libxmlplusplus-general — Open discusions about libXML++

Re: [libxml++] Borland C++/C++ Builder

[Christophe de V. <cde...@al...>]

Le Vendredi 10 Juin 2005 05:08, Darko Miletic a =E9crit=A0:
[...]
>
> There is allso need to change code a bit.
>
[...]
>         ( publicId ? std::string((const char*)publicId) : std::string("")=
),
>         ( systemId ? std::string((const char*)systemId) : std::string("")=
),
>         ( content ? std::string((const char*)content) : std::string("")) =
);

Wouldn't std::string() instead of std::string("") be better ?
Anyway this is something we can include in the 1.0 branch.

Regards

Christophe

Re: [libxml++] Borland C++/C++ Builder

[Darko M. <da...@uv...>]

Simon Reye wrote:
> Does anyone have a guide to building libxml++ under Borland compilers?

Guide? It is indeed simple.

Unfortunately it can not be built from IDE because of the cc extension
but with makefile is just ok.

Attached to this message is the makefile for building libxml++ 1.0.4
under borland free compiler or bcb 5/6. Unpack it to the win32_msvc6
folder and from there run the makefile like this:

for debug: make -f libxml++bcb.mak
for release: make -DBMODE=RELEASE -f libxml++bcb.mak

There is allso need to change code a bit.

1. saxparser.cc line 297
this:
        ( name ? std::string((const char*)name) : ""),
change to this:
        ( name ? std::string((const char*)name) : std::string("")),

allso from line 299 to line 301 simmilar change in the code:
from this:
        ( publicId ? std::string((const char*)publicId) : ""),
        ( systemId ? std::string((const char*)systemId) : ""),
        ( content ? std::string((const char*)content) : "") );
to this:
        ( publicId ? std::string((const char*)publicId) : std::string("")),
        ( systemId ? std::string((const char*)systemId) : std::string("")),
        ( content ? std::string((const char*)content) : std::string("")) );

Simmilar change is needed at line 504 and 505 too.

Aparently borland compiler requires that in ?  opeator both values are
of exactly the same type.

After these changes you can build the library. You will also need
headers fro libxml2 and iconv libraries.

Of course all this applies to libxml++ 1.0.4 . More recent versions of
library require other gtkmm libraries which is just pain to compile. For
that reason I use this older version.

(Attachment is not in this message for the list because it is not 
possible to send message with attachment)
Darko

[libxml++] Borland C++/C++ Builder

[Simon R. <sr...@in...>]

Does anyone have a guide to building libxml++ under Borland compilers?

Simon

Re: [libxml++] NodeList problem. (SOLVED)

[Matthew S. <m....@sy...>]

Darko Miletic wrote:

> m....@sy... wrote:
>
>> Even if I call the function like so:
>>
>>     root->get_children().begin()
>> I get the violation. And no, if I reference to the list, it is not 
>> invalid and .begin() should still be callable. It is the deferenceing 
>> it that will call the problem. The reference is the same as making a 
>> nodelist pointer that points to root->get_children() and calling 
>> ->begin() on it. I don't see how that should invalidate anything. 
>
>
> Now that is strange. I tried this micro code and no exception occured.
>
I seem to have fixed (if not figured out the problem. I am using 
STLPort, and I was compiling with the preprocessor _STLP_DEBUG defined. 
I stopped defineing it (though I am still compiling in debug mode) and 
it works. I think this is a problem with either STLPort or Microsofts 
compiler haveing difficullties with templates. Thanks for your help guys!

Re: [libxml++] NodeList problem.

[Darko M. <da...@uv...>]

m....@sy... wrote:

> Even if I call the function like so:
> 
>     root->get_children().begin() 
> 
> I get the violation. And no, if I reference to the list, it is not invalid and .begin() should still be callable. It is the deferenceing it that will call the problem. The reference is the same as making a nodelist pointer that points to root->get_children() and calling ->begin() on it. I don't see how that should invalidate anything. 

Now that is strange. I tried this micro code and no exception occured.

