are correct.
> Then 3rd and 4th rows are insert by odb code. I don't know why ID looks so
> big and not continuous. And the insert code only works for first time, when
> I run same program for second time, the insert failed with exception
> "
> terminate called after throwing an instance of
> 'odb::object_already_persistent'
>   what():  object already persistent
> Aborted
> "
>
> Here is the code of my main function:
>  std::auto_ptr<odb::database> db(new
> odb::oracle::database("****","****","localhost:1521/XE",0));
>         MEMBER m;
>         m.USER_ID_ = "test12";
>         m.PASSWORD_ = "test12";
>         MEMBER m1;
>         m1.USER_ID_ = "test13";
>         m1.PASSWORD_ = "test13";
>         transaction t(db->begin());
>         db->persist(m);
>         db->persist(m1);
>         t.commit();
>
>
> Can anyone help explain this?
>
>
> Thanks.
> Regards.
> Scott
>
>
>
>
>
>
>
>
>
From Davide.Anastasia at qualitycapital.com  Wed Sep 26 05:47:34 2012
From: Davide.Anastasia at qualitycapital.com (Davide Anastasia)
Date: Wed Sep 26 05:49:36 2012
Subject: [odb-users] View and Object from common struct
Message-ID: <6989CEDF160F9A42A3F5FF832CBAA72C17362A@exbe19.nasstar-t1.net>

Hi,

I am currently working on a database that stores timeseries and events.
In most cases, when I want to perform a query, I want a bunch of this
samples, in a specified order. As far as I understood, the way of doing
this is to create a view of the object with ORDER BY statement. However,
this two objects (the Event and the EventView) are essentially the same
thing.

I wonder whether anything like this is possible then:

 

struct EventData

{

std::string data1_;

int data2_;

};

 

#pragma db object

class EventObj : public EventData

{

 

};

 

#pragma db view object(EventObj)

class EventView : public EventData

{

};

 

 

Best,

 

Davide Anastasia
Analyst, Research & Development


Quality Capital Management Ltd.
QCM House * Horizon Business Village
No. 1 Brooklands Road
Weybridge * Surrey KT13 0TJ
United Kingdom

Tel: +44 (0) 1932 334 400
Fax: +44 (0) 1932 334 415
Email: Davide.Anastasia@QualityCapital.com
<mailto:Davide.Anastasia@QualityCapital.com> 

www.qualitycapital.com <http://www.qualitycapital.com/> 

________________________________

This email and any attachments are confidential and intended solely for
the use of the individual(s) to whom it is addressed. 

Any views or opinions presented are solely those of the author and do
not necessarily represent those of Quality Capital Management Ltd. If
you are not the intended recipient, be advised that you have received
this email in error and that any use, dissemination, printing,
forwarding or copying of this email is strictly prohibited. Please
contact the sender if you have received this email in error. You should
also be aware that emails are susceptible to interference and you should
not assume that the contents of this email originated from the sender
above or that they have been accurately reproduced in their original
form. Quality Capital Management Ltd is authorised and regulated by the
Financial Services Authority in the UK and is a member of the National
Futures Association in the US. 

________________________________

 

From macromarship at gmail.com  Wed Sep 26 05:54:39 2012
From: macromarship at gmail.com (Scott Zhang)
Date: Wed Sep 26 05:54:53 2012
Subject: [odb-users] Re: about use libodb-oracle
In-Reply-To: <CABy3iKycfHi4ZWhT-x7LgTt72tYGrcjTxT1xHtif0styfiY2-w@mail.gmail.com>
References: <CABy3iKyoAEJND7=_sDf+a7O_g5_ZBcW5aB3+8vCJ0UTYKDAxeA@mail.gmail.com>
	<CABy3iKycfHi4ZWhT-x7LgTt72tYGrcjTxT1xHtif0styfiY2-w@mail.gmail.com>
Message-ID: <CABy3iKx0FXkxo9uY_6VWu3WmJJ-webzR-mN2v+rnNGdMX2ZCbA@mail.gmail.com>

ok. Confirmed the issue is:
When odb insert row into database, looks it random insert some value into
the ID field. Then the default trigger created by oracle sqldeveloper only
insert new id from seq when ID is null. So the odb insert the generate
value into db. And next time it tries to insert, it fails.

I have to update oracle's generated sql to fix this.


Regards.
Scott



On Wed, Sep 26, 2012 at 5:44 PM, Scott Zhang <macromarship@gmail.com> wrote:

