RE: [xml] Release of libxml2-2.4.21
- From: "Peter Jacobi" <pj walter-graphtek com>
- To: "Igor Zlatkovic" <izlatkovic daenet de>
- Cc: <xml gnome org>
- Subject: RE: [xml] Release of libxml2-2.4.21
- Date: Thu, 2 May 2002 13:08:36 +0200
Igor,
let my guess once more on this problem.
It is so stupid, that we should only look for the most
stupid solution, and it will work.
My next try:
Use an additional .c file which handles all
the LIBXML_THREAD_ENABLED etc the
other way around than the real code, to define
dummys for all symbols not defined by the main
code.
Now you can use ctags to extract all symbols and to
produce a .def or .def.src file.
Hmmm?!
Peter
Subject: RE: [xml] Release of libxml2-2.4.21
Date sent: Thu, 2 May 2002 10:44:44 +0200
From: "Igor Zlatkovic" <izlatkovic daenet de>
To: "Peter Jacobi" <pj walter-graphtek com>
Copies to: <xml gnome org>
Hello Peter,
Not really, but wouldn't ctags or cscope be a string point?
[...]
For example, the symbol __docbDefaultSAXHandler exists only if both
LIBXML_THREAD_ENABLED and LIBXML_DOCB_ENABLED are #defined.
If ctags could
output this dependence, I would be a happy man.
Then, wouldn't this fit:
1. use the MSFT command line compiler in preprocess only mode
2. process output of step 1. with ctags
Good idea. Unfortunately, that would give me the '.def' file itself, not
the '.def.src'. Something that generates the .def file must be run as a
part of the build process, because the contents of the .def file is
dependent on chosen libxml features. I cannot use ctags there, because
almost no Windows machine has it.
Ciao
Igor
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]