What is the problem you are having with rclone?
OneDrive is case insensitive, so case-collisions aside, restoring from a backup to OneDrive would lose all of the case sensitivity of my local (Linux ext4) filenames.
Will something like filename_encoding = base32768
in my OneDrive configuration resolve this issue?
If so, how can I rename the files in an existing copy on OneDrive without having to re-upload the whole local filesystem?
I have read Convert Filename Encoding but I am not using any crypt options in my config, so not sure how that maps.
Run the command 'rclone version' and share the full output of the command.
rclone 1.60.1
- os/version: fedora 38 (64 bit)
- os/kernel: 6.3.4-201.fc38.x86_64 (x86_64)
- os/type: linux
- os/arch: amd64
- go/version: go1.19.3
- go/linking: dynamic
- go/tags: none
Which cloud storage system are you using? (eg Google Drive)
OneDrive
The command you were trying to run (eg rclone copy /tmp remote:tmp
)
rclone sync $source onedrive:/Backups/$dest \
--local-case-sensitive \
--links \
--delete-excluded \
--exclude-from excludes \
--progress
The rclone config contents with secrets removed.
[onedrive]
type = onedrive
token = [redacted]
drive_id = [redacted]
drive_type = business
No as it is case preserving so case-collisions aside everything will work.
It is crypt
option only
But if I add crypt
to my configuration, I will eliminate the case-collisions?
Does it make sense for the situations of case-collisions for rclone
to be able to encode filenames even when not encrypting?
It should - the best is to test.
It is possible today - use crypt and set no-data-encryption = true
Please note that base32768 is case sensitive though...
opened 10:56AM - 03 Mar 23 UTC
For various reasons I was trying to understand base32768 in details and I have s… uspicion that base32768 encoding rclone is using is not 100% safe on case insensitive remotes like OneDrive.
The reason is that so called safe alphabet used contains both upper and lower case characters.
I have raised the question with @Max-Sum as his go encoding procedure is used in rclone.
https://github.com/Max-Sum/base32768/issues/1
Chances of upper/lower case collisions are extremely low (but not zero) given that 99% of alphabet used does not have upper and lower cases (asian characters) and it is only used for file names. We would need by chance two different files resulting with base32768 encoding results like:
䏠Ҡ䈠䉀䉠䊀䊠䋀䋠䌀䌠䍀䍠䎀䎠䏀
and
䏠ҡ䈠䉀䉠䊀䊠䋀䋠䌀䌠䍀䍠䎀䎠䏀
In this example Ҡ and ҡ are problematic and as I understand it they are both part of alphabet used by rclone.
Maybe I am wrong but I think it would be helpful for somebody more knowledgeable to double check it.
It is maybe theoretical as chances of collision for encrypted names is extremely low but if you want to be 100% safe use base32.
system
(system)
Closed
July 14, 2023, 2:01pm
6
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.