When my users uploads files with filepicker.io and they use characters with accents (like ąłżźð) some of the files uploaded to S3 use NFC while others use NFD unicode normalization form (there is about ~20% of NFD, and sources suggests those would be from Mac OS X users).
In result I cannot simply download NFD files from S3, without trying both forms (the file names I store are always in NFC).
Is there a way to tell filepicker.io to always convert filenames to NFC before uploading to S3? Or do I really have to resort to doing conversion myself in javascript (for example using: https://github.com/walling/unorm )?
For those investigating similar issue, few background reads on the topic:
https://github.com/fog/fog/issues/1294
File.listFiles() mangles unicode names with JDK 6 (Unicode Normalization issues)
