Skip coverage when launching parallel processes

[hadley]

No description provided.

[claude]

Claude finished @hadley's task —— View job

---

Code Review Progress

• Read changed files to understand modifications
• Analyze changes for potential bugs or issues
• Provide review feedback

Review

The changes add skip_on_covr() calls to parallel testing functions that involve spawning processes. This prevents these tests from running during code coverage analysis, which is appropriate since:

1. Process spawning tests can interfere with coverage collection
2. Similar pattern already exists in other parallel tests in the codebase
3. The teardown test was already skipped anyway due to the skip("teardown errors are ignored") call

LGTM - these are defensive test skips that prevent coverage measurement interference without affecting actual test functionality.

---

[hadley]

@claude did I miss any tests?

[claude]

Claude finished @hadley's task —— View job

---

Analysis: Checking for missed parallel tests

• Search for other parallel test files
• Check for process spawning tests
• Review existing skip_on_covr() usage
• Analyze findings and provide recommendations

Findings

Good news - you didn't miss any tests! All parallel test files that spawn processes and could interfere with coverage collection already have skip_on_covr():

Files already protected (before this PR):

• test-parallel-crash.R:3 - ✅ Has skip_on_covr()
• test-parallel-errors.R:2 - ✅ Has skip_on_covr()
• test-parallel-setup.R:2 - ✅ Has skip_on_covr()
• test-parallel-startup.R:2 - ✅ Has skip_on_covr()
• test-parallel-stdout.R:2 - ✅ Has skip_on_covr()

Files covered by this PR:

• test-parallel-outside.R - ✅ Added skip_on_covr()
• test-parallel-teardown.R - ✅ Added skip_on_covr()
• test-parallel.R - ✅ Added skip_on_covr() to process-spawning tests

Non-issues:

• test-reporter-progress.R only has ParallelProgressReporter tests (no process spawning)

All parallel tests that spawn subprocesses via test_local() with TESTTHAT_PARALLEL = "TRUE" are now properly protected from coverage interference.

---

[hadley]

A test-coverage job ran successfully three times in a row, so I think this is the fix.