ISO-8859-1 is not honoured in attachments of text files and file names of attachements are changed.

This seems very similar to a bug I found listed under Confluence at http://jira.atlassian.com/browse/CONF-2487 .

JIRA does not handle attached text files encoded as ISO-8859-1 correctly. JIRA also translates the spaces in file names to + characters.

I attached a file called "Vorschlag Update - Ralf Ehrhard.txt" to a issue. The file is a text file with ISO-8859-1 encoding. The uploading worked perfectly. When clicking on the attachement (because it is a text file) my Firefox browser opened the file and rendered the text as UTF-8, replacing non-ASCII charachers with little 'diamond symbols'.

Trying to download the file, the file name became "Vorschlag+Update+-+Ralf+Ehrhard.txt". Once I opened the downloaded file from my local file system, the file rendered correctly again.

Inspecting the HTTP headers for the attachment, I found it to be

content-disposition: inline; filename*=UTF-8''Vorschlag+Update+-+Ralf+Ehrhard.txt;
Content-type: text/plain;charset=UTF-8
Content-Lenth: 1065

so the encoding of the file has not been honoured. Since we have users from all over the globe and multiple languages and encodings are being used,
it is important that encodings can be honoured in some way. Would it be possible to have an edit functionality where a user can manually correct the encoding (dropdown set up by local administrator perhaps even?) or where JIRA at least correctly senses and records the encoding.