Re: Memory leaks

[Marcin Garski <mgarski post pl>]

Pavel Tsekov wrote:
> Maybe its about time stop playing valgrind and other tools you do not
> understand. In case you want to file a real bug report take your time and
> analyse the suspected code properly. Once you are sure that there is a real
> bug involved post your analisys to the general mailing list - mc at gnome
> dot org.

Do you think that the only "real" bug reports that should be filed is 
the one with analisys of the suspected code? If I just spot the bug (not 
the exact source code line) I shouldn't write about it?

See:
http://bugzilla.gnome.org/show_bug.cgi?id=308546
Only report from valgrind and somehow it's a bug that has been fixed.

Also see:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=169549
This was a trivial bug where I could point bad source line, but a bug.

There are also two guys on Red Hat's bugzilla that has reported a lot of 
bugs, that mostly the only text in bug report was a copy&paste warning 
from gcc about probably uninitialized varibles.
And a lot of this warnings where real bug that has been fixed.

So if someone don't have enough time or knowledge to analyse bug he 
shouldn't mention about it?

Back to my bug report :)
Second leak COULD be caused by not freeing "temp" in panel_operate() 
(file.c), but I'm not shure if it's help because I've added g_free 
(temp) and valgrind still reports leak.
--
Marcin Garski