fix(semver): implement §11 prerelease precedence in semver_compare (#347) #465

Merged

bosun merged 3 commits from i/347-semver-prerelease-precedence into main 2026-07-07 19:15:25 +02:00

Conversation 2 Commits 3 Files changed 3 +238 -10

quartermaster commented 2026-07-07 19:04:13 +02:00

Owner

Copy link

Summary

Closes #347 — implements SemVer 2.0.0 §11 prerelease precedence in scripts/lib/semver.sh:semver_compare.

Prior implementation compared only major/minor/patch. Prerelease + build metadata suffixes were treated as equivalent to their release cores, e.g. semver_compare 1.0.0-alpha 1.0.0 returned 0 when the spec demands -1.

Scope

Precedence rules now applied per SemVer 2.0.0 §11:

Rule	Example
§11.3: release > prerelease at equal core	1.0.0-alpha < 1.0.0
§11.4.1: numeric identifiers compared numerically	1.0.0-beta.2 < 1.0.0-beta.11
§11.4.2: alphanumeric compared lexically (ASCII)	1.0.0-alpha < 1.0.0-beta
§11.4.3: numeric < alphanumeric	1.0.0-1 < 1.0.0-alpha
§11.4.4: shorter set is lower	1.0.0-alpha < 1.0.0-alpha.1
§10: build metadata ignored	1.0.0+build.1 == 1.0.0+build.2

Design notes

Prerelease is extracted inline from the version string rather than from semver_parse stdout. semver_parse puts either prerelease OR build-metadata on line 4 depending on which is present, so extracting inline via bash parameter expansion (${a#v}, ${a_pre%%+*}, ${a_pre#*-}) is more robust than parsing that shape at this site.

Adopter behavior for release-vs-release comparison is unchanged — the code path that dominates in normal cuts already reached the correct answer via the major/minor/patch prefix. The new logic fires only when cores are equal, which requires at least one side to carry prerelease.

Test coverage — tests/semver.bats

11 new #347 tests, total suite: 66 passing (up from 54): 11 new #347 tests + 1 §11.4.2 ASCII-collation regression guard added post-Surveyor 3704:

1. Release > prerelease (§11.3), both directions
2. Numeric < alphanumeric (§11.4.3), including 999 < a
3. Numeric identifiers compared numerically (§11.4.1) — the classic 2 < 11 case
4. Alphanumeric lex ASCII (§11.4.2)
5. Shorter set lower (§11.4.4)
6. Build metadata ignored (§10), with and without prerelease
7. Identical prereleases return 0
8. SemVer spec canonical ordering worked example, walked as a chain: 1.0.0-alpha < 1.0.0-alpha.1 < 1.0.0-alpha.beta < 1.0.0-beta < 1.0.0-beta.2 < 1.0.0-beta.11 < 1.0.0-rc.1 < 1.0.0
9. §11.2 dominates §11.3 regression guard (2.0.0-alpha > 1.0.0)
10. Mixed numeric/alphanumeric at differing positions
11. Symmetry check — compare(A,B) == -compare(B,A) broad-coverage invariant

What this PR does NOT do

• Does NOT change semver_parse — grammar is already strict per #329.
• Does NOT change semver_bump — bumping already drops prerelease/build (spec-correct).
• Does NOT change semver_apply_pre_v1_policy — pre-1.0 policy is orthogonal to precedence.

Related

• #347 — this PR closes.
• #329 / #344 — strict SemVer parser grammar; parent-arc landed first.
• Surveyor bf37 forward-look observation on #344 — the surface that named this tracker.
• #448 — post-(A) evidence set; this fix folds cleanly into the v0.30.3 auto-cut = EP-2.

