From daniel at pfeifer-mail.de  Thu Jul  2 16:26:48 2009
From: daniel at pfeifer-mail.de (Daniel Pfeifer)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Prefix can not be resolved to a URI
Message-ID: <a57cfbd3b22903f90a7d3d991cd8ed1b-EhVcX1dJTQZXRw0RBgwNVDBXdh9XVlpBXkJHHF5ZWC9QUUYPW0F3BlJLXF4wQEcGWV9fRlxVUA==-webmailer2@server04.webmailer.hosteurope.de>


Hello everyone!

I am trying to build an xml-parser from the schema files provided at
http://xmpp.org/schemas/.
When I compile the file jabber-client.xsd, I get the following errors:

jabber-client.xsd:172:49: error: Prefix: 'err' can not be resolved to a URI
jabber-client.xsd:172:49: error: Schema Representation Constraint:
Namespace '' is referenced without <import> declaration
jabber-client.xsd:174:36: error: Prefix: 'err' can not be resolved to a URI
jabber-client.xsd:174:36: error: Schema Representation Constraint:
Namespace '' is referenced without <import> declaration
jabber-client.xsd:174:36: error: Ref element err:text not found in the
Schema

The relating snippet of the file http://xmpp.org/schemas/jabber-client.xsd:

-- snip --
171  <xs:sequence  xmlns:err='urn:ietf:params:xml:ns:xmpp-stanzas'>
172    <xs:group   ref='err:stanzaErrorGroup'/>
173    <xs:element ref='err:text'
174                minOccurs='0'/>
175  </xs:sequence>
-- snap --

Is this a bug in the xsd parser generator, a bug in the schema or is the
problem sitting in front of my computer?

cheers, Daniel


From boris at codesynthesis.com  Thu Jul  2 22:10:36 2009
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Prefix can not be resolved to a URI
In-Reply-To: <a57cfbd3b22903f90a7d3d991cd8ed1b-EhVcX1dJTQZXRw0RBgwNVDBXdh9XVlpBXkJHHF5ZWC9QUUYPW0F3BlJLXF4wQEcGWV9fRlxVUA==-webmailer2@server04.webmailer.hosteurope.de>
References: <a57cfbd3b22903f90a7d3d991cd8ed1b-EhVcX1dJTQZXRw0RBgwNVDBXdh9XVlpBXkJHHF5ZWC9QUUYPW0F3BlJLXF4wQEcGWV9fRlxVUA==-webmailer2@server04.webmailer.hosteurope.de>
Message-ID: <boris.20090703040357@codesynthesis.com>

Hi Daniel,

Daniel Pfeifer <daniel@pfeifer-mail.de> writes:

> When I compile the file jabber-client.xsd, I get the following errors:
> 
> jabber-client.xsd:172:49: error: Prefix: 'err' can not be resolved to a URI
> ...

When I tried to compile jabber-client.xsd I got this error:

jabber-client.xsd: error: 'http://xmpp.org/schemas/stanzaerror.xsd': unable to open in read mode

This error is due to the XSD compiler only supporting local filesystem
paths for security reasons. Perhaps you tried to "fix" this error by
removing the import declaration that caused it and that's how you got
the above error? ;-)

The recommended fix for this is to download stanzaerror.xsd (and xml.xsd)
and then use the --location-map options to re-map URI locations to local
files. In the end I was able to compile jabber-client.xsd with the
following command line:

xsd cxx-tree \
--location-map http://xmpp.org/schemas/stanzaerror.xsd=stanzaerror.xsd \
--location-map http://www.w3.org/2001/03/xml.xsd=xml.xsd \
--namespace-map jabber:client=jabber_client \
--namespace-map urn:ietf:params:xml:ns:xmpp-stanzas=xmpp_stanzas \
jabber-client.xsd

Boris


From ongs_1999 at yahoo.com  Sun Jul  5 13:08:26 2009
From: ongs_1999 at yahoo.com (Steve)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Problem compiling the xsd-3.2.0+dep package
Message-ID: <658484.22849.qm@web32203.mail.mud.yahoo.com>

Hi,

I am trying to compile the xsd-3.2.0+dep package and running into problems. I am compiling it on cygwin which I compiled boost and xerces-c++ successfully. 

Here is the problem

First of all, I made the following modifications to the "build.sh" script:
CPPFLAGS=" -I/opt/boost/include/boost-1_39 -I/opt/xerces/include/xercesc $CPPFLAGS"
LDFLAGS="-L/opt/boost/lib -L/opt/xerces/lib $LDFLAGS"

MAKEFLAGS="install_prefix=/opt verbose=1"

MAKEFLAGS="-I$pwd/install/include $MAKEFLAGS"

...When I ran the "build.sh", I got the following errors:

$ ./build.sh
make: Entering directory `/cygdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/c
ult'
g++? -I/cygdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/cult/arch/i386/i486/
i586/i686 -I/cygdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/cult/arch/i386/
i486/i586 -I/cygdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/cult/arch/i386/
i486 -I/cygdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/cult/arch/i386 -I/cy
gdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/cult/arch/generic -I/cygdrive/
c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4 -I/cygdrive/c/projects/XML/xsd-3.2.0+
dep/libcult-1.4.4? -I/cygdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/../ins
tall/include???? -o /cygdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/cult/rt
ti/type-info.o -c /cygdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/cult/rtti
/type-info.cxx
In file included from /cygdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/cult/
types.hxx:10,
???????????????? from /cygdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/cult/
containers/vector.hxx:9,
???????????????? from /cygdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/cult/
rtti/type-info.hxx:13,
???????????????? from /cygdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/cult/
rtti/type-info.cxx:6:
/cygdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/cult/types/string.hxx: In m
ember function `Cult::Types::StringTemplate<NarrowerChar, typename Cult::Types::
Bits::NarrowerChar<NarrowerChar>::Type> Cult::Types::StringTemplate<C, NarrowerC
har>::to_narrow() const [with C = Cult::Types::Fundamental::WideChar, NarrowerCh
ar = Cult::Types::Fundamental::Char]':
/cygdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/cult/types/string.hxx:373:
error: `wcstombs' is not a member of `std'
/cygdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/cult/types/string.hxx:385:
error: `wcstombs' is not a member of `std'
make: *** [/cygdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/cult/rtti/type-i
nfo.o] Error 1
make: Leaving directory `/cygdrive/c/projects/XML/xsd-3.2.0+dep/libcult-1.4.4/cu
lt'


... Please help me out.

