Observe another fix in 5.2.2 and correct the header.
respect existence of distro CFLAGS and CPPFLAGS
Preserve ENV values for CC and PREFIX in Makefile
Removed giflib-logo target
Ok, sure, thanks for your time
Can youy ship me a version of this patch updated to work on the repository heaf?
Merged by hand, thanks.
Merged by hand, thanks.
I prever to have it there in case I ever want to modify the full-sized logo, so I don't have to think about the extra step of downscaling it for web use.
b) should be removed to finish avoiding an install-time dependency on convert for https://sourceforge.net/p/giflib/feature-requests/10/ I was wrong about this. doc/giflib-logo.gif is enough to work around the problem. It still seems like the giflib-logo target is superfluous though? Your call.
Thanks very much. I think the giflib-logo.gif target is no longer needed after that? I've created https://sourceforge.net/p/giflib/code/merge-requests/13/
Thanks for the commit. I think removing the giflib-logo.gif target is also needed, so I've created https://sourceforge.net/p/giflib/code/merge-requests/13/
Removed giflib-logo target
Great, I am going to take look at the weekend.
ImageMagick dependency on Linux
I've committed the derived giflib.a to the repository. Thanks.
Cause "make clean" to remove giflib.7
Avoid having an install-time dependency on convert(1).
Correct document page install.
giflib head installs xml files instead of manpafes
Applied.
I'd like to see it and review it.
I can't accept this patch. it breaks generration of .so files. Can you ship a conditionalized verrsion that detects mingw and does the right thing withour disturbing the GCC 10 build?
Reassert logical equivalence.
cannot compile under mingw
That should have been expanded to "if and only if" and now is.
Should be fixed now.
Correccyt wrong binary translation is docs.
OK, that is a pretty funny error. I can laugh becaause I didn't do it. Am investigating.
Matthew counting_pine@users.sourceforge.net: Hi Eric. Glad to hear that you are feeling better after being ill. (I saw you posted about it, although I can't now find those posts again. I find SourceForge a bit confusing nowadays.) I noticed your fixes in d565f and 865b8. It now says 010, but I think it should be 100, which is the last three digits of 8C (10001100), and is 4 in binary. Ugh. You're right. Please reopen thisssue. If you can't, file another one. <a href="http://www.catb.org/~esr/">Eric...
Just to clarify: if "iff" was wrong there, was that because of a typo or because of a logic mistake? If colors from ColorIn2 did exist before, could they still be copied? If the file size is not bigger than specified, could it save into the file?
Hi Eric. Glad to hear that you are feeling better after being ill. (I saw you posted about it, although I can't now find those posts again. I find SourceForge a bit confusing nowadays.) I noticed your fixes in d565f and 865b8. It now says 010, but I think it should be 100, which is the last three digits of 8C (10001100), and is 4 in binary.
The GifErrorString error is still there... $ make cc -std=gnu99 -fPIC -Wall -Wno-format-truncation -O2 -I/pjiu/include -c -o dgif_lib.o dgif_lib.c cc -std=gnu99 -fPIC -Wall -Wno-format-truncation -O2 -I/pjiu/include -c -o egif_lib.o egif_lib.c cc -std=gnu99 -fPIC -Wall -Wno-format-truncation -O2 -I/pjiu/include -c -o gifalloc.o gifalloc.c cc -std=gnu99 -fPIC -Wall -Wno-format-truncation -O2 -I/pjiu/include -c -o gif_err.o gif_err.c cc -std=gnu99 -fPIC -Wall -Wno-format-truncation -O2 -I/pjiu/include...
The GifErrorString error is still there... $ make cc -std=gnu99 -fPIC -Wall -Wno-format-truncation -O2 -I/pjiu/include -c -o dgif_lib.o dgif_lib.c cc -std=gnu99 -fPIC -Wall -Wno-format-truncation -O2 -I/pjiu/include -c -o egif_lib.o egif_lib.c cc -std=gnu99 -fPIC -Wall -Wno-format-truncation -O2 -I/pjiu/include -c -o gifalloc.o gifalloc.c cc -std=gnu99 -fPIC -Wall -Wno-format-truncation -O2 -I/pjiu/include -c -o gif_err.o gif_err.c cc -std=gnu99 -fPIC -Wall -Wno-format-truncation -O2 -I/pjiu/include...
