Sun 19 Aug 2007 02:40:58 PM UTC, comment #20:
Apparently, this problem may not occur when all Chinese file names. That did not happen again for my last 2 chinese file downloads. You should ignore my previous comment. If this problem happens again, I will try to give you more information and a reproduction scenario.
Sorry for the inconvenience.
|
Sun 19 Aug 2007 10:43:58 AM UTC, comment #19:
Hello,
I have a quick question. What release of mldonley introduces this fix?
I supposed that was version 2.8.2 but I'm running this version on Ubuntu Festy and I have a similar problem with Chinese file names. I checked the notes of all the releases starting from version 2.8.1 but I could not found any reference to this bug (including in version 2.9).
Thank you.
|
Wed 08 Nov 2006 12:29:26 PM UTC, comment #18:
[13:28:04] <Laibsch> spiralvoice: Japanese named file downloaded successfully. Closing bug.
|
Tue 07 Nov 2006 09:30:38 PM UTC, comment #17:
Please test current CVS
cvs -d:pserver:anonymous@cvs.sv.gnu.org:/sources/mldonkey co -P mldonkey
it should work for you.
If not it should be fixed before 2.8.2 comes out.
|
Sun 22 Oct 2006 07:28:45 PM UTC, comment #16:
OK no matter what i Thanks your help
I hope 2.8.2 fix up the problem somehow.
|
Sun 22 Oct 2006 06:58:02 PM UTC, comment #15:
No, the patch is still in CVS:
http://cvs.savannah.nongnu.org/viewcvs/mldonkey/src/daemon/common/commonInteractive.ml?root=mldonkey&r1=1.76&r2=1.77
|
Sun 22 Oct 2006 04:47:51 PM UTC, comment #14:
may be.. you omit the patch in 2.8.*
$ diff mldonkey-2.8.1-without5272patch/src/daemon/common/commonInteractive.ml mldonkey-2.8.1-with5272patch/src/daemon/common/commonInteractive.ml
87c87
< if not Autoconf.windows then
---
> if Autoconf.windows then
90c90
< Buffer.contents buf (* Windows uses patched Ocaml which always uses Unicode chars *)
---
> Buffer.contents buf
|
Sat 21 Oct 2006 05:14:46 PM UTC, comment #13:
Patch 5272 is already used in MLDonkey since and including 2.8.0
|
Sat 21 Oct 2006 04:04:05 PM UTC, comment #12:
ha.. sorry for my ugly english.
the patch that work for me is http://savannah.nongnu.org/patch/?5272#attached
|
Sat 21 Oct 2006 01:34:53 PM UTC, comment #11:
I do not know about which patch lv yi is talking about but maybe
this one already in CVS:
2006/10/08
5445: Self-test charset conversion, disable conversion if test fails
|
Sat 21 Oct 2006 01:02:07 PM UTC, comment #10:
what patch is that?
|
Sat 21 Oct 2006 12:25:08 PM UTC, comment #9:
Thx spiralvoice your patch work well for me.
with your patch, mlnet recognize my utf-8 locale correctly now.
i think the patch will also work for Japanese characters.
something odd is mlnet 2.7.7 can't recognize my utf-8 locale correctly, but it hand Chinese characters well.
I hope your patch will be adopt in 2.8.2.
thank you again.
|
Sat 21 Oct 2006 09:48:54 AM UTC, comment #8:
Chinese Characters commit also regression.
Chinese commit work well in 2.7.7 . but when 2.8.0 and 2.8.1
committed chinese filename become underscores .
|
Wed 27 Sep 2006 01:54:00 PM UTC, comment #7:
ed2k://|file||51|0285C0C793948296C4EFED81155AFFEB|/
I removed the filename. Hope it works nonetheless.
|
Wed 27 Sep 2006 12:41:43 PM UTC, comment #6:
Please post the complete ed2k link of the testfile, I need at least the filesize to get sources from EDK servers. Searching by hash did not help.
|
Wed 27 Sep 2006 10:21:15 AM UTC, comment #5:
Get the patch here: http://savannah.nongnu.org/patch/?5272#attached
|
Wed 27 Sep 2006 12:06:36 AM UTC, comment #4:
Sorry, but I do not understand how to get that patch. https://savannah.nongnu.org/bugs/download.php?file_id=5272 does not seem to be said patch. I was not able to find any other way to go from a number to an URL in this context. Please be more specific.
|
Mon 25 Sep 2006 05:23:14 PM UTC, comment #3:
OK, I will try and see if I can compile this. Last time I failed.
Just wanted to let you know that German Umlauts seem to be affected as well. So this is really a wide-spread problem.
|
Mon 25 Sep 2006 07:30:29 AM UTC, comment #2:
Please remove this patch and rebuild MLDonkey:
5272: Fix OS-detection bug in Unicode conversion when committing files
- should solve file naming problems on non-UTF8 systems
|
Mon 25 Sep 2006 12:49:13 AM UTC, comment #1:
I prepared a small test file for you to verify this. The hash is 0285C0C793948296C4EFED81155AFFEB and the file name is ใในใ.txt
|
Mon 25 Sep 2006 12:46:42 AM UTC, original submission:
Hi,
when downloading files with Japanese characters, these characters are now being replaced by underscores again. I believe this was still working in version 2.7.7 from which I upgraded less than 48 hourse ago.
Regards
Rolf
|