At 10:24 PM 2/15/01 -0500, Dom Lachowicz wrote:
>>Reorder includes
>>
>> wv/exporter/wvexporter.c
>
>Bad Paul! No donut. All wv changes (no matter how simple) go through me, k?
Eeek. Sorry. How do you want them -- private patches or what?
Paul
PS: The reason I reordered them (in a fashion consistent with the rest of
wv) was to take advantage of the following nifty hack which should never be
checked in:
Index: wv.h
===================================================================
RCS file: /cvsroot/wv/wv.h,v
retrieving revision 1.75
diff -u -r1.75 wv.h
--- wv.h 2001/01/31 05:27:56 1.75
+++ wv.h 2001/02/16 06:33:37
@@ -7,6 +7,7 @@
/* redefs of things that are either in glibc or we have to include them
ourselv
es*/
#if defined(WIN32) && !defined(__MWERKS__)
#define strcasecmp(s1,s2) stricmp(s1,s2)
+#include "ut_types.h" // for leak checking
#else
This one-liner allows the expat library built by Makefile.abi to also use
the VC leak-hunting tools (which track down offending mallocs to the right
line) so that we can isolate leaks during AbiWord runs.
The relevant snippet of code from ut-types.h is here:
/*
The MSVC debug runtime library can track leaks back to the
original allocation via the following black magic.
*/
#if defined(_MSC_VER) && defined(_DEBUG) && defined(_CRTDBG_MAP_ALLOC)
#include <crtdbg.h>
#define UT_DEBUG_NEW new(_NORMAL_BLOCK, __FILE__, __LINE__)
#define new UT_DEBUG_NEW
#endif /* _MSC_VER && _DEBUG && _CRTDBG_MAP_ALLOC */
I doubt that even this snippet belongs in the real wv sources, but to get it
to work, the ordering of includes in the wv sources (wrt wv.h) is important.