Thanks,
Steve






      
From boris at codesynthesis.com  Mon Jul  6 01:24:25 2009
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Problem compiling the xsd-3.2.0+dep package
In-Reply-To: <658484.22849.qm@web32203.mail.mud.yahoo.com>
References: <658484.22849.qm@web32203.mail.mud.yahoo.com>
Message-ID: <boris.20090706071333@codesynthesis.com>

Hi Steve,

Steve <ongs_1999@yahoo.com> writes:

> I am trying to compile the xsd-3.2.0+dep package and running into 
> problems. I am compiling it on cygwin which I compiled boost and 
> xerces-c++ successfully. 
>
> [...]
> 
> error: `wcstombs' is not a member of `std'

I don't think libstdc++ (Standard C++ library implementation that
comes with g++) on Cygwin includes support for wide characters,
similar to MinGW. One way to resolve this is to build and use
STLPort instead. This is covered in the build instructions for
Windows:

http://www.codesynthesis.com/projects/xsd/extras/build-windows.xhtml

I also remember your previous attempt where you used STLPort and
reportedly succeeded:

http://www.codesynthesis.com/pipermail/xsd-users/2009-January/002137.html

You should be able to use STLPort with the xsd+dep way of building
by simply adding the necessary options to the corresponding *FLAGS
variables.

Boris


From boris at codesynthesis.com  Wed Jul  8 15:16:45 2009
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Problem compiling the xsd-3.2.0+dep package
In-Reply-To: <949331.66561.qm@web32208.mail.mud.yahoo.com>
References: <949331.66561.qm@web32208.mail.mud.yahoo.com>
Message-ID: <boris.20090708210551@codesynthesis.com>

Hi Steve,

Steve <ongs_1999@yahoo.com> writes:

> Sorry for reply directly to you, but I don't know how to reply to 
> the email chain in the mailing list.

Simply press "reply all" in your email client. If you don't have
this option, send all emails just to xsd-users@codesynthesis.com
and I will get them via the mailing list. Support with private
communications is only available on the commercial basis.
 

> I am now trying to compile the xsd-3.2.0+dep on the Linux machine (I
> would have to come back to the cygwin later)...
> 
> I've added the "boost_1_39_0" and "xerces-c-3.0.1" (i.e., unzip/untar)
> under "xsd-3.2.0+dep" and compiled them...
> 
> My question at this point is:
> - How can change the 'build.sh' so that it knows about these directories
> in order for it to compile the others?

You don't need to change build.sh. As described in the accompanying
README file, you can put the corresponding headers and libraries in
xsd-3.2.0+dep/install/include and xsd-3.2.0+dep/install/lib, 
respectively. You can either install the packages that you compiled
into xsd-3.2.0+dep/install or you can manually add symlinks to the
corresponding files and directories.

Boris


From ongs_1999 at yahoo.com  Thu Jul  9 16:58:01 2009
From: ongs_1999 at yahoo.com (Steve)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Re: xsd-users Digest, Vol 49, Issue 3
Message-ID: <441377.10146.qm@web32203.mail.mud.yahoo.com>

Hi Boris,

I was able to compiled successfully on the Linux machine. Thanks.

I wonder if I just replace the current xsd load with the STLPort, would it be sufficient to compile on cygwin using the 'build.sh' script? Please let me know.

Furthermore, during the process, I did develop some scripts that automate Boost library and the dependencies build as well as the xerces-c++ library build. If you like, I can share them with you.

Thanks very much for all your helps.

Regards,
Steven



--- On Thu, 7/9/09, xsd-users-request@codesynthesis.com <xsd-users-request@codesynthesis.com> wrote:

From: xsd-users-request@codesynthesis.com <xsd-users-request@codesynthesis.com>
Subject: xsd-users Digest, Vol 49, Issue 3
To: xsd-users@codesynthesis.com
Date: Thursday, July 9, 2009, 11:00 AM

Send xsd-users mailing list submissions to
??? xsd-users@codesynthesis.com

To subscribe or unsubscribe via the World Wide Web, visit
??? http://codesynthesis.com/mailman/listinfo/xsd-users
or, via email, send a message with subject or body 'help' to
??? xsd-users-request@codesynthesis.com

You can reach the person managing the list at
??? xsd-users-owner@codesynthesis.com

When replying, please edit your Subject line so it is more specific
than "Re: Contents of xsd-users digest..."


Today's Topics:

???1. Re: Problem compiling the xsd-3.2.0+dep package (Boris Kolpackov)


----------------------------------------------------------------------

Message: 1
Date: Wed, 8 Jul 2009 21:16:45 +0200
From: Boris Kolpackov <boris@codesynthesis.com>
Subject: Re: [xsd-users] Problem compiling the xsd-3.2.0+dep package
To: Steve <ongs_1999@yahoo.com>
Cc: xsd-users@codesynthesis.com
Message-ID: <boris.20090708210551@codesynthesis.com>
Content-Type: text/plain; charset=us-ascii

Hi Steve,

Steve <ongs_1999@yahoo.com> writes:

> Sorry for reply directly to you, but I don't know how to reply to 
> the email chain in the mailing list.

Simply press "reply all" in your email client. If you don't have
this option, send all emails just to xsd-users@codesynthesis.com
and I will get them via the mailing list. Support with private
communications is only available on the commercial basis.
 

> I am now trying to compile the xsd-3.2.0+dep on the Linux machine (I
> would have to come back to the cygwin later)...
> 
> I've added the "boost_1_39_0" and "xerces-c-3.0.1" (i.e., unzip/untar)
> under "xsd-3.2.0+dep" and compiled them...
> 
> My question at this point is:
> - How can change the 'build.sh' so that it knows about these directories
> in order for it to compile the others?

You don't need to change build.sh. As described in the accompanying
README file, you can put the corresponding headers and libraries in
xsd-3.2.0+dep/install/include and xsd-3.2.0+dep/install/lib, 
respectively. You can either install the packages that you compiled
into xsd-3.2.0+dep/install or you can manually add symlinks to the
corresponding files and directories.

Boris



------------------------------

_______________________________________________
xsd-users mailing list
xsd-users@codesynthesis.com
http://codesynthesis.com/mailman/listinfo/xsd-users


End of xsd-users Digest, Vol 49, Issue 3
****************************************



      
From ongs_1999 at yahoo.com  Sun Jul 12 20:12:23 2009
From: ongs_1999 at yahoo.com (Steve)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Problem compiling the xsd-3.2.0+dep package
Message-ID: <323984.18068.qm@web32204.mail.mud.yahoo.com>

Hi Boris,

I'm trying to compile the "xsd-3.2.0+dep" package on cygwin. Although I compiled the "STLport-5.0.1-mingw" and copied the library into $DIR/xsd-3.2.0+dep/install, I still have problem compiling (see below ...)