> Looks odb-oracle generated sql has some effect. When I import the sql and
> run odb oracle with it. the id is generated successfully.
>
>
>
> On Wed, Sep 26, 2012 at 3:45 PM, Scott Zhang <macromarship@gmail.com>wrote:
>
>> hello.
>>    I am using odb-oracle in my c++ code.
>>    Basically I have a member table in my db, I added this table
>> manually,  which has following fields. ID is autoincrement field.primary key
>> "
>> IDNUMBER(38,0)No - 1 USER_IDVARCHAR2(20)Yes - - FULLNAMEVARCHAR2(50)Yes- -
>> PASSWORDVARCHAR2(100)Yes - - EMAILVARCHAR2(50)Yes - - BIRTHDAYDATEYes - -
>> ADDRESSVARCHAR2(500)Yes - - GENDERVARCHAR2(1)Yes - - PHONEVARCHAR2(20)Yes- -
>> REG_DATETIMEDATEYes - - LAST_ACTIVITY_DATETIMEDATEYes - "
>>    Most of them are operated by php code. In my c++ code, I only need to
>> check USER_ID and PASSWORD, and may update the password field.  So I create
>> a member class
>> #pragma db object
>> class MEMBER
>> {
>> public:
>>         friend class odb::access;
>>         MEMBER() {}
>> #pragma db id auto
>> unsigned long ID_;
>> std::string USER_ID_;
>> std::string PASSWORD_;
>>
>> };
>>
>> I tried to save a record into database and it works. But the data
>> inserted doesn't looks correct.
>>
>> 1    marship    Scott Zhang    c81e728d9d4c2f636f067f89cc14862c
>> macromarship@gmail.com    09/18/2012    xi'an    m    13679178184
>> 09/18/2012    09/18/2012
>> Edit    2    test    -     test    -     -     -     -     -     -     -
>> Edit    6349320    test    -     test    -     -     -     -     -     -
>>     -
>> Edit    6348160    test    -     test
>>
>> First 2 rows are inserted into by PHP code, whose ID are correct.
>> Then 3rd and 4th rows are insert by odb code. I don't know why ID looks
>> so big and not continuous. And the insert code only works for first time,
>> when I run same program for second time, the insert failed with exception
>> "
>> terminate called after throwing an instance of
>> 'odb::object_already_persistent'
>>   what():  object already persistent
>> Aborted
>> "
>>
>> Here is the code of my main function:
>>  std::auto_ptr<odb::database> db(new
>> odb::oracle::database("****","****","localhost:1521/XE",0));
>>         MEMBER m;
>>         m.USER_ID_ = "test12";
>>         m.PASSWORD_ = "test12";
>>         MEMBER m1;
>>         m1.USER_ID_ = "test13";
>>         m1.PASSWORD_ = "test13";
>>         transaction t(db->begin());
>>         db->persist(m);
>>         db->persist(m1);
>>         t.commit();
>>
>>
>> Can anyone help explain this?
>>
>>
>> Thanks.
>> Regards.
>> Scott
>>
>>
>>
>>
>>
>>
>>
>>
>>
>
From macromarship at gmail.com  Wed Sep 26 05:59:47 2012
From: macromarship at gmail.com (Scott Zhang)
Date: Wed Sep 26 05:59:56 2012
Subject: [odb-users] View and Object from common struct
In-Reply-To: <6989CEDF160F9A42A3F5FF832CBAA72C17362A@exbe19.nasstar-t1.net>
References: <6989CEDF160F9A42A3F5FF832CBAA72C17362A@exbe19.nasstar-t1.net>
Message-ID: <CABy3iKwXBLZ1R2QNWy3Ecdrr7WA6dN56OqxRQF2rU-3vG001Ew@mail.gmail.com>

I guess you just need to append "order by" into your query statement. odb
support embed sql.

Regards.
Scott

On Wed, Sep 26, 2012 at 5:47 PM, Davide Anastasia <
Davide.Anastasia@qualitycapital.com> wrote:

> Hi,
>
> I am currently working on a database that stores timeseries and events.
> In most cases, when I want to perform a query, I want a bunch of this
> samples, in a specified order. As far as I understood, the way of doing
> this is to create a view of the object with ORDER BY statement. However,
> this two objects (the Event and the EventView) are essentially the same
> thing.
>
> I wonder whether anything like this is possible then:
>
>
>
> struct EventData
>
> {
>
> std::string data1_;
>
> int data2_;
>
> };
>
>
>
> #pragma db object
>
> class EventObj : public EventData
>
> {
>
>
>
> };
>
>
>
> #pragma db view object(EventObj)
>
> class EventView : public EventData
>
> {
>
> };
>
>
>
>
>
> Best,
>
>
>
> Davide Anastasia
> Analyst, Research & Development
>
>
> Quality Capital Management Ltd.
> QCM House * Horizon Business Village
> No. 1 Brooklands Road
> Weybridge * Surrey KT13 0TJ
> United Kingdom
>
> Tel: +44 (0) 1932 334 400
> Fax: +44 (0) 1932 334 415
> Email: Davide.Anastasia@QualityCapital.com
> <mailto:Davide.Anastasia@QualityCapital.com>
>
> www.qualitycapital.com <http://www.qualitycapital.com/>
>
> ________________________________
>
> This email and any attachments are confidential and intended solely for
> the use of the individual(s) to whom it is addressed.
>
> Any views or opinions presented are solely those of the author and do
> not necessarily represent those of Quality Capital Management Ltd. If
> you are not the intended recipient, be advised that you have received
> this email in error and that any use, dissemination, printing,
> forwarding or copying of this email is strictly prohibited. Please
> contact the sender if you have received this email in error. You should
> also be aware that emails are susceptible to interference and you should
> not assume that the contents of this email originated from the sender
> above or that they have been accurately reproduced in their original
> form. Quality Capital Management Ltd is authorised and regulated by the
> Financial Services Authority in the UK and is a member of the National
> Futures Association in the US.
>
> ________________________________
>
>
>
>
From Davide.Anastasia at qualitycapital.com  Wed Sep 26 06:53:09 2012
From: Davide.Anastasia at qualitycapital.com (Davide Anastasia)
Date: Wed Sep 26 06:53:48 2012
Subject: [odb-users] View and Object from common struct
References: <6989CEDF160F9A42A3F5FF832CBAA72C17362A@exbe19.nasstar-t1.net>
	<CABy3iKwXBLZ1R2QNWy3Ecdrr7WA6dN56OqxRQF2rU-3vG001Ew@mail.gmail.com>
Message-ID: <6989CEDF160F9A42A3F5FF832CBAA72C173641@exbe19.nasstar-t1.net>

Thanks Scott,

After your hint, and a small search, I have found this old post in the
mailing list that gives a quick example:

 

http://www.codesynthesis.com/pipermail/odb-users/2012-January/000431.htm
l

 

Regards,

Davide

 

From: Scott Zhang [mailto:macromarship@gmail.com] 
Sent: 26 September 2012 11:00
To: Davide Anastasia
Cc: odb-users@codesynthesis.com
Subject: Re: [odb-users] View and Object from common struct

 

I guess you just need to append "order by" into your query statement.
odb support embed sql.

Regards.
Scott

On Wed, Sep 26, 2012 at 5:47 PM, Davide Anastasia
<Davide.Anastasia@qualitycapital.com> wrote:

Hi,

I am currently working on a database that stores timeseries and events.
In most cases, when I want to perform a query, I want a bunch of this
samples, in a specified order. As far as I understood, the way of doing
this is to create a view of the object with ORDER BY statement. However,
this two objects (the Event and the EventView) are essentially the same
thing.

