A customer of our product (which uses libcurl) ran into an issue with file:// urls and unicode characters. It appears that when there are any non-ascii characters in the url, curl will not retrieve the file. Now, how does the url need to be encoded for this to work in curl? Or can curl handle this at all? I have the feeling it could be necessary for libcurl to use the _wopen() function internally instead of _open() on windows. If so, is this the right place to submit a bug to get it changed?

Btw, on the mac, this just works, by passing the urls to curl in UTF-8 format. Ideally, we'd like windows to be the same.

Thanks for any information on this,
jonas
-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette: http://curl.haxx.se/mail/etiquette.html
Received on 2010-05-11