$ ./build.sh
make: Entering directory `/projects/xml/xsd-3.2.0+dep/libcult-1.4.4/cult'
c++ /projects/xml/xsd-3.2.0+dep/libcult-1.4.4/cult/rtti/type-info.cxx
In file included from /projects/xml/xsd-3.2.0+dep/libcult-1.4.4/cult/types.hxx:1
0,
???????????????? from /projects/xml/xsd-3.2.0+dep/libcult-1.4.4/cult/containers/
vector.hxx:9,
???????????????? from /projects/xml/xsd-3.2.0+dep/libcult-1.4.4/cult/rtti/type-i
nfo.hxx:13,
???????????????? from /projects/xml/xsd-3.2.0+dep/libcult-1.4.4/cult/rtti/type-i
nfo.cxx:6:
/projects/xml/xsd-3.2.0+dep/libcult-1.4.4/cult/types/string.hxx: In member funct
ion `Cult::Types::StringTemplate<NarrowerChar, typename Cult::Types::Bits::Narro
werChar<NarrowerChar>::Type> Cult::Types::StringTemplate<C, NarrowerChar>::to_na
rrow() const [with C = Cult::Types::Fundamental::WideChar, NarrowerChar = Cult::
Types::Fundamental::Char]':
/projects/xml/xsd-3.2.0+dep/libcult-1.4.4/cult/types/string.hxx:373: error: `wcs
tombs' is not a member of `std'
/projects/xml/xsd-3.2.0+dep/libcult-1.4.4/cult/types/string.hxx:385: error: `wcs
tombs' is not a member of `std'
make: *** [/projects/xml/xsd-3.2.0+dep/libcult-1.4.4/cult/rtti/type-info.o] Erro
r 1
make: Leaving directory `/projects/xml/xsd-3.2.0+dep/libcult-1.4.4/cult'

Thanks in advance for your help.

Regards,
Steven


--- On Wed, 7/8/09, Boris Kolpackov <boris@codesynthesis.com> wrote:

From: Boris Kolpackov <boris@codesynthesis.com>
Subject: Re: [xsd-users] Problem compiling the xsd-3.2.0+dep package
To: "Steve" <ongs_1999@yahoo.com>
Cc: xsd-users@codesynthesis.com
Date: Wednesday, July 8, 2009, 2:16 PM

Hi Steve,

Steve <ongs_1999@yahoo.com> writes:

> Sorry for reply directly to you, but I don't know how to reply to 
> the email chain in the mailing list.

Simply press "reply all" in your email client. If you don't have
this option, send all emails just to xsd-users@codesynthesis.com
and I will get them via the mailing list. Support with private
communications is only available on the commercial basis.
 

> I am now trying to compile the xsd-3.2.0+dep on the Linux machine (I
> would have to come back to the cygwin later)...
> 
> I've added the "boost_1_39_0" and "xerces-c-3.0.1" (i.e., unzip/untar)
> under "xsd-3.2.0+dep" and compiled them...
> 
> My question at this point is:
> - How can change the 'build.sh' so that it knows about these directories
> in order for it to compile the others?

You don't need to change build.sh. As described in the accompanying
README file, you can put the corresponding headers and libraries in
xsd-3.2.0+dep/install/include and xsd-3.2.0+dep/install/lib, 
respectively. You can either install the packages that you compiled
into xsd-3.2.0+dep/install or you can manually add symlinks to the
corresponding files and directories.

Boris



      
From boris at codesynthesis.com  Mon Jul 13 01:49:01 2009
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Problem compiling the xsd-3.2.0+dep package
In-Reply-To: <323984.18068.qm@web32204.mail.mud.yahoo.com>
References: <323984.18068.qm@web32204.mail.mud.yahoo.com>
Message-ID: <boris.20090713074049@codesynthesis.com>

Hi Steve,

Steve <ongs_1999@yahoo.com> writes:
 
> I'm trying to compile the "xsd-3.2.0+dep" package on cygwin. Although
> I compiled the "STLport-5.0.1-mingw" and copied the library into 
> $DIR/xsd-3.2.0+dep/install, I still have problem compiling

You will need to add a number of options to make g++ use STLPort
instead of the default libstdc++. See the build instructions for
Windows for details:

http://codesynthesis.com/projects/xsd/extras/build-windows.xhtml

You will also need to compile both Xerces-C++ and Boost to use
STLPort.

If after doing all this you are still having trouble compiling
xsd-3.2.0+dep, please follow the step-by-step instructions in
the above guide to build the compiler using the normal xsd-3.2.0
package. The xsd-3.2.0+dep package is meant for UNIX-like
platforms and we haven't tested it on Cygwin/MinGW.

Boris


From esodan at gmail.com  Tue Jul 14 22:03:13 2009
From: esodan at gmail.com (Daniel Espinosa)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Code Generation Error
Message-ID: <144160610907141903j4a94b3e8g4f4b7fd07d2e0bb5@mail.gmail.com>

I'm using xsdcxx cxx-tree in order to generate code for an XSD file, version
3.1 for Ubuntu.

I'm using the following code:

xsdcxx cxx-tree --generate-serialization --namespace-map
http://www.iec.ch/61850/2003/SCL=IEC61850::SCL --hxx-suffix .h --cxx-suffix
.cc --generate-comparison SCL_Enums.xsd

The error message is the following:

terminate called after throwing an instance of 'std::ios_base::failure'
  what():  basic_filebuf::_M_convert_to_external conversion error

Even with out any option in the command line the error is the same.

-- 
Trabajar, la mejor arma para tu superaci?n
"de grano en grano, se hace la arena" (R) (en tr?mite, pero para los cuates:
LIBRE)

From boris at codesynthesis.com  Wed Jul 15 01:49:49 2009
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Code Generation Error
In-Reply-To: <144160610907141903j4a94b3e8g4f4b7fd07d2e0bb5@mail.gmail.com>
References: <144160610907141903j4a94b3e8g4f4b7fd07d2e0bb5@mail.gmail.com>
Message-ID: <boris.20090715074045@codesynthesis.com>

Hi Daniel,

Daniel Espinosa <esodan@gmail.com> writes:

> I'm using xsdcxx cxx-tree in order to generate code for an XSD file, version
> 3.1 for Ubuntu.
>
> [...] 
> 
> terminate called after throwing an instance of 'std::ios_base::failure'
>   what():  basic_filebuf::_M_convert_to_external conversion error

I think I remember fixing something with these symptoms. Can you
try the latest version of XSD, which is 3.2.0? There is a package
in the Debian repository.

Boris


From fredrikz at gmail.com  Wed Jul 15 08:23:36 2009
From: fredrikz at gmail.com (Fredrik Zetterman)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Compiling COLLADA 1.5 using XSD v 3.2.0
Message-ID: <op.uw3zdlf4hjnvh2@tatakae.bredbandsbolaget.se>

Hello,

I've been trying to figure out how to compile the COLLADA 1.5 schema using  
C++/Tree, but now I've stumbled on a problem which I cannot solve.

