Indexer breaks if there is unknown characters in a filename

[Jamie Bennett]

** IMPORTED FROM GOOGLE CODE **

Reported by xiretal, Jan 13, 2008

File "/home/stijn/entertainer/devel/src/frontend/medialibrary/music.py",
line 273, in __init__
    for row in cursor:
pysqlite2.dbapi2.OperationalError: Could not decode to UTF-8 column
'filename' with text '/home/stijn/Music/Wir sind Helden/Die Reklamation/Die
Reklamation-02-R?ssel an Schwanz.mp3'

The following error is displayed, when there is an error in my filename.

What should be happening:
Entertainer should continue without loading that file.

Comment 1 by joshuascotton, Jan 15, 2008

Can you confirm what type of characters are causing the problems?

Status: New
Comment 2 by morenoviola, Feb 10, 2008

In that case it must have been the german "ü"
I noticed the same thing with "AC/DC". I suppose the slash caused the error there

Comment 3 by joshuascotton, Mar 13, 2008

The indexer breaks if there is a problem adding the file to the database, which only
supports UTF-8.
Fixed in rev 332
The files in question are now dropped from the database so indexing can continue

Status: Fixed
Comment 4 by rockstar.dev, Mar 19, 2008

Having the files dropped from the database doesn't constitute a fix. It only
constitutes a workaround. We need to implement a fix that will still allow these
files to show up in the database, or risk having a bug filed that says "Files with
unknown characters don't show up in Entertainer"

Status: Started
Owner: joshuascotton
Cc: rockstar.dev

[Matt Layman (mblayman)]

Migrating importance from Google Code.

[Paul Hummer (rockstar)]

We've fixed MOST of this bug, but until we have good unicode support in the database, this bug just can't be fixed. There are currently some good workarounds in main branch, however.

[Paul Hummer (rockstar)]

This bug is closely related to bug 238560 although it's a different manifestation of the same problem. I've got the fix for the image indexer right now, with the the other media indexers coming shortly.