Bug 11435 - Scanner crashes when it encounters corrupt JPEG
: Scanner crashes when it encounters corrupt JPEG
Status: RESOLVED DUPLICATE of bug 4699
Product: Logitech Media Server
Classification: Unclassified
Component: Scanner
: 7.4.0
: PC Windows XP
: -- normal (vote)
: ---
Assigned To: Unassigned bug - please assign me!
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2009-03-21 09:16 UTC by Jordan Hirsch
Modified: 2009-03-21 11:59 UTC (History)
0 users

See Also:
Category: ---


Attachments
Scanner.log showing the error in prcoessCoverArtRequest (71.61 KB, application/x-zip-compressed)
2009-03-21 09:16 UTC, Jordan Hirsch
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Jordan Hirsch 2009-03-21 09:16:06 UTC
Created attachment 4956 [details]
Scanner.log showing the error in prcoessCoverArtRequest

My nightly "new & changed music" rescan crashes every night - and I am able to reproduce the crash by running a manual "new & changed music" rescan.  I enabled scanner-related logging; the error appears to occur when it encounters an apparently corrupted JPEG file.  This error shows up towards the end of the log file:

Slim::Web::Graphics::processCoverArtRequest (346) Unable to process JPEG image using Imager: Unsupported marker type 0xa9: Corrupt JPEG data: premature end of data segment at /<C:\PROGRA~1\SQUEEZ~1\server\scanner.exe>Slim/Web/Graphics.pm line 342.

Rather than crash (by "crash," I mean I get a Windows popup box asking me to either debug of terminate the scanner.exe process), the scanner should be able to move on past this error.
Comment 1 KDF 2009-03-21 11:59:21 UTC
This is bug 4699, and bug 6458

The log output is the result of reporting the error.  It should not be causing the scanner to stop (at least with 7.4 code which is that I'm looking at), as this is simply the reported result of an "eval" around the image processing.  Thus, it reports the problem with that pass, and should then move on.  

The error is still reported because the crash is due to a bug/limitation in the external image processing library.  In the end, some corruptions are corruptions so this flag enables the user to be aware that the image header has a problem.  Re-embedding the image can get you past this error in most cases.

*** This bug has been marked as a duplicate of bug 4699 ***