Directory visibility problems caused by non-UTF-8 filenames

09.02.2016 - 15:37 - 10.02.2016 - 16:37
Description: 

It has come to the IT administration's attention that the new home and group directory NAS server will cut directory listings over NFSv4 short if it encounters a non-UTF-8 filename. This will present to the user as files and directories being apparently missing. The seemingly missing files (at least ones whose names do not use 8-bit legacy encodings) should nevertheless be accessible using their names.

Considering that NFSv4 does require UTF-8 encoding to be used for filenames, that the department's Linux systems have used UTF-8 for years now, and that the legacy encodings are likely to be an usability issue otherwise as well, we are presently preparing to simply transcode all non-UTF-8 compliant filenames from legacy encodings to UTF-8. Consider this fair warning that this will probably happen in the near future. We will try to heuristically determine the original character set as best we can. We will be able to reconstruct the original filenames, if in some cases strange names might result. (All valid UTF-8 names, including all plain 7-bit ASCII names, will remain untouched.)

If you wish to have your home or group directory processed immediately, please make a request at itsupport@cs.helsinki.fi

11.02.2016 - 00:56 Mikko J Rauhala
09.02.2016 - 17:09 Mikko J Rauhala