ref(ci): use sentry-xcodebuild.sh to wrap more xcodebuild usages

[armcknight]

We can simplify some places where there's a long complicated xcodebuild invocation, and instead place that in the existing wrapper script we have for it, where it can benefit from some other tricks we've learned over time.

There's still a few clunky invocations of the wrapper script itself that should improve after #4808

I stared at the one in the fastfile for a while but i'm not immediately sure what to do with it. I kinda feel like that step just shouldn't be done with fastlane, but i'll save that for another PR.

#skip-changelog

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 91.405%. Comparing base (be6ae95) to head (a27134e).

Additional details and impacted files

@@                          Coverage Diff                          @@
##           armcknight/ci/rename-xcode-test     #4818       +/-   ##
=====================================================================
+ Coverage                           91.399%   91.405%   +0.005%     
=====================================================================
  Files                                  627       627               
  Lines                                74569     74569               
  Branches                             26802     26804        +2     
=====================================================================
+ Hits                                 68156     68160        +4     
+ Misses                                6314      6310        -4     
  Partials                                99        99               

see 4 files with indirect coverage changes

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update be6ae95...a27134e. Read the comment docs.

[philprime]

l: why the fallback to false? isn't "" == true also falsy?

[armcknight]

the script fails here if you run it locally if your machine doesn't define the CI environment variable at all

[philprime]

Why does it fail when a variable is not set? shouldn't it just evaluate to an empty string?

[armcknight]

hm, after trying another minimal example it does. i was getting an unbound variable error at some point. i'll fix this.

[philprime]

I am not sure if we can add these env values here, as we are also building signed versions of the sample apps.

[philprime]

m: Correct me if I am wrong, but I think this will not be executed because the entire script will exit if the xcodebuild command fails

[armcknight]

IDK, i moved this from https://github.com/getsentry/sentry-cocoa/pull/4818/files#diff-b3aa5b62d9aab3e854e3fe47cc6aaacd2e4bd86fd6bb7484fb51d698623bf6a6L8-L19

[philprime]

@philipphofmann it looks like you introduced the test-with-thread-sanitizer.sh. Any insights here?

[armcknight]

Could it be that it used to work, but then the addition of && set -o pipefail at the beginning of the pipeline would cause it to no longer propagate the $? the same way? we also already had that at the top of the script, but i think that wouldn't have applied to the pipeline, which is why it was added there?

[philipphofmann]

I can't recall any specific reasons why I added it like that. If it's broken, feel free to fix it.

[armcknight]

honestly every time i make an edit to sentry-xcodebuild.sh at this point, it feels wrong to continue with it as a bash script. rewriting it in swift would be a huge improvement.