@gitkraken/core-gitlens 0.9.4 → 0.9.6

package/CHANGELOG.md

@@ -6,6 +6,32 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/) and this p
 
 ## [Unreleased]
 
+## [0.9.6] - 2026-10-02
+
+### Changed
+
+- **Breaking (git)** — `PullRequestRef.owner` and `repo` are typed `string | undefined`, which they already were at runtime for a pull request whose head repository was deleted: GitHub's rows leave both `undefined`, with `exists` `false`, so a consumer's type check now catches code that assumed a string. provider-apis' converters still fill `''` there, so test for a value rather than for `undefined`. Converting such a row to provider-apis' shape now gives no head repository instead of a blank one (git, plus/integrations)
+
+## [0.9.5] - 2026-10-02
+
+### Added
+
+- Adds `viewer` (`{ id, username? }`) to `PullRequestShape`: the identity a row's `authoredByMe` was matched against, present exactly when that was. Match `reviewRequests` and `latestReviews` entries against it rather than against `getCurrentAccount`, which on Azure DevOps Server is a different id than any collection's. Rows are resolved per row through the new `GitHostIntegration.getPullRequestViewers` (#5916) (git, plus/integrations)
+- Adds field presence, so a consumer can tell a field a read didn't fetch from one that is empty. Every pull request and issue row now carries a `projection` tag naming the read that produced it (`point`, `search`, `search-summary`, `text-search`, `account`, `account-summary`, `repos`, `repos-summary`, `batch`, and `project` for tracker project reads), and `getPullRequestFieldPresence(row)` / `getIssueFieldPresence(row)` answer, per field group (description, reviews, review requests, review decision, assignees, checks, mergeability, diff stats, commit count, comments, reactions, access, authored-by-me, stack), whether it was `fetched`, `not-requested` or `unavailable`. A group is `fetched` only when the read selects it and surfaces it faithfully: provider-apis' converters fill a few fields they never fetched (a Bitbucket `mergeableState` of Mergeable, a null GitLab count as 0, an empty Linear label list), and presence reports those as `not-requested`, so ignore such a value. Every provider has a table (GitHub and GitHub Enterprise, GitLab and self-managed, Azure DevOps and Server, Bitbucket Cloud and Data Center, Jira Cloud and Data Center, Linear, Trello), published in the integrations docs. A review decision outside GitHub is derived from reviewer states rather than read from the host, so it reads as `not-requested` there; derive it from `reviews` and `reviewRequests` when needed (plus/integrations)
+- Adds change detection to `getPullRequestsBatch` and `getIssuesBatch`, so a periodic refresh re-reads only what moved. Every fully read row now carries an opaque `etag` — computed by core from the item's change state, not the provider's HTTP ETag; compare it for equality only. A target that sends its `etag` back is checked cheaply first and, when it still matches, comes back `{ key, unchanged: true, etag }` with nothing else fetched. GitHub/GHE (a minimal aliased document) and GitLab, including self-managed (one query per project, instead of one or two requests per merge request or issue), have the cheap check, as do Azure DevOps (work items batched 200 at a time; one request per pull request instead of two, and since Azure DevOps keeps no update time, a pull request etag also carries a 64-bit hash of the title, description, target branch and reviewers with their votes, so it sees those edits but still not a comment), Bitbucket Cloud (one list request per repository per 50 pull requests instead of one request per pull request; Bitbucket Cloud's own converter's rows, which the batch read returns, now carry `isDraft` and carry no mergeability or check rollup, so `'mergeable'` and `'checks'` add nothing there) and, on the tracker form of `getIssuesBatch`, Jira Cloud (one bulk fetch per site instead of one request per key) and Linear (one query per team per 50 issues instead of one request per issue; an issue the query doesn't return, such as one whose team was renamed, falls through to its own full read); every other host and tracker reads in full and etags the row. `getPullRequestsBatch` also takes `etagIncludes`, a list drawn from `'mergeable'`, `'reviewDecision'` and `'checks'` (the check rollup) that widens the etag to those inputs, since GitHub changes them without moving the update time. Each one costs its own fields in the cheap check, so a caller pays only for the ones it lists; on GitHub the review decision is the costliest. `getIssuesBatch` takes `etagIncludes` too: `'reactions'` widens the etag to the thumbs-up count, which GitHub/GHE and GitLab change without moving the update time, and widens nothing where rows carry no real count (Azure DevOps work items, Jira, Linear). Order and repeats don't matter, and an unknown value refuses the call. A call that sends etags makes up to three integration calls instead of one — the cheap check, a full read of the targets without an etag started alongside it, and a full read of the ones that changed — and a call that sends none makes the same single call as before. A target the cheap check fails falls through to the full read, except on an auth, rate-limit or connection failure, which drops it with that warning and `fetchFailed`. An etag from another scheme or another `etagIncludes` set compares unequal and costs a full read, never a false `unchanged`; only the Azure DevOps hash can collide, at about 2^-64 (plus/integrations)
+
+### Changed
+
+- Changes pull request and issue rows to leave a field their read didn't fetch `undefined` instead of filling a placeholder: `mergeableState` is no longer `Unknown` on GitHub rows whose read didn't select it, GitLab's point reads and free-text search, or Bitbucket Data Center's own converter's rows, nor `Mergeable` on Bitbucket Cloud's own converter's rows; `repository.accessLevel` is no longer Write on Azure DevOps, Bitbucket and GitLab rows or any other row converted from provider-apis' shape; `viewerCanUpdate` is no longer `true` on a GitLab row provider-apis sends no permissions for; a GitHub pull request's `thumbsUpCount` is no longer 0 and a review request's `isCodeOwner` no longer `false` where neither was read; an Azure DevOps work item from `getIssuesBatch` no longer reads 0 reactions; and a row's `repository.id` is `undefined`, not `''`, where provider-apis doesn't know it. Field presence reports each as `not-requested` or `unavailable`; a consumer that wants a default applies its own. GitHub's account-wide and batch reads (`account`, `account-summary`, `batch`) now return GitHub's own rows instead of converting them to provider-apis' shape and back, so they keep values they fetched and used to drop: `filesChanged`, `stack`, each review request's `isCodeOwner`, a `viewerCanUpdate` of `false`, `repository.accessLevel` from the viewer's permission, and the base ref's `isFork`, so their `diffStats`, `viewer` and `stack` groups now read `fetched` (`stack` on github.com only, and `account-summary` gains only `stack`). They also keep GitHub's own review decision: on a repository that requires no review, `reviewDecision` is now `undefined`, as on GitHub's other reads, instead of a `ReviewRequired` derived from a pending review request. A pull request whose head repository was deleted reads as GitHub's other reads have it, with `refs.head.exists` `false` and its owner, repository and URL `undefined`, instead of `exists: true` with blanks. GitHub's other reads now set `number` too, and a deleted (ghost) member's `username` is `ghost`. provider-apis' own placeholders remain, listed in the integrations docs (plus/integrations, plus/git-github)
+- Changes the refusal for an unsupported capability to `kind: 'unsupported'` instead of `'other'`, so a consumer can tell a provider that can't serve a request (a surface such as issues on Bitbucket or pull requests on an issue tracker, a batch read, org or repository discovery, a current-account lookup, or a filter, sort, state, search criterion or option combination it can't express) from a transient or unclassified failure without matching `message`. The `ProviderWarningKind` union widened, so an exhaustive `switch` on it needs a new case. `message` wording is unchanged, and malformed or contradictory input (a duplicate batch key, an unusable scope) and incompleteness warnings stay `'other'` (plus/integrations)
+
+### Fixed
+
+- Fixes Azure DevOps and Azure DevOps Server reviewer reads missing every pull request whose reviewer is a group the user is a member of: a team (the usual shape when a branch policy requires one), a security group such as `[project]\Contributors`, a group reached through another, or a group of another project. Azure's `searchCriteria.reviewerId` only matches the identity it names, so the account-wide read (`listPullRequestsPage` and sweeps with `ReviewRequested` or `Assignee`), Azure DevOps Server's filtered search and count, and the legacy `getMyPullRequests` read (Launchpad, which now goes through the account-wide read and so drains every page) read the user's groups once per organization (`_apis/identities` with expanded membership, then `_apis/identitybatch`: two requests for up to 100 groups, kept for five minutes and dropped on re-sync) and read each project's open pull requests once without a reviewer filter, keeping those that name the user, or name one of the user's groups and weren't written by the user. Closed and merged pull requests are still read through Azure's filter for the user, so a group's request is only followed while a pull request is open. A request to one of those groups carries the new `PullRequestReviewer.isMyGroup` on `reviewRequests` (and `latestReviews`), so a consumer can tell a group review from one asked of the user by name. When the groups can't be read (e.g. a personal access token without the Identity scope), the reads keep what names the user and report the organization or collection as a scoped failure (warning, and the search and its count reported incomplete), and the legacy read keeps the user's own reviews ([#5917](https://github.com/gitkraken/vscode-gitlens/issues/5917)) (plus/integrations)
+- Fixes Azure DevOps Server pull requests never being `authoredByMe`. Azure DevOps Server gives one person a different identity id in each collection than at the server level, and the rows were matched against the server-level account, so even the user's own pull requests came back `authoredByMe: false`. `sweepPullRequests`, `listPullRequestsPage` (account-wide and repo-scoped), `getPullRequestsBatch` and `getPullRequestsForBranches` now match each row against the user's id in that row's collection, reusing the lookup the creator/reviewer filters already cache. A collection whose id can't be read leaves its rows' `authoredByMe` (and `viewer`) unset instead of `false`, without failing the read (#5916) (plus/integrations)
+- Fixes a Linear issue that doesn't exist failing `getIssuesBatch` and every other Linear issue read instead of reading as a proven absence. Linear answers an identifier it can't resolve with HTTP 200 and the single GraphQL error `Entity not found: Issue`, and only provider-apis' own `Linear issue not found` message was recognized, so a deleted issue's link could never settle (plus/integrations)
+- Fixes `getPullRequestsForBranches` on GitHub and GitHub Enterprise reporting `truncated` for every branch whose name other forks share (`main`, `master`, `patch-1`, `develop` in a fork-heavy repository), even when every matching pull request was returned, so a consumer that waits for a proven answer never settled. The GraphQL query matches the branch name across every fork and filters by head repository afterwards, so it counted pull requests that could never match. A target it reports truncated now costs one more REST request (`GET /repos/{owner}/{repo}/pulls?head=owner:branch`, the new `GitHubApi.getPullRequestNumbersForBranch`), which filters by head owner on the server; any match the first request missed is resolved through `getPullRequestsBatch`'s own read, so its row and `projection` are the same as the others, and the target is `truncated` only when more than 10 truly match. A target whose REST request fails is dropped with `fetchFailed`, never answered from the first request alone, and a target that wasn't truncated costs nothing more. A pull request from a deleted fork still matches its old owner as `headOwner`, but the REST request can't see one, so on a truncated target such a pull request beyond the first request's page can be missing from an answer that isn't `truncated` (plus/integrations, plus/git-github)
+
 ## [0.9.4] - 2026-09-30
 
