[ric-oliv]

Part 1 of 6 for #5806. Every job installed or wrote things nothing uses. This PR stops that, one commit per change.

Stop installing Mono. Every Linux and macOS job installed Mono, but nothing runs it. The only test that did, Ensure System.Reflection.Metadata is not missing, has been Windows-only since #4461. On Ubuntu the install downloads 53 MB from a package mirror, and on October 23 that took 15 to 21 minutes in 4 linux-x64 legs and an Android job. In #5794, Android Device Tests took 27.4 minutes against a median of 45.8 over the 10 runs before.

Stop writing detailed text build logs. The Build step wrote a text log next to the binlog, 1.9 GB on macOS, and uploaded both. Nothing reads it, and the binlog has the same content: a text log regenerated from a macOS binlog has the same 216 warning lines. If someone wants the text, dotnet msbuild build.binlog -flp:logfile=build.log writes it again in a few seconds. In #5792, Upload build logs went from 10 to 16 seconds per leg to 2 or 3, and the log artifacts from 54 to 119 MB to 13 to 33 MB.

Install only the workloads each job needs. Setup Environment ran dotnet workload restore without a project, so every job installed the workloads for all of Sentry.slnx.

• The environment action now requires projects, the solution filters or project files the job builds, and restores only their workloads. Jobs that build everything pass Sentry.slnx.
• dotnet workload restore doesn't take a solution filter, so the action passes the projects in it. A job that passes nothing, or a filter with no projects, fails with an error that says what to pass.
• The Android test app is restored only for jobs that build the tests that build it, and the legs without mobile targets skip Java and the Android SDK.
• CodeQL, format-code and verify-api no longer run a second, unfiltered workload restore after the action.

In #5784 and the throwaway #5804, the Linux legs without mobile targets stopped installing workloads at all. Setup Environment went from about 190 to 89 seconds in the trim job, and from 170 to 180 seconds to 71 to 78 in the Android device test build.

On this PR's own CI

2 runs, against main at b41b96c. Setup Environment went from 3.2 to 1.3 and 1.6 minutes on linux-arm64, from 3.3 to 1.9 and 2.3 on linux-x64, from 6.5 to 3.0 and 4.0 on win-x64, and from 3.1 to 1.9 in the trim job. The Android net10.0 build's setup went from 3.1 to 1.4 and 1.2. On main the net9.0 build's setup took 12.7, but that was a slow download of the Android command-line tools, which this PR doesn't touch. The whole stack's numbers are in #5806.

#skip-changelog

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.12%. Comparing base (b41b96c) to head (b13062d).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #5807      +/-   ##
==========================================
- Coverage   76.13%   76.12%   -0.02%     
==========================================
  Files         528      528              
  Lines       19579    19579              
  Branches     3829     3829              
==========================================
- Hits        14906    14904       -2     
- Misses       3791     3792       +1     
- Partials      882      883       +1     

☔ View full report in Codecov by Harness.
📢 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.