feat: read v-prefixed tags and keep their prefix on the next tag (#93)

Author: lesnik512

CONTEXT.md

@@ -39,12 +39,19 @@ and nothing else — no network, no tag history, no commit range. A strategy nev
 version; the use-case does that around it.
 
 **Latest tag**:
-The bump baseline: the **highest by SemVer precedence** among the repo's SemVer-parseable tags — not
-the most recently created one, and not the tag reachable from HEAD. `_select_latest_semver_tag`
-sorts by `semver.Version` and takes the maximum.
+The bump baseline: the **highest by SemVer precedence** among the repo's SemVer-form tags, bare
+(`1.2.0`) or **`v`-prefixed** (`v1.2.0`) — not the most recently created one, and not the tag
+reachable from HEAD. `_select_latest_semver_tag` sorts by `semver.Version` and takes the maximum.
+The new tag inherits the latest tag's **tag prefix**: after `v1.2.0` comes `v1.3.0`, after `1.2.0`
+comes `1.3.0`.
 _Avoid_: last tag, most recent tag. Both read as "newest by date", which names a different tag the
 moment a patch on an older line is pushed after a newer minor.
 
+**Tag prefix**:
+What precedes the SemVer version in a tag name: lowercase `v`, or nothing. Any other prefix
+(`V1.2.0`, `release-1.2.0`) makes the tag non-SemVer-form, so it is never a latest-tag candidate.
+_Avoid_: bare "prefix". The `branch-prefix` strategy owns that word for `feature/`, `bugfix/`, ….
+
 **Outcome**:
 What a run did, as the closed sum `Created | DryRun | NoTags | AlreadyTagged | NoBump` in
 `semvertag/_outcome.py`. It is internal and free to grow — the renderers `match` it exhaustively, so
@@ -61,7 +68,8 @@ is the form semvertag publishes *itself* under, and what `release.yml` means by
 which form you mean.
 
 **Release tag** / **floating major tag**:
-A release tag is **bare semver, no `v`** (`0.4.0`): what the tool emits, what `just publish` feeds to
+This repo's own release tags are **bare semver, no `v`** (`0.4.0`): what semvertag emits here because
+its latest tag is bare, what `just publish` feeds to
 `uv version $GITHUB_REF_NAME`, and what `release.yml` triggers on. The floating major tag is
 **`v`-prefixed** (`v0`): a single moving ref `release.yml` force-updates so action consumers can pin
 `uses: modern-python/semvertag@v0`. Two conventions, one repo; a `v` on a release tag breaks