 ### Added
@@ -491,7 +517,9 @@ 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/v0.9.4...HEAD
+[unreleased]: https://github.com/gitkraken/vscode-gitlens/compare/releases/core/v0.9.6...HEAD
+[0.9.6]: https://github.com/gitkraken/vscode-gitlens/compare/releases/core/v0.9.5...gitkraken:releases/core/v0.9.6
+[0.9.5]: https://github.com/gitkraken/vscode-gitlens/compare/releases/core/v0.9.4...gitkraken:releases/core/v0.9.5
 [0.9.4]: https://github.com/gitkraken/vscode-gitlens/compare/releases/core/v0.9.3...gitkraken:releases/core/v0.9.4
 [0.9.3]: https://github.com/gitkraken/vscode-gitlens/compare/releases/core/v0.9.2...gitkraken:releases/core/v0.9.3
 [0.9.2]: https://github.com/gitkraken/vscode-gitlens/compare/releases/core/v0.9.1...gitkraken:releases/core/v0.9.2

package/dist/git/models/issue.d.ts

@@ -153,6 +153,23 @@ export interface IssueSearchCapabilities {
    */
   sorts: IssueSorting[];
 }
+/**
+ * The read that produced an issue row, as a key for `getIssueFieldPresence` (`@gitkraken/core-gitlens/plus/integrations`). It names a
+ * READ, not a field set: the same tag can carry different fields on different providers.
+ */
+export type IssueProjection =
+/** A single-item read. */
+'point' |
+/** The filtered search. */
+'search' |
+/** The account-wide list, and the trackers' "my issues". */
+'account' |
+/** A repository-scoped list. */
+'repos' |
+/** A tracker project's issues (`listIssueTrackerIssuesPage`, `getIssuesForProject`). */
+'project' |
+/** The batch read. */
+'batch';
 export interface IssueShape extends IssueOrPullRequest {
   /** `undefined` when the provider can't resolve the author, e.g. a deleted GitHub account */
   author: IssueMember | undefined;
@@ -162,12 +179,17 @@ export interface IssueShape extends IssueOrPullRequest {
   repository?: IssueRepository;
   labels?: IssueLabel[];
   body?: string;
-  /** The syntax used by `body`. When omitted, consumers should treat `body` as Markdown. */
+  /**
+   * The syntax used by `body`. When omitted, consumers should treat `body` as Markdown — except that Azure DevOps
+   * work items and Bitbucket Cloud issues carry HTML with no `bodyFormat` set.
+   */
   bodyFormat?: IssueBodyFormat;
   project?: IssueProject;
   issueType?: string;
   /** An issue can belong to several Jira sprints; Azure reports one iteration. */
   iterations?: IssueIteration[];
+  /** The read that produced this row. Absent means unknown: a legacy row, or a provider or read not yet tagged. */
+  readonly projection?: IssueProjection;
 }
 export declare class Issue implements IssueShape {
   readonly provider: ProviderReference;
@@ -193,8 +215,9 @@ export declare class Issue implements IssueShape {
   readonly providerState?: IssueProviderState | undefined;
   readonly bodyFormat?: IssueBodyFormat | undefined;
   readonly iterations?: IssueIteration[] | undefined;
+  readonly projection?: IssueProjection | undefined;
   readonly type = "issue";
-  constructor(provider: ProviderReference, id: string, nodeId: string | undefined, title: string, url: string, createdDate: Date, updatedDate: Date, closed: boolean, state: IssueOrPullRequestState, author: IssueMember | undefined, assignees: IssueMember[], repository?: IssueRepository | undefined, closedDate?: Date | undefined, labels?: IssueLabel[] | undefined, commentsCount?: number | undefined, thumbsUpCount?: number | undefined, body?: string | undefined, project?: IssueProject | undefined, number?: string | undefined, issueType?: string | undefined, providerState?: IssueProviderState | undefined, bodyFormat?: IssueBodyFormat | undefined, iterations?: IssueIteration[] | undefined);
+  constructor(provider: ProviderReference, id: string, nodeId: string | undefined, title: string, url: string, createdDate: Date, updatedDate: Date, closed: boolean, state: IssueOrPullRequestState, author: IssueMember | undefined, assignees: IssueMember[], repository?: IssueRepository | undefined, closedDate?: Date | undefined, labels?: IssueLabel[] | undefined, commentsCount?: number | undefined, thumbsUpCount?: number | undefined, body?: string | undefined, project?: IssueProject | undefined, number?: string | undefined, issueType?: string | undefined, providerState?: IssueProviderState | undefined, bodyFormat?: IssueBodyFormat | undefined, iterations?: IssueIteration[] | undefined, projection?: IssueProjection | undefined);
   static is(issue: unknown): issue is Issue;
 }
 export type IssueBodyFormat = 'markdown' | 'jira-wiki';

package/dist/git/models/issue.js

@@ -43,8 +43,9 @@ let Issue = _Issue = class Issue {
 	providerState;
 	bodyFormat;
 	iterations;
+	projection;
 	type = "issue";
-	constructor(provider, id, nodeId, title, url, createdDate, updatedDate, closed, state, author, assignees, repository, closedDate, labels, commentsCount, thumbsUpCount, body, project, number, issueType, providerState, bodyFormat, iterations) {
+	constructor(provider, id, nodeId, title, url, createdDate, updatedDate, closed, state, author, assignees, repository, closedDate, labels, commentsCount, thumbsUpCount, body, project, number, issueType, providerState, bodyFormat, iterations, projection) {
 		this.provider = provider;
 		this.id = id;
 		this.nodeId = nodeId;
@@ -68,6 +69,7 @@ let Issue = _Issue = class Issue {
 		this.providerState = providerState;
 		this.bodyFormat = bodyFormat;
 		this.iterations = iterations;
+		this.projection = projection;
 	}
 	static is(issue) {
 		return issue instanceof _Issue;