I wonder whether anything like this is possible then:



struct EventData

{

std::string data1_;

int data2_;

};



#pragma db object

class EventObj : public EventData

{



};



#pragma db view object(EventObj)

class EventView : public EventData

{

};





Best,



Davide Anastasia
Analyst, Research & Development


Quality Capital Management Ltd.
QCM House * Horizon Business Village
No. 1 Brooklands Road
Weybridge * Surrey KT13 0TJ
United Kingdom

Tel: +44 (0) 1932 334 400
Fax: +44 (0) 1932 334 415
Email: Davide.Anastasia@QualityCapital.com
<mailto:Davide.Anastasia@QualityCapital.com>

www.qualitycapital.com <http://www.qualitycapital.com/>

________________________________

This email and any attachments are confidential and intended solely for
the use of the individual(s) to whom it is addressed.

Any views or opinions presented are solely those of the author and do
not necessarily represent those of Quality Capital Management Ltd. If
you are not the intended recipient, be advised that you have received
this email in error and that any use, dissemination, printing,
forwarding or copying of this email is strictly prohibited. Please
contact the sender if you have received this email in error. You should
also be aware that emails are susceptible to interference and you should
not assume that the contents of this email originated from the sender
above or that they have been accurately reproduced in their original
form. Quality Capital Management Ltd is authorised and regulated by the
Financial Services Authority in the UK and is a member of the National
Futures Association in the US.

________________________________




 

From boris at codesynthesis.com  Wed Sep 26 07:43:24 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Wed Sep 26 07:22:03 2012
Subject: [odb-users] Re: about use libodb-oracle
In-Reply-To: <CABy3iKx0FXkxo9uY_6VWu3WmJJ-webzR-mN2v+rnNGdMX2ZCbA@mail.gmail.com>
References: <CABy3iKyoAEJND7=_sDf+a7O_g5_ZBcW5aB3+8vCJ0UTYKDAxeA@mail.gmail.com>
	<CABy3iKycfHi4ZWhT-x7LgTt72tYGrcjTxT1xHtif0styfiY2-w@mail.gmail.com>
	<CABy3iKx0FXkxo9uY_6VWu3WmJJ-webzR-mN2v+rnNGdMX2ZCbA@mail.gmail.com>
Message-ID: <boris.20120926133622@codesynthesis.com>

Hi Scott,

Scott Zhang <macromarship@gmail.com> writes:

> When odb insert row into database, looks it random insert some value into
> the ID field.

Prior to 2.1.0 ODB used a sequence and a trigger combo to implement
auto object ids. That trigger would use the next sequence value 
regardless of the value specified in the INSERT statement.

In 2.1.0 we simplified this by getting rid of the trigger and just
using the sequence directly. So now the generated INSERT statement
looks like this:

INSERT ... VALUES (MEMBER_seq.nextval, ...)

The name of the sequence is derived by appending _seq to the class
name.

Boris

From macromarship at gmail.com  Wed Sep 26 07:30:59 2012
From: macromarship at gmail.com (Scott Zhang)
Date: Wed Sep 26 07:31:09 2012
Subject: [odb-users] Re: about use libodb-oracle
In-Reply-To: <boris.20120926133622@codesynthesis.com>
References: <CABy3iKyoAEJND7=_sDf+a7O_g5_ZBcW5aB3+8vCJ0UTYKDAxeA@mail.gmail.com>
	<CABy3iKycfHi4ZWhT-x7LgTt72tYGrcjTxT1xHtif0styfiY2-w@mail.gmail.com>
	<CABy3iKx0FXkxo9uY_6VWu3WmJJ-webzR-mN2v+rnNGdMX2ZCbA@mail.gmail.com>
	<boris.20120926133622@codesynthesis.com>
Message-ID: <CABy3iKy3GcdkCY0fYonKKK20d4Tp5r7DAiFwYmWHQ29RbHNa6w@mail.gmail.com>

Hi. Boris.
    Well. Your new way will need me to change the name of existing seq. But
anyway it works. We just need to keep the sequence name following the odb
rules. And your way is compatible with the default trigger way (which php
code needs) so it is ok.


Regards.
Scott

On Wed, Sep 26, 2012 at 7:43 PM, Boris Kolpackov <boris@codesynthesis.com>wrote:

> Hi Scott,
>
> Scott Zhang <macromarship@gmail.com> writes:
>
> > When odb insert row into database, looks it random insert some value into
> > the ID field.
>
> Prior to 2.1.0 ODB used a sequence and a trigger combo to implement
> auto object ids. That trigger would use the next sequence value
> regardless of the value specified in the INSERT statement.
>
> In 2.1.0 we simplified this by getting rid of the trigger and just
> using the sequence directly. So now the generated INSERT statement
> looks like this:
>
> INSERT ... VALUES (MEMBER_seq.nextval, ...)
>
> The name of the sequence is derived by appending _seq to the class
> name.
>
> Boris
>
From ststrou at sandia.gov  Wed Sep 26 13:08:11 2012
From: ststrou at sandia.gov (Stroud, Sean T)
Date: Wed Sep 26 13:08:48 2012
Subject: [odb-users] Error trying to create schema with polymorphic pragma
Message-ID: <8EDFC8F388A1E44BB67C9BAA57091F9E285C9A82@EXMB03.srn.sandia.gov>

Hi,

I get the following runtime error trying to create schema via the schema_catalog:

   Error: 1005 (HY000): Can't create table 'TEST_DB.Sensor_Product' (errno: 150)

I can't figure out what I am doing wrong.

Here are some details:
- I'm on RHEL 5.8
- I'm using ODB 2.1.0
- I have two classes Product and Sensor_Product (see files below).
- Sensor_Product inherits from Product.  Both use the "polymorphic" pragma.
- If I remove the "polymorphic" pragma, I no longer get the error.

Here is how I invoke odb on product.h and sensor_product.h:

  % which odb
  /home/ststrou/ODB2.1/odb-2.1.0-x86_64-linux-gnu/bin/odb

  % odb -d mysql -q -I /home/ststrou/ODB2.1_install/include --generate-schema --schema-format embedded product.h sensor_product.h

The "/home/ststrou/ODB2.1_install/" directory is where I've installed the odb 2.1.0 includes and libraries.