#include <iostream>
#include <string>
#include <libxml++/libxml++.h>

int main(int argc, char* argv[]) {

  try {
     if (argc > 1) {
       xmlpp::DomParser parser(argv[1], false);
       xmlpp::Element* root = parser.get_document()->get_root_node();
       if (NULL != root) {
         const xmlpp::Node::NodeList& nlist = root->get_children();
         xmlpp::Node::NodeList::const_iterator itr = nlist.begin();
       }
       else {
         std::cout << "No root node!" << std::endl;
       }
     }
     else {
       std::cout << "Specify filename please" << std::endl;
     }
   }
   catch (const std::exception& e){
     std::cerr << e.what() << std::endl;
     return -1;
   }
   return 0;
}

And using sample xml as input parameter with this content

<?xml version="1.0"?>
   <!DOCTYPE root>
<root>
   <child>
     <grandchildren>
     </grandchildren>
   </child>
</root>

I tested this with Borland Free Compiler 5.5.1 and with vc 2005 beta 1 
compilers.
libxml++  version 1.0.4

Compilers where using multithreaded dynamic RTL.

Could you please verify that this does or does not work in your case?

Darko

Re: [libxml++] NodeList problem.

[Darko M. <da...@uv...>]

m....@sy... wrote:
> Even if I call the function like so:
> 
>     root->get_children().begin() 
> 
> I get the violation. And no, if I reference to the list, it is not invalid and .begin() should still be callable. It is the deferenceing it that will call the problem. The reference is the same as making a nodelist pointer that points to root->get_children() and calling ->begin() on it. I don't see how that should invalidate anything. 

Now that is strange. I tried this micro code and no exception occured.

#include <iostream>
#include <string>
#include <libxml++/libxml++.h>

int main(int argc, char* argv[]) {

  try {
     if (argc > 1) {
       xmlpp::DomParser parser(argv[1], false);
       xmlpp::Element* root = parser.get_document()->get_root_node();
       if (NULL != root) {
         const xmlpp::Node::NodeList& nlist = root->get_children();
         xmlpp::Node::NodeList::const_iterator itr = nlist.begin();
       }
       else {
         std::cout << "No root node!" << std::endl;
       }
     }
     else {
       std::cout << "Specify filename please" << std::endl;
     }
   }
   catch (const std::exception& e){
     std::cerr << e.what() << std::endl;
     return -1;
   }
   return 0;
}

And using sample xml as input parameter with this content

<?xml version="1.0"?>
   <!DOCTYPE root>
<root>
   <child>
     <grandchildren>
     </grandchildren>
   </child>
</root>

I tested this with Borland Free Compiler 5.5.1 and with vc 2005 beta 1 
compilers.
libxml++  version 1.0.4

Compilers where using multithreaded dynamic RTL.

Could you please verify that this does or does not work in your case?

Darko

Re: [libxml++] NodeList problem.

[Jonathan W. <co...@co...>]

On Thu, Jun 09, 2005 at 01:53:03PM -0300, Darko Miletic wrote:

> Jonathan Wakely wrote:
> 
> >Surely not?!?!
> >
> >If the container is empty calling begin() will return an iterator that
> >compares equal to end().  _dereferencing_ that iterator would be an
> >error, but not simply copying it.
> 
> It will be error if you use (as you did) reference to NodeList. You 
> wrote the code like this:
> 
> const xmlpp::Node::NodeList& m_nodeList = root->get_children();
> 
> that means that you take reference to the temporary created nodelist 
> object. It is invalid after that line is executed an hence the access 

Not true, the standard requires that a temporary object bound to a const
reference is kept in scope as long as the reference.

> violation. Therefore you need to copy it. Instead of const 
> xmlpp::Node::NodeList& use just xmlpp::Node::NodeList that way the list 
> will be copied as it should because the declaration of 
> xmlpp::Node::get_children is like this:
> 
> NodeList get_children(const std::string& name = std::string());
> 
> So no reference is returned here.
> 
> The code should be like this:
> 
> xmlpp::Node::NodeList m_nodeList = root->get_children();
> xmlpp::Node::NodeList::const_iterator itr = m_nodeList.begin();

Not true, the original code should work AFAICT.

jon

--
"An alcoholic is someone you don't like who drinks as much as you do."
	- Dylan Thomas

Re: Re: [libxml++] NodeList problem.

