leaks

From: David Abrahams (david.abrahams_at_[hidden])
Date: 2002-07-04 09:16:01

I just fixed a major one, which cut the memory use for my test case by more
than 50%, from 65M to 25M!

I'm a little baffled by this one, though it's small:

==30721== 60 bytes in 5 blocks are possibly lost in loss record 2 of 25
==30721== at 0x40043BDB: malloc (vg_clientfuncs.c:100)
==30721== by 0x804E02A: addsettings (rules.c:235)
==30721== by 0x8049EDA: compile_local (compile.c:461)
==30721== by 0x804DDBC: parse_evaluate (parse.c:138)
==30721== by 0x8049F19: compile_local (compile.c:470)
==30721== by 0x804DDBC: parse_evaluate (parse.c:138)
==30721== by 0x8049F19: compile_local (compile.c:470)
==30721==

After all, freesettings() is getting called in compile_local, and
donerules() does clean up the settings_freelist. Can anyone explain this?

-Dave

+---------------------------------------------------------------+
David Abrahams
C++ Booster (http://www.boost.org) O__ ==
Pythonista (http://www.python.org) c/ /'_ ==
resume: http://users.rcn.com/abrahams/resume.html (*) \(*) ==
email: david.abrahams_at_[hidden]
+---------------------------------------------------------------+