Here are the files I am using:

------------ file main.cc -----------------
#include <odb/database.hxx>
#include <odb/transaction.hxx>
#include <odb/mysql/database.hxx>
#include <odb/schema-catalog.hxx>

#include <iostream>
#include <stdlib.h>

using namespace std;
using namespace odb::core;

int main(int argc, char *argv[])
{
  try {
    auto_ptr<database> db (new odb::mysql::database (argc, argv));

    transaction t (db->begin ());
    schema_catalog::create_schema (*db);
    t.commit ();
  } catch ( odb::exception &ex) {
    cout << "Error: " << ex.what() << endl;
    exit(1);
  }
}

------------ file product.h -----------------
#ifndef Product_H
#define Product_H

#include <odb/core.hxx>

#pragma db object polymorphic
class Product {
  public:
    Product();

    virtual ~Product();

  protected:
    friend class odb::access;

    #pragma db id
    int  m_product_id;
};

#endif

------------ file product.cc -----------------
#include "product.h"

Product::Product()
{
  m_product_id = -1;
}

Product::~Product()
{
}

------------ file sensor_product.h -----------------
#ifndef Sensor_Product_H
#define Sensor_Product_H

#include "product.h"
#include <vector>

#include <odb/core.hxx>

#pragma db object polymorphic
class Sensor_Product : public Product {
  public:
    Sensor_Product();

    virtual ~Sensor_Product();

  private:
    friend class odb::access;

    #pragma db type("MEDIUMBLOB")
    std::vector<char> m_sensor_data;
};

#endif

------------ file sensor_product.cc -----------------
#include "sensor_product.h"

Sensor_Product::Sensor_Product()
{
}

Sensor_Product::~Sensor_Product()
{
}



From boris at codesynthesis.com  Wed Sep 26 14:16:05 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Wed Sep 26 13:54:43 2012
Subject: [odb-users] Error trying to create schema with polymorphic pragma
In-Reply-To: <8EDFC8F388A1E44BB67C9BAA57091F9E285C9A82@EXMB03.srn.sandia.gov>
References: <8EDFC8F388A1E44BB67C9BAA57091F9E285C9A82@EXMB03.srn.sandia.gov>
Message-ID: <boris.20120926200813@codesynthesis.com>

Hi Sean,

Stroud, Sean T <ststrou@sandia.gov> writes:

> I get the following runtime error trying to create schema via the 
> schema_catalog:
> 
>    Error: 1005 (HY000): Can't create table 'TEST_DB.Sensor_Product' (errno: 150)

I tried your test and I get the same error. It appears we don't account
for the order of table creation when adding a foreign key constraint. So
this is a bug in ODB.

If you would like, I can build you a binary with the fix. 

Alternatively, there is a fairly easy workaround. The idea is to assign
different names to product and sensor_product schemas and then create
them in the correct order:

odb ... --schema-name base product.h
odb ... --schema-name derived sensor_product.h

And then in main():

schema_catalog::create_schema (*db, "base");
schema_catalog::create_schema (*db, "derived");

Thanks for reporting this and providing the test case!

Boris

From ststrou at sandia.gov  Wed Sep 26 14:01:22 2012
From: ststrou at sandia.gov (Stroud, Sean T)
Date: Wed Sep 26 14:02:01 2012
Subject: [EXTERNAL] Re: [odb-users] Error trying to create schema
	with polymorphic pragma
In-Reply-To: <boris.20120926200813@codesynthesis.com>
References: <8EDFC8F388A1E44BB67C9BAA57091F9E285C9A82@EXMB03.srn.sandia.gov>
	<boris.20120926200813@codesynthesis.com>
Message-ID: <8EDFC8F388A1E44BB67C9BAA57091F9E285C9AB5@EXMB03.srn.sandia.gov>

Hi Boris,

No need to build a special binary - I can just use the workaround you provided for now.

Thanks for the quick response!

Sean

-----Original Message-----
From: Boris Kolpackov [mailto:boris@codesynthesis.com] 
Sent: Wednesday, September 26, 2012 12:16 PM
To: Stroud, Sean T
Cc: odb-users@codesynthesis.com
Subject: [EXTERNAL] Re: [odb-users] Error trying to create schema with polymorphic pragma

Hi Sean,

Stroud, Sean T <ststrou@sandia.gov> writes:

> I get the following runtime error trying to create schema via the
> schema_catalog:
> 
>    Error: 1005 (HY000): Can't create table 'TEST_DB.Sensor_Product' 
> (errno: 150)

I tried your test and I get the same error. It appears we don't account for the order of table creation when adding a foreign key constraint. So this is a bug in ODB.

If you would like, I can build you a binary with the fix. 

Alternatively, there is a fairly easy workaround. The idea is to assign different names to product and sensor_product schemas and then create them in the correct order:

odb ... --schema-name base product.h
odb ... --schema-name derived sensor_product.h

And then in main():

schema_catalog::create_schema (*db, "base"); schema_catalog::create_schema (*db, "derived");

Thanks for reporting this and providing the test case!

Boris



From macromarship at gmail.com  Thu Sep 27 03:28:56 2012
From: macromarship at gmail.com (Scott Zhang)
Date: Thu Sep 27 03:29:01 2012
Subject: [odb-users] how to get the result size of oracle query
Message-ID: <CABy3iKyggFr4OAOVtXH05vYGvFMRDnsffV68gYmdUdWmftKQ3w@mail.gmail.com>

Hello.
   I am trying to query from oracle. Then I tried to use result.size() to
get the total number. Unfortunately, it raises "result_not_cached". Then I
tried to call "cache" manually but still get such exception.
   I have to look into the libodb-oracle code. And see the cache and size
function is actually not implemented.
   So how do you get the count of rows?

Thanks.
Regards.
Scott
From boris at codesynthesis.com  Thu Sep 27 05:20:23 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Thu Sep 27 04:58:45 2012
Subject: [odb-users] how to get the result size of oracle query
In-Reply-To: <CABy3iKyggFr4OAOVtXH05vYGvFMRDnsffV68gYmdUdWmftKQ3w@mail.gmail.com>
References: <CABy3iKyggFr4OAOVtXH05vYGvFMRDnsffV68gYmdUdWmftKQ3w@mail.gmail.com>
Message-ID: <boris.20120927111202@codesynthesis.com>

Hi Scott,

Scott Zhang <macromarship@gmail.com> writes:

>    I am trying to query from oracle. Then I tried to use result.size() to
> get the total number. Unfortunately, it raises "result_not_cached". Then I
> tried to call "cache" manually but still get such exception.
>    I have to look into the libodb-oracle code. And see the cache and size
> function is actually not implemented.

