@gitkraken/core-gitlens 0.11.0 → 0.12.0

package/CHANGELOG.md

@@ -6,6 +6,44 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/) and this p
 
 ## [Unreleased]
 
+## [0.12.0] - 2026-10-16
+
+### Added
+
+- Adds `orgId` to `ProviderOrganization`: the parent's stable id where `org` is only its label. `listProjects` sets it on Jira projects (the site's id, or the Data Center instance's domain), which is what `listIssueTrackerIssuesPage` takes as `org` for Jira (plus/integrations)
+- Adds `isTerminalIssueStateCategory`, whether an issue state category ends the issue: `DONE` or `CANCELED` (git)
+- Adds the workflow state and its category to Azure DevOps work items (`providerState`, and `issueType` on the native point and search reads), with `Removed` reported as `CANCELED`. The category is read from the work item type's states endpoint (`api-version=6.0-preview.1`, which Azure DevOps Server 2020 requires) and cached per host, project, type and credential for 5 minutes (GKDEV-3655) (plus/integrations)
+
+### Changed
+
+- Changes `listProjects` to publish a Jira project's key as its `name`, where it published the display name, so the item names what the project reads take; the display name is no longer on the item (plus/integrations)
+- Changes `listIssueTrackerIssuesPage`'s `org` and `project` to select Jira sites and projects only by their canonical key (a site's id, a project's key), no longer also by a site's label or a project's name or id: a remembered selection whose site or project was removed selected another one that carried the old value as a name or id. Linear and Trello still match any identifier (plus/integrations)
+- Changes `IssueProviderState.category` to also be `CANCELED`, for abandoned work (canceled, duplicate, removed): `toIssueShape` and `fromProviderIssue` close it as they close `DONE`. Linear's canceled and duplicate states now report `CANCELED` where they reported `DONE`, so a consumer that tests `category === 'DONE'` for a closed issue should test `isTerminalIssueStateCategory` instead (git, plus/integrations)
+- Changes a user-scoped Jira Cloud read of a single project to read that project on its own instead of through the multi-project search, at the same cost, so a failure partway through its pages is attributed to the project rather than only to the site, and its truncation is reported per project (plus/integrations)
+- Changes Azure DevOps work items to close by their state's category as well as their closed date, on the full reads and the cheap etag check alike: a work item in a `Completed`, `Resolved` or `Removed` state with no closed date now reads closed, and its etag changes once. Every work item read now needs its type's states: when that read fails, a project list keeps its rows (the failed type's without a category) as a partial collection with a failure scoped to the project, while a read by id (`getIssuesBatch`) and the cheap etag check fail. A project list keeps only the work items in the states it asked for: provider-apis selects open ones by their dates, so a `Removed` work item with no closed or resolved date is now left out of an open list rather than listed closed, and a closed-only list still doesn't select it. The native work item path asks for the states endpoint with the same `api-version`. Snapshots cached before this change carry no category and can't be reclassified until they are refreshed (GKDEV-3655) (plus/integrations)
+- Changes Jira Data Center projects to carry their project key, which `listProjects` publishes as `name` and `listIssueTrackerIssuesPage` selects by, where they carried the display name (provider-apis now reports the key of `/rest/api/2/project`, GKDEV-3654); a project without one falls back to its numeric id. A Data Center project selection saved as its display name must be picked again, and a page cursor from before this change is best restarted: it re-reads the projects it had completed and no longer recognizes the ones it was retrying (plus/integrations)
+
+### Fixed
+
+- Fixes Jira and Jira Data Center issues in a done status category reading as open (`closed: false`, `state: 'opened'`) on every read but the batch read by key, with no `providerState.category`, so a closed lookup missed them and an open one returned them (gitkraken/kepler#4010). `@gitkraken/provider-apis` 0.65.1 maps the category from the status category's stable `key` rather than its localized `name`, and no longer reads an unrecognized one as `DONE` (GKDEV-3656), so `toIssueShape` now trusts it for Jira as for every other tracker: an issue closes when its category is `done`, and stays open in any other, Jira's "No Category" included, whatever language the instance uses. Every Jira read now reports `providerState.category`, and a Jira Data Center issue in a done category changes its `getIssuesBatch` etag once (plus/integrations)
+- Fixes Jira and Jira Data Center searches failing a whole page when one issue's type hides a field such as Fix versions or Comment, so Stories and Sub-tasks vanished from issue lists while Epics showed: such an issue is now read, one that still can't be read is dropped and reported (next entry) instead of failing its page, and a page whose every issue is unreadable comes back partial instead of ending the drain. By bumping `@gitkraken/provider-apis` to 0.65.0, which also brings the Linear and Jira Data Center changes above (GKDEV-3652, GKDEV-3653) (plus/integrations)
+- Fixes Jira and Jira Data Center issue reads discarding the collection metadata of each search page, so an issue `@gitkraken/provider-apis` drops because it can't read it (GKDEV-3652, e.g. one whose type hides a field the normalizer needed) vanished with no warning. Project reads (`listIssueTrackerIssuesPage`, `IssuesIntegration.getIssuesForProject*`) keep its siblings and now report the drop as a `provider` failure scoped to its project: `fetchFailed`, a warning, and the project `truncated` with recovery `none`. A user-scoped search across projects that hits one is read project by project; on Jira Data Center a search that names a single project is read as that project from the start, so a dropped issue costs no second read. A run of pages that dropped issues is reported once per scope, naming the first page's keys and how many more pages dropped issues, rather than once per page. A page whose every issue was unreadable no longer loses its failure to a later page error, and a project read the page backstop stopped after such a page also reports the backstop. `searchMyIssuesWithTruncationResult` (account-wide, which no facade read reaches for these trackers) carries the same metadata on both hosts; on Jira Cloud each site is drained on its own, so a site whose read failed keeps what it read and is reported as that site's failure (a refused one going through the scoped auth check) instead of being only logged, and a site the page backstop stopped makes the read `truncated`. The metadata comes from `@gitkraken/provider-apis` 0.65.0 ([#5992](https://github.com/gitkraken/vscode-gitlens/issues/5992)) (plus/integrations)
+- Fixes Jira Cloud's per-project issue read addressing the project by its display name, which JQL resolves against project keys first, so a project named like another project's key read that project's issues; it now reads by key, as the multi-project search does, and attributes its failures to the key (plus/integrations)
+- Fixes Azure DevOps repository-scoped issue reads (`listIssuesPage({ repos })`, `getMyIssuesForRepos`) dropping a project read's own collection metadata and truncation, so a project read that came back incomplete was reported as complete (plus/integrations)
+- Fixes the Linear cheap etag check treating a `duplicate` issue as open while the full read closes it, which made the two compute different etags (plus/integrations)
+
+## [0.11.1] - 2026-10-16
+
+### Added
+
+- Adds tip commits to `WorktreeSummary`: `sha` is HEAD's commit (`undefined` before the first commit) and `base.sha` is the base ref's tip, so a host can key on commits rather than counts. A commit and a fast-forward of the base that leave every count the same now change the summary and fire. `getAheadOfBase` answers the tip as `base.sha`, resolved from the cached ref list in git's own order (a tag before a same-named branch), and its `dependsOn` now names that tag too, so the tag moving recomputes the base. Both come from reads the summary already does, except that a recompute of only the base re-reads the shared ref list when the ref move that triggered it cleared that list (git, git-cli)
+- Adds `baseDependsOn` to `WorktreeSummary`: the full ref names whose movement can change the base answer, as `getAheadOfBase` reports them, taken from the summary's own base read and given with or without a base. A host can narrow a watch subscription's `relevantRefs` to it without reading the base again, and a change to the list alone fires (git)
+- Adds `targetKeys` to `ProviderWarning`: a batch read's warning (`getPullRequestsBatch`, `getIssuesBatch`, `getPullRequestsForBranches`) names the keys of the targets it concerns — the target that was refused or failed, or every target of an integration call that failed as a whole — so a caller that merges several callers' targets into one call can hand each failure only to its own targets. A warning without it refused the whole call before any target was read. Equal warnings merge their keys rather than one being dropped (plus/integrations)
+
+### Fixed
+
+- Fixes `getAheadOfBase` answering a stale count against a tag-named base after the tag moved: a tag change now clears the cached commit counts and merge bases, which are keyed by ref name, as a branch move already did (git)
+
 ## [0.11.0] - 2026-10-15
 
