Zotero losing files with non-english characters

I was trying to figure out why some of my attachments are missing, and I think I found something worth fixing in Zotero. Specifically, the getValidFilename function in file.js (which is called by getFileBaseNameFromItem) does not remove non-english characters like accented vowels and such.

So for an article by "Müller", the auto-generated file name contains a "ü", and somehow in the process of syncing with the Zotero server, this becomes "M├╝ller" with that weird character. Files with names that have that weird character somehow don't get linked into the zotero database correctly, so on my other computers, even though this file downloaded (can search in Windows and find it), Zotero can't find it.

Furthermore, when I tried to compress my zotero storage for backup, I found that the compression program in Windows won't compress files with strange characters like in "M├╝ller", even though it is fine with "Müller".

-Keith
  • edited November 17, 2010
    We need more details than this, since saving files with Unicode characters works fine for most users. Reverting to the lowest common denominator of ASCII filenames just because some systems don't properly support Unicode characters isn't necessarily the best solution.

    What OSes, filesystems, and Firefox versions are you using on the original computer and the other computers? What's an example site you're saving from? When exactly is the mangling happening?
Sign In or Register to comment.