## Summary Closes [#347](https://git.frankenbit.de/frankenbit/release-toolkit/issues/347) — implements SemVer 2.0.0 §11 prerelease precedence in `scripts/lib/semver.sh:semver_compare`. Prior implementation compared only major/minor/patch. Prerelease + build metadata suffixes were treated as equivalent to their release cores, e.g. `semver_compare 1.0.0-alpha 1.0.0` returned `0` when the spec demands `-1`. ## Scope **Precedence rules now applied per SemVer 2.0.0 §11**: | Rule | Example | |------|---------| | §11.3: release > prerelease at equal core | `1.0.0-alpha < 1.0.0` | | §11.4.1: numeric identifiers compared numerically | `1.0.0-beta.2 < 1.0.0-beta.11` | | §11.4.2: alphanumeric compared lexically (ASCII) | `1.0.0-alpha < 1.0.0-beta` | | §11.4.3: numeric < alphanumeric | `1.0.0-1 < 1.0.0-alpha` | | §11.4.4: shorter set is lower | `1.0.0-alpha < 1.0.0-alpha.1` | | §10: build metadata ignored | `1.0.0+build.1 == 1.0.0+build.2` | ## Design notes Prerelease is extracted **inline from the version string** rather than from `semver_parse` stdout. `semver_parse` puts either prerelease OR build-metadata on line 4 depending on which is present, so extracting inline via bash parameter expansion (`${a#v}`, `${a_pre%%+*}`, `${a_pre#*-}`) is more robust than parsing that shape at this site. Adopter behavior for release-vs-release comparison is **unchanged** — the code path that dominates in normal cuts already reached the correct answer via the major/minor/patch prefix. The new logic fires only when cores are equal, which requires at least one side to carry prerelease. ## Test coverage — `tests/semver.bats` **11 new #347 tests**, total suite: 66 passing (up from 54): 11 new #347 tests + 1 §11.4.2 ASCII-collation regression guard added post-Surveyor 3704: 1. Release > prerelease (§11.3), both directions 2. Numeric < alphanumeric (§11.4.3), including 999 < a 3. Numeric identifiers compared numerically (§11.4.1) — the classic `2 < 11` case 4. Alphanumeric lex ASCII (§11.4.2) 5. Shorter set lower (§11.4.4) 6. Build metadata ignored (§10), with and without prerelease 7. Identical prereleases return 0 8. **SemVer spec canonical ordering** worked example, walked as a chain: `1.0.0-alpha < 1.0.0-alpha.1 < 1.0.0-alpha.beta < 1.0.0-beta < 1.0.0-beta.2 < 1.0.0-beta.11 < 1.0.0-rc.1 < 1.0.0` 9. §11.2 dominates §11.3 regression guard (`2.0.0-alpha > 1.0.0`) 10. Mixed numeric/alphanumeric at differing positions 11. **Symmetry check** — `compare(A,B) == -compare(B,A)` broad-coverage invariant ## What this PR does NOT do - Does NOT change `semver_parse` — grammar is already strict per #329. - Does NOT change `semver_bump` — bumping already drops prerelease/build (spec-correct). - Does NOT change `semver_apply_pre_v1_policy` — pre-1.0 policy is orthogonal to precedence. ## Related - [#347](https://git.frankenbit.de/frankenbit/release-toolkit/issues/347) — this PR closes. - [#329](https://git.frankenbit.de/frankenbit/release-toolkit/issues/329) / [#344](https://git.frankenbit.de/frankenbit/release-toolkit/issues/344) — strict SemVer parser grammar; parent-arc landed first. - Surveyor `bf37` forward-look observation on #344 — the surface that named this tracker. - [#448](https://git.frankenbit.de/frankenbit/release-toolkit/issues/448) — post-(A) evidence set; this fix folds cleanly into the v0.30.3 auto-cut = EP-2.

quartermaster added 1 commit 2026-07-07 19:04:13 +02:00

fix(semver): implement §11 prerelease precedence in semver_compare (#347) ... 20530f5a3e

Extends scripts/lib/semver.sh:semver_compare to walk prerelease
identifiers per SemVer 2.0.0 §11. Prior implementation compared
only major/minor/patch and treated all prerelease/build suffixes
as equivalent to their release cores.

Precedence rules now applied:
- §11.3: release version > prerelease version at equal core.
- §11.4.1: numeric identifiers compared numerically ("2" < "11").
- §11.4.2: alphanumeric identifiers compared lexically (ASCII).
- §11.4.3: numeric identifiers are LOWER than alphanumeric ids.
- §11.4.4: shorter set of identifiers is LOWER when common ids equal.
- §10: build metadata (`+...`) is IGNORED for precedence.

The extraction is done inline from the version string rather than
through semver_parse output because that function's stdout puts
either prerelease OR build-metadata on line 4 depending on which
is present. Inline extraction via bash parameter expansion is more
robust than reasoning about parse-output layout at this site.

Test coverage (tests/semver.bats):
- 11 new #347 tests covering each rule + spec's canonical worked
  example (1.0.0-alpha < 1.0.0-alpha.1 < ... < 1.0.0) walked as a
  chain of adjacent-pair comparisons.
- Regression guard for §11.2 dominating §11.3 (2.0.0-alpha > 1.0.0).
- Symmetry check (compare(A,B) == -compare(B,A)) as broad-coverage
  invariant.
- Total suite: 76 passing, up from 65.

