Appreciation

[Chris Cheung <ch...@clc.cuhk.edu.hk>]

Hi Jason,

  Also, just want to let you know that I appreciate your voluntary work. 
We are actual using Xerces-Perl in a commercial project, which will be put 
into production in 1 or 2 months. 

Xerces-Perl is good as it enables Perl applications directly access the
Xerces-C ability to perform validation against schema, apart from parsing
and serialization. (It seems that libxml2, which also has a Perl wrapper,
does not do validation so well.) However, the memory leak problem (during
parsing even if validation is turned off, and during serialization) is 
hindering the project to use Xerces in a more large-scale way.

-- 
Best Regards,

Chris Cheung


---------------------------------------------------------------------
To unsubscribe, e-mail: xerces-p-dev-unsubscribe@xml.apache.org
For additional commands, e-mail: xerces-p-dev-help@xml.apache.org

Re: Appreciation

["Jason E. Stewart" <ja...@openinformatics.com>]

"Steven N. Hirsch" <hi...@btv.ibm.com> writes:

> It would also be great to find out why exception handling does not operate 
> reliably under AIX.  It is not a showstopper for daily use (or, at least, 
> has not been to date) but results in chronic regression test failures.
>
> I have taken several time-consuming stabs at isolating the problem, but 
> have run into serious difficulties in getting the AIX debugger to load and 
> trace test execution.  If anyone from IBM compiler tools feels motivated, 
> I think this needs someone a bit more clueful than I working at it.

Ok, I've just sent another note to the Xerces-C group at IBM in
toronto. I'll keep after them until I hear something positive.

Steve, ping me on this every week or so if you don't hear back, and
I'll keep checking with them.

> Otherwise, I'll add my voice to those thanking Jason for his tireless 
> work.  A hearty "Way to Go!".

Thanks, I wish I could say my work is tireless, but mostly I feel lazy
for not putting more time into the project.

Cheers,
jas.

---------------------------------------------------------------------
To unsubscribe, e-mail: xerces-p-dev-unsubscribe@xml.apache.org
For additional commands, e-mail: xerces-p-dev-help@xml.apache.org

Re: Appreciation

["Steven N. Hirsch" <hi...@btv.ibm.com>]

On Thu, 17 Jun 2004, Jason E. Stewart wrote:

> There are a few issues. This is the long term list:
> 
> 1) Get 2.5 building out of the box. Right now there are two methods
>    being wrapped by SWIG that won't compile - I need to add an %ignore
>    directive so that SWIG skips those two methods. This *should* be
>    simple, but I wasn't able to do it in the short time I had
>    available.
> 
> 2) Fix the *huge* memory leaks. XML-Xerces needs to be usable for
>    long-running processes like mod_perl, and there are huge leaks
>    coming from somewhere. If someone with valgrind or purify can
>    locate these I will patch them.
> 
> 3) Fix the horrible icky multiple inheritance hack I added for 2.0 to
>    make the DOMNode hierarchy work properly. I need the help of
>    someone who is more C++ savvy than I am to debug why the SWIG
>    wrapping of the multiply inherited DOMNode methods are breaking
>    unless I do an icky pointer cast from Perl.
> 
> 4) Document the system. The only documents are the Doxygen C++
>    docs. Most Perl programmers shouldn't be forced to learn C++ to
>    fully understand the system.

It would also be great to find out why exception handling does not operate 
reliably under AIX.  It is not a showstopper for daily use (or, at least, 
has not been to date) but results in chronic regression test failures.

I have taken several time-consuming stabs at isolating the problem, but 
have run into serious difficulties in getting the AIX debugger to load and 
trace test execution.  If anyone from IBM compiler tools feels motivated, 
I think this needs someone a bit more clueful than I working at it.

Otherwise, I'll add my voice to those thanking Jason for his tireless 
work.  A hearty "Way to Go!".

Steve


-- 

--------------------------------------------------------------------
Opinions expressed in this message are mine personally, not those of 
my employer.
--------------------------------------------------------------------


---------------------------------------------------------------------
To unsubscribe, e-mail: xerces-p-dev-unsubscribe@xml.apache.org
For additional commands, e-mail: xerces-p-dev-help@xml.apache.org

Re: Appreciation

["Jason E. Stewart" <ja...@openinformatics.com>]

"Lincoln A. Baxter" <la...@lincolnbaxter.com> writes:

> Since we are appreciating... let me add my appreciation.
>
> At work (A major Financial Institution) we are using XML-Xerces (where
> before we had used expat), because we wanted a _validating_ parser, for
> a particular project.  It worked great. It was only a minor annoyance
> that the current version is 'old' (2.3).

Hi Lincoln,

Thanks! It's great to know that it is being used! 

Honestly, it is dificult for me at times to continue with the project
- I feel like I am the only person interested in it - I know this is
*not* true - without help from many people we would not have the
system working on Windows or OS X, for example.

So knowing that people use XML-Xerces and want to see it continued
really helps.

> At home, I created and submitted a Gentoo Linux portage ebuild for
> XML-Xerces, but by the time they got around to looking at it, their
> standard Xerces (had moved on to 2.4), and it would nolonger build
> without a downgrade, so they rejected it with the note I that I could
> reopen the bugzilla ticket, if I had an ebuild that would work against
> 2.4. or 2.5.  I had to hack my system to keep portage from upgrading
> Xerces (and breaking XML-Xerces).  So I would love to see a version
> built against a more recent xerces-c.  I promise that I will upgrade the
> ebuild, and resubmit, if an new version is made available.

In the meantime, I *did* release a version of XML-Xerces for 2.4, and
I put it in the nightly builds section of the download page.

