RE: [xml] Creating Win32 wrapper function for exported variables

["Igor Zlatkovic" <izlatkovic daenet de>]

Hello there.

>   What I would really like to see is one stable maintainer 
> for Windows.
> It is really difficult for me to judge what is appropriate and what is
> not w.r.t. handling the Makefiles and specificities of this 
> environment.
> When I get conflicting opinions I can hardly do anything better than
> keeping the status quo, it's not good. I would also like to 
> see prepackaged
> libraries and binaries for libxml and libxslt. Ideally the maintainer
> should also use CVS to check things regulary before releases.

Right you are, my Lord. A maintainer is needed. If you still think it
apropriate, and noone objects, I would be honoured to take the task.

  "By my honour, I, Igor of the Zlatkovic, heir to the throne of the
Zlatkovic, proclaim herewith that I shall take pains to maintain the
Win32/MSVC port of libxml2, for at least one year."
  -- 19. July 2001, Igor Zlatkovic

That was the official statement :-) 

I shall use the CVS version on a regular basis and patch the MSVC
project files and the surrounding stuff as I go. Further I shall
provide:
- Project files which can be used to build the tools which come with the
library, such as xmllint.
- Binary vesion of the whole stuff. This is preferred to the source code
by most of the Win32 developers and it saves a lot of support of the
kind 'why does this not compile and what are those warnings'.

Ideally, the source downloaded from xmlsoft.org should always work.
Unfortulately, that shall not always be possible, because Daniel has no
way to test Win32/MSVC before he releases and from time to time
something shall slip through, that we cannot avoid. For that reason, I
shall maintain a web site with the last-known-good release for
Win32/MSVC and perhaps some additional info about the matters at hand.

Oh, and apply all that to libxslt as well.

If this isn't enough, read on for reasons why I haven't done this
earlier.

So far, I was afraid not to be able to spend required time with the
library. In their wisdom, the decision makers in the company I work for
have decided to use xerces in all future projects because it has a C++
interface (preferred by most, if not all colleagues) and is quite
source-similar to what we have in java. I was left with two small
projects that use libxml2 and could receive an order to adapt them to
xerces any time. This meant that I could invest my worktime in libxml2
only as long these projects stayed unchanged. Had they terminated, or
got converted, I would have been able to tend libxml2 only during my
free hours, which I rather spend in a pub, disco, beach, party than in
front of the computer screen.

Things have changed. In the meantime I have a different position here
and can decide the one thing or another myself. The one new colleague or
another tried to serialize the memory-built DOM on the disk using xerces
and came running into my office, gasping for breath, asking where is
libxml2, cursing at xerces. Further, chief folk in here are too busy
with MS .NET to notice their coffee is getting cold, so being
source-similar to java is nothing they argue about anymore. Shortly: I
am keeping my projects and they are keeping libxml2.

Ciao
Igor