Unicode Errors result in always re-download
Opened this issue · 3 comments
There seems to be a bug with Unicode detecting the same filename on the disk and the server differently.
So it re-downloads the whole file every time.
Added 'Übungsmaterial/Folien_Übung2.pdf'
Deleted 'Übungsmaterial/Folien_Übung2.pdf'
Exact copied output.
You can see it here https://www.soscisurvey.de/tools/view-chars.php
Redownload is on default
OS: Windows 10
Version: PFERD 3.4.3
Sounds like the file name is not unicode normalized on ilias, but at some point after downloading the file it is unicode normalized to NFC/NFKC. What file system are you running this on?
After further research, I could pinpoint the root cause of the issue.
It's a combination of PFERD, ILIAS and synchronization tools. A lot of the tools for synchronization normalize the file-/foldernames. In my case, it's syncthing.
There isn't probably a good solution in PFERD for this type of issue.
A setting (disabled by default) to normalize paths before applying transforms could maybe fix this.