[FFmpeg-devel] [PATCH v10 0/3] avutil/log: Add log flag to control printing of memory addresses

ffmpegagent ffmpegagent at gmail.com
Thu Apr 10 03:38:56 EEST 2025


--and disable by default in fftools. The benefits are:

 * Smaller log file sizes
 * Makes log files better readable
 * Allows comparing and viewing log file diffs without almost every line
   being different due to those addresses


Before
======

[hevc @ 0000018e72a89cc0] nal_unit_type:
[hevc @ 0000018e72a89cc0] Decoding PPS
[hevc @ 0000018e72a89cc0] nal_unit_type: 39(SEI_P.. [hevc @
0000018e72a89cc0] Decoding SEI
[mp4 @ 0000018e72a8e240] All [mp4 @ 0000018e72a8e240] Afte [hevc @
0000018e742f6b40] Decoded frame with POC .. detected 16 logical cores
[Parsed_scale_0 @ 0000018e74382f40] Setting 'w' t.. [Parsed_scale_0 @
0000018e74382f40] Setting 'h' t.. [Parsed_scale_1 @ 0000018e74382440]
Setting 'w' t.. [mjpeg @ 0000018e743210c0] Forcing thread count t.. [mjpeg @
0000018e743210c0] intra_quant_bias = 96


After
=====

[hevc] nal_unit_type: [hevc] Decoding PPS
[hevc] nal_unit_type: 39(SEI_P.. [hevc] Decoding SEI
[mp4] All info found
[mp4] After avformat_find_ [hevc] Decoded frame with POC 2.
[Parsed_scale_0] Setting 'w' t.. [Parsed_scale_0] Setting 'h' t..
[Parsed_scale_1] Setting 'w' t.. [mjpeg] Forcing thread count t.. [mjpeg]
intra_quant_bias = 96


Versions
========


V2
==

 * Added log flag for optionally restoring the previous behavior (as
   requested by Gyan)


V3
==

 * Externalize the prefix formatting with a prefix_format callback


V4
==

 * Implement a custom logging callback function for fftools instead of the
   prefix formatting callback (as suggested by Hendrik Leppkes)


V5
==

 * Remove unused var
 * Add missing include to fix build error on PPC (thanks, Michael)


V6
==

 * No more changes to avutil involved
 * Let fftools have its own management of log level and flags (as figured to
   be most likely what Nicolas George was alluding to)


V7
==

 * Minimal version without "simple id" substitution
 * Defaults for printing mem addresses:
   * fftools: off
   * avutil: on


V8
==

 * Negated flag logic
 * Use singular naming
 * Fix flag doc text (thanks, Andreas!)


V9
==

 * Rename 'memaddress' to 'mem' for CLI (as suggested by Gyan)


V10
===

 * Print 'mem' flag in help outrput (thanks, Gyan)

softworkz (3):
  avutil/log: Add log flag AV_LOG_PRINT_MEMADDRESSES
  fftools: add mem log flag and disable printing addresses by default
  doc/fftools-common-opts: document mem log flag

 doc/APIchanges               | 3 +++
 doc/fftools-common-opts.texi | 2 ++
 fftools/ffmpeg.c             | 2 +-
 fftools/ffplay.c             | 2 +-
 fftools/ffprobe.c            | 2 +-
 fftools/opt_common.c         | 7 +++++++
 libavutil/log.c              | 6 ++++--
 libavutil/log.h              | 5 +++++
 libavutil/version.h          | 2 +-
 9 files changed, 25 insertions(+), 6 deletions(-)


base-commit: 02eda84bf2fcf0db7793872204b0f564a6557232
Published-As: https://github.com/ffstaging/FFmpeg/releases/tag/pr-ffstaging-59%2Fsoftworkz%2Fsubmit_logaddresses-v10
Fetch-It-Via: git fetch https://github.com/ffstaging/FFmpeg pr-ffstaging-59/softworkz/submit_logaddresses-v10
Pull-Request: https://github.com/ffstaging/FFmpeg/pull/59

Range-diff vs v9:

 1:  0ce5bd11d7 = 1:  0ce5bd11d7 avutil/log: Add log flag AV_LOG_PRINT_MEMADDRESSES
 2:  0780dd080c ! 2:  207dedbff6 fftools: add mem log flag and disable printing addresses by default
     @@ fftools/opt_common.c: int opt_loglevel(void *optctx, const char *opt, const char
               } else {
                   break;
               }
     +@@ fftools/opt_common.c: int opt_loglevel(void *optctx, const char *opt, const char *arg)
     +         av_log(NULL, AV_LOG_FATAL, "\"level\"\n");
     +         av_log(NULL, AV_LOG_FATAL, "\"time\"\n");
     +         av_log(NULL, AV_LOG_FATAL, "\"datetime\"\n");
     ++        av_log(NULL, AV_LOG_FATAL, "\"mem\"\n");
     +         return AVERROR(EINVAL);
     +     }
     + 
 3:  bfe2e51860 = 3:  dc334f3eeb doc/fftools-common-opts: document mem log flag

-- 
ffmpeg-codebot


More information about the ffmpeg-devel mailing list