package/dist/git/models/issue.js.map

@@ -1 +1 @@
-{"version":3,"file":"issue.js","names":[],"sources":["../../../../git/src/models/issue.ts"],"sourcesContent":["import { loggable } from '@gitlens/utils/decorators/log.js';\nimport type { RequireSome, RequireSomeWithProps } from '@gitlens/utils/types.js';\nimport type { IssueOrPullRequest, IssueOrPullRequestState } from './issueOrPullRequest.js';\nimport type { ProviderReference } from './remoteProvider.js';\nimport type { RepositoryIdentityDescriptor } from './repositoryIdentities.js';\n\n/** Selects which issue states a read should include. `all` covers open + closed. */\nexport type IssueStateFilter = 'open' | 'closed' | 'all';\n\n/**\n * One relationship a filtered issue search can constrain on.\n *\n * The last two are USER-INDEPENDENT — they describe the issue, not the caller — and neither is a SCOPE: either\n * one alone matches every such issue on the host, so a search using one still needs a repository/org scope.\n */\nexport type IssueSearchRelationship =\n\t/** Authored by the current user. */\n\t| 'authored'\n\t/** Assigned to the current user. */\n\t| 'assigned'\n\t/** Mentions the current user. */\n\t| 'mentioned'\n\t/** Assigned to ANYONE. Excludes unassigned issues — see {@link IssueSearchCriteria.relationships}. */\n\t| 'any-assignee'\n\t/** Assigned to nobody. The complement of `any-assignee`, not a broadening of it. */\n\t| 'unassigned';\n\n/**\n * A field an issue read can be ordered by, provider-neutral. No provider expresses all of them: each read\n * validates the requested key against the capability its provider declares and refuses one it can't express\n * server-side, rather than serving a differently-ordered list under the name of the one that was asked for.\n *\n * `closed` and `resolved` are deliberately SEPARATE, not two names for the same date. Azure DevOps tracks both\n * and they differ (an issue can be closed without being resolved and resolved without being closed), and Jira\n * has no close date at all — it models the equivalent as `resolutiondate`. Collapsing them would force one of\n * those two providers to lie about which date it ordered by.\n */\nexport type IssueSortField =\n\t/** Creation date. */\n\t| 'created'\n\t/** Last activity. */\n\t| 'updated'\n\t/** Close date. Only where the provider orders by it server-side — notably NOT GitHub. */\n\t| 'closed'\n\t/** Resolution date (Jira `resolutiondate`, Azure `ResolvedDate`). Distinct from `closed`. */\n\t| 'resolved'\n\t/** Number of comments. */\n\t| 'comments'\n\t/** Reactions/upvotes. */\n\t| 'reactions'\n\t/** The provider's own priority field. */\n\t| 'priority'\n\t/** Due date. */\n\t| 'dueDate'\n\t/** Alphabetical by title. */\n\t| 'title';\n\n/**\n * How an issue read is ordered, as `field:direction` — the same shape `BranchSorting`/`TagSorting` already use,\n * so it is one serializable string a consumer can persist, compare, or bind straight to a setting or a dropdown.\n */\nexport type IssueSorting = `${IssueSortField}:asc` | `${IssueSortField}:desc`;\n\n/**\n * The order every issue read that HAS a default applies when the caller asks for none: most recently updated\n * first, which is what those reads served before ordering was an option.\n *\n * One definition rather than one per layer, because it is a promise made in several places at once — the criteria\n * model documents it, the GitHub query emits it, and the result-ceiling warning quotes it — and three copies would\n * be free to disagree about what \"the default\" is.\n *\n * Note it is not a default every read has, and the two layers differ. The API client's account-wide \"my issues\"\n * search (`GitHubApi.searchMyIssues`) never requested an order, so a direct caller omitting `sort` still gets\n * GitHub's relevance order — the default is opt-in there, to keep that read's already-shipped results unchanged.\n * The FACADE does apply it: GitHub declares `updated:desc` among its account-wide keys, so\n * `readAccountWideIssuesPage` resolves the omission to this default and forwards it, and that read's emitted query\n * gained a `sort:updated` qualifier it did not have before. A read whose surface cannot express even this key is\n * the only one the facade leaves in the provider's own order.\n */\nexport const defaultIssueSort: IssueSorting = 'updated:desc';\n\n/**\n * What a filtered issue search narrows on: a provider-neutral criteria set, translated to each provider's own\n * query language by its integration.\n *\n * Structured rather than a provider query string so a consumer can render filter chips and hide the ones a\n * provider can't express, and so the same criteria mean the same thing on a provider whose query language looks\n * nothing like GitHub's search syntax. Every field is validated against the provider's declared capability\n * BEFORE the read runs, all-or-nothing: a criterion the provider can't express server-side refuses the whole\n * read rather than serving a list that was never narrowed. Client-side narrowing is not an option — the dropped\n * items still counted toward the page and cursor the provider produced, so a post-filter would leave the items\n * describing a different result set than the paging that came with them.\n */\nexport interface IssueSearchCriteria {\n\t/**\n\t * Relationships to constrain on, OR-ed together (each becomes its own provider query, unioned and deduped).\n\t * Omitted means NO relationship constraint at all — every issue in scope, assigned or not — which requires a\n\t * repository/org scope.\n\t *\n\t * `any-assignee` and `unassigned` partition the scope between them and are mutually exclusive; passing both is\n\t * refused rather than silently dropping one. Note that \"all visible issues\" is the OMITTED case, NOT\n\t * `any-assignee`, which excludes unassigned issues.\n\t */\n\trelationships?: IssueSearchRelationship[];\n\t/**\n\t * Free text, matched by the provider's own relevance rules — NOT a substring match, and not a way to smuggle\n\t * qualifiers: a token that would read as one is dropped, as are quotes and control characters, so text can\n\t * never re-scope the search. The structured criteria are the qualifier channel.\n\t */\n\ttext?: string;\n\t/** Issue states to include. Omitted reads open issues only. */\n\tstate?: IssueStateFilter;\n\t/** Labels the issue must carry — AND-ed, matching GitHub's own `label:` semantics. */\n\tlabels?: string[];\n\tmilestone?: string;\n\t/** ISO date (`YYYY-MM-DD`). The most effective narrowing criterion by far on a large scope. */\n\tupdatedAfter?: string;\n\t/** ISO date (`YYYY-MM-DD`). */\n\tcreatedAfter?: string;\n\t/** Issues with no linked pull request. */\n\twithoutLinkedPullRequest?: boolean;\n\t/** Includes issues in archived repositories, which are excluded by default. */\n\tincludeArchived?: boolean;\n\t/**\n\t * How to order the results. Omitted means `updated:desc`, which is what this search has always served.\n\t *\n\t * Validated like every other criterion — all-or-nothing against {@link IssueSearchCapabilities.sorts}: a key\n\t * the provider can't express server-side refuses the WHOLE read rather than falling back to the default.\n\t * Serving another order is not a smaller error than serving a wider result: combined with the provider's\n\t * result ceiling it returns a different subset than was asked for, and the paging that comes with it describes\n\t * that other subset.