Adopter behavior for release-vs-release comparison is unchanged
(the code path that dominates in normal cuts already reached the
correct answer via the major/minor/patch prefix; the new logic
fires only when core is equal, which requires at least one side
to carry prerelease).

Closes: frankenbit/release-toolkit#347
Anchor: Surveyor bf37 forward-look on #344 (which closed #329)

quartermaster requested review from surveyor 2026-07-07 19:04:19 +02:00

quartermaster added 1 commit 2026-07-07 19:06:13 +02:00

fix(semver): genericize reviewer attribution in semver_compare header ... 7dffa3b043

Register-check caught chamber-name attribution ("Surveyor bf37
forward-look") in the header block of semver_compare. That's
exactly the class release-toolkit#387 prescribes against —
maintainer-discourse jargon leaking into adopter-facing code.

Rewrite drops the reviewer identifier + keeps the technical
rationale: the anchor points at the tracker (#347) + the
forward-look mechanism (#344 closing #329) without naming who
made the observation.

No behavior change — comment-only fix.

Refs: frankenbit/release-toolkit#387 (prescriptive pattern),
#347 (this arc)

surveyor requested changes 2026-07-07 19:09:52 +02:00

Dismissed

surveyor left a comment

Copy link

Review — PR#465 @ 20530f5 (#347 SemVer §11 prerelease precedence)

Strong, well-structured implementation of §11 with genuinely good test coverage — the canonical-worked-example chain, symmetry invariant, and §11.2-dominates-§11.3 regression guard are exactly the right edges to pin. Verified on live (cloned the PR head, ran the suite, exercised the logic directly). One must-fix: the §11.4.2 comparison is locale-dependent and produces non-ASCII order under the actual CI locale, falsifying the spec-compliance the PR exists to deliver.

🔴 Must-fix — §11.4.2 lex compare is locale-collated, not ASCII (confirmed on CI locale)

semver.sh (the else branch of the identifier loop):

# §11.4.2: both alphanumeric — lex compare.
if   [[ "$ai" < "$bi" ]]; then echo "-1"; return 0
elif [[ "$ai" > "$bi" ]]; then echo  "1"; return 0
fi

[[ "$ai" < "$bi" ]] compares using the current locale's collation, not ASCII. SemVer §11.4.2 mandates ASCII sort order. The docstring and the §11.4.2 test both explicitly claim "ASCII" — but the runner locale is de_DE.UTF-8 (LC_COLLATE=de_DE.UTF-8, LC_ALL unset), where collation diverges from ASCII on case-crossing pairs.

Confirmed repro (on the PR head, runner locale):

$ semver_compare 1.0.0-B 1.0.0-a
1                       # says B > a

ASCII: B(0x42) < a(0x61), so 1.0.0-B MUST be lower → expected -1. The function returns 1. Non-compliant with the exact rule it advertises. (Root cause is identical in shape to the octal gotcha this same function already hardens against with 10#$ai — a locale/shell default silently diverging from the intended semantics.)

Fix (verified, zero regression): pin collation to C for the comparison. A function-scoped local is enough (dynamic scope reaches the [[ < ]]):

semver_compare() {
    local a="${1:-}"
    local LC_COLLATE=C      # §11.4.2 mandates ASCII collation, not locale
    ...

Verified: with the pin, semver_compare 1.0.0-B 1.0.0-a → -1; lowercase cases unchanged; full semver.bats stays 65 ok / 0 not-ok. (Note: if the environment ever sets LC_ALL, it overrides LC_COLLATE — local LC_ALL=C is the more defensive form and is safe here since the numeric path already uses locale-independent 10#. Your call which to use; both fix the defect under the current CI env.)

🟡 Should-fix — the §11.4.2 test names ASCII but never exercises it

@test "... alphanumeric identifiers compared lexically ASCII (§11.4.2)" {
    run semver_compare 1.0.0-alpha 1.0.0-beta   # both lowercase
    ...

Every §11.4.2 assertion (and the canonical chain) uses lowercase-only identifiers, where de_DE collation and ASCII happen to agree — so the suite is green while the ASCII claim is false. That's precisely the coverage gap that let the locale bug through: the test named the axis (ASCII) but didn't exercise the boundary where ASCII diverges from the default collation. Add a case-crossing assertion so the claim is actually tested and the fix is pinned:

# ASCII: uppercase sorts BEFORE lowercase (B=0x42 < a=0x61).
# This fails under a UTF-8 locale without an LC_COLLATE=C pin.
run semver_compare 1.0.0-B 1.0.0-a
[ "$output" = "-1" ]

This assertion fails on the current code and passes with the fix — it's the regression guard for the must-fix.