If you search the mailing list archives you should find a few messages
with the URL if you can't find it on your own (my 'net connection is
down at the moment, so I can't find it for you).

> What is left to do?

There are a few issues. This is the long term list:

1) Get 2.5 building out of the box. Right now there are two methods
   being wrapped by SWIG that won't compile - I need to add an %ignore
   directive so that SWIG skips those two methods. This *should* be
   simple, but I wasn't able to do it in the short time I had
   available.

2) Fix the *huge* memory leaks. XML-Xerces needs to be usable for
   long-running processes like mod_perl, and there are huge leaks
   coming from somewhere. If someone with valgrind or purify can
   locate these I will patch them.

3) Fix the horrible icky multiple inheritance hack I added for 2.0 to
   make the DOMNode hierarchy work properly. I need the help of
   someone who is more C++ savvy than I am to debug why the SWIG
   wrapping of the multiply inherited DOMNode methods are breaking
   unless I do an icky pointer cast from Perl.

4) Document the system. The only documents are the Doxygen C++
   docs. Most Perl programmers shouldn't be forced to learn C++ to
   fully understand the system.


For the short term, only 1) needs to be done. For the medium term 1)
and 2) need to be done.

Cheers,
jas.

---------------------------------------------------------------------
To unsubscribe, e-mail: xerces-p-dev-unsubscribe@xml.apache.org
For additional commands, e-mail: xerces-p-dev-help@xml.apache.org

Re: Appreciation

["Lincoln A. Baxter" <la...@lincolnbaxter.com>]

Since we are appreciating... let me add my appreciation.

At work (A major Financial Institution) we are using XML-Xerces (where
before we had used expat), because we wanted a _validating_ parser, for
a particular project.  It worked great. It was only a minor annoyance
that the current version is 'old' (2.3).

At home, I created and submitted a Gentoo Linux portage ebuild for
XML-Xerces, but by the time they got around to looking at it, their
standard Xerces (had moved on to 2.4), and it would nolonger build
without a downgrade, so they rejected it with the note I that I could
reopen the bugzilla ticket, if I had an ebuild that would work against
2.4. or 2.5.  I had to hack my system to keep portage from upgrading
Xerces (and breaking XML-Xerces).  So I would love to see a version
built against a more recent xerces-c.  I promise that I will upgrade the
ebuild, and resubmit, if an new version is made available.

What is left to do?

On Tue, 2004-06-15 at 21:51, Jason E. Stewart wrote:
> Chris Cheung <ch...@clc.cuhk.edu.hk> writes:
> 
> > Also, just want to let you know that I appreciate your voluntary work. 
> > We are actual using Xerces-Perl in a commercial project, which will be put 
> > into production in 1 or 2 months. 
> 
> Hi Chris!
> 
> Thanks for the note of appreciation, it helps emmensely. I started
> working on this project because I needed a good XML parser for Perl,
> and I've continued working on it, mostly alone, for that same reason.
> 
> Having a user community is something that really inspires me, so
> thanks again for the note.
> 
> > Xerces-Perl is good as it enables Perl applications directly access the
> > Xerces-C ability to perform validation against schema, apart from parsing
> > and serialization. (It seems that libxml2, which also has a Perl wrapper,
> > does not do validation so well.) However, the memory leak problem (during
> > parsing even if validation is turned off, and during serialization) is 
> > hindering the project to use Xerces in a more large-scale way.
> 
> I agree. I debugged this as much as I was able and came up with a big
> zero. It is *not* a Xerces-C problem, so it is a XML-Xerces problem -
> but given the code we are executing I am not able to figure out where
> the problem lies.
> 
> I could really use some help here.
> 
> If anyone has some time to run a test program through purify or
> valgrind and identify exactly where the leaks are happening I'll be
> happy to fix it. 
> 
> Cheers,
> jas.
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: xerces-p-dev-unsubscribe@xml.apache.org
> For additional commands, e-mail: xerces-p-dev-help@xml.apache.org
> 


---------------------------------------------------------------------
To unsubscribe, e-mail: xerces-p-dev-unsubscribe@xml.apache.org
For additional commands, e-mail: xerces-p-dev-help@xml.apache.org

Re: Appreciation

["Jason E. Stewart" <ja...@openinformatics.com>]

Chris Cheung <ch...@clc.cuhk.edu.hk> writes:

> Also, just want to let you know that I appreciate your voluntary work. 
> We are actual using Xerces-Perl in a commercial project, which will be put 
> into production in 1 or 2 months. 

Hi Chris!

Thanks for the note of appreciation, it helps emmensely. I started
working on this project because I needed a good XML parser for Perl,
and I've continued working on it, mostly alone, for that same reason.

Having a user community is something that really inspires me, so
thanks again for the note.

> Xerces-Perl is good as it enables Perl applications directly access the
> Xerces-C ability to perform validation against schema, apart from parsing
> and serialization. (It seems that libxml2, which also has a Perl wrapper,
> does not do validation so well.) However, the memory leak problem (during
> parsing even if validation is turned off, and during serialization) is 
> hindering the project to use Xerces in a more large-scale way.

I agree. I debugged this as much as I was able and came up with a big
zero. It is *not* a Xerces-C problem, so it is a XML-Xerces problem -
but given the code we are executing I am not able to figure out where
the problem lies.

I could really use some help here.

If anyone has some time to run a test program through purify or
valgrind and identify exactly where the leaks are happening I'll be
happy to fix it. 

Cheers,
jas.

---------------------------------------------------------------------
To unsubscribe, e-mail: xerces-p-dev-unsubscribe@xml.apache.org
For additional commands, e-mail: xerces-p-dev-help@xml.apache.org