[<m....@sy...>]

> It will be error if you use (as you did) reference to NodeList. You 
> wrote the code like this:
> 
> const xmlpp::Node::NodeList& m_nodeList = root->get_children();
> 

Even if I call the function like so:

    root->get_children().begin() 

I get the violation. And no, if I reference to the list, it is not invalid and .begin() should still be callable. It is the deferenceing it that will call the problem. The reference is the same as making a nodelist pointer that points to root->get_children() and calling ->begin() on it. I don't see how that should invalidate anything. 

Re: Re: [libxml++] NodeList problem.

[<m....@sy...>]

> It will be error if you use (as you did) reference to NodeList. You 
> wrote the code like this:
> 
> const xmlpp::Node::NodeList& m_nodeList = root->get_children();
> 

Even if I call the function like so:

    root->get_children().begin() I get the violation. And no, if I reference to the list, it is not invalid and .begin() should still be callable. It is the deferenceing it that will call the problem. The reference is the same as making a nodelist pointer that points to root->get_children() and calling ->begin() on it. I don't see how that should invalidate anything. 

Re: [libxml++] NodeList problem.

[Darko M. <da...@uv...>]

Jonathan Wakely wrote:

> Surely not?!?!
> 
> If the container is empty calling begin() will return an iterator that
> compares equal to end().  _dereferencing_ that iterator would be an
> error, but not simply copying it.

It will be error if you use (as you did) reference to NodeList. You 
wrote the code like this:

const xmlpp::Node::NodeList& m_nodeList = root->get_children();

that means that you take reference to the temporary created nodelist 
object. It is invalid after that line is executed an hence the access 
violation. Therefore you need to copy it. Instead of const 
xmlpp::Node::NodeList& use just xmlpp::Node::NodeList that way the list 
will be copied as it should because the declaration of 
xmlpp::Node::get_children is like this:

NodeList get_children(const std::string& name = std::string());

So no reference is returned here.

The code should be like this:

xmlpp::Node::NodeList m_nodeList = root->get_children();
xmlpp::Node::NodeList::const_iterator itr = m_nodeList.begin();

Darko

Re: [libxml++] NodeList problem.

[Jonathan W. <co...@co...>]

On Tue, Jun 07, 2005 at 04:14:41PM -0300, Darko Miletic wrote:

> m....@sy... wrote:
> >Hi, I have am trying to use your API with MSVC71. Most things seem to work 
> >okay, but when I have a line of code like the following:
> >
> >    const xmlpp::Node::NodeList& m_nodeList = root->get_children();
> >    xmlpp::Node::NodeList::const_iterator itr = m_nodeList.begin();
> >
> >After running my program through a debugger, I found that it's when 
> >m_nodeList.begin() is called that the problem arises. I can make calls to 
> >m_nodeList.size() and empty() without a problem, it's just the begin one. 
> >The error I get is "The application has requested the runtime to terminate 
> >in an unusual way..."
> >
> >Does anyone know what I am doing wrong? Thanks in advance for any help,
> 
> You do it wrong. get_children return a list container that should 
> contain all children elements. However if there is no children of that 
> node get_children returns empty list. You must check if list contains 
> something. You can do it like this:
> 
>     xmlpp::Node::NodeList nl = node->get_children();
>     if ( !nl.empty() ) {
>       xmlpp::Node::NodeList::iterator it = nl.begin();
>       //other code here
>     }

Surely not?!?!

If the container is empty calling begin() will return an iterator that
compares equal to end().  _dereferencing_ that iterator would be an
error, but not simply copying it.

jon

-- 
"What I tell you three times is true"
	- The Hunting of the Snark

Re: [libxml++] NodeList problem.