\n\t *\n\t * Do not change it mid-pagination. A cursor carries the sort it was produced under, and threading it under a\n\t * different key is refused rather than serving a sequence with gaps and repeats: drop the cursor instead.\n\t */\n\tsort?: IssueSorting;\n}\n\n/**\n * Which {@link IssueSearchCriteria} fields a provider's filtered issue search can express server-side.\n *\n * An empty `relationships` with every flag false means the provider has NO filtered issue search — the read is\n * refused outright, so a consumer should offer neither the surface nor its chips.\n */\nexport interface IssueSearchCapabilities {\n\t/** Relationships the search can constrain on. Empty means the search itself is unsupported. */\n\trelationships: IssueSearchRelationship[];\n\ttext: boolean;\n\tlabels: boolean;\n\tmilestone: boolean;\n\tupdatedAfter: boolean;\n\tcreatedAfter: boolean;\n\twithoutLinkedPullRequest: boolean;\n\t/** Whether {@link IssueSearchCriteria.state} can select anything other than the provider's default (open). */\n\tstates: boolean;\n\t/**\n\t * Sort keys the search can express IN SERVER. Always contains at least `updated:desc` — the historical default\n\t * — when the search exists at all, so it is never empty for a usable surface; a provider without a filtered\n\t * search reports an empty `relationships`, which is already the signal that there is no surface to order.\n\t */\n\tsorts: IssueSorting[];\n}\n\nexport interface IssueShape extends IssueOrPullRequest {\n\t/** `undefined` when the provider can't resolve the author, e.g. a deleted GitHub account */\n\tauthor: IssueMember | undefined;\n\tassignees: IssueMember[];\n\t/** The provider's own workflow state, when it exposes one beyond the normalized open/closed state. */\n\tproviderState?: IssueProviderState;\n\trepository?: IssueRepository;\n\tlabels?: IssueLabel[];\n\tbody?: string;\n\t/** The syntax used by `body`. When omitted, consumers should treat `body` as Markdown. */\n\tbodyFormat?: IssueBodyFormat;\n\tproject?: IssueProject;\n\tissueType?: string;\n\t/** An issue can belong to several Jira sprints; Azure reports one iteration. */\n\titerations?: IssueIteration[];\n}\n\n@loggable(i => i.id)\nexport class Issue implements IssueShape {\n\treadonly type = 'issue';\n\n\tconstructor(\n\t\tpublic readonly provider: ProviderReference,\n\t\tpublic readonly id: string,\n\t\tpublic readonly nodeId: string | undefined,\n\t\tpublic readonly title: string,\n\t\tpublic readonly url: string,\n\t\tpublic readonly createdDate: Date,\n\t\tpublic readonly updatedDate: Date,\n\t\tpublic readonly closed: boolean,\n\t\tpublic readonly state: IssueOrPullRequestState,\n\t\tpublic readonly author: IssueMember | undefined,\n\t\tpublic readonly assignees: IssueMember[],\n\t\tpublic readonly repository?: IssueRepository,\n\t\tpublic readonly closedDate?: Date,\n\t\tpublic readonly labels?: IssueLabel[],\n\t\tpublic readonly commentsCount?: number,\n\t\tpublic readonly thumbsUpCount?: number,\n\t\tpublic readonly body?: string,\n\t\tpublic readonly project?: IssueProject,\n\t\tpublic readonly number?: string,\n\t\tpublic readonly issueType?: string,\n\t\tpublic readonly providerState?: IssueProviderState,\n\t\tpublic readonly bodyFormat?: IssueBodyFormat,\n\t\tpublic readonly iterations?: IssueIteration[],\n\t) {}\n\n\tstatic is(issue: unknown): issue is Issue {\n\t\treturn issue instanceof Issue;\n\t}\n}\n\nexport type IssueBodyFormat = 'markdown' | 'jira-wiki';\n\nexport const enum RepositoryAccessLevel {\n\tAdmin = 100,\n\tMaintain = 40,\n\tWrite = 30,\n\tTriage = 20,\n\tRead = 10,\n\tNone = 0,\n}\n\nexport interface IssueLabel {\n\tcolor?: string;\n\tname: string;\n}\n\n/**\n * Sprint metadata is optional because some providers only report an iteration path. An absent `isActive`\n * means unknown, not inactive, so consumers must not use it to exclude those iterations by default.\n */\nexport interface IssueIteration {\n\t/**\n\t * A sprint id on Jira, the verbatim iteration path on Azure — so it is only meaningful within the provider's\n\t * project, never across providers. Unique at read time but NOT durable: Azure rewrites the path when an\n\t * iteration node is renamed or re-parented, so don't persist it as a long-lived id.\n\t */\n\tid: string;\n\tname: string;\n\tisActive?: boolean;\n\tstartDate?: Date;\n\tendDate?: Date;\n}\n\nexport interface IssueMember {\n\tid: string;\n\t/**\n\t * Absent when the provider exposes no display name for the member. Optional on purpose: a fallback string\n\t * invented here (`'unknown'`, `''`) can't be told apart from a real name downstream, so it can't be undone by a\n\t * consumer that needs to render something else — or nothing at all, as when the member feeds an AI prompt.\n\t * Each consumer picks its own presentation.\n\t */\n\tname?: string;\n\t/** The provider's handle for this person, when it has one — GitHub's login, Azure's `uniqueName` (a UPN,\n\t *  so an email), Bitbucket's mutable `nickname`. Display/labelling only: it is neither guaranteed present\n\t *  (GitLab's native mapper has none) nor a stable identity, so never key a match off it. */\n\tusername?: string;\n\tavatarUrl?: string;\n\turl?: string;\n}\n\nexport type IssueProviderState = {\n\tid?: string;\n\tname: string;\n\tcolor?: string;\n\tcategory?: 'TO_DO' | 'IN_PROGRESS' | 'DONE';\n};\n\nexport interface IssueProject {\n\tid: string;\n\tname: string;\n\tresourceId: string;\n\tresourceName: string;\n}\n\nexport interface IssueRepository {\n\towner: string;\n\trepo: string;\n\taccessLevel?: RepositoryAccessLevel;\n\turl?: string;\n\tid?: string;\n}\n\nexport type IssueRepositoryIdentityDescriptor = RequireSomeWithProps<\n\tRequireSome<RepositoryIdentityDescriptor<string>, 'provider'>,\n\t'provider',\n\t'id' | 'domain' | 'repoDomain' | 'repoName'\n> &\n\tRequireSomeWithProps<RequireSome<RepositoryIdentityDescriptor<string>, 'remote'>, 'remote', 'domain'>;\n"],"mappings":";;;;;;;;;;;;;;;;;;;;AA+EA,MAAa,mBAAiC;AAsGvC,IAAM,QAAA,SAAN,MAAM,MAA4B;CAIvB;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CAzBjB,OAAgB;CAEhB,YACC,UACA,IACA,QACA,OACA,KACA,aACA,aACA,QACA,OACA,QACA,WACA,YACA,YACA,QACA,eACA,eACA,MACA,SACA,QACA,WACA,eACA,YACA,YACC;EAvBe,KAAA,WAAA;EACA,KAAA,KAAA;EACA,KAAA,SAAA;EACA,KAAA,QAAA;EACA,KAAA,MAAA;EACA,KAAA,cAAA;EACA,KAAA,cAAA;EACA,KAAA,SAAA;EACA,KAAA,QAAA;EACA,KAAA,SAAA;EACA,KAAA,YAAA;EACA,KAAA,aAAA;EACA,KAAA,aAAA;EACA,KAAA,SAAA;EACA,KAAA,gBAAA;EACA,KAAA,gBAAA;EACA,KAAA,OAAA;EACA,KAAA,UAAA;EACA,KAAA,SAAA;EACA,KAAA,YAAA;EACA,KAAA,gBAAA;EACA,KAAA,aAAA;EACA,KAAA,aAAA;CACd;CAEH,OAAO,GAAG,OAAgC;EACzC,OAAO,iBAAA;CACR;AACD;AAjCC,QAAA,SAAA,WAAA,CAAA,UAAS,MAAK,EAAE,EAAE,CAAA,GAAA,KAAA;AAqCnB,IAAkB,wBAAX,yBAAA,uBAAA;CACN,sBAAA,sBAAA,WAAA,OAAA;CACA,sBAAA,sBAAA,cAAA,MAAA;CACA,sBAAA,sBAAA,WAAA,MAAA;CACA,sBAAA,sBAAA,YAAA,MAAA;CACA,sBAAA,sBAAA,UAAA,MAAA;CACA,sBAAA,sBAAA,UAAA,KAAA;;AACD,EAAA,CAAA,CAAA"}
