Auto-parse filename characters to interoperate between Unix and Windows

[narcisgarcia]

Is your feature request related to a problem? Please describe.
Example: From GNU/Linux a screenshot is saved with a name like "Screensot 2022-03-22 13:48:25.png". Nextcloud desktop client uploads it to server. Nextcloud Windows client tries to download it and complains: "Following characters are not allowed on system: * " | & ? , ; : \ / ~ < >"

Describe the solution you'd like
Both client app and server software should have configuration parameters to allow choosing automatic behaviour:

• Server side: What to do with not interoperative characters: a) Store as is and serve translation only if needed by client (default) b) Store with a character replacement (keep original name in metadata) c) Store with character replacement and make source client to do so.
• Client side: a) Upload as is b) Upload with character replacement c) Rename to interoperative name and upload. d) Upload only files with compatible names and signal others.

Describe alternatives you've considered
Alternatives should be put on sysop hands.

Additional context
I haven't designed a solution to resolve conflicts with trasnlated names:

• File 1 at Unix client: "a:b.png" (uploads)
• File 2 at Unix client: "a b.png" (uploads)
• File 1 translated to Windows client: "a b.png" (downloads)
• File 2 translated to Windows client: "a b.png" (repeated name)

[joshtrichards]

Related issue: #38614
Related just merged PR: #40585