egghelp/eggheads community

UTF-8 input being garbled [solved]

2016-06-15T13:48:51-04:00

I ended up getting the latest eggdrop 1.8 from github, which doesn't require the script above.

Everything is working well now.

Statistics: Posted by artaslove — Wed Jun 15, 2016 1:48 pm

UTF-8 input being garbled [solved]

2016-06-14T22:26:50-04:00

Upon further investigation, while the script posted above did solve the issue with utf-8 input being garbled, it introduced some other problems with utf-8 output that I am still working out.

Statistics: Posted by artaslove — Tue Jun 14, 2016 10:26 pm

UTF-8 input being garbled [solved]

2016-06-14T21:31:28-04:00

Solved by http://forum.egghelp.org/viewtopic.php?t=18879

I was trying to use the solution at http://eggwiki.org/Bugs/Utf-8 which did not work in my case.

Statistics: Posted by artaslove — Tue Jun 14, 2016 9:31 pm

UTF-8 input being garbled [solved]

2016-06-14T21:32:28-04:00

Hello,

I'm running eggdrop 1.8, patched to include UTF-8 support. My system LANG is set to UTF-8. My chat client supports UTF-8. The server I'm connecting to supports UTF-8. I have no troubles at all sending any sort of valid UTF-8 character to the channel my bot is connected to.

I have a simple script that connects to Microsoft translate and returns the translated result. It is working for many languages already. However when translating from Russian, Japanese, Chinese and possibly others it encounters a problem.

in my proc I experimented with "putlog $text" immediately at the beginning of the proc, where $text is everything the bot thinks the user has entered for that binding.

For example:

User enters: "!trans ru|en очень хорошо"
logfile shows: "ru|en >G5=L E>@>H>"

Naturally this does not work out to be the correct translation.

If I urlencode очень хорошо to %D0%BE%D1%87%D0%B5%D0%BD%D1%8C%20%D1%85%D0%BE%D1%80%D0%BE%D1%88%D0%BE and send it to the translator I get the expected result of "very well".

For some reason some UTF-8 characters work for example:
User enters: "!trans fr|en très bien"
logfile shows: "fr|en tr▒s bien"

However the character is correctly percent-encoded to %C8%A8 in that case and I get the expected result of "very well".

Thanks for any insight.

Statistics: Posted by artaslove — Tue Jun 14, 2016 3:58 pm