+{"version":3,"file":"issue.js","names":[],"sources":["../../../../git/src/models/issue.ts"],"sourcesContent":["import { loggable } from '@gitlens/utils/decorators/log.js';\nimport type { RequireSome, RequireSomeWithProps } from '@gitlens/utils/types.js';\nimport type { IssueOrPullRequest, IssueOrPullRequestState } from './issueOrPullRequest.js';\nimport type { ProviderReference } from './remoteProvider.js';\nimport type { RepositoryIdentityDescriptor } from './repositoryIdentities.js';\n\n/** Selects which issue states a read should include. `all` covers open + closed. */\nexport type IssueStateFilter = 'open' | 'closed' | 'all';\n\n/**\n * One relationship a filtered issue search can constrain on.\n *\n * The last two are USER-INDEPENDENT — they describe the issue, not the caller — and neither is a SCOPE: either\n * one alone matches every such issue on the host, so a search using one still needs a repository/org scope.\n */\nexport type IssueSearchRelationship =\n\t/** Authored by the current user. */\n\t| 'authored'\n\t/** Assigned to the current user. */\n\t| 'assigned'\n\t/** Mentions the current user. */\n\t| 'mentioned'\n\t/** Assigned to ANYONE. Excludes unassigned issues — see {@link IssueSearchCriteria.relationships}. */\n\t| 'any-assignee'\n\t/** Assigned to nobody. The complement of `any-assignee`, not a broadening of it. */\n\t| 'unassigned';\n\n/**\n * A field an issue read can be ordered by, provider-neutral. No provider expresses all of them: each read\n * validates the requested key against the capability its provider declares and refuses one it can't express\n * server-side, rather than serving a differently-ordered list under the name of the one that was asked for.\n *\n * `closed` and `resolved` are deliberately SEPARATE, not two names for the same date. Azure DevOps tracks both\n * and they differ (an issue can be closed without being resolved and resolved without being closed), and Jira\n * has no close date at all — it models the equivalent as `resolutiondate`. Collapsing them would force one of\n * those two providers to lie about which date it ordered by.\n */\nexport type IssueSortField =\n\t/** Creation date. */\n\t| 'created'\n\t/** Last activity. */\n\t| 'updated'\n\t/** Close date. Only where the provider orders by it server-side — notably NOT GitHub. */\n\t| 'closed'\n\t/** Resolution date (Jira `resolutiondate`, Azure `ResolvedDate`). Distinct from `closed`. */\n\t| 'resolved'\n\t/** Number of comments. */\n\t| 'comments'\n\t/** Reactions/upvotes. */\n\t| 'reactions'\n\t/** The provider's own priority field. */\n\t| 'priority'\n\t/** Due date. */\n\t| 'dueDate'\n\t/** Alphabetical by title. */\n\t| 'title';\n\n/**\n * How an issue read is ordered, as `field:direction` — the same shape `BranchSorting`/`TagSorting` already use,\n * so it is one serializable string a consumer can persist, compare, or bind straight to a setting or a dropdown.\n */\nexport type IssueSorting = `${IssueSortField}:asc` | `${IssueSortField}:desc`;\n\n/**\n * The order every issue read that HAS a default applies when the caller asks for none: most recently updated\n * first, which is what those reads served before ordering was an option.\n *\n * One definition rather than one per layer, because it is a promise made in several places at once — the criteria\n * model documents it, the GitHub query emits it, and the result-ceiling warning quotes it — and three copies would\n * be free to disagree about what \"the default\" is.\n *\n * Note it is not a default every read has, and the two layers differ. The API client's account-wide \"my issues\"\n * search (`GitHubApi.searchMyIssues`) never requested an order, so a direct caller omitting `sort` still gets\n * GitHub's relevance order — the default is opt-in there, to keep that read's already-shipped results unchanged.\n * The FACADE does apply it: GitHub declares `updated:desc` among its account-wide keys, so\n * `readAccountWideIssuesPage` resolves the omission to this default and forwards it, and that read's emitted query\n * gained a `sort:updated` qualifier it did not have before. A read whose surface cannot express even this key is\n * the only one the facade leaves in the provider's own order.\n */\nexport const defaultIssueSort: IssueSorting = 'updated:desc';\n\n/**\n * What a filtered issue search narrows on: a provider-neutral criteria set, translated to each provider's own\n * query language by its integration.\n *\n * Structured rather than a provider query string so a consumer can render filter chips and hide the ones a\n * provider can't express, and so the same criteria mean the same thing on a provider whose query language looks\n * nothing like GitHub's search syntax. Every field is validated against the provider's declared capability\n * BEFORE the read runs, all-or-nothing: a criterion the provider can't express server-side refuses the whole\n * read rather than serving a list that was never narrowed. Client-side narrowing is not an option — the dropped\n * items still counted toward the page and cursor the provider produced, so a post-filter would leave the items\n * describing a different result set than the paging that came with them.\n */\nexport interface IssueSearchCriteria {\n\t/**\n\t * Relationships to constrain on, OR-ed together (each becomes its own provider query, unioned and deduped).\n\t * Omitted means NO relationship constraint at all — every issue in scope, assigned or not — which requires a\n\t * repository/org scope.\n\t *\n\t * `any-assignee` and `unassigned` partition the scope between them and are mutually exclusive; passing both is\n\t * refused rather than silently dropping one. Note that \"all visible issues\" is the OMITTED case, NOT\n\t * `any-assignee`, which excludes unassigned issues.\n\t */\n\trelationships?: IssueSearchRelationship[];\n\t/**\n\t * Free text, matched by the provider's own relevance rules — NOT a substring match, and not a way to smuggle\n\t * qualifiers: a token that would read as one is dropped, as are quotes and control characters, so text can\n\t * never re-scope the search. The structured criteria are the qualifier channel.\n\t */\n\ttext?: string;\n\t/** Issue states to include. Omitted reads open issues only. */\n\tstate?: IssueStateFilter;\n\t/** Labels the issue must carry — AND-ed, matching GitHub's own `label:` semantics. */\n\tlabels?: string[];\n\tmilestone?: string;\n\t/** ISO date (`YYYY-MM-DD`). The most effective narrowing criterion by far on a large scope. */\n\tupdatedAfter?: string;\n\t/** ISO date (`YYYY-MM-DD`). */\n\tcreatedAfter?: string;\n\t/** Issues with no linked pull request. */\n\twithoutLinkedPullRequest?: boolean;\n\t/** Includes issues in archived repositories, which are excluded by default. */\n\tincludeArchived?: boolean;\n\t/**\n\t * How to order the results. Omitted means `updated:desc`, which is what this search has always served.\n\t *\n\t * Validated like every other criterion — all-or-nothing against {@link IssueSearchCapabilities.sorts}: a key\n\t * the provider can't express server-side refuses the WHOLE read rather than falling back to the default.\n\t * Serving another order is not a smaller error than serving a wider result: combined with the provider's\n\t * result ceiling it returns a different subset than was asked for, and the paging that comes with it describes\n\t * that other subset.\n\t *\n\t * Do not change it mid-pagination. A cursor carries the sort it was produced under, and threading it under a\n\t * different key is refused rather than serving a sequence with gaps and repeats: drop the cursor instead.\n\t */\n\tsort?: IssueSorting;\n}\n\n/**\n * Which {@link IssueSearchCriteria} fields a provider's filtered issue search can express server-side.