Ticket #234 (new task)
search results with non-ascii characters are displayed incorrectly
| Reported by: | anonymous | Owned by: | daelstorm |
|---|---|---|---|
| Priority: | minor | Milestone: | Release 1.2.11 |
| Component: | nicotine | Version: | 1.2.10 |
| Keywords: | Cc: |
Description
Search results from Soulseek clients (not Nicotine+ clients, tested it on myself) that have funny characters (e.g. characters with umlauts) in them display incorrectly: the funny character and two normal ones following it are replaced by a single color-inverted question mark ("täna" becomes "t�"). I believe this happens because Soulseek sends the results in iso8859-something(15?), but Nicotine+ expects utf8. I believe so because I see many �'s in the estonian "eesti mehed" chat room when the encoding is set utf8, but not when it's iso8859-15.
I understand the benefits of utf8 and I know that the proper way to fix the problem is to make Soulseek use unicode, but until then (is it ever going to happen?) we need a work-around. If it's possible to automatically identify the charset used or whether the client that sent the search result is unicode-capable, then problem solved. If not, I think an option to manually set the charset of different search tabs (just like in chat tabs) would be sufficient, for now.
Attachments
Change History
comment:1 Changed 4 years ago by offhand
- Priority changed from major to minor
- Version set to 1.2.10
- Milestone set to Release 1.2.11
comment:2 Changed 3 years ago by nachtigall@…
I have the same problem with ubuntu 10.4 and nicotine+ 1.2.12
comment:3 Changed 3 years ago by anonymous
Yes, alas. As far as I understand the network we cannot actually fix this properly. The protocol doesn't use a fixed encoding, so people send out replies in their native encoding, which could be anything.
What I can do is wrap the same code used for the chat decoding with search results. This will try a few different encodings and use the one with the least amount of problems.
