[Libexif-devel] Bugfix for exif/exif/main.c

[Sebastian W. <se...@se...>]

Hi,

exif kept on dumping core on me, so I decided to go searching for the
culprit and here it is (BTW: valgrind really rocks):

Index: exif/main.c
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
RCS file: /cvsroot/libexif/exif/exif/main.c,v
retrieving revision 1.37
diff -u -r1.37 main.c
--- exif/main.c 30 Sep 2003 22:43:00 -0000      1.37
+++ exif/main.c 28 Jan 2004 20:35:27 -0000
@@ -109,7 +109,7 @@
         */
         if (e->format =3D=3D EXIF_FORMAT_ASCII) {
                if (e->data) free (e->data);
-               e->components =3D strlen (e->data) + 1;
+               e->components =3D strlen (set_value) + 1;
                e->size =3D strlen (set_value) + 1;
                e->data =3D malloc (sizeof (char) * e->size);
                 if (!e->data) {

What a trivial bug, but so hard to find.

And before I forget: Thanks for libexif.

Ah, and another remark: sizeof (char) =3D=3D 1. Always. Thats guaranteed by
the standard. That libexif (and most probably 90% of the existing
software which reads/writes binary data from/to files etc.) would stop
working with a char size of more then 8 bit is quite obvious.

Bye,
Sebastian
--=20
Sebastian Wilhelmi                 |            h=E4r ovanf=F6r alla molnen
mailto:se...@se...              |     =E4r himmlen s=E5 f=F6runderligt b=
l=E5
http://seppi.de                    |