[Darko M. <da...@uv...>]

m....@sy... wrote:
> Hi, I have am trying to use your API with MSVC71. Most things seem to work okay, but when I have a line of code like the following:
> 
>     const xmlpp::Node::NodeList& m_nodeList = root->get_children();
>     xmlpp::Node::NodeList::const_iterator itr = m_nodeList.begin();
> 
> After running my program through a debugger, I found that it's when m_nodeList.begin() is called that the problem arises. I can make calls to m_nodeList.size() and empty() without a problem, it's just the begin one. The error I get is "The application has requested the runtime to terminate in an unusual way..."
> 
> Does anyone know what I am doing wrong? Thanks in advance for any help,

You do it wrong. get_children return a list container that should 
contain all children elements. However if there is no children of that 
node get_children returns empty list. You must check if list contains 
something. You can do it like this:

     xmlpp::Node::NodeList nl = node->get_children();
     if ( !nl.empty() ) {
       xmlpp::Node::NodeList::iterator it = nl.begin();
       //other code here
     }

Darko

[libxml++] NodeList problem.

[<m....@sy...>]

Hi, I have am trying to use your API with MSVC71. Most things seem to work okay, but when I have a line of code like the following:

    const xmlpp::Node::NodeList& m_nodeList = root->get_children();
    xmlpp::Node::NodeList::const_iterator itr = m_nodeList.begin();

After running my program through a debugger, I found that it's when m_nodeList.begin() is called that the problem arises. I can make calls to m_nodeList.size() and empty() without a problem, it's just the begin one. The error I get is "The application has requested the runtime to terminate in an unusual way..."

Does anyone know what I am doing wrong? Thanks in advance for any help,

Matt Smart <m....@sy...>

[libxml++] Global initialization code

[Thomas J. <tho...@in...>]

Hi,

libxml++ uses global initialization code like this:

Document::Init::Init()
{
   xmlInitParser(); //Not always necessary, but necessary for thread safety.
   xmlRegisterNodeDefault(on_libxml_construct);
   xmlDeregisterNodeDefault(on_libxml_destruct);
   xmlThrDefRegisterNodeDefault(on_libxml_construct);
   xmlThrDefDeregisterNodeDefault(on_libxml_destruct);
}

Document::Init Document::init_;

We use PHP as apache module and have another module linking libxml++.
The above piece of code totally breaks PHP's XML parser
just by loading the second module.

Changing the code is not easy as it would break existing applications.
One possible solution would be to add a check to every constructor if the 
library was already initialized. On the destruction of the last libxml++ 
class if would have to be uninitialized.

A more clean way would be to have an init()/deinit() function.
Perhaps the use of this mode could be a compile time option?

Please let me know what you think about this.

Cheers,
Thomas

[libxml++] How can I get unresolved predefined entities from ContentNode::get_content()?

[Mario K. <mar...@bo...>]

Hi all,

does one of you know a way to get the predefined entities unresolved from 
ContentNode::get_content()?

The behaviour looks odd to me here since I can use 
Parser::set_substitute_entities() to substitute other entities, but this 
does not affect the predefined ones (actually when I turn it off, which also 
is the default)!?

Thank you.

Mario

Re: [libxml++] problems with xmlpp

[Darko M. <da...@uv...>]

Daniel Holbach wrote:
> Hi Darko,
> 
> Am Montag, den 30.05.2005, 14:23 -0300 schrieb Darko Miletic:
> 

> I used it, before I used RelaxNG in favour of XML Schemas. I used the
> underlying libxml2 code - have a look over here:
> http://cvs.berlios.de/cgi-bin/viewcvs.cgi/*checkout*/protosquared/protosquared/src/ConfigFile.cc?rev=1.3
> 
> And the lines I changed for Relax NG:
> http://cvs.berlios.de/cgi-bin/viewcvs.cgi/protosquared/protosquared/src/ConfigFile.cc.diff?r1=1.3&r2=1.4