\n *\n * An empty `relationships` with every flag false means the provider has NO filtered issue search — the read is\n * refused outright, so a consumer should offer neither the surface nor its chips.\n */\nexport interface IssueSearchCapabilities {\n\t/** Relationships the search can constrain on. Empty means the search itself is unsupported. */\n\trelationships: IssueSearchRelationship[];\n\ttext: boolean;\n\tlabels: boolean;\n\tmilestone: boolean;\n\tupdatedAfter: boolean;\n\tcreatedAfter: boolean;\n\twithoutLinkedPullRequest: boolean;\n\t/** Whether {@link IssueSearchCriteria.state} can select anything other than the provider's default (open). */\n\tstates: boolean;\n\t/**\n\t * Sort keys the search can express IN SERVER. Always contains at least `updated:desc` — the historical default\n\t * — when the search exists at all, so it is never empty for a usable surface; a provider without a filtered\n\t * search reports an empty `relationships`, which is already the signal that there is no surface to order.\n\t */\n\tsorts: IssueSorting[];\n}\n\n/**\n * The read that produced an issue row, as a key for `getIssueFieldPresence` (`@gitlens/integrations`). It names a\n * READ, not a field set: the same tag can carry different fields on different providers.\n */\nexport type IssueProjection =\n\t/** A single-item read. */\n\t| 'point'\n\t/** The filtered search. */\n\t| 'search'\n\t/** The account-wide list, and the trackers' \"my issues\". */\n\t| 'account'\n\t/** A repository-scoped list. */\n\t| 'repos'\n\t/** A tracker project's issues (`listIssueTrackerIssuesPage`, `getIssuesForProject`). */\n\t| 'project'\n\t/** The batch read. */\n\t| 'batch';\n\nexport interface IssueShape extends IssueOrPullRequest {\n\t/** `undefined` when the provider can't resolve the author, e.g. a deleted GitHub account */\n\tauthor: IssueMember | undefined;\n\tassignees: IssueMember[];\n\t/** The provider's own workflow state, when it exposes one beyond the normalized open/closed state. */\n\tproviderState?: IssueProviderState;\n\trepository?: IssueRepository;\n\tlabels?: IssueLabel[];\n\tbody?: string;\n\t/**\n\t * The syntax used by `body`. When omitted, consumers should treat `body` as Markdown — except that Azure DevOps\n\t * work items and Bitbucket Cloud issues carry HTML with no `bodyFormat` set.\n\t */\n\tbodyFormat?: IssueBodyFormat;\n\tproject?: IssueProject;\n\tissueType?: string;\n\t/** An issue can belong to several Jira sprints; Azure reports one iteration. */\n\titerations?: IssueIteration[];\n\t/** The read that produced this row. Absent means unknown: a legacy row, or a provider or read not yet tagged. */\n\treadonly projection?: IssueProjection;\n}\n\n@loggable(i => i.id)\nexport class Issue implements IssueShape {\n\treadonly type = 'issue';\n\n\tconstructor(\n\t\tpublic readonly provider: ProviderReference,\n\t\tpublic readonly id: string,\n\t\tpublic readonly nodeId: string | undefined,\n\t\tpublic readonly title: string,\n\t\tpublic readonly url: string,\n\t\tpublic readonly createdDate: Date,\n\t\tpublic readonly updatedDate: Date,\n\t\tpublic readonly closed: boolean,\n\t\tpublic readonly state: IssueOrPullRequestState,\n\t\tpublic readonly author: IssueMember | undefined,\n\t\tpublic readonly assignees: IssueMember[],\n\t\tpublic readonly repository?: IssueRepository,\n\t\tpublic readonly closedDate?: Date,\n\t\tpublic readonly labels?: IssueLabel[],\n\t\tpublic readonly commentsCount?: number,\n\t\tpublic readonly thumbsUpCount?: number,\n\t\tpublic readonly body?: string,\n\t\tpublic readonly project?: IssueProject,\n\t\tpublic readonly number?: string,\n\t\tpublic readonly issueType?: string,\n\t\tpublic readonly providerState?: IssueProviderState,\n\t\tpublic readonly bodyFormat?: IssueBodyFormat,\n\t\tpublic readonly iterations?: IssueIteration[],\n\t\tpublic readonly projection?: IssueProjection,\n\t) {}\n\n\tstatic is(issue: unknown): issue is Issue {\n\t\treturn issue instanceof Issue;\n\t}\n}\n\nexport type IssueBodyFormat = 'markdown' | 'jira-wiki';\n\nexport const enum RepositoryAccessLevel {\n\tAdmin = 100,\n\tMaintain = 40,\n\tWrite = 30,\n\tTriage = 20,\n\tRead = 10,\n\tNone = 0,\n}\n\nexport interface IssueLabel {\n\tcolor?: string;\n\tname: string;\n}\n\n/**\n * Sprint metadata is optional because some providers only report an iteration path. An absent `isActive`\n * means unknown, not inactive, so consumers must not use it to exclude those iterations by default.\n */\nexport interface IssueIteration {\n\t/**\n\t * A sprint id on Jira, the verbatim iteration path on Azure — so it is only meaningful within the provider's\n\t * project, never across providers. Unique at read time but NOT durable: Azure rewrites the path when an\n\t * iteration node is renamed or re-parented, so don't persist it as a long-lived id.\n\t */\n\tid: string;\n\tname: string;\n\tisActive?: boolean;\n\tstartDate?: Date;\n\tendDate?: Date;\n}\n\nexport interface IssueMember {\n\tid: string;\n\t/**\n\t * Absent when the provider exposes no display name for the member. Optional on purpose: a fallback string\n\t * invented here (`'unknown'`, `''`) can't be told apart from a real name downstream, so it can't be undone by a\n\t * consumer that needs to render something else — or nothing at all, as when the member feeds an AI prompt.\n\t * Each consumer picks its own presentation.\n\t */\n\tname?: string;\n\t/** The provider's handle for this person, when it has one — GitHub's login, Azure's `uniqueName` (a UPN,\n\t *  so an email), Bitbucket's mutable `nickname`. Display/labelling only: it is neither guaranteed present\n\t *  (GitLab's native mapper has none) nor a stable identity, so never key a match off it. */\n\tusername?: string;\n\tavatarUrl?: string;\n\turl?: string;\n}\n\nexport type IssueProviderState = {\n\tid?: string;\n\tname: string;\n\tcolor?: string;\n\tcategory?: 'TO_DO' | 'IN_PROGRESS' | 'DONE';\n};\n\nexport interface IssueProject {\n\tid: string;\n\tname: string;\n\tresourceId: string;\n\tresourceName: string;\n}\n\nexport interface IssueRepository {\n\towner: string;\n\trepo: string;\n\taccessLevel?: RepositoryAccessLevel;\n\turl?: string;\n\tid?: string;\n}\n\nexport type IssueRepositoryIdentityDescriptor = RequireSomeWithProps<\n\tRequireSome<RepositoryIdentityDescriptor<string>, 'provider'>,\n\t'provider',\n\t'id' | 'domain' | 'repoDomain' | 'repoName'\n> &\n\tRequireSomeWithProps<RequireSome<RepositoryIdentityDescriptor<string>, 'remote'>, 'remote', 'domain'>;\n"],"mappings":";;;;;;;;;;;;;;;;;;;;AA+EA,MAAa,mBAAiC;AA6HvC,IAAM,QAAA,SAAN,MAAM,MAA4B;CAIvB;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CA1BjB,OAAgB;CAEhB,YACC,UACA,IACA,QACA,OACA,KACA,aACA,aACA,QACA,OACA,QACA,WACA,YACA,YACA,QACA,eACA,eACA,MACA,SACA,QACA,WACA,eACA,YACA,YACA,YACC;EAxBe,KAAA,WAAA;EACA,KAAA,KAAA;EACA,KAAA,SAAA;EACA,KAAA,QAAA;EACA,KAAA,MAAA;EACA,KAAA,cAAA;EACA,KAAA,cAAA;EACA,KAAA,SAAA;EACA,KAAA,QAAA;EACA,KAAA,SAAA;EACA,KAAA,YAAA;EACA,KAAA,aAAA;EACA,KAAA,aAAA;EACA,KAAA,SAAA;EACA,KAAA,gBAAA;EACA,KAAA,gBAAA;EACA,KAAA,OAAA;EACA,KAAA,UAAA;EACA,KAAA,SAAA;EACA,KAAA,YAAA;EACA,KAAA,gBAAA;EACA,KAAA,aAAA;EACA,KAAA,aAAA;EACA,KAAA,aAAA;CACd;CAEH,OAAO,GAAG,OAAgC;EACzC,OAAO,iBAAA;CACR;AACD;AAlCC,QAAA,SAAA,WAAA,CAAA,UAAS,MAAK,EAAE,EAAE,CAAA,GAAA,KAAA;AAsCnB,IAAkB,wBAAX,yBAAA,uBAAA;CACN,sBAAA,sBAAA,WAAA,OAAA;CACA,sBAAA,sBAAA,cAAA,MAAA;CACA,sBAAA,sBAAA,WAAA,MAAA;CACA,sBAAA,sBAAA,YAAA,MAAA;CACA,sBAAA,sBAAA,UAAA,MAAA;CACA,sBAAA,sBAAA,UAAA,KAAA;;AACD,EAAA,CAAA,CAAA"}