Query result caching is not supported for Oracle, as discussed in Section
16.5.2, "Query Result Caching" in the ODB manual.


> So how do you get the count of rows?

Oracle streams query results so the result size is not known. You can
either iterate until the end of the result set (potentially saving the
loaded items) counting the entries. Or, you can run SELECT COUNT(*);
the Views chapter has an example of this.

Note also that both of these approaches have drawbacks. In fact, the
best solution is to re-design your application not to need the result
size in the first place. This is also what's recommended by Oracle.

Boris

From chezgi at yahoo.com  Thu Sep 27 05:21:25 2012
From: chezgi at yahoo.com (abbas ali chezgi)
Date: Thu Sep 27 06:02:56 2012
Subject: [odb-users] bug in generating schema for QList<QString>
Message-ID: <1348737685.66244.YahooMailNeo@web121306.mail.ne1.yahoo.com>

if we define QList<QString> in "value" struct ==> generated code dosnt contain any schema for that member. and nothing will be saved.

Group::members field in this code:

------------------------- sample.h

#include <QString>
#include <QList> 

#pragma db value
class Group
{
public:
??? QString name;
??? QList<QString> members;
};

#pragma db object
class Global
{
public:
??? #pragma db id auto
??? long long id_;
??? QList<Group> groups;
};

--------------------------------
From boris at codesynthesis.com  Thu Sep 27 06:28:58 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Thu Sep 27 06:07:17 2012
Subject: [odb-users] bug in generating schema for QList<QString>
In-Reply-To: <1348737685.66244.YahooMailNeo@web121306.mail.ne1.yahoo.com>
References: <1348737685.66244.YahooMailNeo@web121306.mail.ne1.yahoo.com>
Message-ID: <boris.20120927122541@codesynthesis.com>

Hi,

abbas ali chezgi <chezgi@yahoo.com> writes:

> if we define QList<QString> in "value" struct ==> generated code dosnt
> contain any schema for that member. and nothing will be saved.
> 
> Group::members field in this code:
> 
> #pragma db value
> class Group
> {
> public:
>     QString name;
>     QList<QString> members;
> };
> 
> #pragma db object
> class Global
> {
> public:
>     #pragma db id auto
>     long long id_;
>     QList<Group> groups;
> };

ODB (as most other ORMs) does not support containers of containers. In
this case, Group contains a container and is itself used as an element
of a container. See Chapter 5, "Containers" in the ODB manual for details.

Boris

From magnus.granqvist at tailormade.se  Thu Sep 27 03:06:52 2012
From: magnus.granqvist at tailormade.se (Magnus Granqvist)
Date: Thu Sep 27 07:00:04 2012
Subject: [odb-users] Write-on-commit support
Message-ID: <90C6E2CB-80D8-4DA9-8FFC-4DE7D86ED37B@tailormade.se>

Hi,
 
Does ODB has support for "write-on-commit" like for example the Wt::Dbo library?
"The session also keeps track of objects that have been modified and which need to be flushed (using SQL statements) to the database. Flushing happens automatically when committing the transaction".
 
I like to avoid to explicit call the update() method to execute the update statement to DB, instead like the session (cache) and the transaction handle this when calling the commit() method.
 
Kind regards,
Magnus
___________________________________  
Magnus Granqvist 
System Developer
 
TailorMade 
Mobile: +46 70 449 54 40
Direct: + 46 8 5624 54 40
E-mail: magnus.granqvist @tailormade.se
Web: www.tailormadeglobal.com

Esplanaden 3A, SE-172 67 Sundbyberg, Sweden 
Fax: +46 8 29 36 56 Phone: +46  8 562 454 00
 

 
TailorMade develop and implement billing solutions for the telco and utility industries. We have 15 years of experience with installations in 15 countries.  We focus on delivering innovative billing solutions that can increase flexibility and add new capabilities to the existing Billing / CIS systems. 
 
P  Please consider the environment before printing this e-mail
 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 10461 bytes
Desc: not available
Url : http://codesynthesis.com/pipermail/odb-users/attachments/20120927/32435a94/image001-0001.png
From boris at codesynthesis.com  Thu Sep 27 07:53:13 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Thu Sep 27 07:31:34 2012
Subject: [odb-users] Write-on-commit support
In-Reply-To: <90C6E2CB-80D8-4DA9-8FFC-4DE7D86ED37B@tailormade.se>
References: <90C6E2CB-80D8-4DA9-8FFC-4DE7D86ED37B@tailormade.se>
Message-ID: <boris.20120927133422@codesynthesis.com>

Hi Magnus,

Magnus Granqvist <magnus.granqvist@tailormade.se> writes:

> Does ODB has support for "write-on-commit" like for example the Wt::Dbo
> library? "The session also keeps track of objects that have been modified
> and which need to be flushed (using SQL statements) to the database.
> Flushing happens automatically when committing the transaction".

No, ODB does not (yet) support this.

Generally, with ODB, we try to avoid the situations where you have an
innocent looking call that can result in a large and surprising number
of database operations. We believe that this kind of "high-level" 
abstractions are the reason why people often find themselves in a mess
with ORMs like Hibernate (and also why some people think that ORMs are
evil). As a result, we try to stick with the "one call translates to
one statement execution" principle as much as possible.

In the case of this particular feature, there is also the problem of
detecting object modifications (aka dirty checking). The only way to
implement it in a general and robust way is to keep a complete copy
of the object to compare against later. Again, this can be a huge,
hidden, and non-obvious overhead. At the same time, an application
developer can probably provide a much more efficient implementation
(e.g., store a modification flag in the object).


> I like to avoid to explicit call the update() method to execute the
> update statement to DB, instead like the session (cache) and the
> transaction handle this when calling the commit() method.

