Uploads break on Unicode filenames

Bug #786654 reported by Matt Giuca
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
MUGLE
Triaged
Low
Matt Giuca

Bug Description

If a file is uploaded (inside a Zip file) with a Unicode filename, any non-Latin-1 characters will show up as "?". Latin-1 characters appear to work. "café" shows as "café", whereas "一個文本文件" shows as "??????.txt". Interestingly, it has the correct number of characters, so I don't know what the deal is.

Running the exact same Zip code from the command-line, reading from stdin, it works perfectly, and gives well-formed Unicode strings. So I suppose it has something to do with the encoding of the input stream. (But I thought InputStream was a binary stream and didn't have any encoding.)

Tags: unicode
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.