Bug #1367
closedkernel libiconv update
Added by polachok over 15 years ago. Updated about 15 years ago.
0%
Description
Hi.
I'm porting msdosfs to use kernel libiconv, but it needs functions not available
in current version. Current msdosfs implementation with conversion tables loaded
with mount_msdos cannot work well with UTF8 system locale.
I'll post msdosfs diffs if this one is imported.
Files
klibiconv.tgz (7.68 KB) klibiconv.tgz | polachok, 05/18/2009 06:09 PM | ||
msdos-libiconv.tar.gz (74.3 KB) msdos-libiconv.tar.gz | polachok, 06/04/2009 08:55 PM | ||
msdosfs-patches.tar.gz (26.5 KB) msdosfs-patches.tar.gz | polachok, 06/05/2009 02:37 PM |
Updated by corecode over 15 years ago
I went over it and fixed some porting errors. In the future, please don't drop
all local DragonFly changes.
Could you please have a look and try it?
http://gitweb.dragonflybsd.org/~corecode/dragonfly.git/shortlog/refs/heads/kiconv
Updated by polachok over 15 years ago
Could you please have a look and try it?
Works for me.
msdosfs with libiconv support attached.
Updated by corecode over 15 years ago
Alexander Polakov (via DragonFly issue tracker) wrote:
msdosfs with libiconv support attached.
Thanks! Could you prepare this as a git repo or as a git diff of a repo
(using git format-patch)? That way it is much easier to review and apply.
Thanks,
simon
Updated by polachok over 15 years ago
Could you prepare this as a git repo or as a git diff of a repo
(using git format-patch)?
Okay. I attached git diffs against a local branch.
Updated by polachok over 15 years ago
Any progress on this? I ported cd9660 too:
http://github.com/polachok/dragonflybsd/commits/cd9660/
Updated by russiane39 over 15 years ago
I've reminded this to corecode, I hope he will respond soon.
Updated by corecode over 15 years ago
could you check whether
<http://gitweb.dragonflybsd.org/~corecode/dragonfly.git/kiconv> works? I had to
rebase and merge some commits. Now that you're so active, maybe you could
prepare your submissions commit-ready? :)
Also I don't like that mount_msdosfs needs to be linked dynamically. Probably
libkiconv will have to be statically linked with the citrus libs, is that correct?
It would be good if you could verify that the code works and maybe clean up the
commit messages to conform to the subsystem: summary standard.
Thanks for the submission!
Updated by polachok over 15 years ago
could you check whether
<http://gitweb.dragonflybsd.org/~corecode/dragonfly.git/kiconv> works? I had to
rebase and merge some commits.
I cherry-picked 2 fixes from my main branch, look here
http://gitweb.dragonflybsd.org/~polachok/dragonfly.git/kiconv2
I haven't built world (it's about 2.5 hours here), just kernel and mount tools
and it looks like working okay.
Also I don't like that mount_msdosfs needs to be linked dynamically. Probably
libkiconv will have to be statically linked with the citrus libs, is that correct?
I wonder whether it can be linked statically, look here
http://gitweb.dragonflybsd.org/~polachok/dragonfly.git/blob/48fa07367833c3907f2e5ae3502ad91901a55fea:/lib/libkiconv/xlat16_iconv.c
Updated by corecode over 15 years ago
did we commit this? Can we commit part of it without linking mount_msdos
dynamically?
Updated by polachok over 15 years ago
did we commit this?
no
Can we commit part of it without linking mount_msdos
dynamically?
We can commit libiconv, it shouldn't break anything, as I can say. Committing
msdos will break mount_msdos -L and -W options. Committing cd9660 shouldn't
break mount_cd9660.
Updated by polachok about 15 years ago
I'm going to commit http://gitweb.dragonflybsd.org/~polachok/dragonfly.git/i18n
today, so if there any objections please tell me NOW.
And we don't need to link anything dynamically anymore ;)