Bugzilla – Bug 9126
--charset=utf8 does not pick up accented characters
Last modified: 2009-09-08 09:24:16 UTC
+++ This bug was initially created as a clone of Bug #5339 +++ In bug 5339 a fix is made for iTunes songs containing filepaths with accented characters on a UTF8 encoded filesystem, regardless if the filesystem is in composed or decomposed form. This works if the filesystem is UTF8 and the characterset is also UTF8 (in specific cases, which is this bug). This bug addresses the following: if the locale is for example en_GB.UTF-8 ($LC_CTYPE=en_GB.UTF-8) the accented characters are picked up correctly and the scanner picks up files with accented characters correctly. The status line reports the following: SqueezeCenter Version: 7.2 - 22491 @ Sun Aug 10 02:00:35 PDT 2008 - SUSE - EN - utf8 Setting --charset=en_GB.UTF-8 or --charset=utf8 does not have the same behaviour and files with accented characters cannot be found by the scanner, causing the songs to be skipped The first setting yields the following status line: SqueezeCenter Version: 7.2 - 22491 @ Sun Aug 10 02:00:35 PDT 2008 - SUSE - EN - en_GB.UTF-8 Setting --charset=utf8 yields the status line: SqueezeCenter Version: 7.2 - 22491 @ Sun Aug 10 02:00:35 PDT 2008 - SUSE - EN - utf8 which is the same as the $LC_CTYPE setting and would lead you to believe it would deliver the same results and scan correctly and this bug wouldn't exist. Bottom line: setting –-charset=utf8 causes the scanner to miss iTunes songs that have filepaths with accented characters on a UTF-8 filesystem.
Please note there's no such thing as the "--charset=utf8" parameter. There's a charset parameter, and you have to set it to the value which matches your system. You neither have to use it nor is it limited to utf8. What I don't understand about your report is whether scanning does work in one of the permutations you list. Does it?
Created attachment 3794 [details] /etc/sysconfig/squeezecenter on SuSE linux
Is this the configuration which is fine for you?
the --charset is an argument to the startup parameters to SqueezeCenter on linux, see the attached configuration file "/etc/sysconfig/squeezecenter on SuSE linux" Scanning only works correctly in the permutation where the characterset is configured via the environment variable LC_CTYPE. All the --charset settings do not work. It is somewhat important as in some cases you cannot set the LC_CTYPE of the user account used to run squeezecenter.
So the issue here is that the additional startup setting "--charset=utf8", which is set by default, does break scanning on your system? Once it's removed you're fine?
Yes, exactly. The --charset parameter does not work as advertised ;=) t
Michael, is this a bug in the --charset option that can be fixed for 7.2? If not, let me know and I'll undo the change that makes --charset=utf8 the default for the RPM and deb.
> Michael, is this a bug in the --charset option that can be fixed for 7.2? If I'm still trying to understand why it doesn't work. But unlikely to be fixed in 7.2. > not, let me know and I'll undo the change that makes --charset=utf8 the default > for the RPM and deb. Is there a way to make it conditional? If LC_TYPE is already something utf8, then don't add it, otherwise do? I'm not sure whether it breaks more installations than it helps. I'd leave it in for now and see how many complaints we're getting. Dean - objections?
(In reply to comment #8) > Is there a way to make it conditional? If LC_TYPE is already something utf8, > then don't add it, otherwise do? Well, when we first discussed this, we talked about checking LC_TYPE and setting --charset to match it. I never bothered to do this because, AFAIK, all major Linux distros ship with UTF8 as the default. If we really want to leave --charset unset for UTF8, we might as well just remove it from the RPM and Deb. > I'm not sure whether it breaks more installations than it helps. I'd leave it > in for now and see how many complaints we're getting. Dean - objections? OK.
I don't have an opinion on the specific issue, but I do know that I don't want to break existing working installs. Do we _need_ to do anything here for 7.2?
Moving 7.4 bugs to 8.0.