 ### Added
@@ -637,7 +675,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/) and this p
 
 - Initial release. Bundles `@gitlens/utils`, `@gitlens/git`, `@gitlens/git-cli`, `@gitlens/ai`, and `@gitlens/git-github` into a single core npm package with subpath exports.
 
-[unreleased]: https://github.com/gitkraken/vscode-gitlens/compare/releases/core/v0.11.0...HEAD
+[unreleased]: https://github.com/gitkraken/vscode-gitlens/compare/releases/core/v0.12.0...HEAD
+[0.12.0]: https://github.com/gitkraken/vscode-gitlens/compare/releases/core/v0.11.1...gitkraken:releases/core/v0.12.0
+[0.11.1]: https://github.com/gitkraken/vscode-gitlens/compare/releases/core/v0.11.0...gitkraken:releases/core/v0.11.1
 [0.11.0]: https://github.com/gitkraken/vscode-gitlens/compare/releases/core/v0.10.7...gitkraken:releases/core/v0.11.0
 [0.10.7]: https://github.com/gitkraken/vscode-gitlens/compare/releases/core/v0.10.6...gitkraken:releases/core/v0.10.7
 [0.10.6]: https://github.com/gitkraken/vscode-gitlens/compare/releases/core/v0.10.5...gitkraken:releases/core/v0.10.6

package/dist/git/cache.js

@@ -453,6 +453,8 @@ var Cache = class Cache {
 				keysToClear.add("commitCount");
 				keysToClear.add("logShas");
 				keysToClear.add("resolvedRevisions");
+				keysToClear.add("leftRightCommitCount");
+				keysToClear.add("mergeBase");
 			}
 			if (types.includes("tracking")) {
 				keysToClear.add("fileExistence");