From omar.medina at web.de  Thu Dec  8 19:16:51 2016
From: omar.medina at web.de (Omar Medina)
Date: Fri Dec  9 10:10:07 2016
Subject: [studxml-users] Suggestion to extend the attribute template per
	compiler definiton
Message-ID: <54b7b25d-db56-aaa0-49cb-9dce94115647@web.de>

Hello dear Studxml-Users,

my suggestion (below) involves a long known use-case, that impacts 
developers, when they are handling the read hexadecimal values from 
XML-data. They hex-string (0x...) needs to be previously processed 
before it will be converted. The same problems occurs with the provided 
attribute template of libstudxml, which can not handle directly hex-string

The following code will thrown an exception, if the read string-value is 
described in the XML-file as a hex-string.
Code
	XML
p.next_expect(xml::parser::start_element,"drm");
p.content(xml::parser::content_type::complex);
do
{
   p.next_expect(xml::parser::start_element);
   p.content(xml::parser::content_type::complex);
   p.attribute<int>("min"),
   p.attribute<int>("max"),
   p.attribute<int>("mean",0),
   p.attribute<int>("factor",1),
   p.attribute<int>("divisor",1),
}
p.next_expect(xml::parser::end_element);
	<drm>
   <range ID="2" min="1" max="0x5"/>
</drm>


My suggestion: Either a compiler-definiton or an Api-Interface makes 
available the following code in the file "value-traits.txx"
Current
	Suggestion
-
	  /* simple string trim function */
    static bool is_not_whitespace(int c)
     {
       if (
          (c == ' ')  || /* whitespace */
          (c == '\t') || /* tab */
          (c == '\n') || /* linefeed */
          (c == '\r') || /* carriage-return */
          (c == '\f') || /* formfeed */
          (c == '\v')    /* vertical-tab */
          ) return false;
       return true;
     }

     // trim from end
     static inline std::string &trimr(std::string &s)
     {
       s.erase(std::find_if(s.rbegin(), s.rend(), 
is_not_whitespace).base(), s.end());
       return s;
     }

     // trim from begin
     static inline std::string &triml(std::string &s)
     {
       s.erase(s.begin(), std::find_if(s.begin(), s.end(), 
is_not_whitespace));
       return s;
     }

     // trim from both ends
     static inline std::string &trim(std::string &s)
     {
       return trimr(triml(s));
     }
   template <typename T>
   T default_value_traits<T>::
   parse (std::string s, const parser& p)
   {
     T r;
     std::istringstream is (s);
     if (!(is >> r && is.eof ()) )
       throw parsing (p, "invalid value '" + s + "'");
     return r;
   }
	  template <typename T>
   T default_value_traits<T>::
   parse (std::string s, const parser& p)
   {
     bool h =false;
     T r;
     std::string ss = s;
     triml(ss);
     if((ss.length() > 2) && ((ss[0]=='0') && ((ss[1]=='x')||(ss[1]=='X'))))
     {
       ss[0]=' ';
       ss[1]=' ';
       h = true;
     }

     std::istringstream is (((h == true)? ss : s));
     if (h == true)
     {
       if (!((is >> std::hex >> r) && is.eof ()) )
       {
         throw parsing (p, "invalid value '" + s + "'");
       }
     }
     else
     {
       if (!((is >> r) && is.eof ()) )
       {
         throw parsing (p, "invalid value '" + s + "'");
       }
     }
     return r;
   }


By the way i want to take the opportunity to say thanks for the libstudxml.

Best regards
Omar

-- -------------------------- Omar Martin Medina Izaguirre 
--------------------------

From boris at codesynthesis.com  Mon Dec 12 11:16:08 2016
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Mon Dec 12 11:16:17 2016
Subject: [studxml-users] Suggestion to extend the attribute template per
	compiler definiton
In-Reply-To: <54b7b25d-db56-aaa0-49cb-9dce94115647@web.de>
References: <54b7b25d-db56-aaa0-49cb-9dce94115647@web.de>
Message-ID: <boris.20161212181140@codesynthesis.com>

Hi Omar,

Omar Medina <omar.medina@web.de> writes:

> my suggestion (below) involves a long known use-case, that impacts
> developers, when they are handling the read hexadecimal values from
> XML-data.

I've been dealing with XML parsing for over 10 years now and it's
the first time I encounter this "long known use-case" ;-).

Besides, the problem with your approach is that not all strings
that start with 0x are necessarily hex-encoded numbers. So we
cannot embed such an assumption inside libstudxml.

The good news is, the library allows you to provide a custom parsing
specialization (via value_traits). So if in your application you can
assume that anything that starts with 0x is a hex-number, then you
can do it.

Boris

From omar.medina at web.de  Tue Dec 13 18:25:26 2016
From: omar.medina at web.de (Omar Medina)
Date: Wed Dec 14 11:55:00 2016
Subject: [studxml-users] Is possible to skip the current xml-element to the
	next
In-Reply-To: <boris.20161212181140@codesynthesis.com>
References: <54b7b25d-db56-aaa0-49cb-9dce94115647@web.de>
	<boris.20161212181140@codesynthesis.com>
Message-ID: <05dcfdc4-7043-2e11-991b-b8f7db917156@web.de>

Hello Mr. Boris Kolpackov,

i am wondering if you could response my question about your library 
libstudxml.

Do I have the possibility with the library libstudxml to skip the 
current element, also to the next element, for the case the the current 
element is not expected or it can not be processed.

Suppose please in the small xml-example below  : The application is only 
waiting for b-elements. The element <m> is valid for the file but not 
for the application. Also, i want to skip the element <m> to the next 
element.
<a>
   <b x="f" ... />
   <b x="f" ... />
....
   <m x="f" .... />
   <b x="f" .../>
....
   <b x="f" ... />
....
</a>

The intention is to continue with the reading of he XML-data and at the 
end of the reading all not used elements will be indicated.

Thank in advance for your replay.

best regard

Omar

-- 

--------------------------
Omar Medina
--------------------------

From boris at codesynthesis.com  Thu Dec 15 05:40:53 2016
From: boris at codesynthesis.com (Boris Kolpackov)
Date: Thu Dec 15 05:41:02 2016
Subject: [studxml-users] Re: Is possible to skip the current xml-element to
	the next
In-Reply-To: <05dcfdc4-7043-2e11-991b-b8f7db917156@web.de>
References: <54b7b25d-db56-aaa0-49cb-9dce94115647@web.de>
	<boris.20161212181140@codesynthesis.com>
	<05dcfdc4-7043-2e11-991b-b8f7db917156@web.de>
Message-ID: <boris.20161215123757@codesynthesis.com>

Hi Omar,

Omar Medina <omar.medina@web.de> writes:

> Do I have the possibility with the library libstudxml to skip the current
> element, also to the next element, for the case the the current element is
> not expected or it can not be processed.

There is nothing automatic so you will have to handle skipping of unexpected
elements in the application. Which should not be hard: just keep reading
and ignoring elements while keeping track of the depth. Once you are back
at depth 0, you resume processing. Or you could also do this with recursion.

Boris