I'm using the following command:
xsd cxx-tree --options-file settings collada_schema_1_5.xsd

With the file 'settings' looking like this:
--generate-polymorphic
--root-element COLLADA
--namespace-map http://www.w3.org/1998/Math/MathML=mathML
--location-map  
http://www.w3.org/Math/XMLSchema/mathml2/mathml2.xsd=mathml2/mathml2.xsd
--location-map http://www.w3.org/2001/xml.xsd=xml/xml.xsd
--generate-forward
--location-regex-trace

This gives me the following output:

location: 'common/math.xsd'
location: 'common/common-attribs.xsd'
location: 'xlink-href.xsd'
location: 'presentation/common-types.xsd'
location: 'presentation/common-attribs.xsd'
location: 'presentation/characters.xsd'
location: 'presentation/tokens.xsd'
location: 'presentation/scripts.xsd'
location: 'presentation/space.xsd'
location: 'presentation/layout.xsd'
location: 'presentation/table.xsd'
location: 'presentation/style.xsd'
location: 'presentation/error.xsd'
location: 'presentation/action.xsd'
location: 'content/common-attrib.xsd'
location: 'content/tokens.xsd'
location: 'content/arith.xsd'
location: 'content/functions.xsd'
location: 'content/logic.xsd'
location: 'content/constructs.xsd'
location: 'content/constants.xsd'
location: 'content/elementary-functions.xsd'
location: 'content/relations.xsd'
location: 'content/semantics.xsd'
location: 'content/sets.xsd'
location: 'content/linear-algebra.xsd'
location: 'content/calculus.xsd'
location: 'content/vector-calculus.xsd'
location: 'content/statistics.xsd'
location: ''
C:\<snip>\collada\mathml2\common: error:  
'C:\<snip>\collada\mathml2\common': unable to open in read mode
C:\<snip>\collada\collada_schema_1_5.xsd:1938:41: warning F001: element  
'unnormalized' is implicitly of anyType
C:\<snip>\collada\collada_schema_1_5.xsd:1938:41: info: did you forget to  
specify 'type' attribute?

It seems (to me) that I somehow get a location which refers to '' and it  
somehow gets mapped to a folder(mathml2\common)...
I cannot see how this is triggered since it seems to be able to parse  
mathml2.xsd completely beforehand (if I only remove --location-map  
http://www.w3.org/2001/xml.xsd=xml/xml.xsd from the file, I get an error  
that it cannot be found. The reference to xml.xsd is _after_ mathml2.xsd,  
so therefore mathml2.xsd should already be parsed at that point...)

Is there a way for me to see where this location reference is triggered?  
(I haven't found an option for it, did I miss it?)

Has anyone tried to compile COLLADA 1.5 and successfully done so? Any  
tips/hints on what I could be doing wrong?

Kind regards,
Fredrik Z


From Tim.Muirhead at tibra.com  Wed Jul 15 11:27:34 2009
From: Tim.Muirhead at tibra.com (Tim Muirhead)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Problem getting code to read data from files including
 using XInclude
Message-ID: <7025D5D72EC4424F9456258BD1E6E52901B7C151D3@lcinf01.tibra.com>

Hi,
I've been trying to get a simple example using xinclude to work and am wondering if someone can assist.  I'm using xsd-3.2.0 and xerces 3.0.1,

I've created an xsd schema that allows a choice between including a section of config directly, or including it from an another file.  The example program (xmlval.cpp) then tries to read and print data.  When I try and use an included file, it currently fails.  It knows that that the file has been included, but perhaps I missing a step (like explicitly parsing the included file?)  I haven't been able to find any examples on the web that show how xinclude works from the xsd to the code stage - so I created 4 simple files to demonstrate the usage of xinclude:

These are
test.xsd
test.xml - the main xml
testRef.xml - the 'included' xml
xmlval.cpp - trys to read and print the data.

(You can uncomment the section in test.xml to run the case when it is specifying it directly and not using xinclude)

Any help would be greatly appreciated.

Best regards
Tim Muirhead


The contents of this email including any attachments are confidential. If you have received this email in error, please advise the sender by return email and delete this email. Any unauthorised use of the contents of the email is prohibited and you must not disseminate, copy or distribute the message or use the information contained in the email or its attachments in any way.

The views or opinions expressed are the author's own and may not reflect the views or opinions of Tibra. Tibra does not guarantee the integrity of any emails or attached files. E-mails may be interfered with, may contain computer viruses or other defects. Under no circumstances do we accept liability for any loss or damage which may result from your receipt of this message or any attachments.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: xsdTest.zip
Type: application/x-zip-compressed
Size: 2472 bytes
Desc: xsdTest.zip
Url : http://codesynthesis.com/pipermail/xsd-users/attachments/20090715/7008c15d/xsdTest.bin
From esodan at gmail.com  Wed Jul 15 15:57:04 2009
From: esodan at gmail.com (Daniel Espinosa)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Code Generation Error
In-Reply-To: <boris.20090715074045@codesynthesis.com>
References: <144160610907141903j4a94b3e8g4f4b7fd07d2e0bb5@mail.gmail.com>
	<boris.20090715074045@codesynthesis.com>
Message-ID: <144160610907151257o441c648ft999af3f07c000237@mail.gmail.com>

Thanks is resolved in 3.2. One issue is that it has conflic with xsdcxx
package from Ubuntu and some with mono; I've uninstalled them.

2009/7/15 Boris Kolpackov <boris@codesynthesis.com>

> Hi Daniel,
>
> Daniel Espinosa <esodan@gmail.com> writes:
>
> > I'm using xsdcxx cxx-tree in order to generate code for an XSD file,
> version
> > 3.1 for Ubuntu.
> >
> > [...]
> >
> > terminate called after throwing an instance of 'std::ios_base::failure'
> >   what():  basic_filebuf::_M_convert_to_external conversion error
>
> I think I remember fixing something with these symptoms. Can you
> try the latest version of XSD, which is 3.2.0? There is a package
> in the Debian repository.
>
> Boris
>



-- 
Trabajar, la mejor arma para tu superaci?n
"de grano en grano, se hace la arena" (R) (en tr?mite, pero para los cuates:
LIBRE)

From Markus.Hummel at gmx.de  Wed Jul 15 16:32:50 2009
From: Markus.Hummel at gmx.de (Markus Hummel)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Visitor pattern automatically generated/ and/or included
Message-ID: <200907152232.50490.Markus.Hummel@gmx.de>

Hello everyone,

I am currently working with an XML schema having about 300 classes. Everything 
works fine so far. I can access all the information I need.

But I am requiring some kind of a "traverser" to get a simplified but predefined 
access to  the different information.
Is it possible on using 
  > XSD cxx-tree --generate-serialization --generated-visitor-pattern
to add the visitor pattern to all the generated classes? 
An according empty default Visitor would also be very helpful.

If the according Visit()s and Accept()s are generated by the cxx-tree it would 
be very simple to implement any traverser on the model.

