After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 746256 - Too many warnings from tracker-extract in the journal
Too many warnings from tracker-extract in the journal
Status: RESOLVED FIXED
Product: tracker
Classification: Core
Component: Extractor
1.2.x
Other Linux
: High major
: ---
Assigned To: tracker-extractor
tracker-extractor
Depends on:
Blocks:
 
 
Reported: 2015-03-15 17:43 UTC by Flittermice
Modified: 2015-03-20 19:59 UTC
See Also:
GNOME target: ---
GNOME version: 3.13/3.14


Attachments
tracker-extract-gstreamer,msoffice: Improve warning message (1.75 KB, patch)
2015-03-17 20:54 UTC, Carlos Garnacho
committed Details | Review
tracker-extract: Do not pass lesser errors as processing errors (1.39 KB, patch)
2015-03-17 20:54 UTC, Carlos Garnacho
committed Details | Review
tracker-extract-gstreamer: Lower message severity (1.11 KB, patch)
2015-03-17 20:54 UTC, Carlos Garnacho
committed Details | Review

Description Flittermice 2015-03-15 17:43:51 UTC
In https://bbs.archlinux.org/viewtopic.php?pid=1510303#p1510303 garnacho told me to file a bug, so I'll go on:

I regularly find my journal (it's Arch Linux) flooded with tracker-extract warnings. Many of the come from Microsoft file formats (e.g. mdb) or lesser known file formats which tracker-extract is not able to read. Even more lines (regularly over 300) result from failed SQL inserts (short excerpt here: http://pastebin.com/wQ6hkEY1).

Is it possible to reduce the verbosity of tracker-extract? 
I like to use tracker, but the information which files which could not be indexed is unimportant for me.
Comment 1 Martyn Russell 2015-03-17 19:36:08 UTC
(In reply to Flittermice from comment #0)
> In https://bbs.archlinux.org/viewtopic.php?pid=1510303#p1510303 garnacho
> told me to file a bug, so I'll go on:
> 
> I regularly find my journal (it's Arch Linux) flooded with tracker-extract
> warnings. Many of the come from Microsoft file formats (e.g. mdb) or lesser
> known file formats which tracker-extract is not able to read. Even more
> lines (regularly over 300) result from failed SQL inserts (short excerpt
> here: http://pastebin.com/wQ6hkEY1).
> 
> Is it possible to reduce the verbosity of tracker-extract? 
> I like to use tracker, but the information which files which could not be
> indexed is unimportant for me.

Hi, thanks for the bug report. So 2 things:

1. These sorts of bugs are actually really bad. IT means: we know about the metadata, but when we try to insert it, our database OR (usually) extractor assumes a schema which is incorrect.

2. These sorts of errors should _NEVER_ occur.

It's usually a broken extractor module that is using an older ontology that's no longer supported.

About the verbosity, yes it is, but not to the syslog I don't think. This is a warning/error and will always be reported, because it's quite serious.

The bug looks like it is with the EBook extractor.
Comment 2 Carlos Garnacho 2015-03-17 20:54:20 UTC
Thanks for the bug report! The fixes will be available in Tracker 1.3.6

The following fixes have been pushed:
a36cf90 tracker-extract-gstreamer,msoffice: Improve warning message
d8c1fb0 tracker-extract: Do not pass lesser errors as processing errors
1aa7bdd tracker-extract-gstreamer: Lower message severity
Comment 3 Carlos Garnacho 2015-03-17 20:54:25 UTC
Created attachment 299645 [details] [review]
tracker-extract-gstreamer,msoffice: Improve warning message

If we give the uri there, it's possible to know the file that issued the
warning without verbosity>1 logs.
Comment 4 Carlos Garnacho 2015-03-17 20:54:30 UTC
Created attachment 299646 [details] [review]
tracker-extract: Do not pass lesser errors as processing errors

The errors sent back to the decorator are meant to be sort of critical/
disrupting, things that fall upon the normal mode of operation that
tracker-extract should be prepared for (eg. files we don't have
extractors for) shouldn't be issued as criticals.
Comment 5 Carlos Garnacho 2015-03-17 20:54:34 UTC
Created attachment 299647 [details] [review]
tracker-extract-gstreamer: Lower message severity

A missing plugin can hardly be warning worthy, we can't warn on user
choices.
Comment 6 Carlos Garnacho 2015-03-18 13:47:09 UTC
FYI, I filed bug #746401 for the epub problem.
Comment 7 Martyn Russell 2015-03-20 19:58:06 UTC
Review of attachment 299645 [details] [review]:

You know I love debugging :)
Comment 8 Martyn Russell 2015-03-20 19:58:52 UTC
Review of attachment 299646 [details] [review]:

Nice!
Comment 9 Martyn Russell 2015-03-20 19:59:21 UTC
Review of attachment 299647 [details] [review]:

Probably a good idea.