fix: upload linked PDBs for iOS

[jpnurmi]

Upload linked/*.pdb (instead of just linked/$(AssemblyName).pdb added for Mac Catalyst in #4503) to make missing symbols available on iOS with EnableAssemblyILStripping=false as reported in:

• iOS <EnableAssemblyILStripping>false</EnableAssemblyILStripping> not working: symbols missing #4482

Close: #4482

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 73.35%. Comparing base (4581417) to head (20d121f).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4527      +/-   ##
==========================================
- Coverage   73.36%   73.35%   -0.02%     
==========================================
  Files         479      479              
  Lines       17505    17505              
  Branches     3445     3445              
==========================================
- Hits        12843    12840       -3     
- Misses       3782     3785       +3     
  Partials      880      880              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[jpnurmi]

I'm having trouble ^[*] verifying the fix but the integration test shows that it uploads the symbol file reported missing in #4482. Is that a sufficient confirmation, or how should we proceed with this?

[*] I have a brand new iPhone Air running iOS 26, and an old iPhone 6S stuck on iOS 15. Neither of them seems good for verifying this fix because the former is too new i.e. not yet fully supported, and the latter might just be too old as even xcrun xctrace list devices shows it in offline status. 🙁

[Flash0ver]

I'm having trouble [*] verifying the fix but the integration test shows that it uploads the symbol file reported missing in #4482. Is that a sufficient confirmation, or how should we proceed with this?

We could do a pre-release. On the other hand, if an Integration Test is covering and asserting the scenario, it LGTM.

Could this cause any unintended regression? If it might be "dangerous" without further testing, I am leaning towards a pre-release. If it cannot break existing scenarios, but may not actually fix all affected scenarios, I am leaning towards merging.

What do you prefer?

[jpnurmi]

I'm leaning towards merging. All we do is pass an extra wildcard argument to sentry-cli debug-files upload to instruct it to upload more debug files. The CLI tool ignores duplicates and existing files, so I doubt the change is in any way dangerous. The only potential harm is that it might waste quota by uploading unnecessary .pdb files.

Although, as @jamescrosswell brought up in #4482, it's a bit unclear why these files are not available on the Microsoft symbol server that the user had enabled:

OK so it's not available on any of the symbol servers that Sentry checked.

@jpnurmi it looks like this one was in a "linked" directory:

Any chance that fix you rolled out for MacCatalyst is related here?

It is odd that we can't pull this from the Microsoft Symbol servers though.

[jpnurmi]

The only potential harm is that it might waste quota by uploading unnecessary .pdb files.

Actually, https://sentry.zendesk.com/hc/en-us/articles/25269209638555-How-much-does-it-cost-to-upload-Debug-Files

Sentry does not charge for debug information files (DIF) uploads. There is no limit on how many files you can upload and files are retained for 90 days without being used.

[jamescrosswell]

No reason not to merge this I don't think.

Thanks @jpnurmi !