diff options
author | Dave Love <fx@gnu.org> | 2000-10-26 22:08:28 +0000 |
---|---|---|
committer | Dave Love <fx@gnu.org> | 2000-10-26 22:08:28 +0000 |
commit | 96e5116dcb8b2d14c8a2f959ef78cf738a5d11f5 (patch) | |
tree | 4e533127e74fac114e7ead7e1bfeab0595a588fc /src/s | |
parent | 47afdebfc15c99c250115d379e00558727e72159 (diff) | |
download | emacs-96e5116dcb8b2d14c8a2f959ef78cf738a5d11f5.tar.gz |
Don't define SYSTEM_MALLOC so that we can find out
when it's necessary.
Diffstat (limited to 'src/s')
-rw-r--r-- | src/s/sol2-5.h | 8 |
1 files changed, 7 insertions, 1 deletions
diff --git a/src/s/sol2-5.h b/src/s/sol2-5.h index d80c22b37ba..b3b0470cc80 100644 --- a/src/s/sol2-5.h +++ b/src/s/sol2-5.h @@ -19,7 +19,13 @@ /* This is the only known way to avoid some crashes that seem to relate to screwed up malloc data after deleting a frame. */ -#define SYSTEM_MALLOC +/* rms: I think the problems using ralloc had to do with system + libraries that called the system malloc even if we linked in the + GNU malloc. I could not see any way to fix the problem except to + have just one malloc and that had to be the system one. */ +/* This is not always necessary. Turned off at present for testers to + identify any problems with gmalloc more accurately. */ +/* #define SYSTEM_MALLOC */ #if 0 /* A recent patch in unexelf.c should eliminate the need for this. */ /* Don't use the shared libraries for -lXt and -lXaw, |