Thanks Daniel, but my problem is more complex. I have xml file that 
needs to be validated with 5 schema files.
here is the root element of the file
<manifest xmlns="http://www.imsglobal.org/xsd/imscp_v1p1" 
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" 
xmlns:adlcp="http://www.adlnet.org/xsd/adlcp_v1p3" 
xmlns:imsss="http://www.imsglobal.org/xsd/imsss" 
xmlns:adlseq="http://www.adlnet.org/xsd/adlseq_v1p3" 
xmlns:adlnav="http://www.adlnet.org/xsd/adlnav_v1p3" 
identifier="MANIFEST-6CF68835-9183-7DD5-93E9-46A5952A16A6" 
xsi:schemaLocation="http://www.imsglobal.org/xsd/imscp_v1p1 
imscp_v1p1.xsd http://www.adlnet.org/xsd/adlcp_v1p3 adlcp_v1p3.xsd 
http://www.imsglobal.org/xsd/imsss imsss_v1p0.xsd 
http://www.adlnet.org/xsd/adlseq_v1p3 adlseq_v1p3.xsd 
http://www.adlnet.org/xsd/adlnav_v1p3 adlnav_v1p3.xsd" version="1.3">

If I use default validation of xsi:schemaLocation I get internal error : 
like this

Unimplemented block at ..\xmlschemastypes.c:2156
file:///C%3A/reload/test_scorm/imsmanifest.xml:21: element controlMode: 
Schemas
validity error : Element 'imsss:controlMode', [strict WC]: No matching 
global de
claration available.

I noticed that if file passes validation xpath can be used. In other 
case no.

Of course file is valid, it appears to be limitation of libxml2.

The code I used for automatic validation is this:

http://mail.gnome.org/archives/xml/2005-January/msg00190.html

I use libxml++ 1.0.04 and libxml2 2.6.19

Darko

Re: [libxml++] problems with xmlpp

[Daniel H. <dh...@ma...>]

Hi Darko,

Am Montag, den 30.05.2005, 14:23 -0300 schrieb Darko Miletic:
> If xml file has only specified schema files and not DTD files xpath can=20
> not be used. I there a workaround for this or something else that can be=20
> done?

I used it, before I used RelaxNG in favour of XML Schemas. I used the
underlying libxml2 code - have a look over here:
http://cvs.berlios.de/cgi-bin/viewcvs.cgi/*checkout*/protosquared/protosqua=
red/src/ConfigFile.cc?rev=3D1.3

And the lines I changed for Relax NG:
http://cvs.berlios.de/cgi-bin/viewcvs.cgi/protosquared/protosquared/src/Con=
figFile.cc.diff?r1=3D1.3&r2=3D1.4

Have a nice day,
 Daniel

[libxml++] problems with xmlpp

[Darko M. <da...@uv...>]

Hi,

Recently I started to use libxml++ for more complex xml files and 
stumbled upon problem.
Problem description is:

If xml file has only specified schema files and not DTD files xpath can 
not be used. I there a workaround for this or something else that can be 
done?

Darko

[libxml++] [ libxmlplusplus-Patches-1200735 ] better parsing error handling in domparser

[SourceForge.net <no...@so...>]

Patches item #1200735, was opened at 2005-05-12 15:07
Message generated for change (Tracker Item Submitted) made by Item Submitter
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=312999&aid=1200735&group_id=12999

Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: virgile devaux (vdevaux)
Assigned to: Nobody/Anonymous (nobody)
Summary: better parsing error handling in domparser

Initial Comment:
to obtain a message more precise than "Document not
well formed." if a parsing error occured in DomParser i
slightly modified the parse_context function of the
pbject an dadd a fonction to format an error message
given a xmlErrorPtr.
so there are two files in this patche,
domparser.cc.patch and domparser.h.patch.