This would be pretty easy to implement with a bit of custom code if
there was a way to iterate over objects in a session. While there
is no such support currently, it will be pretty easy to add.

Also, implementing this feature shouldn't be that hard provided the
user does dirty checking. We can use the callback machinery to ask for
this information. So if you really need this feature and would be
willing to try it and give feedback, then I can go ahead and implement
it as an option.

Boris

From pstath at axxcelera.com  Thu Sep 27 09:38:09 2012
From: pstath at axxcelera.com (Stath, Paul)
Date: Thu Sep 27 09:38:15 2012
Subject: [odb-users] Prepared statement feature
In-Reply-To: <201209271102.q8RB2NqV019699@codesynthesis.com>
References: <201209271102.q8RB2NqV019699@codesynthesis.com>
Message-ID: <3233D27CC5658E4598557F8521F6B07E2146550D06@RIC-MS01.abw.int>

It there a timeline for when prepared statements as described in http://www.codesynthesis.com/pipermail/odb-users/2012-January/000434.html
might be available/released?

Right now I don't see any way to create a prepared statement, or to obtain the prepared statement from an executed query.
(Please correct me if I'm just missing it.)

I have a table that repeatedly uses a query, and I would prefer to store the prepared statement for reuse.

Database:  SQLite

-- Paul

From ststrou at sandia.gov  Thu Sep 27 14:27:24 2012
From: ststrou at sandia.gov (Stroud, Sean T)
Date: Thu Sep 27 14:27:50 2012
Subject: [odb-users] Odd behavior with 'results not cached' exception with
	mysql
Message-ID: <8EDFC8F388A1E44BB67C9BAA57091F9E285C9D67@EXMB03.srn.sandia.gov>

Hi,

I am seeing some odd behavior, which the code below illustrates:

    bool make_it_fail = true;   // set this to "false" to make the exception occur

    odb::result<My_Class> r( db->query<My_Class>(false) );

    r.cache();

    if ( make_it_fail )
      for ( odb::result<My_Class>::iterator iter(r.begin()); iter != r.end(); ++iter );

    // If the for-loop above is executed, then this line will throw a "results not cached" exception.
    // Otherwise no exception will be thrown
    cout << r.size() << endl;

Is this the intended behavior?  Seems odd that merely iterating over the result set would cause it to become no longer cached.  The example class "My_Class" is very simple - just a couple of ints and std::strings (no pointers, containters, inheritance, etc).

I am using ODB 2.1.0 on RHEL 5.8 with mysql 5.5.27.

Sean


From Davide.Anastasia at qualitycapital.com  Fri Sep 28 05:46:53 2012
From: Davide.Anastasia at qualitycapital.com (Davide Anastasia)
Date: Fri Sep 28 05:46:52 2012
Subject: [odb-users] Query problem
Message-ID: <6989CEDF160F9A42A3F5FF832CBAA72C173789@exbe19.nasstar-t1.net>

Hi,

I've hit a problem with retrieval of persisted objects that I cannot
solve! L

 

I have a query statement, that runs fine: the tracer reports the correct
query. I can run it on the pgadmin3 interface without problems and it
gives the expected output. If I ask the size of the iterator, it will
report the right number, but as soon as I deference the iterator with
operator*, I get an segmentation fault (traced to boost::shared_ptr).

 

My code is basically the for loop across the element of a query result
that you can find listed in the manual dozens of times, which I have
used myself many times as well before now.

 

Do you have any idea of what I should look at to understand the problem?

 

Best,

 

Davide Anastasia
Analyst, Research & Development


Quality Capital Management Ltd.
QCM House * Horizon Business Village
No. 1 Brooklands Road
Weybridge * Surrey KT13 0TJ
United Kingdom

Tel: +44 (0) 1932 334 400
Fax: +44 (0) 1932 334 415
Email: Davide.Anastasia@QualityCapital.com
<mailto:Davide.Anastasia@QualityCapital.com> 

www.qualitycapital.com <http://www.qualitycapital.com/> 

________________________________

This email and any attachments are confidential and intended solely for
the use of the individual(s) to whom it is addressed. 

Any views or opinions presented are solely those of the author and do
not necessarily represent those of Quality Capital Management Ltd. If
you are not the intended recipient, be advised that you have received
this email in error and that any use, dissemination, printing,
forwarding or copying of this email is strictly prohibited. Please
contact the sender if you have received this email in error. You should
also be aware that emails are susceptible to interference and you should
not assume that the contents of this email originated from the sender
above or that they have been accurately reproduced in their original
form. Quality Capital Management Ltd is authorised and regulated by the
Financial Services Authority in the UK and is a member of the National
Futures Association in the US. 

________________________________

 

From boris at codesynthesis.com  Fri Sep 28 06:45:00 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Fri Sep 28 06:23:03 2012
Subject: [odb-users] Query problem
In-Reply-To: <6989CEDF160F9A42A3F5FF832CBAA72C173789@exbe19.nasstar-t1.net>
References: <6989CEDF160F9A42A3F5FF832CBAA72C173789@exbe19.nasstar-t1.net>
Message-ID: <boris.20120928124105@codesynthesis.com>

Hi Davide,

Davide Anastasia <Davide.Anastasia@qualitycapital.com> writes:

> I have a query statement, that runs fine: the tracer reports the correct
> query. I can run it on the pgadmin3 interface without problems and it
> gives the expected output. If I ask the size of the iterator, it will
> report the right number, but as soon as I deference the iterator with
> operator*, I get an segmentation fault (traced to boost::shared_ptr).

Seeing a stack trace would be helpful. Without any further information,
my guess would be you have a relationship cycle which requires a session
to load. So the first thing I would try is add a session and see if it
helps.

BTW, ODB 2.0.0 and later detect this problem and throw the session_required
exception.

