Re: valgrind

[jdonner]

Robert Anderson wrote:
> Has anyone ever looked into why emacs doesn't play well
> with valgrind? It dies immediately with:
> ..

Joe Buehler
> My guess would be undump() -- the build process creates
> the final emacs binary by starting emacs, loading up some
> LISP files, then writing all its memory out to a
> new executable.

I get the same crash when using valgrind on a recent
CVS build; is Joe Buehler's explanation correct?
Is anyone using valgrind or purify on emacs?

Here are the errors again, for reference:
...
==10911== Invalid free() / delete / delete[]
==10911==    at 0x40029CF7: realloc (in 
/usr/lib/valgrind/vgskin_memcheck.so)
==10911==    by 0x8162E97: xrealloc (alloc.c:547)
==10911==    by 0x815B13B: regex_compile (regex.c:2716)
==10911==    by 0x816204B: re_compile_pattern (regex.c:5994)
==10911==    Address 0x8457848 is not stack'd, malloc'd or free'd
==10911==
==10911== Invalid free() / delete / delete[]
==10911==    at 0x400299D9: free (in 
/usr/lib/valgrind/vgskin_memcheck.so)
==10911==    by 0x8162D86: memory_full (alloc.c:478)
==10911==    by 0x8162ED8: xrealloc (alloc.c:552)
==10911==    by 0x815B13B: regex_compile (regex.c:2716)
==10911==    Address 0x82FDBE0 is not stack'd, malloc'd or free'd
emacs: Memory exhausted--use M-x save-some-buffers then exit and 
restart Emacs
==10911==
...

Jeff Donner