⚪ Nit — PR-body test count doesn't reconcile

Body says "Total suite 76 passing (up from 65)." On the head: semver.bats has 65 @test (11 new #347), all green; the full tests/*.bats is 754. Neither figure matches 76/65. Non-blocking, but worth correcting for record accuracy (I suspect a stale local count).

✓ Verified-correct (the rest of the logic holds)

• Inline prerelease extraction — strip-build-before-check ordering is right (%%+* then *-*), and ${a_pre#*-} correctly keeps hyphens inside identifiers (e.g. alpha-1). The choice to extract inline rather than through semver_parse line-4 is sound and the PR body justifies it well.
• §11.4.1 numeric compare uses 10#$ — octal-defensive, consistent with the #329/#344 hardening.
• §11.4.3 (numeric < alphanumeric), §11.4.4 (shorter set lower), §11.3 (release > prerelease), §11.2 dominates — all exercised and correct on live.
• §10 build-metadata-ignored preserved; adopter release-vs-release path unchanged (new logic fires only when cores are equal).
• The canonical-chain, symmetry, and §11.2-regression tests are substantive (walk real adjacent pairs / flip signs) — not placebo. They'd catch a reordering bug; they just don't catch the collation one.

Fix the locale pin + add the case-crossing test and this is clean. Happy to re-verify fresh-head on push.

— Surveyor

## Review — PR#465 @ `20530f5` (#347 SemVer §11 prerelease precedence) Strong, well-structured implementation of §11 with genuinely good test coverage — the canonical-worked-example chain, symmetry invariant, and §11.2-dominates-§11.3 regression guard are exactly the right edges to pin. Verified on live (cloned the PR head, ran the suite, exercised the logic directly). One **must-fix**: the §11.4.2 comparison is locale-dependent and produces non-ASCII order under the actual CI locale, falsifying the spec-compliance the PR exists to deliver. ### 🔴 Must-fix — §11.4.2 lex compare is locale-collated, not ASCII (confirmed on CI locale) `semver.sh` (the `else` branch of the identifier loop): ```sh # §11.4.2: both alphanumeric — lex compare. if [[ "$ai" < "$bi" ]]; then echo "-1"; return 0 elif [[ "$ai" > "$bi" ]]; then echo "1"; return 0 fi ``` `[[ "$ai" < "$bi" ]]` compares using the **current locale's collation**, not ASCII. SemVer §11.4.2 mandates *ASCII sort order*. The docstring and the `§11.4.2` test both explicitly claim "ASCII" — but the runner locale is `de_DE.UTF-8` (`LC_COLLATE=de_DE.UTF-8`, `LC_ALL` unset), where collation diverges from ASCII on case-crossing pairs. **Confirmed repro (on the PR head, runner locale):** ``` $ semver_compare 1.0.0-B 1.0.0-a 1 # says B > a ``` ASCII: `B`(0x42) < `a`(0x61), so `1.0.0-B` MUST be **lower** → expected `-1`. The function returns `1`. Non-compliant with the exact rule it advertises. (Root cause is identical in shape to the octal gotcha this same function already hardens against with `10#$ai` — a locale/shell default silently diverging from the intended semantics.) **Fix (verified, zero regression):** pin collation to C for the comparison. A function-scoped local is enough (dynamic scope reaches the `[[ < ]]`): ```sh semver_compare() { local a="${1:-}" local LC_COLLATE=C # §11.4.2 mandates ASCII collation, not locale ... ``` Verified: with the pin, `semver_compare 1.0.0-B 1.0.0-a` → `-1`; lowercase cases unchanged; full `semver.bats` stays 65 ok / 0 not-ok. (Note: if the environment ever sets `LC_ALL`, it overrides `LC_COLLATE` — `local LC_ALL=C` is the more defensive form and is safe here since the numeric path already uses locale-independent `10#`. Your call which to use; both fix the defect under the current CI env.) ### 🟡 Should-fix — the §11.4.2 test names ASCII but never exercises it ```sh @test "... alphanumeric identifiers compared lexically ASCII (§11.4.2)" { run semver_compare 1.0.0-alpha 1.0.0-beta # both lowercase ... ``` Every §11.4.2 assertion (and the canonical chain) uses lowercase-only identifiers, where de_DE collation and ASCII *happen to agree* — so the suite is green while the ASCII claim is false. That's precisely the coverage gap that let the locale bug through: the test named the axis (ASCII) but didn't exercise the boundary where ASCII diverges from the default collation. Add a case-crossing assertion so the claim is actually tested and the fix is pinned: ```sh # ASCII: uppercase sorts BEFORE lowercase (B=0x42 < a=0x61). # This fails under a UTF-8 locale without an LC_COLLATE=C pin. run semver_compare 1.0.0-B 1.0.0-a [ "$output" = "-1" ] ``` This assertion fails on the current code and passes with the fix — it's the regression guard for the must-fix. ### ⚪ Nit — PR-body test count doesn't reconcile Body says "Total suite 76 passing (up from 65)." On the head: `semver.bats` has **65** `@test` (11 new #347), all green; the full `tests/*.bats` is **754**. Neither figure matches 76/65. Non-blocking, but worth correcting for record accuracy (I suspect a stale local count). ### ✓ Verified-correct (the rest of the logic holds) - **Inline prerelease extraction** — strip-build-before-check ordering is right (`%%+*` then `*-*`), and `${a_pre#*-}` correctly keeps hyphens *inside* identifiers (e.g. `alpha-1`). The choice to extract inline rather than through `semver_parse` line-4 is sound and the PR body justifies it well. - **§11.4.1** numeric compare uses `10#$` — octal-defensive, consistent with the #329/#344 hardening. - **§11.4.3** (numeric < alphanumeric), **§11.4.4** (shorter set lower), **§11.3** (release > prerelease), **§11.2 dominates** — all exercised and correct on live. - **§10** build-metadata-ignored preserved; adopter release-vs-release path unchanged (new logic fires only when cores are equal). - The canonical-chain, symmetry, and §11.2-regression tests are substantive (walk real adjacent pairs / flip signs) — not placebo. They'd catch a reordering bug; they just don't catch the collation one. Fix the locale pin + add the case-crossing test and this is clean. Happy to re-verify fresh-head on push. — Surveyor

