[hafqa] [hafqa] [Bug 3779] ke-recv should use locale-dependent encoding for mmc mounting

[From: bugzilla-daemon at maemo.org]

https://bugs.maemo.org/show_bug.cgi?id=3779





------- Comment #4 from anpaza at mail.ru  2008-10-13 01:04 GMT+3 -------
>Hm. I just tested this:
>- /media/mmc1# mkdir ööää
>- /media/mmc1# mkdir ÖÖÄÄ
>mkdir: cannot create directory 'ÖÖÄÄ': File exists
>I.e. this works also for Finnish characters in the device itself.
That's because cp437 contains latin glyphs with all kinds of umlauts, so cp437
works fine for Finnish (I don't know what codepage is the default in DOS for
Finnish, but could be cp437).

>As already discussed, the short names shouldn't be really an issue (nobody uses
>FAT anymore).  However, above doesn't explain why Busybox doesn't complain that
>the file already exists for cyrillic, so I started to wonder what could be the
>reason... Broken Busybox (wouldn't be anything new, but in this case, complain
>to upstream), broken Russian collation tables in C-library, some issue in
>kernel?
It can't be a bug anywhere higher than kernel, because these kinds of errors
are detected by the filesystem itself. When you create a file, the filesystem
checks if such a file already exists - according to the rules of that file
system, e.g. Unix file systems are case-sensitive, DOS filesystems are case
insensitive.

The comparison for duplicate fails because vfat checks first because there's a
bug in vfat itself (I've reported it already on bugzilla.kernel.org), and
second because it's mounted with the wrong codepage option.

cp437 simply does not contain any cyrillic characters. If I remount with cp866,
I cannot create anymore duplicate files.

Remounting with cp866 manually is not a solution, unless there are some hooks
to do that automatically.


-- 
Configure bugmail: https://bugs.maemo.org/userprefs.cgi?tab=email
Replies to this email are NOT read, instead please add comments at
https://bugs.maemo.org/show_bug.cgi?id=3779
------- You are receiving this mail because: -------
You are the QA contact for the bug, or are watching the QA contact.