Boris

From boris at codesynthesis.com  Fri Sep 28 06:56:59 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Fri Sep 28 06:35:03 2012
Subject: [odb-users] Prepared statement feature
In-Reply-To: <3233D27CC5658E4598557F8521F6B07E2146550D06@RIC-MS01.abw.int>
References: <201209271102.q8RB2NqV019699@codesynthesis.com>
	<3233D27CC5658E4598557F8521F6B07E2146550D06@RIC-MS01.abw.int>
Message-ID: <boris.20120928125126@codesynthesis.com>

Hi Paul,

Stath, Paul <pstath@axxcelera.com> writes:

> It there a timeline for when prepared statements as described in 
> http://www.codesynthesis.com/pipermail/odb-users/2012-January/000434.html
> might be available/released?

It is on our TODO list but there is no definite timeline. Generally, we
assign higher priority to features that someone is willing to start using
immediately (i.e., test a pre-release, give feedback, etc). So if you are
interested, we can implement something for you to try.

And speaking of feedback, what do you think about the usage scenario
shown in that email? Will that fit with what you are trying to do? To
me personally it feels a bit awkward.

Boris

From boris at codesynthesis.com  Fri Sep 28 08:54:49 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Fri Sep 28 08:32:51 2012
Subject: [odb-users] Odd behavior with 'results not cached' exception
	with mysql
In-Reply-To: <8EDFC8F388A1E44BB67C9BAA57091F9E285C9D67@EXMB03.srn.sandia.gov>
References: <8EDFC8F388A1E44BB67C9BAA57091F9E285C9D67@EXMB03.srn.sandia.gov>
Message-ID: <boris.20120928144639@codesynthesis.com>

Hi Sean,

Stroud, Sean T <ststrou@sandia.gov> writes:

>     odb::result<My_Class> r( db->query<My_Class>(false) );
>     r.cache();
>
>     if ( make_it_fail )
>       for ( odb::result<My_Class>::iterator iter(r.begin()); iter != r.end(); ++iter );
> 
>     // If the for-loop above is executed, then this line will throw a 
>     // "results not cached" exception. Otherwise no exception will be
>     // thrown.
>     cout << r.size() << endl;

That's a bug (you seem to be particularly good at uncovering them ;-)).

What happens is we try to release the underlying result set as soon as
possible in order not to hold on to database resources unnecessarily.
So in this case, once we hit the end during iteration, we release the
result set immediately but with that we also loose the size information.

I've committed a fix and can package you libodb-mysql-2.0.1 if you would
like. There is also an easy work-around: simply query the size before
iteration:

size_t n (r.size ());

for ( odb::result<My_Class>::iterator iter(r.begin()); iter != r.end(); ++iter );

cout << n << endl;

I guess it is somewhat unusual to query the result set size after the
iteration. That's probably why we haven't heard about anyone running
into this bug before.

Boris

From Davide.Anastasia at qualitycapital.com  Fri Sep 28 08:36:28 2012
From: Davide.Anastasia at qualitycapital.com (Davide Anastasia)
Date: Fri Sep 28 08:38:35 2012
Subject: [odb-users] Query problem
References: <6989CEDF160F9A42A3F5FF832CBAA72C173789@exbe19.nasstar-t1.net>
	<boris.20120928124105@codesynthesis.com>
Message-ID: <6989CEDF160F9A42A3F5FF832CBAA72C1737BE@exbe19.nasstar-t1.net>

This is my backtrace:

#0  0x00000038ca8328a5 in raise () from /lib64/libc.so.6
#1  0x00000038ca834085 in abort () from /lib64/libc.so.6
#2  0x00000038ca82ba1e in __assert_fail_base () from /lib64/libc.so.6
#3  0x00000038ca82bae0 in __assert_fail () from /lib64/libc.so.6
#4  0x0000000000468813 in boost::shared_ptr<QCM::ODB::Quote>::operator*
(this=0x6c4128) at /usr/include/boost/smart_ptr/shared_ptr.hpp:412
#5  0x0000000000467fd9 in
odb::pointer_traits<boost::shared_ptr<QCM::ODB::Quote> >::get_ref
(p=...) at /opt/odb/include/odb/boost/smart-ptr/pointer-traits.hxx:45
#6  0x0000000000467307 in odb::result_iterator<QCM::ODB::Quote,
(odb::class_kind)0>::operator* (this=0x7fffffffd920) at
/opt/odb/include/odb/object-result.hxx:98
#7  0x000000000046673d in QCM::reader::OdbReaderImpl::next
(this=0x6b8e80, outputTick=...) at
/home/davide/DatabasePerfToolTickData/trunk/include/reader/OdbReader.cpp
:102
#8  0x0000000000466803 in QCM::reader::OdbReaderImpl::next
(this=0x6b8e80, outputTick=...) at
/home/davide/DatabasePerfToolTickData/trunk/include/reader/OdbReader.cpp
:112
#9  0x00000000004657ef in QCM::reader::OdbReader::next
(this=0x7fffffffdbb0, outputTick=...) at
/home/davide/DatabasePerfToolTickData/trunk/include/reader/OdbReader.cpp
:198
#10 0x000000000044fb3a in TestOdbReader_CheckOrdering_Test::TestBody
(this=0x6aec00) at
/home/davide/DatabasePerfToolTickData/trunk/test/TestOdbReader.cpp:160

I am currently using ODB 2.2.0.

Best,
Davide

-----Original Message-----
From: Boris Kolpackov [mailto:boris@codesynthesis.com] 
Sent: 28 September 2012 11:45
To: Davide Anastasia
Cc: odb-users@codesynthesis.com
Subject: Re: [odb-users] Query problem

Hi Davide,

Davide Anastasia <Davide.Anastasia@qualitycapital.com> writes:

> I have a query statement, that runs fine: the tracer reports the 
> correct query. I can run it on the pgadmin3 interface without problems

> and it gives the expected output. If I ask the size of the iterator, 
> it will report the right number, but as soon as I deference the 
> iterator with operator*, I get an segmentation fault (traced to
boost::shared_ptr).

Seeing a stack trace would be helpful. Without any further information,
my guess would be you have a relationship cycle which requires a session
to load. So the first thing I would try is add a session and see if it
helps.

BTW, ODB 2.0.0 and later detect this problem and throw the
session_required exception.