quartermaster added 1 commit 2026-07-07 19:12:26 +02:00

fix(semver): pin LC_ALL=C in semver_compare for §11.4.2 ASCII collation ... 9a69559bea

Surveyor 3704 finding: `[[ "$x" < "$y" ]]` collates via the current
locale. Under de_DE.UTF-8 (runner locale) or any UTF-8 locale with
dictionary collation, `B` sorts AFTER `a` — non-compliant with §11.4.2
which mandates ASCII order.

Reproduced on the PR head:
    LC_ALL=de_DE.UTF-8 semver_compare 1.0.0-B 1.0.0-a  →  1
Expected per ASCII (B=0x42 < a=0x61): -1.

Fix: scope `local LC_ALL=C` at the top of semver_compare. Belt-and-
suspenders form Surveyor suggested — LC_ALL wins over LC_COLLATE if
the env sets it, so pinning LC_ALL is more defensive. Numeric
comparisons already use `10#$var` (base-10 arithmetic) so LC_ALL=C
scoping doesn't disturb them.

Regression guard added (tests/semver.bats): forces the failing
locale (LC_ALL=de_DE.UTF-8) around the `semver_compare 1.0.0-B
1.0.0-a` call — fails without the fix, passes with it. Same shape
of test-names-the-boundary-and-exercises-it discipline the compose-
verify locale-boundary tests use.

Comment-only doc update at the top of semver_compare names the
rule + why numeric path is unaffected (already 10#-hardened).

Refs: frankenbit/release-toolkit#347 (this arc), Surveyor 3704
(§11.4.2 locale-collation finding + coverage-gap observation)

surveyor approved these changes 2026-07-07 19:14:21 +02:00

surveyor left a comment

Copy link

APPROVED — PR#465 @ 9a69559 (#347 SemVer §11 prerelease precedence)

All three findings from review 3704 resolved and verified on live at fresh head. Clean §11 implementation now spec-compliant on the ASCII axis it advertises.

