Repository navigation
canary: cmuxTests full-recompile probe (do not merge) - #14320
teamleaderleo wants to merge 7 commits into
Conversation
…heme build (do not merge) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Important Draft PR not reviewedDraft PRs are not automatically reviewed by default.
To automatically review draft PRs, update your CodeRabbit configuration: reviews:
auto_review:
drafts: trueThanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
|
All contributors have signed the CLA ✍️ ✅ |
… not merge) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Data point for this probe: test-only canary #14269, run 36077356582, 6vcpu, j6 seed at distance 0. The cmux scheme did 0 SwiftDriver, 0 SwiftEmitModule and 0 SwiftCompile tasks. cmux-unit then did SwiftCompile 43 tasks (1,205 CPU s) and emitted cmuxTests. So the full cmuxTests recompile happens with no cmux module change at all. Log: job 107898790966. |
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… (do not merge) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…not merge) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Cause of the cached full cmuxTests rebuild, measured on #14249 (run 36081880621, 12vcpu-macos-26, Xcode 26.6): with |
|
Closing: the full cmuxTests recompile this canary measured (swift-driver rewriting cmuxTests' ChainedBridgingHeader.h on every run, 1,116 files per one-line test edit) is what #14349 fixed by building cmuxTests without the compilation cache. This branch predated #14349. -no-auto-bridging-header-chaining and SWIFT_PRECOMPILE_BRIDGING_HEADER=NO did not stop the rewrite with the cache on. |
Measurement canary. Do not merge; closing once the admission reports.
Every PR compile admission recompiles all ~1,100
cmuxTestsfiles (800 to 1,000 CPU s, about 200 s wall on 12vcpu), even whencmuxcompiled 1 to 7 CPU s (jobs 107856043911, 107863037290, 107853852596). That includes the owned minis with their own persistent DerivedData, so it is not seed transport. #14235 measured acmux-unitnull rebuild at 29 s. Admission builds thecmuxscheme first, and that build andcmux-uniteach relinkcmuxand about 13 package frameworks.After admission, this probe runs:
cmux-unittwice (null)cmuxalone, then lists what it rewrote under Build/cmux-unitagain with XCBuild build debugging on, then lists what it rewroteIt also diffs the cmux and cmuxTests link and driver arguments between the two schemes and uploads the traces as
canary-cmuxtests-rebuild-*. It includes a comment-only test change, so it is a test-only PR admission.🤖 Generated with Claude Code
Need help on this PR? Tag
@codesmith-botwith what you need. Autofix is disabled.Summary by cubic
Measurement canary (do not merge) that probes why the compile admission rebuilds all ~1,100
cmuxTestsfiles (800–1,000 CPU s) whencmuxitself compiles in 1–7 CPU s and acmux-unitnull rebuild takes ~29 s.cmux-unitwith-driver-show-incrementalafter the canary test edit and again as a null rebuild, captures swiftdeps and ChainedBridgingHeader write times, then repeats both withSWIFT_PRECOMPILE_BRIDGING_HEADER=NOto test whether the precompiled bridging header causes the full rebuild.CMUX_CI_BUILD_SCHEMESto filter schemes and passesCMUX_CI_EXTRA_SWIFT_FLAGS/CMUX_CI_EXTRA_BUILD_SETTINGthrough toxcodebuild.canary-cmuxtests-rebuild-*artifacts; the comment-only test change keeps this a test-only admission.Written for commit 3be3d2a. Summary will update on new commits.