Fix uploading zipped tiff (should set needsConversion=true)

[fm3]

Fixes two bugs in dataset upload

• The frontend failed to set needsConversion=true in case of zipped data (e.g. a zip with tiffs), because the return statement exited not only the for loop but the whole function
• The backend detected N5 data too eagerly (return value of containsMatchingFile was not used)

Steps to test:

• yarn enable-jobs (no actual worker needed though)
• In dataset upload view, upload a zip with tiff files
• After the upload, a conversion job should be started
• Upload bare tiffs, upload wkw, these other cases should still behave as expected

Issues:

• fixes uploading a zip with tiff files does not set needsConversion=true #7855

---

• Updated changelog
• Considered common edge cases
• Needs datastore update after deployment

[hotzenklotz]

Suggested change

	// We return here since not more than 1 zip archive is supported anyway. This is guarded
	// We break here since not more than 1 zip archive is supported anyway. This is guarded

[hotzenklotz]

Maybe rephrase entire comment

[fm3]

good point, done :)

[MichaelBuessemeyer]

Works 🎉

I tested all tiff formats included in: https://www.notion.so/scalableminds/Test-Datasets-c0563be9c4a4499dae4e16d9b2497cfb?pvs=4#7db7f26b5e494352a8153ab073487ab3 as well as wkw.

Just left one question :)

[MichaelBuessemeyer]

What is the advantage of this change?
To me the code semantically reads the same. Just that the testing for whether dataSources has atleast a one item is more explicit. Or am I misunderstanding something here?

[fm3]

head on an empty list raises an exception, which bypasses some of our error handling. headOption.toFox creates a Fox.failure that will yield nicer error messages in our context.
I’d say head is always a codesmell in our codebase, unless it is directly wrapped in a tryo (but then headOption should always be possible instead) or we are absolutely certain that the list will not be empty.