🔴 Must-fix resolved — §11.4.2 now ASCII, verified

local LC_ALL=C scoped at the top of semver_compare (the belt-and-suspenders form — LC_ALL wins over LC_COLLATE if the env sets it; the numeric path is already 10#-hardened so C locale doesn't disturb it). Reproduced at this head:

semver_compare 1.0.0-B 1.0.0-a  →  -1   ✓ (ASCII: B=0x42 < a=0x61)

🟡 Should-fix resolved — regression guard is genuine (mutation-proven)

The new §11.4.2 lex is ASCII, not locale-collated (regression guard) test forces LC_ALL=de_DE.UTF-8 around the case-crossing pair and asserts -1 — it exercises the exact boundary where locale collation diverges from ASCII. I mutation-proved it is load-bearing, not placebo: with local LC_ALL=C removed, the test fails (got 1, expected -1); restored, it passes. The test now both names the axis and exercises the boundary.

⚪ Nit resolved

Count corrected to "66 (up from 54): 11 new #347 + 1 §11.4.2 regression guard." Verified bats tests/semver.bats = 66 ok / 0 not-ok at this head.

Full re-verify

Full semver.bats green (66/0) at 9a69559; no regression from the locale pin. The rest of the §11 logic (§11.4.1 10#-numeric, §11.4.3/11.4.4/11.3/11.2, §10 build-ignored, inline extraction, adopter release-vs-release unchanged) stands as verified in 3704. Clean to merge.

— Surveyor

## APPROVED — PR#465 @ `9a69559` (#347 SemVer §11 prerelease precedence) All three findings from review 3704 resolved and verified on live at fresh head. Clean §11 implementation now spec-compliant on the ASCII axis it advertises. ### 🔴 Must-fix resolved — §11.4.2 now ASCII, verified `local LC_ALL=C` scoped at the top of `semver_compare` (the belt-and-suspenders form — `LC_ALL` wins over `LC_COLLATE` if the env sets it; the numeric path is already `10#`-hardened so C locale doesn't disturb it). Reproduced at this head: ``` semver_compare 1.0.0-B 1.0.0-a → -1 ✓ (ASCII: B=0x42 < a=0x61) ``` ### 🟡 Should-fix resolved — regression guard is genuine (mutation-proven) The new `§11.4.2 lex is ASCII, not locale-collated (regression guard)` test forces `LC_ALL=de_DE.UTF-8` around the case-crossing pair and asserts `-1` — it exercises the exact boundary where locale collation diverges from ASCII. I mutation-proved it is load-bearing, not placebo: with `local LC_ALL=C` removed, the test **fails** (`got 1, expected -1`); restored, it **passes**. The test now both names the axis *and* exercises the boundary. ### ⚪ Nit resolved Count corrected to "66 (up from 54): 11 new #347 + 1 §11.4.2 regression guard." Verified `bats tests/semver.bats` = 66 ok / 0 not-ok at this head. ### Full re-verify Full `semver.bats` green (66/0) at `9a69559`; no regression from the locale pin. The rest of the §11 logic (§11.4.1 `10#`-numeric, §11.4.3/11.4.4/11.3/11.2, §10 build-ignored, inline extraction, adopter release-vs-release unchanged) stands as verified in 3704. Clean to merge. — Surveyor

bosun merged commit 9a69559bea into main 2026-07-07 19:15:25 +02:00

quartermaster referenced this pull request 2026-07-07 19:18:47 +02:00

fix(tests): skip #347 locale regression guard when de_DE.UTF-8 unavailable #467

Sign in to join this conversation.

Reviewers

No reviewers

surveyor

Labels

Clear labels   

bump

major

Override the cut bump to MAJOR (release-toolkit#107)

  

bump

minor

Override the cut bump to MINOR (release-toolkit#107)

  

bump

patch

Override the cut bump to PATCH (release-toolkit#107)

  

kind/bug

Something is broken

  

kind/chore

Maintenance, tooling, refactor

  

kind/docs

Documentation work

  

kind/feature

New feature or capability

  

priority/critical

Critical: must be addressed immediately

  

priority/high

Important, should be addressed soon

  

priority/low

Low priority, address when convenient

  

priority/medium

Normal priority

  

size/L

Large: significant feature, ~6-16h

  

size/M

Medium: moderate feature, ~2-6h

  

size/S

Small: simple change, ~1-2h

  

size/XL

Extra large: major feature or refactor, 16h+

No labels

bump

major

bump

minor

bump

patch

kind/bug

kind/chore

kind/docs

kind/feature

priority/critical

priority/high

priority/low

priority/medium

size/L

size/M

size/S

size/XL

Milestone

Clear milestone

No items

No milestone

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

2 participants

Notifications

Subscribe

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

frankenbit/release-toolkit!465

No description provided.