I can try to refresh this against current git head if the basic idea is OK. Please note that adding symbol versioning also breaks the ABI, so this must not be done if the soname is not bumped anyway.
giflib head installs xml files instead of manpafes
ImageMagick dependency on Linux
Ready to ship 5.2.2
Ready to ship 5.2.1
Re-reflow so the resylt will be stable under clang-format...
Bloc braces for single lines everywhere, with uncrustify.
Fix up cppcheck suppressions after reflow.
1TBS reflow.
Finish up cppcheck suppressions.
Improve cppcheck line
Quiet down cppcheck.
cppcheck cleanup - more const declarations.
cppcheck cleanup - remove unneeded guard.
cpp lint - remove unneeded guard.
Question on DGifSlurp API
Thatr depends omnn tghe last error condition reported in the Error membor of the GifType structure. Usually that will tell you what went wrong. Sorry about the long-delated reply. I don'yt check support requesrs often (most of my projects are on GitLab) and I was sidelined with serious medical issues for abour two years.
I think that patch already exists as https://sourceforge.net/p/giflib/code/merge-requests/8/ tho it may be out of date after the flurry of recent commits.
Fix various trivial cppcheck issues.
Banish a GCC 10 warning.
cppcheck: Apply const declarations where indicated.
Incredible .. ticket opened to years ago and now is promise to move was from sf.net. So hopefully this promise will be materialised in 2026 ..
NEWS update: explain the point release.
Apply SF patch #31: Wrong position of gif version mark
Need help with gif animation implementation
Wrong position of gif version mark
Applied, thanks.
EGifSpew leaks GifFileOut->SColorMap
The fix I aspplied for this turns out to be incorrect, revealing a deeper issue about the design of the library that will take some further work.
Prevent integer overflow when handing gif2rgb -s arument.
Add bounds check againsts too many rgb entries.
Prevent null-pointer dereference on failed image read.
Address #138: Documentation for obsolete utilities still installed.
Fix typo in patch from issue #133 - SOEXTENION needed to be SOEXTENSION
Address SF issue #146: Limit installed man pages to binaries, move giflib to section 7
Register some bug fixes in README.
Fix heap-buffer overflow (CVE-2022-28506)
Add a warning that gif2rgb is unstable when fed malformed GIFs.
Warn about #148: Out of Bound Read in gif2rgb.c:98 DumpScreen2RGB
Fix from #139: ypo in "LZW image data" page ("110_2 = 4_10")
Conditionalize as suggested in #144: build breaks under Windows.
Apply ppatch in #145: Remove manual pages installation...
Fix issue #140: Typo in "LZW image data" page ("LWZ")
Apply patch from #143: Does not compile on macOS anymore
Prevent ineger overflow when handing gif2rgb -s arument.
Add bounds check againsts too many rgb entries.
Prevent null-pointer dereference on failed image read.
Address #138: Documentation for obsolete utilities still installed.
Fix typo in patch from issue #133 - SOEXTENION needed to be SOEXTENSION
Address SF issue #146: Limit installed man pages to binaries, move giflib to section 7
Register some bug fixes in README.
Fix heap-buffer overflow (CVE-2022-28506)
Move project to github or gitlab
I'm aware of tge problem. I have most of my stuff on GitLLab these days. I do plan to movem buut I want to clear the SourceForge issue list first. I've got 11 leeft to handle.
Your Makefile also contains system paths, so that's not much of a reason. All other projects that do include a pkg-config file do so in the main codebase. Distribution packages generally refuse to the pkg-config file because they consider it to be upstream responsibility. My request to add a .pc file to the giflib package on Arch was denied for that reason.
I apologize for the long delay. I had stomach cancer and was sidelined for nearly two years. I'll take a patch to do a soname bump.
detected memory leaks in openbsd_reallocarray giflib/openbsd-reallocarray.c
I can't fix this. Your memory leak is going to be inside the system realloc function, which my code is a trivial wrapper around.
Add pkg-config file
Because this file contains system paths, I think I would prefere to leave this detail to distribution packages.
Fix of spoilt image counter
Applied, thanks.
detected memory leaks in GifMakeMapObject giflib/gifalloc.c
Thanks, this prroblem has been fixed - see issue #165,
EGifSpew leaks GifFileOut->SColorMap