package/dist/git/models/pullRequest.d.ts

@@ -4,6 +4,29 @@ import { ProviderReference } from "./remoteProvider.js";
 import { IssueOrPullRequest, IssueOrPullRequestState } from "./issueOrPullRequest.js";
 import { IssueProject, IssueRepository } from "./issue.js";
 //#region ../git/src/models/pullRequest.d.ts
+/**
+ * The read that produced a pull request row, as a key for `getPullRequestFieldPresence` (`@gitkraken/core-gitlens/plus/integrations`). It
+ * names a READ, not a field set: the same tag can carry different fields on different providers.
+ */
+export type PullRequestProjection =
+/** A single-item read: by number, for a branch, or for a commit. */
+'point' |
+/** The filtered search with the full projection, and the host's "my pull requests" search. */
+'search' |
+/** The filtered search with `summary`. */
+'search-summary' |
+/** The free-text search (`searchPullRequests`). */
+'text-search' |
+/** An account-wide sweep with the full projection (`includeReviews`). */
+'account' |
+/** An account-wide list or sweep with the lightweight projection. */
+'account-summary' |
+/** A repository-scoped list or sweep. */
+'repos' |
+/** A repository-scoped list with `summary`. */
+'repos-summary' |
+/** The coordinate and branch batch reads. */
+'batch';
 export interface PullRequestShape extends IssueOrPullRequest {
   readonly author: PullRequestMember;
   /**
@@ -24,10 +47,9 @@ export interface PullRequestShape extends IssueOrPullRequest {
   readonly reviewRequests?: PullRequestReviewer[];
   /**
    * Reviews already submitted, as opposed to the still-pending {@link reviewRequests}. Absent means the read
-   * didn't fetch reviews; empty means nobody has reviewed. Those are different answers and neither substitutes
-   * for the other, so which one you get is provider- and read-dependent: GitHub/GHE carry reviews on the
-   * repo-scoped read and the filtered search, and on the account-wide read only where `includeReviews` opts
-   * into the full projection, while GitLab never populates the field at all.
+   * didn't fetch reviews; empty means nobody has reviewed — but only where the read actually fetched them, which
+   * is provider- and read-dependent: ask `getPullRequestFieldPresence` (`@gitkraken/core-gitlens/plus/integrations`) for the row's
+   * `reviews` group rather than trusting an empty array.
    * {@link PullRequestReviewer.commitOid} is the narrower promise: only the full GitHub/GHE projection has it.
    */
   readonly latestReviews?: PullRequestReviewer[];
@@ -36,6 +58,19 @@ export interface PullRequestShape extends IssueOrPullRequest {
   readonly stack?: PullRequestStackInfo;
   readonly number?: number;
   readonly authoredByMe?: boolean;
+  /**
+   * The current account as this pull request's members identify it: the identity {@link authoredByMe} was matched
+   * against, present exactly when that was. A consumer matching {@link reviewRequests} or {@link latestReviews}
+   * entries to the current user compares against this, not the account `getCurrentAccount` returns: Azure DevOps
+   * Server gives one person a different id in each collection than at the server level, and the members carry the
+   * collection's. Core matches `username` too only on GitHub and GitHub Enterprise, whose rows key people by login.
+   */
+  readonly viewer?: {
+    readonly id: string;
+    readonly username?: string;
+  };
+  /** The read that produced this row. Absent means unknown: a legacy row, or a provider or read not yet tagged. */
+  readonly projection?: PullRequestProjection;
 }
 export declare class PullRequest implements PullRequestShape {
   readonly provider: ProviderReference;
@@ -74,13 +109,21 @@ export declare class PullRequest implements PullRequestShape {
   readonly body?: string | undefined;
   readonly number?: number | undefined;
   readonly authoredByMe?: boolean | undefined;
+  readonly projection?: PullRequestProjection | undefined;
+  readonly viewer?: {
+    readonly id: string;
+    readonly username?: string;
+  } | undefined;
   readonly type = "pullrequest";
   constructor(provider: ProviderReference, author: PullRequestMember, id: string, nodeId: string | undefined, title: string, url: string, repository: IssueRepository, state: IssueOrPullRequestState, createdDate: Date, updatedDate: Date, closedDate?: Date | undefined, mergedDate?: Date | undefined, mergeableState?: PullRequestMergeableState | undefined, viewerCanUpdate?: boolean | undefined, refs?: PullRequestRefs | undefined, isDraft?: boolean | undefined, additions?: number | undefined, deletions?: number | undefined, commentsCount?: number | undefined, thumbsUpCount?: number | undefined, reviewDecision?: PullRequestReviewDecision | undefined, reviewRequests?: PullRequestReviewer[] | undefined, latestReviews?: PullRequestReviewer[] | undefined, assignees?: PullRequestMember[] | undefined, statusCheckRollupState?: PullRequestStatusCheckRollupState | undefined, project?: IssueProject | undefined, version?: number | undefined,
   /** Commits on the head that aren't on the base. Appended rather than slotted with the other counts:
    *  this constructor is positional and every mapper passes through it. */
   commitCount?: number | undefined, stack?: PullRequestStackInfo | undefined, filesChanged?: number | undefined,
   /** The pull request's description, as markdown. */
-  body?: string | undefined, number?: number | undefined, authoredByMe?: boolean | undefined);
+  body?: string | undefined, number?: number | undefined, authoredByMe?: boolean | undefined, projection?: PullRequestProjection | undefined, viewer?: {
+    readonly id: string;
+    readonly username?: string;
+  } | undefined);
   get closed(): boolean;
   static is(pr: unknown): pr is PullRequest;
   static formatDate(pr: PullRequestShape, format?: string | null): string;
@@ -260,8 +303,13 @@ export interface PullRequestSearchCapabilities {
   sorts: PullRequestSorting[];
 }
 export interface PullRequestRef {
-  owner: string;
-  repo: string;
+  /**
+   * `undefined` when the ref's repository was deleted (a removed fork): GitHub's rows say so, with `exists` false.
+   * provider-apis' converters fill `''` instead, so test for a value rather than for `undefined`.
+   */
+  owner: string | undefined;
+  /** `undefined` (or `''`) when the ref's repository was deleted, as {@link owner}. */
+  repo: string | undefined;
   branch: string;
   sha: string;
   exists: boolean;
@@ -315,6 +363,13 @@ export interface PullRequestReviewer {
    * repo-scoped read carries natively.
    */
   commitOid?: string;
+  /**
+   * The reviewer is a group the current user is a member of (a team or another group, directly or through a group
+   * nested in it) rather than a person: a request to it is one the user can answer as a member, not one addressed to
+   * them by name. Only the Azure DevOps reads that resolve the user's groups set it; absent elsewhere, including for
+   * groups a read could not tell the user belongs to.
+   */
+  isMyGroup?: boolean;
 }
 export type PullRequestRepositoryIdentityDescriptor = RequireSomeWithProps<RequireSome<RepositoryIdentityDescriptor<string>, 'provider'>, 'provider', 'id' | 'domain' | 'repoDomain' | 'repoName'> & RequireSomeWithProps<RequireSome<RepositoryIdentityDescriptor<string>, 'remote'>, 'remote', 'domain'>;
 //#endregion

package/dist/git/models/pullRequest.js

@@ -38,8 +38,10 @@ let PullRequest = _PullRequest = class PullRequest {
 	body;
 	number;
 	authoredByMe;
+	projection;
+	viewer;
 	type = "pullrequest";
-	constructor(provider, author, id, nodeId, title, url, repository, state, createdDate, updatedDate, closedDate, mergedDate, mergeableState, viewerCanUpdate, refs, isDraft, additions, deletions, commentsCount, thumbsUpCount, reviewDecision, reviewRequests, latestReviews, assignees, statusCheckRollupState, project, version, commitCount, stack, filesChanged, body, number, authoredByMe) {
+	constructor(provider, author, id, nodeId, title, url, repository, state, createdDate, updatedDate, closedDate, mergedDate, mergeableState, viewerCanUpdate, refs, isDraft, additions, deletions, commentsCount, thumbsUpCount, reviewDecision, reviewRequests, latestReviews, assignees, statusCheckRollupState, project, version, commitCount, stack, filesChanged, body, number, authoredByMe, projection, viewer) {
 		this.provider = provider;
 		this.author = author;
 		this.id = id;
@@ -73,6 +75,8 @@ let PullRequest = _PullRequest = class PullRequest {
 		this.body = body;
 		this.number = number;
 		this.authoredByMe = authoredByMe;
+		this.projection = projection;
+		this.viewer = viewer;
 	}
 	get closed() {
 		return this.state === "closed";

