Repository navigation
doc(stela): rename the third version segment to patch - #67
Conversation
verStela's third segment is `patch` now, matching semver and the PATCH that build-common already parses out of CHANGELOGS.md. This document was the last place still calling it `fix`, and its two verified examples called asFix(). The version section also gains the optional fourth segment, `revision`, and the reason it is a real segment rather than semver's `+build` metadata: the spec requires that metadata be ignored when comparing, and a revision has to compare. Two things in the same document had gone stale against the code and are corrected here rather than left to contradict it: the tree-API example built a bare `stela`, which stopped compiling when stela became an ADT, and the class diagram predates defStela and rootStela. Both examples were run against the stela test harness before this commit.
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 122f792d0f
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| ASSERT_EQ(ver.asMajor(), 1); | ||
| ASSERT_EQ(ver.asMinor(), 0); | ||
| ASSERT_EQ(ver.asFix(), 8); | ||
| ASSERT_EQ(ver.asPatch(), 8); |
There was a problem hiding this comment.
Update the stela pin before calling asPatch
The core build still fetches byeolang/stela at GIT_TAG ade6414 in module/core/CMakeLists.txt, while this commit's own message says asPatch() exists only in the pending stela#10 change. Until the dependency is released and this repository's pin is advanced, this newly marked verified example (and the identical one below) does not compile against the version Byeol actually builds and documents.
Useful? React with 👍 / 👎.
Codecov Report✅ All modified and coverable lines are covered by tests. 🚀 New features to boost your workflow:
|
Why
verStela's third version segment is being renamedfix->patchinbyeolang/stela#10. This document is the last place still calling it
fix, andits two
verifiedC++ examples callasFix().patchis what semver and every package manager downstream of it call thatslot, and this project already writes PATCH everywhere else —
build-common'sby_parse_verexportsBY_VER_PATCHfrom the same number inCHANGELOGS.md.Only the C++ API and this page said
fix.What changed
The version section now documents
major.minor.patchplus the optionalfourth segment,
revision, with the reason it is a real segment rather thansemver's
+buildmetadata: the spec requires build metadata to be ignoredwhen comparing versions, and a revision has to compare. The names follow .NET's
major.minor.build.revisionand Debian'supstream_version-debian_revision, which agree on both the meaning and therule that the fourth segment resets when anything above it moves.
Two
verifiedexamples now callasPatch().Two staleness fixes in the same document
Not part of the rename, but both are this page contradicting the code it
documents, and both are one-liners:
stela root("root");. That stoppedcompiling when
stelabecame an ADT (refactor(ast): split the def block out of stela into defStela stela#6); a document is arootStelanow.defStelaandrootStela, and drewstelaasan instantiable class that
stelaParsercreates directly.On the
verifiedbadgeThe document convention says C++ blocks get
verifiedwithout an automatedcheck, since there is no online way to run them. So I ran both changed
examples against stela's own test harness (
byeolTest, on the branch forbyeolang/stela#10) before committing — they compile and pass. That is also how
the broken
stela root("root")line surfaced.Ordering
This depends on byeolang/stela#10, which carries the actual rename. Merge that
first, or this page will describe an API that does not exist yet.