README.md

@@ -27,8 +27,9 @@ uvx semvertag tag
 ```
 
 semvertag bumps from the highest existing semver tag and never creates
-the first one. Before the first run, create a plain semver tag such as
-`0.1.0` (a `v` prefix does not parse and is ignored).
+the first one. Before the first run, create a semver tag such as
+`0.1.0`, or `v0.1.0` if you want `v`-prefixed tags: each new tag keeps
+the prefix of the one it bumps from.
 
 ## Use it in GitLab CI
 

action.yml

@@ -22,7 +22,7 @@ inputs:
 
 outputs:
   tag:
-    description: 'The created tag (e.g. 1.2.3), or empty string if no bump was warranted.'
+    description: 'The created tag (e.g. 1.2.3, or v1.2.3 in a v-prefixed repo), or empty string if no bump was warranted.'
     value: ${{ steps.run.outputs.tag }}
   bump:
     description: 'The computed bump: none | patch | minor | major.'

docs/adr/0003-semver-form-tags-only.md

@@ -1,13 +1,14 @@
 # The bump baseline is SemVer-form only, and a prerelease baseline finalizes
 
-`_select_latest_semver_tag` keeps only tags `semver.Version.parse` accepts, strips build metadata,
-and takes the maximum by SemVer precedence; `_compute_new_version` then calls `Version.next_version`
+`_select_latest_semver_tag` keeps only tags `semver.Version.parse` accepts once a leading `v` is
+stripped, strips build metadata, and takes the maximum by SemVer precedence; `_compute_new_version`
+then calls `Version.next_version`
 rather than `bump_*`, so a prerelease baseline finalizes (`1.0.0-rc.1` plus a patch bump gives
 `1.0.0`, not `1.0.1`). That is the correct release-ramp semantics and is identical to `bump_*` on
 every stable baseline, so it changed no existing behaviour. PEP 440 prereleases such as `0.9.0rc1`
 are deliberately skipped: python-semver cannot parse them, its `coerce` recipe discards the `rc1` and
 makes a prerelease masquerade as final, and honest support means running `packaging` alongside
-`semver` to consume a form a SemVer tagger should not have to. A leading `v` is skipped too, which is
-a deferral rather than a rejection, since it is a one-line strip but would make semvertag consume a
-convention it does not emit; the cost until then is that a repo whose history is entirely
-`v`-prefixed reports `NoTags` and never bumps.
+`semver` to consume a form a SemVer tagger should not have to. A leading lowercase `v` is different:
+it is a prefix on a SemVer-form version, not another version grammar, so it is stripped before
+parsing and carried onto the next tag; a repo keeps whichever convention its latest tag uses (a `v`
+tag wins an exact precedence tie). Any other prefix, `V` included, is skipped.

docs/index.md

@@ -18,8 +18,9 @@ git tag.
 ## Quick start
 
 semvertag bumps from the highest existing semver tag and never creates
-the first one. Before the first run, create a plain semver tag such as
-`0.1.0` (a `v` prefix does not parse and is ignored).
+the first one. Before the first run, create a semver tag such as
+`0.1.0`, or `v0.1.0` if you want `v`-prefixed tags: each new tag keeps
+the prefix of the one it bumps from.
 
 ### GitHub Actions
 

docs/providers/github.md

@@ -41,10 +41,18 @@ ref via the GitHub API. If no bump is warranted, the job exits 0
 without pushing.
 
 > semvertag bumps from the highest existing semver tag
-> and never creates the first one. It reads only plain semver tags such
-> as `0.1.0`; a `v` prefix (`v0.1.0`) does not parse and is ignored. Until
-> one exists, every run reports `no_tags` and exits 0. Push one with
-> `git tag 0.1.0 && git push origin 0.1.0`.
+> and never creates the first one. It reads plain semver tags such as
+> `0.1.0` and `v`-prefixed ones such as `v0.1.0`, and each new tag keeps
+> the prefix of the one it bumps from. Other forms (`V0.1.0`,
+> `release-0.1.0`) are ignored. Until a readable tag exists, every run
+> reports `no_tags` and exits 0. Push one with
+> `git tag 0.1.0 && git push origin 0.1.0` (or `v0.1.0` for `v` tags).
+
+> **Upgrading from a release that ignored `v` tags:** if your history is
+> `v`-prefixed and you seeded a bare tag such as `0.1.0` to get started,
+> semvertag now bumps from whichever tag has the highest version, usually
+> your latest `v` tag (after `v1.4.0` comes `v1.5.0`, not `0.2.0`). When
+> the same version exists in both forms, the `v` tag wins.
 
 > semvertag detects GitHub Actions from the `GITHUB_ACTIONS=true` env
 > var that GHA sets automatically, so the `--provider` flag is optional
@@ -97,7 +105,7 @@ When you give the step an `id:`, downstream steps can read three outputs:
 
 | Output | Value |
 |---|---|
-| `tag` | The created tag (e.g. `1.2.3`), or empty string when `status` is `no-bump`. |
+| `tag` | The created tag (e.g. `1.2.3`, or `v1.2.3` in a `v`-prefixed repo), or empty string when `status` is `no-bump`. |
 | `bump` | `none` \| `patch` \| `minor` \| `major`. |
