Hello List,
I just played with valgrind 1.0.0 a bit. It found a few minor memleaks.
I've fixed one of them. Attached are
- fix (I'm not really happy with it as it fixes the symptom, the real
cause is a not so nice contstruct of allocating memory in one function
and freeing it its calling function - or not)
- the command line I used to run valgrind
- a bzip2ed output file (stderr) of the valgrind run
Ciao
Jörg
--
Joerg Mayer <jmayer@xxxxxxxxx>
I found out that "pro" means "instead of" (as in proconsul). Now I know
what proactive means.
Changelog: <jmayer@xxxxxxxxx>
gtk/colors.c: Fix a memleak found by valgrind
(Note that there are similar bugs in other functions that
use the same construct with path!)
Index: ethereal/gtk/colors.c
===================================================================
RCS file: /cvsroot/ethereal/gtk/colors.c,v
retrieving revision 1.20
diff -u -r1.20 colors.c
--- colors.c 2002/05/27 22:00:37 1.20
+++ colors.c 2002/07/29 00:13:04
@@ -154,7 +154,7 @@
guint16 fg_r, fg_g, fg_b, bg_r, bg_g, bg_b;
GdkColor fg_color, bg_color;
color_filter_t *colorf;
- const gchar *path;
+ gchar *path;
FILE *f;
dfilter_t *temp_dfilter;
@@ -166,8 +166,11 @@
"Could not open filter file\n\"%s\": %s.", path,
strerror(errno));
}
+ g_free((gchar *)path);
return FALSE;
}
+ g_free((gchar *)path);
+ path = NULL;
do {
if (fgets(buf,sizeof buf, f) == NULL)
Attachment:
runvalgrind.sh
Description: Bourne shell script
Attachment:
stderr.bz2
Description: application/bunzip2