Regards,

Markus


From boris at codesynthesis.com  Thu Jul 16 03:35:30 2009
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Compiling COLLADA 1.5 using XSD v 3.2.0
In-Reply-To: <op.uw3zdlf4hjnvh2@tatakae.bredbandsbolaget.se>
References: <op.uw3zdlf4hjnvh2@tatakae.bredbandsbolaget.se>
Message-ID: <boris.20090716092749@codesynthesis.com>

Hi Fredrik,

Fredrik Zetterman <fredrikz@gmail.com> writes:

> location: ''
> C:\<snip>\collada\mathml2\common: error:  
> 'C:\<snip>\collada\mathml2\common': unable to open in read mode
>
> It seems (to me) that I somehow get a location which refers to '' and it  
> somehow gets mapped to a folder(mathml2\common)...

This is caused by an empty import declaration in 
mathml2/common/common-attribs.xsd. Some of strangest things you see
in schemas from W3C...

Once I commented this import, the error disappeared. I've added an
item to my TODO list to implement ignoring such imports.


> Has anyone tried to compile COLLADA 1.5 and successfully done so? Any  
> tips/hints on what I could be doing wrong?

After the location error is gone I got an unstable conflict error
which I resolved by adding the following option to your settings
file:

--anonymous-regex %.* .* (.+)/(.+)%$1_$2%

After that the schema compiles ok. Let us know if you run into
any other problems.

Boris


From boris at codesynthesis.com  Thu Jul 16 04:03:56 2009
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Problem getting code to read data from files
	including using XInclude
In-Reply-To: <7025D5D72EC4424F9456258BD1E6E52901B7C151D3@lcinf01.tibra.com>
References: <7025D5D72EC4424F9456258BD1E6E52901B7C151D3@lcinf01.tibra.com>
Message-ID: <boris.20090716094007@codesynthesis.com>

Hi Tim,

Tim Muirhead <Tim.Muirhead@tibra.com> writes:

> I've created an xsd schema that allows a choice between including a 
> section of config directly, or including it from an another file. 
> The example program (xmlval.cpp) then tries to read and print data.
> When I try and use an included file, it currently fails.  It knows
> that that the file has been included, but perhaps I missing a step
> (like explicitly parsing the included file?)

XInclude is handled by the underlying XML parser (Xerces-C++) and
is disabled by default. If you want to enable it, you will need
to set up your own XML-to-DOM and then DOM-to-object model stages.
For an example on how to do this, see, for instance, the multiroot
example in examples/cxx/tree. In your case you will need to add 
something like:

conf->setParameter(XMLUni::fgXercesDoXInclude, true);

in the dom-parse.cxx file. See the XInclude example in Xerces-C++
for more information.

The one big drawback of the XInclude implementation in Xerces-C++
is that you don't have control over the configuration of the
parser that is used to parser the included XML file. For example,
the schema validation setting is set to "auto" which means the
document will be validated if there is a schema available. There
is also no way to specify the schema location other than in the
included XML file.

If you want more control over the way the included file is parsed
then you will need to handle the inclusion yourself. You can do
it on the DOM level. However, it could be easier to do it on the
object model level. For this you will need to customize the type(s)
which need support for inclusion (TestComplexType in your case).
In the parsing constructor of this type you will need to detect
whether there is inclusion or the inline content. If it is the
inline content that you do nothing, just call the base c-tor and
it will handle everything. If there is inclusion then you load
the included fragment into a separate object model and then set
it as if it was the inline content. This way it will be complete
transparent to the users of your object model. 

For an example on how to customize the type and override its
parsing c-tor, see, for instance the wildcard example in the
examples/cxx/tree/custom/ directory. There is also the C++/Tree
Mapping Customization Guide:

http://wiki.codesynthesis.com/Tree/Customization_guide

Boris


From boris at codesynthesis.com  Sat Jul 18 01:56:46 2009
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Visitor pattern automatically generated/ and/or
	included
In-Reply-To: <200907152232.50490.Markus.Hummel@gmx.de>
References: <200907152232.50490.Markus.Hummel@gmx.de>
Message-ID: <boris.20090718074611@codesynthesis.com>

Hi Markus,

Markus Hummel <Markus.Hummel@gmx.de> writes:

> But I am requiring some kind of a "traverser" to get a simplified but 
> predefined access to  the different information.
>
> Is it possible on using 
>   > XSD cxx-tree --generate-serialization --generated-visitor-pattern
> to add the visitor pattern to all the generated classes?
> An according empty default Visitor would also be very helpful.
> 
> If the according Visit()s and Accept()s are generated by the cxx-tree 
> it would be very simple to implement any traverser on the model.

The problem with automatically generating the Visitor mechanism is
that the Visitor class needs to define the visit() function for each
type in the object model. When you have separate compilation of
schemas that are included or imported, this becomes a very tricky
task.

The other part of the Visitor mechanism, the accept() functions,
should be fairly easy to generate.

So it looks like the best bet to get this automated would be to
define the Visitor class manually and then customize the code
generator in XSD to emit the accept() function implementation
for each object model class. If you would like, I can give you
some pointers on how to do the customization.

BTW, is anybody else interested in this functionality? 

Boris


From jnw at xs4all.nl  Sat Jul 18 01:52:10 2009
From: jnw at xs4all.nl (Jeroen N. Witmond [Bahco])
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Visitor pattern automatically generated/ and/or 
	included
Message-ID: <4035.80.101.201.227.1247896330.squirrel@webmail.xs4all.nl>

Boris Kolpackov writes:
> Hi Markus,
>
> Markus Hummel <Markus.Hummel@gmx.de> writes:
>
>> But I am requiring some kind of a "traverser" to get a simplified but
>> predefined access to  the different information.
>>
[snip]
>> If the according Visit()s and Accept()s are generated by the cxx-tree
>> it would be very simple to implement any traverser on the model.
>
[snip]
> BTW, is anybody else interested in this functionality?

Little bit off-topic, but I'm interested in learning more about this
functionality. Where can I find an introduction on this subject?

Jeroen.


From boris at codesynthesis.com  Sat Jul 18 20:42:42 2009
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Visitor pattern automatically generated/ and/or
	included
In-Reply-To: <4035.80.101.201.227.1247896330.squirrel@webmail.xs4all.nl>
References: <4035.80.101.201.227.1247896330.squirrel@webmail.xs4all.nl>
Message-ID: <boris.20090719024110@codesynthesis.com>

Hi Jeroen,

Jeroen N. Witmond [Bahco] <jnw@xs4all.nl> writes:

> Little bit off-topic, but I'm interested in learning more about this
> functionality. Where can I find an introduction on this subject?