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

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=312999&aid=1200735&group_id=12999

[libxml++] derive DomParser Class

[Virgile D. <vd...@je...>]

Hi,
i wanted libxmlpp to return me parse errors as libxml2 wrties it to the
standard error. For example, when there is a parse error like a
redefined attributes, all i can see when i catch the exception is
"Document not well-formed.". It would be interesting to see the same
message as the one libxml2 outputs to stderr.
So first i tried by modifiing the libxmlpp fonction parse_context in
domparser.cc. It works by just always checking errno instead of
returning when the wellFormed flag is false.
But to be more clean, i try to create a new class that inherites from
the DomParser one. As parse_context is a virtual function, all i intend
to do is to modify the error handling. But i've got the following
problem: the constructor "Document(_xmlDoc* doc);" (that is used by
parse_context) of the class Document is private and is only available in
parse_context because of the "friend class DomParser;" in this class.
So i can't find a way to do what i want without modifying libxml++
source code.
Can anyone gives me a hint about this wanted feature?

Re: [libxml++] display html file

[Daniel V. <vei...@re...>]

On Tue, May 10, 2005 at 04:16:37PM +0200, Christophe de VIENNE wrote:
> Le Mardi 10 Mai 2005 16:00, Aaron Walker a =E9crit=A0:
> > The xml part is possible with libxml++.  I doubt it does (x)html but =
I
> > could be mistaken.
>=20
> xhtml should be ok. An html parser wrapping the HtmlParser of libxml wa=
s=20
> proposed as a patch, but it was too buggy to be merged and I did not ha=
ve=20
> time to do it. It should be fairly easy to implement though (if someone=
 is=20
> willing to do it...)

  you should be able to reuse most of the XML side wrapping as an example
as the HTML parser APIs of libxml2 are modelled after the XML ones.

Daniel

--=20
Daniel Veillard      | Red Hat Desktop team http://redhat.com/
vei...@re...  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/

Re: [libxml++] display html file

[Christophe de V. <cde...@al...>]

Le Mardi 10 Mai 2005 16:00, Aaron Walker a =C3=A9crit=C2=A0:
> The xml part is possible with libxml++.  I doubt it does (x)html but I
> could be mistaken.

xhtml should be ok. An html parser wrapping the HtmlParser of libxml was=20
proposed as a patch, but it was too buggy to be merged and I did not have=20
time to do it. It should be fairly easy to implement though (if someone is=
=20
willing to do it...)

Regards,

Christophe

Re: [libxml++] display html file

[Aaron W. <ka...@ge...>]

jua...@ug... wrote:
> hello:
> 	I'm newbie programmer working with this library ang i'd like to
> 	learn about how to display a html file, xml file or xhtml file
> 	on an gtk widget.

Can't help with the html, xhtml, or gtk. xml I can do.  The gtk or gtkmm list
would probably be a more appropriate place for gtk discussion.

> 
> 	I'd like to parse a html/xml file and display the inside text,
> 	or table,...

Those are two totally different problems.  I'd say concentrate on the latter
part first since you'll need the data to display for the former.

> 
> 	Is this posible using libxml++2.6?

The xml part is possible with libxml++.  I doubt it does (x)html but I could be
mistaken.

Cheers
-- 
When the cup is full, carry it level.

Aaron Walker <ka...@ge...>
[ BSD | cron | forensics | shell-tools | commonbox | netmon | vim | web-apps ]

[libxml++] display html file

[<jua...@ug...>]

hello:
	I'm newbie programmer working with this library ang i'd like to
	learn about how to display a html file, xml file or xhtml file
	on an gtk widget.

	I'd like to parse a html/xml file and display the inside text,
	or table,...

	Is this posible using libxml++2.6?

	thank you in advance.

Re: [libxml++] on_characters not being callled

[Christophe de V. <cde...@ne...>]

Le Jeudi 5 Mai 2005 08:40, John Harpour a =E9crit=A0:
>
> Looking at the sax_parser example it appears the on_characters function
> is not being called during the document parsing?
>
> I am probably missing something obvious here.....but understanding why
> would be helpful.

You're not missing anything, it's a know bug :
http://bugzilla.gnome.org/show_bug.cgi?id=3D301712

It's fixed in the cvs, and will be in the next release.

Regards,

Christophe