Secret File Drop: Uploads with same name don't get renamed to file (1).txt

[tbsbdr]

Steps to reproduce

1. Create a Secret File Drop MyFiledropFolder
2. Upload file.txt to MyFiledropFolder via the sharing link
3. Repeat: Again upload file.txt via the sharing link
4. MyFiledropFolder contains

 ├── file.txt

Expected behavior

• MyFiledropFolder contains 2 files file.txt and file (1).txt

 ├── file.txt
 └── file (1).txt

Actual behavior

• MyFiledropFolder contains

 ├── file.txt

[tbsbdr]

@micbar lets discuss the prio
After discussion with @kulmann

• collision-check must happen in backend (client does not know about existing files because there are no read permissions)

Open questions:

• possible to pass a header value to the upload requests that overwrites should be forbidden? Then the backend would not need to check on its own if the upload happened via a secret file drop link.

[ScharfViktor]

we have api tests for that which are in the expected failures file with issue tag #1267

[phil-davis]

And that links to discussion in cs3org/reva#2480
Which links to discussion in cs3org/reva#2510
with what I thought was the end decision in comment cs3org/reva#2510 (comment) on Feb 21, 2022 - nearly 2 years ago.

But nothing happened after that.

[phil-davis]

And more related discussion about what the behavior should be in #5074

[butonic]

@tbsbdr what naming scheme should be used?
0) pm: file.txt, file (1).txt, file (2).txt

1. oc10 & windows: file.txt, file (2).txt, file (3).txt (humans count the file.txt as 1)
2. phil: file.txt, file_fs93e.txt, file_pa4io.txt
3. CERN: {uuid} file.txt
4. timestamp: file.txt, file (2024-02-01 14:56:54.003Z).txt, file (2024-02-01 14:56:55.004Z).txt

I highly recommend not deviating from what other systems like windows do. Endusers might expect this behavior. Also ... our oc10 tests already expect option 1)

[tbsbdr]

option 1 like in Windows
file.txt, file (2).txt, file (3).txt

[AlexAndBear]

@tbsbdr could you check that with the current implementation in web? IHMO we go f, f(1) (you skipped that). Maybe we need to adjust web then

[tbsbdr]

I would just align it if the effort is < 30 mins as I bet no one will notice it (except for the pain we experience by being aware of that inconsistency)

[dragonchaser]

@tbsbdr should we close this here and track this on web?

[tbsbdr]

fixed via #8340