I found this description which is not bad:

http://www.swe.uni-linz.ac.at/research/deco/designpatterns/visitor/visitor.defaultroles.c++.html

Boris


From boris at codesynthesis.com  Mon Jul 20 23:19:52 2009
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Visitor pattern automatically generated/ and/or
	included
In-Reply-To: <30304.192.12.88.7.1247974395.squirrel@www.dre.vanderbilt.edu>
References: <200907152232.50490.Markus.Hummel@gmx.de>
	<boris.20090718074611@codesynthesis.com>
	<30304.192.12.88.7.1247974395.squirrel@www.dre.vanderbilt.edu>
Message-ID: <boris.20090721051000@codesynthesis.com>

Hi Sumant,

Sumant Tambe <sutambe@dre.vanderbilt.edu> writes:
 
> > The problem with automatically generating the Visitor mechanism is
> > that the Visitor class needs to define the visit() function for each
> > type in the object model. When you have separate compilation of
> > schemas that are included or imported, this becomes a very tricky
> > task.
>
> Can you please elaborate the difficulty here? What do you mean by separate
> compilation of schemas?

Say you have schema a.xsd which defines type A and schema b.xsd which
defines type B. The resulting visitor class for this object model would
look like this:

struct Visitor
{
  virtual void visit (A&) = 0;
  virtual void visit (B&) = 0;
};

Because when the XSD compiler is running on a.xsd it doesn't know 
anything about b.xsd (and vice versa), it is not clear when and 
how the above interface could be generated.

We could do some tricks like generating a visitor class per schema
(and, perhaps, inherit it from other visitor classes that correspond
to included schemas), but it gets quite hairy, as you can see.

 
> > So it looks like the best bet to get this automated would be to
> > define the Visitor class manually and then customize the code
> > generator in XSD to emit the accept() function implementation
> > for each object model class. If you would like, I can give you
> > some pointers on how to do the customization.
>
> Is such customization possible without recompiling the code generator?

No, you would need to customize the code generator to generate the
accept() functions.


> Anyways, I'm interested in learning more on how to do this.

In a nutshell, you would add the generation of the accept()
declarations in tree-header.cxx. The implementations can be
generated in tree-source.cxx. You can also add an option,
say --generate-visitor <visitor-name>, which would trigger
the generation as well as provide the visitor class name.
You can look at existing options for an example on how to
do this. The header that defines the visitor class can be
included in to the generated code using the --hxx-prologue
option.


> Finally, I' extremely interested in the functionality! 

Are you interested in it from a practical or purely academic
perspective? ;-)

Boris


From christopher.fuhrman at etsmtl.ca  Mon Jul 20 23:04:26 2009
From: christopher.fuhrman at etsmtl.ca (Christopher Fuhrman)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Visitor pattern automatically generated/ and/or 
	included
In-Reply-To: <boris.20090721051000@codesynthesis.com>
References: <200907152232.50490.Markus.Hummel@gmx.de>
	<boris.20090718074611@codesynthesis.com>
	<30304.192.12.88.7.1247974395.squirrel@www.dre.vanderbilt.edu>
	<boris.20090721051000@codesynthesis.com>
Message-ID: <955f8a1b0907202004s2bf5d350x24b9185e2114daf0@mail.gmail.com>

Hello,

On Mon, Jul 20, 2009 at 11:19 PM, Boris Kolpackov
<boris@codesynthesis.com>wrote:

> Say you have schema a.xsd which defines type A and schema b.xsd which
> defines type B. The resulting visitor class for this object model would
> look like this:
>
> struct Visitor
> {
>  virtual void visit (A&) = 0;
>  virtual void visit (B&) = 0;
> };
>
> Because when the XSD compiler is running on a.xsd it doesn't know
> anything about b.xsd (and vice versa), it is not clear when and
> how the above interface could be generated.
>

One of the assumptions about the applicability of the Visitor design pattern
is that the data structure is not likely to change. The third point in the
"Applicability" section of
http://www.oodesign.com/visitor-pattern.html(which is similar to the
original GoF reference for Visitor):

The object structure is not likely to be changed but is very probable to
> have new operations which have to be added. Since the pattern separates the
> visitor (representing operations, algorithms, behaviors) from the object
> structure it's very easy to add new visitors as long as the structure
> remains unchanged.


Automatically generated object structures seem *very* likely to change. I
don't think the traditional Visitor makes sense for that reason.

The original question mentions getting an automatically generated traverser,
which is similar to two variants of Visitor mentioned in GoF: traversing
logic in an Iterator or in the Composite structure. It might make sense in a
code-generation scenario to generate this traversing logic, but without
using a Visitor that knows about all elements. Maybe it would make sense to
hand-code a Visitor in this case once (if?) the data structure is stable?

-- 
Christopher Fuhrman
Professor
Department of Software and IT Engineering
?cole de technologie sup?rieure (ETS)
http://profs.logti.etsmtl.ca/cfuhrman/
+1 514 396 8638

From Tim.Muirhead at tibra.com  Tue Jul 21 14:46:04 2009
From: Tim.Muirhead at tibra.com (Tim Muirhead)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] RE: Problem getting code to read data from files
 including using XInclude
Message-ID: <7025D5D72EC4424F9456258BD1E6E52901B7C15AA8@lcinf01.tibra.com>

Hi,
I have followed Boris's advice and modified the sample program taken from examples/cxx/tree/multiroot

It now compiles but when I run it with the following set to true

conf->setParameter(XMLUni::fgXercesDoXInclude, true);

It will give a runtime error
test.xml:9:53 error: no scheme found in URI

This only occurs when trying to include the file
  <TestComplex>
    <xin:include href="testRef.xml" />
  </TestComplex>

It is definitely picking up the correct file - as if you put in a invalid file name you get an 'unable to include resource 'bogusfile.xml'
Also, the testRef.xml file appears valid.

Any help would be greatly appreciated.

Example files included as xsdTest2.zip

Thanks
Tim Muirhead

From: Tim Muirhead
Sent: Wednesday, 15 July 2009 16:28
To: 'xsd-users@codesynthesis.com'
Subject: Problem getting code to read data from files including using XInclude

Hi,
I've been trying to get a simple example using xinclude to work and am wondering if someone can assist.  I'm using xsd-3.2.0 and xerces 3.0.1,

I've created an xsd schema that allows a choice between including a section of config directly, or including it from an another file.  The example program (xmlval.cpp) then tries to read and print data.  When I try and use an included file, it currently fails.  It knows that that the file has been included, but perhaps I missing a step (like explicitly parsing the included file?)  I haven't been able to find any examples on the web that show how xinclude works from the xsd to the code stage - so I created 4 simple files to demonstrate the usage of xinclude:

These are
test.xsd
test.xml - the main xml
testRef.xml - the 'included' xml
xmlval.cpp - trys to read and print the data.

(You can uncomment the section in test.xml to run the case when it is specifying it directly and not using xinclude)

Any help would be greatly appreciated.

Best regards
Tim Muirhead


The contents of this email including any attachments are confidential. If you have received this email in error, please advise the sender by return email and delete this email. Any unauthorised use of the contents of the email is prohibited and you must not disseminate, copy or distribute the message or use the information contained in the email or its attachments in any way.