package/dist/git/models/pullRequest.js.map

@@ -1 +1 @@
-{"version":3,"file":"pullRequest.js","names":[],"sources":["../../../../git/src/models/pullRequest.ts"],"sourcesContent":["import { formatDate, fromNow } from '@gitlens/utils/date.js';\nimport { loggable } from '@gitlens/utils/decorators/log.js';\nimport { serializable } from '@gitlens/utils/decorators/serializable.js';\nimport type { RequireSome, RequireSomeWithProps } from '@gitlens/utils/types.js';\nimport type { IssueProject, IssueRepository } from './issue.js';\nimport type { IssueOrPullRequest, IssueOrPullRequestState as PullRequestState } from './issueOrPullRequest.js';\nimport type { ProviderReference } from './remoteProvider.js';\nimport type { RepositoryIdentityDescriptor } from './repositoryIdentities.js';\n\nexport type { PullRequestState };\n\nexport interface PullRequestShape extends IssueOrPullRequest {\n\treadonly author: PullRequestMember;\n\t/**\n\t * Canonical base-repository identity returned by the provider. Some providers\n\t * scope pull-request ids to a repository, so consumers must not have to infer\n\t * identity from optional refs or URLs.\n\t */\n\treadonly repository?: IssueRepository;\n\treadonly mergedDate?: Date;\n\treadonly refs?: PullRequestRefs;\n\treadonly isDraft?: boolean;\n\treadonly additions?: number;\n\treadonly deletions?: number;\n\treadonly filesChanged?: number;\n\treadonly body?: string;\n\treadonly mergeableState?: PullRequestMergeableState;\n\treadonly reviewDecision?: PullRequestReviewDecision;\n\treadonly reviewRequests?: PullRequestReviewer[];\n\t/**\n\t * Reviews already submitted, as opposed to the still-pending {@link reviewRequests}. Absent means the read\n\t * didn't fetch reviews; empty means nobody has reviewed. Those are different answers and neither substitutes\n\t * for the other, so which one you get is provider- and read-dependent: GitHub/GHE carry reviews on the\n\t * repo-scoped read and the filtered search, and on the account-wide read only where `includeReviews` opts\n\t * into the full projection, while GitLab never populates the field at all.\n\t * {@link PullRequestReviewer.commitOid} is the narrower promise: only the full GitHub/GHE projection has it.\n\t */\n\treadonly latestReviews?: PullRequestReviewer[];\n\treadonly assignees?: PullRequestMember[];\n\treadonly project?: IssueProject;\n\treadonly stack?: PullRequestStackInfo;\n\treadonly number?: number;\n\treadonly authoredByMe?: boolean;\n}\n\n@loggable(i => i.id)\n@serializable\nexport class PullRequest implements PullRequestShape {\n\treadonly type = 'pullrequest';\n\n\tconstructor(\n\t\tpublic readonly provider: ProviderReference,\n\t\tpublic readonly author: PullRequestMember,\n\t\tpublic readonly id: string,\n\t\tpublic readonly nodeId: string | undefined,\n\t\tpublic readonly title: string,\n\t\tpublic readonly url: string,\n\t\tpublic readonly repository: IssueRepository,\n\t\tpublic readonly state: PullRequestState,\n\t\tpublic readonly createdDate: Date,\n\t\tpublic readonly updatedDate: Date,\n\t\tpublic readonly closedDate?: Date,\n\t\tpublic readonly mergedDate?: Date,\n\t\tpublic readonly mergeableState?: PullRequestMergeableState,\n\t\tpublic readonly viewerCanUpdate?: boolean,\n\t\tpublic readonly refs?: PullRequestRefs,\n\t\tpublic readonly isDraft?: boolean,\n\t\tpublic readonly additions?: number,\n\t\tpublic readonly deletions?: number,\n\t\tpublic readonly commentsCount?: number,\n\t\tpublic readonly thumbsUpCount?: number,\n\t\tpublic readonly reviewDecision?: PullRequestReviewDecision,\n\t\tpublic readonly reviewRequests?: PullRequestReviewer[],\n\t\tpublic readonly latestReviews?: PullRequestReviewer[],\n\t\tpublic readonly assignees?: PullRequestMember[],\n\t\tpublic readonly statusCheckRollupState?: PullRequestStatusCheckRollupState,\n\t\tpublic readonly project?: IssueProject,\n\t\tpublic readonly version?: number,\n\t\t/** Commits on the head that aren't on the base. Appended rather than slotted with the other counts:\n\t\t *  this constructor is positional and every mapper passes through it. */\n\t\tpublic readonly commitCount?: number,\n\t\tpublic readonly stack?: PullRequestStackInfo,\n\t\tpublic readonly filesChanged?: number,\n\t\t/** The pull request's description, as markdown. */\n\t\tpublic readonly body?: string,\n\t\tpublic readonly number?: number,\n\t\tpublic readonly authoredByMe?: boolean,\n\t) {}\n\n\tget closed(): boolean {\n\t\treturn this.state === 'closed';\n\t}\n\n\tstatic is(pr: unknown): pr is PullRequest {\n\t\treturn pr instanceof PullRequest;\n\t}\n\n\tstatic formatDate(pr: PullRequestShape, format?: string | null): string {\n\t\treturn formatDate(pr.mergedDate ?? pr.closedDate ?? pr.updatedDate, format ?? 'MMMM Do, YYYY h:mma');\n\t}\n\n\tstatic formatDateFromNow(pr: PullRequestShape): string {\n\t\treturn fromNow(pr.mergedDate ?? pr.closedDate ?? pr.updatedDate);\n\t}\n\n\tstatic formatDateWithStyle(\n\t\tpr: PullRequestShape,\n\t\tformatting: { dateStyle: string; dateFormat: string | null },\n\t): string {\n\t\treturn formatting.dateStyle === 'absolute'\n\t\t\t? PullRequest.formatDate(pr, formatting.dateFormat)\n\t\t\t: PullRequest.formatDateFromNow(pr);\n\t}\n}\n\nexport const enum PullRequestReviewDecision {\n\tApproved = 'Approved',\n\tChangesRequested = 'ChangesRequested',\n\tReviewRequired = 'ReviewRequired',\n}\n\nexport const enum PullRequestMergeableState {\n\tUnknown = 'Unknown',\n\tMergeable = 'Mergeable',\n\tConflicting = 'Conflicting',\n\tFailingChecks = 'FailingChecks',\n\tBlockedByPolicy = 'BlockedByPolicy',\n}\n\nexport const enum PullRequestStatusCheckRollupState {\n\tSuccess = 'success',\n\tPending = 'pending',\n\tFailed = 'failed',\n}\n\nexport const enum PullRequestMergeMethod {\n\tMerge = 'merge',\n\tSquash = 'squash',\n\tRebase = 'rebase',\n}\n\nexport const enum PullRequestReviewState {\n\tApproved = 'APPROVED',\n\tChangesRequested = 'CHANGES_REQUESTED',\n\tCommented = 'COMMENTED',\n\tDismissed = 'DISMISSED',\n\tPending = 'PENDING',\n\tReviewRequested = 'REVIEW_REQUESTED',\n}\n\nexport interface PullRequestComparisonRefs {\n\trepoPath: string;\n\tbase: { ref: string; label: string };\n\thead: { ref: string; label: string };\n}\n\nexport interface PullRequestMember {\n\tid: string;\n\t/**\n\t * Absent when the provider exposes no display name for the member — same contract as `IssueMember.name`: a\n\t * fallback invented here can't be told apart from a real name downstream, so each consumer picks its own\n\t * presentation.\n\t */\n\tname?: string;\n\t/** The provider's handle for this person, when it has one — GitHub's login, Azure's `uniqueName` (a UPN,\n\t *  so an email), Bitbucket's mutable `nickname`. Display/labelling only: it is neither guaranteed present\n\t *  (GitLab's native mapper has none) nor a stable identity, so never key a match off it. */\n\tusername?: string;\n\tavatarUrl?: string;\n\turl?: string;\n}\n\n/** Provider-neutral relationship facets shared by PR listings and filtered PR search. */\nexport enum PullRequestFilter {\n\tAuthor = 'author',\n\tAssignee = 'assignee',\n\tReviewRequested = 'review-requested',\n\t/**\n\t * PRs the user has already reviewed (GitHub `reviewed-by:@me`), as opposed to `ReviewRequested` (a\n\t * still-pending request). Distinguishes \"waiting on the author after my review\" from \"waiting for my\n\t * review\". Not expressible on the repo-scoped listing — no provider SDK exposes a reviewed-by axis there —\n\t * so it is advertised for the account-wide read and for the filtered search (which can still bound itself\n\t * to repositories or an org), never under `supportedPullRequestFilters`.\n\t */\n\tReviewed = 'reviewed-by',\n\tMention = 'mention',\n}\n\n/** Selects which pull request states a read should include. `all` covers open + closed + merged. */\nexport type PullRequestStateFilter = 'open' | 'closed' | 'merged' | 'all';\n\n/**\n * A field a pull-request read can be ordered by, provider-neutral.\n *\n * Only two, and deliberately so — unlike {@link IssueSortField} this is the INTERSECTION of what GitHub's PR\n * search can order by server-side and what a {@link PullRequestShape} carries, because the filtered PR search\n * always merges its relationship × state facets in the facade and so must re-order the union itself (see\n * `getPullRequestComparator`). `created` and `updated` are both. A field GitHub cannot order PRs by, or one a\n * merged page can't reproduce, is left out rather than advertised and then approximated: relevance is unstable\n * and has no comparable value on a merge, and priority is not a pull-request concept on any provider that has PRs.\n */\nexport type PullRequestSortField =\n\t/** Creation date. */\n\t| 'created'\n\t/** Last activity. */\n\t| 'updated';\n\n/**\n * How a pull-request read is ordered, as `field:direction` — the same serializable shape {@link IssueSorting}\n * and `BranchSorting` use, so a consumer can persist it, compare it, or bind it straight to a setting or dropdown.\n */\nexport type PullRequestSorting = `${PullRequestSortField}:asc` | `${PullRequestSortField}:desc`;\n\n/**\n * The order the filtered pull-request search applies when the caller asks for none: most recently updated first,\n * which is what it served before ordering was an option. Named once rather than per layer — the criteria model\n * documents it, the GitHub query emits it, and the result-ceiling warning quotes it — so the three cannot\n * disagree about what \"the default\" is.\n */\nexport const defaultPullRequestSort: PullRequestSorting = 'updated:desc';\n\n/**\n * What the filtered pull-request search narrows on.\n *\n * Structured rather than a provider query string so callers can send free text without giving it a qualifier\n * channel, and can check {@link PullRequestSearchCapabilities} before issuing a read. Every requested criterion\n * is validated all-or-nothing before the provider runs; unsupported criteria never fall through to a wider\n * result set whose paging would no longer describe the returned items.\n *\n * Relationships are OR-ed, as are states. Omitting `relationships` removes the current-user constraint, which is\n * only safe when the manager call supplies a repository or organization scope. Omitting `states` reads open PRs.\n */\nexport interface PullRequestSearchCriteria {\n\t/**\n\t * Free text matched by the provider's own relevance rules. Tokens that look like provider qualifiers are\n\t * removed before the query is built, so text cannot change the search's scope or state.\n\t */\n\ttext?: string;\n\t/** Current-user relationship facets to union. Empty/omitted means no relationship constraint. */\n\trelationships?: PullRequestFilter[];\n\t/** