theoraenc: Segfault when running under valgrind
Hi, libtheora segfaults during encoding when running under valgrind. Without valgrind everything works as expected. This is with latest libtheora release on Debian/unstable and x86-64
(I was told something similar happens during Vorbis encoding).
==15859== Command: /home/slomo/tmp/libtheora-1.1.1+dfsg.1/examples/.libs/lt-encoder_example -o test.ogv test.y4m
==15859==
File test.y4m is 320x240 30.00 fps 420 video.
Compressing....
/==15859== Invalid write of size 2
==15859== at 0x4049DA0: oc_enc_tokenize_ac (tokenize.c:636)
==15859== by 0x402B1A0: oc_enc_block_transform_quantize (analyze.c:770)
==15859== by 0x402E3D7: oc_enc_mb_transform_quantize_luma (analyze.c:889)
==15859== by 0x40340F3: oc_enc_analyze_intra (analyze.c:1282)
==15859== by 0x403EF74: oc_enc_compress_keyframe (encode.c:1161)
==15859== by 0x403F215: th_encode_ycbcr_in (encode.c:1549)
==15859== by 0x40446B: fetch_and_process_video_packet (encoder_example.c:1143)
==15859== by 0x405895: main (encoder_example.c:1197)
==15859== Address 0x0 is not stack'd, malloc'd or (recently) free'd
==15859==
==15859==
==15859== Process terminating with default action of signal 11 (SIGSEGV)
==15859== Access not within mapped region at address 0x0
==15859== at 0x4049DA0: oc_enc_tokenize_ac (tokenize.c:636)
==15859== by 0x402B1A0: oc_enc_block_transform_quantize (analyze.c:770)
==15859== by 0x402E3D7: oc_enc_mb_transform_quantize_luma (analyze.c:889)
==15859== by 0x40340F3: oc_enc_analyze_intra (analyze.c:1282)
==15859== by 0x403EF74: oc_enc_compress_keyframe (encode.c:1161)
==15859== by 0x403F215: th_encode_ycbcr_in (encode.c:1549)
==15859== by 0x40446B: fetch_and_process_video_packet (encoder_example.c:1143)
==15859== by 0x405895: main (encoder_example.c:1197)
==15859== If you believe this happened as a result of a stack
==15859== overflow in your program's main thread (unlikely but
==15859== possible), you can try to increase the size of the
==15859== main thread stack using the --main-stacksize= flag.
==15859== The main thread stack size used in this run was 8388608.