The views or opinions expressed are the author's own and may not reflect the views or opinions of Tibra. Tibra does not guarantee the integrity of any emails or attached files. E-mails may be interfered with, may contain computer viruses or other defects. Under no circumstances do we accept liability for any loss or damage which may result from your receipt of this message or any attachments.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: xsdTest2.zip
Type: application/x-zip-compressed
Size: 5136 bytes
Desc: xsdTest2.zip
Url : http://codesynthesis.com/pipermail/xsd-users/attachments/20090721/64a39882/xsdTest2.bin
From bpringle at sympatico.ca  Tue Jul 21 17:35:01 2009
From: bpringle at sympatico.ca (Bill Pringlemeir)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] --enumerator-regex usage.
Message-ID: <873a8pafei.fsf@sympatico.ca>


With the following schema,

[sample.xsd]

  <xs:simpleType name="bar">
    <xs:restriction base="xs:string">
      <xs:enumeration value="foo" />
      <xs:enumeration value="blip" />
    </xs:restriction>
  </xs:simpleType>

  <xs:simpleType name="baz">
    <xs:restriction base="xs:string">
      <xs:enumeration value="foo" />
      <xs:enumeration value="blonk" />
    </xs:restriction>
  </xs:simpleType>

[/sample.xsd]

I turned on --name-regex-trace and the output looks like this for
enums,

...
enumerator name: 'foo'
...

I was hoping for,

enumerator name: 'bar foo'
enumerator name: 'baz foo'

or something like that.  Is it manditory that each simpleType have a
seperate file to apply the transform rules on an enum?  I don't know
if many people use the --enumerator-regex option.  I would suggest a
--qualified-enumerator-regex or updating 'enumerator name:' text to
inlude the simpleType.

fwiw,
Bill Pringlemeir.


From boris at codesynthesis.com  Wed Jul 22 02:29:34 2009
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] --enumerator-regex usage.
In-Reply-To: <873a8pafei.fsf@sympatico.ca>
References: <873a8pafei.fsf@sympatico.ca>
Message-ID: <boris.20090722082137@codesynthesis.com>

Hi Bill,

Bill Pringlemeir <bpringle@sympatico.ca> writes:

>   <xs:simpleType name="bar">
>     <xs:restriction base="xs:string">
>       <xs:enumeration value="foo" />
>       <xs:enumeration value="blip" />
>     </xs:restriction>
>   </xs:simpleType>
> 
>   <xs:simpleType name="baz">
>     <xs:restriction base="xs:string">
>       <xs:enumeration value="foo" />
>       <xs:enumeration value="blonk" />
>     </xs:restriction>
>   </xs:simpleType>
> 
> I turned on --name-regex-trace and the output looks like this for
> enums,
> 
> ...
> enumerator name: 'foo'
> ...
> 
> I was hoping for,
> 
> enumerator name: 'bar foo'
> enumerator name: 'baz foo'

Yes, the --enumerator-regex is applied on local, unqualified names.
It was initially meant to allow changing the naming convention (e.g.,
making all enumerators upper-case) rather than mangling individual
enumerator values, though people seem to be using them for this as 
well.

Probably the best way to address this would be to add an option
like --qualified-regex-names which makes all the --*-regex options
work on qualified names. I've added this to my TODO list.

Boris


From boris at codesynthesis.com  Wed Jul 22 13:58:21 2009
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] RE: Problem getting code to read data from files
	including using XInclude
In-Reply-To: <7025D5D72EC4424F9456258BD1E6E52901B7C15AA8@lcinf01.tibra.com>
References: <7025D5D72EC4424F9456258BD1E6E52901B7C15AA8@lcinf01.tibra.com>
Message-ID: <boris.20090722193849@codesynthesis.com>

Hi Tim,

Tim Muirhead <Tim.Muirhead@tibra.com> writes:

> It will give a runtime error
> test.xml:9:53 error: no scheme found in URI
> 
> This only occurs when trying to include the file
>   <TestComplex>
>     <xin:include href="testRef.xml" />
>   </TestComplex>
> 
> It is definitely picking up the correct file - as if you put in a 
> invalid file name you get an 'unable to include resource 'bogusfile.xml'
> Also, the testRef.xml file appears valid.

I investigated this a bit and it appears this is not about the validity
of the file but rather about the URI scheme (e.g., http://, file://, etc).
For some reason, XInclude implementation in Xerces-C++ does not like it
when you have a relative path as system id. That is, if I try to use
/tmp/test.xml, everything works but if I use just test.xml, I get the
above error. I have filed a bug report about this:

https://issues.apache.org/jira/browse/XERCESC-1879

One way to work around this problem is to use the LocalFileInputSource
class provided by Xerces-C++ instead of fstream. Internally this class
makes all relative paths absolute. I have updated your test to show how
to do this:

http://www.codesynthesis.com/~boris/tmp/xsdTest2.zip

Boris


From boris at codesynthesis.com  Wed Jul 22 14:26:25 2009
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Visitor pattern automatically generated/ and/or
	included
In-Reply-To: <955f8a1b0907202004s2bf5d350x24b9185e2114daf0@mail.gmail.com>
References: <200907152232.50490.Markus.Hummel@gmx.de>
	<boris.20090718074611@codesynthesis.com>
	<30304.192.12.88.7.1247974395.squirrel@www.dre.vanderbilt.edu>
	<boris.20090721051000@codesynthesis.com>
	<955f8a1b0907202004s2bf5d350x24b9185e2114daf0@mail.gmail.com>
Message-ID: <boris.20090722201225@codesynthesis.com>

Hi Christopher,

Christopher Fuhrman <christopher.fuhrman@etsmtl.ca> writes:

> Automatically generated object structures seem *very* likely to change.

Hm, I don't know. If the schema is mature and stable, I don't see a 
reason why the object structure has to change very often.


> The original question mentions getting an automatically generated traverser,
> which is similar to two variants of Visitor mentioned in GoF: traversing
> logic in an Iterator or in the Composite structure. It might make sense in a
> code-generation scenario to generate this traversing logic, but without
> using a Visitor that knows about all elements. 

There are a number of ways to improve the Visitor design so that there
is no central Visitor class. For example, in the XSD compiler itself
we use a traversal technique which has a separate traverser class for
each object model type. These traversers can then be collected at 
runtime and dispatched on a object. The main drawback of such
approaches is the speed since they rely on (custom) RTTI and dynamic
type casting. The big advantage of the original visitor design is that
it costs just two virtual function calls. 


> Maybe it would make sense to hand-code a Visitor in this case once 
> (if?) the data structure is stable?

The problem is with implementing the accept() functions. They need
to be part of the object model and the only two ways to achieve this
without modifying the code generator is to modify the generated code
or customize each type in the object model.

I agree though that providing a hand-written Visitor interface is
probably not a big deal in this case.

