| Summary: | media-plugins/gst-plugins-ffmpeg contains "array subscript is below|above array bounds" warnings | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | David J Cozatt <djcozatt> |
| Component: | Current packages | Assignee: | GStreamer package maintainers <gstreamer> |
| Status: | RESOLVED DUPLICATE | ||
| Severity: | normal | CC: | gnome |
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: | gst-plugins-ffmpeg build.log | ||
|
Description
David J Cozatt
2010-09-14 01:13:28 UTC
oh yeah this might be relevant ;-( media-plugins/gst-plugins-ffmpeg-0.10.9 - 09/13/2010 haven't tried other versions I'm assuming you are reporting primarily due to the text relocation, so marking as duplicate of an existing bug about that. *** This bug has been marked as a duplicate of bug 294931 *** This was NOt filed due to text relocations it was filed for libavutil/aes.c:133: warning: array subscript is above array bounds │ │libavutil/aes.c:187: warning: array subscript is above array bounds libavutil/aes.c:187: warning: array subscript is above array bounds │ │libavutil/aes.c:187: warning: array subscript is above array bounds │ │libavutil/aes.c:187: warning: array subscript is above array bounds │ │libavutil/aes.c:187: warning: array subscript is above array bounds │ │libavutil/aes.c:187: warning: array subscript is above array bounds I assume since you are closing that these are trivial errors? this is why you need to post *full* build logs as attachments instead of random snippets in your report None of these are actual bugs (it might not be strictly valid C, however no compiler has any issues). However, aes.c in that version has very serious aliasing issues which cause it to not work at all. Latest FFmpeg upstream contains a warning- and issue-free aes.c. (In reply to comment #4) > this is why you need to post *full* build logs as attachments instead of random > snippets in your report > Ummm all of the above errors are in the first set of comments and are from elogv. Full build logs require changing FEATURES to noclean and re-emerging the package but oh well if I must to report...I must. /var/tmp/portage/* can get rather weighty without regular cleaning. ... or set PORT_LOGDIR=/var/log/portage in your make.conf and the logs will always be saved automatically (In reply to comment #5) > None of these are actual bugs (it might not be strictly valid C, however no > compiler has any issues). > However, aes.c in that version has very serious aliasing issues which cause it > to not work at all. > Latest FFmpeg upstream contains a warning- and issue-free aes.c. > good and the other errors? ---this repeated many times then libavcodec/dca.c:1010: warning: array subscript is above array bounds │ │libavcodec/dca.c:1010: warning: array subscript is above array bounds │ │QA Notice: Package has poor programming practices which may compile │ │ fine but exhibit random runtime failures. │ │libavformat/metadata_compat.c:129: warning: the address of 'number' will │ │always evaluate as 'true' │ │libavformat/metadata_compat.c:130: warning: the address of 'number' will │ │always evaluate as 'true' I'll post the full build log as soon as the next emerge completes and I re-run the build an hour or so at most. Created attachment 247358 [details]
gst-plugins-ffmpeg build.log
PORT_LOG set thanks
build attached
|
