I have used encoder-enabled builds on my personal cluster for a couple of years now, and haven’t seen any issues. Let me know if anyone is keen enough to try out these builds too.
I’d consider adding all that as a separate section under the “Advanced” tab though (especially since some of these are very specialised and shouldn’t really be used except for very specific needs ).
How are the files with illegal characters encoded? (Share an example?)
If I rename a source file that has been encoded on the destination I presume the destination filename becomes unencoded assuming the offending characters are removed.
what happens if I rename the encoded filename on the destination side?
I have a few directories where this feature would be helpful and I would probably use it.
Yes the encoder is only on one side. The other nodes don’t seeing anything different on their end. In fact, they don’t even know a node (or more properly: folder) is an encoding node.
The FAT encoder uses the same encoding scheme used by WSL, Git-Bash, etc. So the filename a-pipe| is stored on disk as a-pipe\xf07c. See issue #9539 for details.
Renames/deletes work as expected as encoded filenames are decoded before being transmitted on the wire.
As I mentioned, users may have issues with handling encoded filenames in Windows Explorer and on the command line. But Windows 11 appears much more forgiving than Windows 10 was in handling these encoded filenames.
Yes, the FAT encoder is for any FAT-style partition, no matter the OS, so NTFS, exFAT, FAT32, and ReFS formatted partitions would all benefit from this encoder.
Per @tomasz86 ‘s suggestion, I will move the Filename Encoder setting to the Advanced tab. The other settings, I would still like to hide for non-Windows users, but I’m not sure how.
.
).
