RE: [xml] Compiling 2.3.7

[Igor Zlatkovic <izlatkovic daenet de>]

Hello there.

> To complete Igor's EXCELENT (thank you very much!) piece of 
> work with the Win32 version [...]

Stop it, you bring tears to my eyes. Even if I use Win2k, I am not
completely insensitive to the flattery :-) Thanx.

> 1) Occurrences of "xmlXPathFreeNodeSetList" that are undefined when
> "LIBXML_XPATH_ENABLED" on debugxml.c

Undefined? If there is a problem, you bet it shall be no more, no later than
when I find the time to take a second glance on 2.3.7 :-)

> 2) Change "struct stat stat_buffer" to "struct _stat stat_buffer" on
> xmlio.c.

This we cannot do. The code is shared between many platforms and struct
_stat is specific to MS C runtime. I know this issues a warning, but it is
harmless because the two structures are binary identical. Let's leave it as
it is for now.

> 3) An undefined on strio.c. NDEBUG is already defined on 
> non-debug builds.

It is defined only under MSCC, and that only if you use some sort of
sorcerer's apprentice... ehem... wizard to create the project file. Other
platforms do not define NDEBUG (if you don't insist). The result in our
world is a multiple macro definition, redefinition being the same as the
definition. This is harmless as well.

> I did those changes but (since I'm a poor non-windows 
> programmer and I do
> not know how to use the patch methods that the *nix guys use) 
> I do not know
> how to supply those patches.

That's easy. First you install Services For UNIX, as you get it with MSDN
and the CD is handy. Very soon you shall figure out that diff from that
package only works with CRLF-style files properly. To compensate, you shall
keep a laptop around, install Linux or Solaris on it and use it for these
things. Not long after that, you shall discover that you can do almost
everything more efficiently with that laptop, despite it's being by far
weaker in hardware than the main machine. When you finally start editing
company's source code with Emacs, the boss is going to knock at your door
and ask if he could get some of your free time, for you obviously have too
much of it. By the time you battle to keep your laptop, your Emacs and your
job, making those patches shall be your smallest problem.

> [...] It's 
> possible to remove
> the HTML parsing support? If "yes" then how?

It should be possible, at least that was my intention :-) Perhaps something
changed and if so, it shall be possible again as soon someone finds the time
to fix it. 

If noone proves to be faster, I shall post the patches sometime next week.

Ciao
Igor