Boris


From sutambe at dre.vanderbilt.edu  Sat Jul 18 23:33:15 2009
From: sutambe at dre.vanderbilt.edu (Sumant Tambe)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Visitor pattern automatically generated/ and/or 
	included
In-Reply-To: <boris.20090718074611@codesynthesis.com>
References: <200907152232.50490.Markus.Hummel@gmx.de>
	<boris.20090718074611@codesynthesis.com>
Message-ID: <30304.192.12.88.7.1247974395.squirrel@www.dre.vanderbilt.edu>

Hi Boris,

> The problem with automatically generating the Visitor mechanism is
> that the Visitor class needs to define the visit() function for each
> type in the object model. When you have separate compilation of
> schemas that are included or imported, this becomes a very tricky
> task.
Can you please elaborate the difficulty here? What do you mean by separate
compilation of schemas?

> So it looks like the best bet to get this automated would be to
> define the Visitor class manually and then customize the code
> generator in XSD to emit the accept() function implementation
> for each object model class. If you would like, I can give you
> some pointers on how to do the customization.
Is such customization possible without recompiling the code generator?
Anyways, I'm interested in learning more on how to do this.

> BTW, is anybody else interested in this functionality?
Finally, I' extremely interested in the functionality! In fact, part of my
research work in Vanderbilt University has focussed on improving such
traversals in a type-safe way. You and the xsd-users subscribers might be
interested in a research paper that I recently presented at a
domain-specific languages conference. I blogged about it here.
http://cpptruths.blogspot.com/2009/03/leesa-new-way-of-xml-programming-in-c.html
Interested readers will find a lot of pointers to existing work on
traversal programming in the paper.

I thought this thread would be a prefect place to introduce this paper.
I've heard comments from the researchers. I'm interested in gaining
feedback from C++ developers who deal with schemas and traversals on a
day-to-day basis. Critical and constructive comments are welcome.

Thanks,

Sumant.

----------------------------------------------
Sumant Tambe
Research Assistant
ISIS, Vanderbilt University
Email: sutambe@dre.vanderbilt.edu


From boris at codesynthesis.com  Tue Jul 28 03:08:09 2009
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Re: invalidated iterators
In-Reply-To: <661f825d0907270129y77727938o30b698404c53c176@mail.gmail.com>
References: <661f825d0907270129y77727938o30b698404c53c176@mail.gmail.com>
Message-ID: <boris.20090728085728@codesynthesis.com>

Hi Brian,

In the future please send technical questions like these to the
xsd-users mailing list (which I've CC'ed) instead of to me directly.
For more information, see the Posting Guidelines:

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


B Hart <bhartsb@gmail.com> writes:

> Hi Boris,
> 
> In my XSD code it seems that when I use a parsing function to create an
> object model of my XML doc, and then subsequently change some element value
> and re save the XML doc, if I have a sequence iterator it becomes
> invalidated.  If I change the name of the XML file that I save then the seq
> iterator does not seem to be invalidated.  However, would prefer to not have
> to create a second file.  Any ideas???

The object model is passed as a const reference and is not modified by
serialization functions. So I don't see how serializing the object model
can invalidate any iterators that point to it.

However, the following scenario can lead to the behavior that you are
observing: You obtain an iterator for a sequence somewhere in the 
object model. In another part of your program you add or remove some
elements from this sequence. In this case, the iterator that you
first obtained will be invalid.

If you can't find the source of the problem, feel free to send a test
case that reproduces this and I will take a look.


Boris


[The rest of the email follows for reference.]
 
> Thanks,
> 
> Brian Hart
> 
> Here is a bit of the code:
> 
> 
> LA_Config_T::InputProperties_sequence& ips(la->InputProperties());
> 
> ...
> 
> for(LA_Config_T::InputProperties_iterator LA_iIt(ips.begin()); LA_iIt!=
> ips.end(); LA_iIt++){
> ...
> 
> xml_schema::namespace_infomap map;
> map[""].name = "";
> map[""].schema = "LA_Config.xsd";
> std::ofstream ofs(la->LAConfigXMLFile().c_str());
> LA_Config(ofs, *la, map);
> }  //error happens here when for loop iterator is accessed


From bhartsb at gmail.com  Tue Jul 28 05:20:59 2009
From: bhartsb at gmail.com (B Hart)
Date: Sun Oct 11 15:34:10 2009
Subject: [xsd-users] Re: invalidated iterators
In-Reply-To: <boris.20090728085728@codesynthesis.com>
References: <661f825d0907270129y77727938o30b698404c53c176@mail.gmail.com>
	<boris.20090728085728@codesynthesis.com>
Message-ID: <661f825d0907280220xa3890beudb460ac3501248ba@mail.gmail.com>

Hi Boris,

Yes you were right on, I was forgetting about another thread.

Thanks for your insight.

Brian

On Tue, Jul 28, 2009 at 12:08 AM, Boris Kolpackov
<boris@codesynthesis.com>wrote:

> Hi Brian,
>
> In the future please send technical questions like these to the
> xsd-users mailing list (which I've CC'ed) instead of to me directly.
> For more information, see the Posting Guidelines:
>
> http://www.codesynthesis.com/support/posting-guidelines.xhtml
>
>
> B Hart <bhartsb@gmail.com> writes:
>
> > Hi Boris,
> >
> > In my XSD code it seems that when I use a parsing function to create an
> > object model of my XML doc, and then subsequently change some element
> value
> > and re save the XML doc, if I have a sequence iterator it becomes
> > invalidated.  If I change the name of the XML file that I save then the
> seq
> > iterator does not seem to be invalidated.  However, would prefer to not
> have
> > to create a second file.  Any ideas???
>
> The object model is passed as a const reference and is not modified by
> serialization functions. So I don't see how serializing the object model
> can invalidate any iterators that point to it.
>
> However, the following scenario can lead to the behavior that you are
> observing: You obtain an iterator for a sequence somewhere in the
> object model. In another part of your program you add or remove some
> elements from this sequence. In this case, the iterator that you
> first obtained will be invalid.
>
> If you can't find the source of the problem, feel free to send a test
> case that reproduces this and I will take a look.
>
>
> Boris
>
>
> [The rest of the email follows for reference.]
>
> > Thanks,
> >
> > Brian Hart
> >
> > Here is a bit of the code:
> >
> >
> > LA_Config_T::InputProperties_sequence& ips(la->InputProperties());
> >
> > ...
> >
> > for(LA_Config_T::InputProperties_iterator LA_iIt(ips.begin()); LA_iIt!=
> > ips.end(); LA_iIt++){
> > ...
> >
> > xml_schema::namespace_infomap map;
> > map[""].name = "";
> > map[""].schema = "LA_Config.xsd";
> > std::ofstream ofs(la->LAConfigXMLFile().c_str());
> > LA_Config(ofs, *la, map);
> > }  //error happens here when for loop iterator is accessed
>