docs(quickstart): declare the Path B compile timing as machine-variable - #260
Merged
Merged
Conversation
`content/docs/quickstart.mdx`'s Path B block publishes `✓ Build complete (72ms)`. Every other visible byte in both blocks is deterministic; that number is whatever the transcribing machine measured — `117ms` at CLI 17.1.0, `72ms` at 17.2.0, with no behavioural change behind the difference. The block is a faithful transcript, so the number stays exactly as captured. What changes is the declaration sentence that already tells the reader which parts of their own output will differ: it named the expanded project paths but not the timing, so a reader whose scaffold compiles in 300ms had no way to know that is expected rather than a symptom. The timing now sits in that same list. Second effect: the recurring re-transcription of these blocks at each CLI publish no longer has to justify a one-line timing change as drift, because the value is declared rather than asserted. Path A checked for the same shape and needs nothing: its published lines carry no measured value. Verified against `@objectstack/cli` 17.2.0 source rather than by re-booting — `printServerReady` (the banner both paths share) interpolates no duration, `start.ts` prints no timer at all, and the only `Build complete` timer in the CLI is on the compile path Path B exercises. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01ChPQM8jamxLUfUAxwFpJ8S
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixes #249
Path B's sample block publishes
✓ Build complete (72ms). Every other visible byte in both blocks is deterministic; that one is whatever the transcribing machine measured —117mswhen the block was transcribed at CLI 17.1.0,72mswhen it was re-transcribed at 17.2.0, with no behavioural change behind the difference.What changed
One clause in Path B's existing declaration sentence — the sentence that already tells the reader which parts of their own output will differ.
Before:
After:
The paragraph is re-wrapped to absorb the clause; nothing else in it changed. Diff is 5 insertions, 4 deletions, one file.
What deliberately did not change
The captured line is byte-for-byte as captured. The number is not deleted, not rounded, and not replaced with a placeholder. The block stays a faithful transcript, which is the property the recurring re-transcription card depends on. Mechanical check: the diff adds and removes zero lines inside either sample block (
git diff | grep -cover the block's line prefixes returns 0), andquickstart.mdx:164still reads✓ Build complete (72ms).Neither block was re-transcribed. That belongs to the refresh card and was just done at 17.2.0.
No locale sibling was touched. The ownership gate states it: "This PR touches 0 translation artifact(s) and 1 other file(s)."
Path A was checked and needs nothing
The card asked for Path A to be checked for the same shape rather than assumed. It carries no machine-variable published value, verified at source rather than by re-booting (re-transcription is out of scope here), against a
packages/clitree declaring version 17.2.0 — the same version both blocks were transcribed from:printServerReadyinpackages/cli/src/utils/format.ts— the ready banner both paths share, covering every line from✓ Server is readythroughPress Ctrl+C to stop— interpolates no measured duration. Its variable values are paths, mode, driver label, tenancy posture and plugin count/names, all already declared or version-determined.packages/cli/src/commands/start.tsmatches nothing fortimer,Date.now,elapsed,durationor amsinterpolation. Path A's header rows areprintKVcalls carrying paths and ids only.Build completetimer exists in the CLI source, atpackages/cli/src/commands/compile.ts:812, on the compile path that only Path B exercises.So this card is Path B only, and no Path A edit was manufactured for symmetry.
Gates
Every job CI runs for this diff was run locally at
55bff71, no narrowing. Each gate's own conclusion:check-translation-ownership.mjscheck-translations.mjs(freshness)check-translation-output.mjs --self-testgen-zh-hant.mjs --checkcheck-node-floor.mjsturbo run type-checkturbo run buildcheck-locale-surface.mjssitemap.xml409 URLs, 0 unexpected, 0 missingturbo run test--force; the first pass was a legitimate cache hit, the task's declared inputs excludecontent/docs/**)Exit codes were captured before any pipe. The substantive uncached run was taken on the tree that is now
55bff71withgit diff HEADempty, and the union was re-run after the commit at that sha.No changeset: this repo has no changeset flow (no
.changeset/, nothing published from a package).🤖 Generated with Claude Code
Generated by Claude Code
Generated by Claude Code