Logfile spam playing faulty files

[mixxxbot]

Reported by: daschuer
Date: 2015-06-30T08:31:36Z
Status: Fix Released
Importance: Undecided
Launchpad Issue: lp1470010

---

This was tested in current master using Ubuntu Trusty:

If you play this flac file: https://github.com/Excito/libaudio-scan-perl/blob/master/t/flac/id3tagged.flac

The log files is flooded with error messages. It does not even stop if you pause the track.

Debug [Engine]: WARNING: READER THREAD RECEIVED INVALID CHUNK READ 
Debug [Engine]: WARNING: READER THREAD RECEIVED INVALID CHUNK READ 
Warning [CachingReaderWorker 1]: Seek error at 450560 in "/home/daniel/workspace/t/libaudio-scan-perl-master/t/flac/id3tagged.flac" 
Warning [CachingReaderWorker 1]: Failed to seek chunk position 35615160 <> 450560 
Warning [CachingReaderWorker 1]: Seek error at 458752 in "/home/daniel/workspace/t/libaudio-scan-perl-master/t/flac/id3tagged.flac" 
Warning [CachingReaderWorker 1]: Failed to seek chunk position 35615160 <> 458752

[mixxxbot]

Commented by: daschuer
Date: 2015-06-30T09:38:18Z

---

The same happens for m4a files:

https://github.com/Excito/libaudio-scan-perl/blob/master/t/mp4/882-sample-rate.m4a

Debug [Engine]: WARNING: READER THREAD RECEIVED INVALID CHUNK READ 
Debug [Engine]: WARNING: READER THREAD RECEIVED INVALID CHUNK READ 
SetPosition: seek failed: errno: 0 (src/mp4file_io.cpp,57)
Warning [CachingReaderWorker 1]: Failed to read MP4 input data for sample block 3622 ( min = 1 , max = 15419 ) 
Warning [CachingReaderWorker 1]: Failed to prefetch sample data while seeking 0 < 3072 
Warning [CachingReaderWorker 1]: Failed to seek chunk position 3707904 <> 3710976 
SetPosition: seek failed: errno: 0 (src/mp4file_io.cpp,57)
Warning [CachingReaderWorker 1]: Failed to read MP4 input data for sample block 3630 ( min = 1 , max = 15419 ) 
Warning [CachingReaderWorker 1]: Failed to prefetch sample data while seeking 0 < 3072 
Warning [CachingReaderWorker 1]: Failed to seek chunk position 3716096 <> 3719168 

[mixxxbot]

Commented by: uklotzde
Date: 2015-06-30T12:41:26Z

---

Might be related to the following PR that fixes some logging issues for faulty MP3 files (reported by synthor in the forum):

#626

I will do some testing, thanks for the example files!

[mixxxbot]

Commented by: uklotzde
Date: 2015-06-30T16:22:29Z

---

"Log file spam" still generated by #626

But what shall we do? Suppress justified error messages? Each seek operation fails. We don't keep a history of preceding seek operations that would allow us to suppress subsequent failures of the same kind.

An idea that I already thought about some time ago: Keep track of the maximum successful seek/read position in CachingReader and cut the length of the audio as soon as the first unrecoverable error occurs. But I suspect that this has a lot of implications that we need to think about carefully.

[mixxxbot]

Commented by: uklotzde
Date: 2015-06-30T17:15:21Z

---

The continuing retries after failed seek operations even when the track has already been stopped is an issue of the CachingReader/CachingReaderWorker duo. It already happens in 1.12 and is not an issue of the new SoundSource API on master.

[mixxxbot]

Commented by: uklotzde
Date: 2015-07-18T12:11:06Z

---

#652

[mixxxbot]

Issue closed with status Fix Released.