chore(release): 0.11.1 - #361
Conversation
Signed-off-by: client-software-ci <129794699+client-software-ci@users.noreply.github.com>
| @@ -1,3 +1,18 @@ | |||
| ## 0.11.1 (2026-08-26) | |||
There was a problem hiding this comment.
Version regression: 0.11.1 is inserted above the existing 0.12.0 entry, and it drops the breaking change.
The base branch already has a ## 0.12.0 (2026-08-26) section at the top of this file (now line 16). This PR prepends ## 0.11.1 (2026-08-26) above it, which:
-
Breaks the descending-version ordering of the changelog —
0.11.1<0.12.0, so the newest-first invariant no longer holds, and the top-of-file entry is no longer the release being cut. Sincetag_format = "{version}"and the latest tag is0.11.0, tagging0.11.1after0.12.0notes already exist would also make the tag order and the changelog disagree. -
Mislabels a breaking change as a patch release. This new section lists exactly the same commit (
457a364) as0.12.0, but omits0.12.0s⚠ BREAKING CHANGESblock. That block is real in this tree:extra_let_bindingsno longer exists anywhere insrc/(grep -rn extra_let_bindings src/returns nothing) — it was replaced byresolved_symtabin_session.py:763/:1045and_v1/_session.py:256/:282/:301. Publishing that removal as0.11.1tells downstream consumers a patch bump is safe when it will break any caller passingextra_let_bindings=, which was public in0.11.0.
If the intent is to release these changes, this should stay 0.12.0 (with the BREAKING CHANGES block) rather than adding a second, lower-versioned section for the same commit. If 0.12.0 was added in error and needs replacing, that section should be removed in this PR rather than shadowed.
|
|
||
| ### Features | ||
| * accept a resolved symbol table on the v0 session (#357) ([`457a364`](https://github.com/OpenJobDescription/openjd-sessions-for-python/commit/457a364371cb44bf5a6253e612a451e5678621d3)) | ||
| * deliver step-scope let bindings to run_task ([`457a364`](https://github.com/OpenJobDescription/openjd-sessions-for-python/commit/457a364371cb44bf5a6253e612a451e5678621d3)) |
There was a problem hiding this comment.
This entry advertises an API that does not exist in the released tree. extra_let_bindings was added to run_task and then removed by the same commit range — the 0.12.0 section below (line 25) correctly annotates it with "added extra_let_bindings to run_task, then superseded within this same release; see BREAKING CHANGES above", but that qualifier is dropped here. As written, a reader of 0.11.1 release notes would go looking for an extra_let_bindings parameter on run_task that isn't there.
Relatedly, lines 5 and 7 are the same feature listed twice, differing only in capitalization ("accept a resolved symbol table on the v0 session" / "Accept resolved symbol table on v0 session"), both pointing at commit 457a364. Worth de-duplicating.
0.11.1 (2026-08-26)
Features
457a364)457a364)457a364)Bug Fixes
457a364)457a364)457a364)457a364)