Repository navigation
test(server): cut the local server suite from ~63 to ~12-16 minutes - #58
Merged
Merged
Conversation
The full apps/server suite ran files one at a time, and most of its plugin integration time was real-time waiting rather than work: - The pipe-backed PTY stand-in never showed a line-editor prompt, so every setup script waited out ProjectSetupScriptRunner's 3s quiet fallback and its 1s prompt-return timeout. It now echoes input and prints a prompt the way an interactive terminal does. - Every test server boot spawned the user's login shell synchronously (~1s, host-dependent). Test configs now mark the shell environment as prepared. - Test servers posted analytics to the real ingest host and waited on the shutdown flush. Server tests now disable telemetry. - Files run in parallel at 75% of cores. Two workflow tests lifted their simulated outage while a scheduler-tick drain could still hold a launch validated against the old provider or skill; they now reconcile (drains are serialized) before lifting it. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The full local
apps/serversuite took about 63 minutes. It ran test files one at a time, and most of the time went to plugin integration tests waiting on real timers.PluginHost.testkit.ts). It never printed a shell prompt, so every setup script waited outProjectSetupScriptRunner's 3s fallback and its 1s prompt wait. It now echoes input and prints a prompt the way a real terminal does. A plugin launch goes from 4.3s to 0.35s.shellEnvironmentPrepared: true.telemetry.setup.tsturns telemetry off for server tests.maxWorkers: "75%"replacesfileParallelism: false. Two Workflows*-directtests turned off their simulated outage while a background drain could still hold a launch checked against the old provider or skill. They now runreconcilefirst, which waits on the drain lock, before turning the outage off.Results on one 8-core machine:
origin/main)All runs had only the 12 known failures. CI runs only the smoke files, so CI time doesn't change. An independent review agent read the diff and the code around it and ran the affected integration tests. It approved; its one nit, where a comment sat, is fixed.
🤖 Generated with Claude Code (Claude Opus 5.5)