 | `status` | `created` (tag pushed) \| `no-bump` (nothing to tag: no prior tag, already tagged, no merge commit, or non-conforming commit). On CLI error the action itself exits non-zero and this output is not written. |
 

docs/providers/gitlab.md

@@ -44,10 +44,18 @@ project's `origin`. If no bump is warranted, the job exits 0 without
 pushing.
 
 > semvertag bumps from the highest existing semver tag
-> and never creates the first one. It reads only plain semver tags such
-> as `0.1.0`; a `v` prefix (`v0.1.0`) does not parse and is ignored. Until
-> one exists, every run reports `no_tags` and exits 0. Push one with
-> `git tag 0.1.0 && git push origin 0.1.0`.
+> and never creates the first one. It reads plain semver tags such as
+> `0.1.0` and `v`-prefixed ones such as `v0.1.0`, and each new tag keeps
+> the prefix of the one it bumps from. Other forms (`V0.1.0`,
+> `release-0.1.0`) are ignored. Until a readable tag exists, every run
+> reports `no_tags` and exits 0. Push one with
+> `git tag 0.1.0 && git push origin 0.1.0` (or `v0.1.0` for `v` tags).
+
+> **Upgrading from a release that ignored `v` tags:** if your history is
+> `v`-prefixed and you seeded a bare tag such as `0.1.0` to get started,
+> semvertag now bumps from whichever tag has the highest version, usually
+> your latest `v` tag (after `v1.4.0` comes `v1.5.0`, not `0.2.0`). When
+> the same version exists in both forms, the `v` tag wins.
 
 > `resource_group: semvertag` makes GitLab serialize concurrent
 > `semvertag` jobs across pipelines on the same project, so

semvertag/_outcome.py

@@ -8,7 +8,8 @@
 # words NoTags/AlreadyTagged differently on purpose — edit both if you change the
 # message for one audience.
 _NO_TAGS_REASON: typing.Final = (
-    "No prior semver-conforming tags found; create an initial tag such as 0.1.0 on a default-branch commit."
+    "No prior semver-conforming tags found; create an initial tag such as 0.1.0 "
+    "(or v0.1.0 for v-prefixed tags) on a default-branch commit."
 )
 _ALREADY_TAGGED_REASON: typing.Final = "Latest commit already tagged."
 

semvertag/_output.py

@@ -65,7 +65,7 @@ def _format_outcome(outcome: Outcome, *, strategy: str) -> str:
         case NoTags():
             return (
                 "No tag created — no prior semver-conforming tag to bump from; "
-                "create an initial tag such as 0.1.0 on a default-branch commit."
+                "create an initial tag such as 0.1.0 (or v0.1.0 for v-prefixed tags) on a default-branch commit."
             )
         case AlreadyTagged(tag=tag):
             return f"No tag created — latest commit is already tagged {tag}."

semvertag/_use_case.py

@@ -10,6 +10,9 @@
 from semvertag.strategies._base import BumpStrategy
 
 
+_V_PREFIX: typing.Final = "v"
+
+
 @dataclasses.dataclass(frozen=True, slots=True, kw_only=True)
 class SemvertagUseCase:
     provider: Provider
@@ -39,30 +42,34 @@ def __call__(self, *, output: Output, dry_run: bool = False) -> Outcome:
                 NoBump(status=self.strategy.no_bump_status, reason=self.strategy.no_bump_reason, commit=commit.sha),
             )
 
-        new_version: typing.Final = _compute_new_version(latest_version, bump)
+        new_tag: typing.Final = _tag_prefix(latest_tag) + _compute_new_version(latest_version, bump)
         if dry_run:
-            return self._emit(output, DryRun(tag=new_version, bump=bump, commit=commit.sha))
+            return self._emit(output, DryRun(tag=new_tag, bump=bump, commit=commit.sha))
 
-        output.progress(f"Creating tag {new_version}...")
-        self.provider.create_tag(name=new_version, commit_sha=commit.sha)
-        return self._emit(output, Created(tag=new_version, bump=bump, commit=commit.sha))
+        output.progress(f"Creating tag {new_tag}...")
+        self.provider.create_tag(name=new_tag, commit_sha=commit.sha)
+        return self._emit(output, Created(tag=new_tag, bump=bump, commit=commit.sha))
 
     def _emit(self, output: Output, outcome: Outcome) -> Outcome:
         output.emit(outcome, strategy=self.strategy.name)
         return outcome
 
 
+def _tag_prefix(tag: Tag) -> str:
+    return _V_PREFIX if tag.name.startswith(_V_PREFIX) else ""
+
+
 def _select_latest_semver_tag(tags: list[Tag]) -> tuple[Tag, semver.Version] | None:
     parsed: list[tuple[semver.Version, Tag]] = []
     for tag in tags:
         try:
-            version = semver.Version.parse(tag.name).replace(build=None)
+            version = semver.Version.parse(tag.name.removeprefix(_V_PREFIX)).replace(build=None)
         except ValueError:
             continue
         parsed.append((version, tag))
     if not parsed:
         return None
-    parsed.sort(key=lambda item: item[0])
+    parsed.sort(key=lambda item: (item[0], item[1].name.startswith(_V_PREFIX)))
     version, tag = parsed[-1]
     return tag, version