Package: libimlib2 Version: 1.3.0.0debian1 I find that the loader_pnm.c file in the source package has bugs which manifest themselves while reading ascii PNM files. To be specific the data is read using a char buffer and fgets(). When the buffer ends in the middle of a number which is more than 1 character long, the parser splits that number into two.
I have corrected two instances f this mistake in loader_pnm.c.mine I have attached the output of diff imlib2-1.3.0.0debian1/src/modules/loaders/loader_pnm.c imlib2-1.3.0.0debian1/src/modules/loaders/loader_pnm.c.mine This is an important bug as applications like feh use this to load images and PNM format images are used extensively in research. Thanks, Aravind.
244,245c244,261
< while (buf[i] && !isspace(buf[i]))
< buf2[j++] = buf[i++];
---
> while (!isspace(buf[i]))
> {
> if (buf[i])
> {
> buf2[j++] =
> buf[i++];
> }
> else
> {
> if
> (!fgets(buf, 255, f))
> {
> break;
> }
> else
> {
> i = 0;
> }
> }
> }
338,339c354,371
< while (buf[i] && !isspace(buf[i]))
< buf2[j++] = buf[i++];
---
> while (!isspace(buf[i]))
> {
> if (buf[i])
> {
> buf2[j++] =
> buf[i++];
> }
> else
> {
> if
> (!fgets(buf, 255, f))
> {
> break;
> }
> else
> {
> i = 0;
> }
> }
> }