Boris

From pstath at axxcelera.com  Fri Sep 28 10:04:18 2012
From: pstath at axxcelera.com (Stath, Paul)
Date: Fri Sep 28 10:04:15 2012
Subject: [odb-users] Prepared statement feature
In-Reply-To: <boris.20120928125126@codesynthesis.com>
References: <201209271102.q8RB2NqV019699@codesynthesis.com>
	<3233D27CC5658E4598557F8521F6B07E2146550D06@RIC-MS01.abw.int>
	<boris.20120928125126@codesynthesis.com>
Message-ID: <3233D27CC5658E4598557F8521F6B07E2146550D92@RIC-MS01.abw.int>

I am interested, however I'm also not wild about the usage scenario from that previous email.

You state the two issues with caching queries is storing the query condition as well as any
by-reference parameters.

I'm imagining an abstract ODB prepared query class that can be extended by the programmer with
the query parameter fields.

The odb::pquery would provide an abstract base class that would have an init_query() method.
The init_query() method would be responsible for specifying the condition of the query, and would
be called during construction.

The odb::database would have a factory method for creating pquery objects, and the db->query() method
would be overloaded to take both odb::query and odb::pquery objects.
By creating the pquery object via a factory method, you can store the statement directly in the
pquery object and also save a pointer to the database so that when the deconstructor for the pquery
object is called, it can remove the prepared query from the database session.

The prepared query would typically be defined in the same translation unit as the object the
prepared query returns.  (i.e. person.hxx)

class age_pquery : public odb::pquery<person>
{
    public:
    unsigned short age_;

    private:
    init_query() {
        query (query::age < query::_ref(age_));
    }
}

In the program, I would create the prepared query as follows:

   age_pquery apq (odb::pquery<age_pquery>();

Later, I would execute the query by setting the parameter values and
passing the pquery to the db.

   apq.age = 34;
   result r (db->query<person> (age_pquery));

-- Paul

> -----Original Message-----
> From: Boris Kolpackov [mailto:boris@codesynthesis.com]
> Sent: Friday, September 28, 2012 6:57 AM
> To: Stath, Paul
> Cc: odb-users@codesynthesis.com
> Subject: Re: [odb-users] Prepared statement feature
> 
> Hi Paul,
> 
> Stath, Paul <pstath@axxcelera.com> writes:
> 
> > It there a timeline for when prepared statements as described in
> > http://www.codesynthesis.com/pipermail/odb-users/2012-
> January/000434.html
> > might be available/released?
> 
> It is on our TODO list but there is no definite timeline. Generally, we
> assign higher priority to features that someone is willing to start
> using
> immediately (i.e., test a pre-release, give feedback, etc). So if you
> are
> interested, we can implement something for you to try.
> 
> And speaking of feedback, what do you think about the usage scenario
> shown in that email? Will that fit with what you are trying to do? To
> me personally it feels a bit awkward.
> 
> Boris

From boris at codesynthesis.com  Fri Sep 28 10:26:54 2012
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Fri Sep 28 10:04:56 2012
Subject: [odb-users] Query problem
In-Reply-To: <6989CEDF160F9A42A3F5FF832CBAA72C1737BE@exbe19.nasstar-t1.net>
References: <6989CEDF160F9A42A3F5FF832CBAA72C173789@exbe19.nasstar-t1.net>
	<boris.20120928124105@codesynthesis.com>
	<6989CEDF160F9A42A3F5FF832CBAA72C1737BE@exbe19.nasstar-t1.net>
Message-ID: <boris.20120928162213@codesynthesis.com>

Hi Davide,

Davide Anastasia <Davide.Anastasia@qualitycapital.com> writes:

> This is my backtrace:
> 
> [...]

Is it possible that you are dereferencing an iterator when it reached
the end? Can you show the relevant code in:

#7  0x000000000046673d in QCM::reader::OdbReaderImpl::next
(this=0x6b8e80, outputTick=...) at
/home/davide/DatabasePerfToolTickData/trunk/include/reader/OdbReader.cpp:102

Boris

From ststrou at sandia.gov  Fri Sep 28 10:25:59 2012
From: ststrou at sandia.gov (Stroud, Sean T)
Date: Fri Sep 28 10:26:25 2012
Subject: [EXTERNAL] Re: [odb-users] Odd behavior with 'results not
	cached' exception with mysql
In-Reply-To: <boris.20120928144639@codesynthesis.com>
References: <8EDFC8F388A1E44BB67C9BAA57091F9E285C9D67@EXMB03.srn.sandia.gov>
	<boris.20120928144639@codesynthesis.com>
Message-ID: <8EDFC8F388A1E44BB67C9BAA57091F9E285C9E78@EXMB03.srn.sandia.gov>

Hi Boris,

Thanks for the explanation - makes sense.  No need for a special package, the workaround you suggested is good enough for what I'm doing.

Thanks!

Sean

-----Original Message-----
From: Boris Kolpackov [mailto:boris@codesynthesis.com] 
Sent: Friday, September 28, 2012 6:55 AM
To: Stroud, Sean T
Cc: odb-users@codesynthesis.com
Subject: [EXTERNAL] Re: [odb-users] Odd behavior with 'results not cached' exception with mysql

Hi Sean,

Stroud, Sean T <ststrou@sandia.gov> writes:

>     odb::result<My_Class> r( db->query<My_Class>(false) );
>     r.cache();
>
>     if ( make_it_fail )
>       for ( odb::result<My_Class>::iterator iter(r.begin()); iter != 
> r.end(); ++iter );
> 
>     // If the for-loop above is executed, then this line will throw a 
>     // "results not cached" exception. Otherwise no exception will be
>     // thrown.
>     cout << r.size() << endl;

That's a bug (you seem to be particularly good at uncovering them ;-)).

What happens is we try to release the underlying result set as soon as possible in order not to hold on to database resources unnecessarily.
So in this case, once we hit the end during iteration, we release the result set immediately but with that we also loose the size information.

I've committed a fix and can package you libodb-mysql-2.0.1 if you would like. There is also an easy work-around: simply query the size before
iteration:

size_t n (r.size ());

for ( odb::result<My_Class>::iterator iter(r.begin()); iter != r.end(); ++iter );

cout << n << endl;

I guess it is somewhat unusual to query the result set size after the iteration. That's probably why we haven't heard about anyone running into this bug before.

Boris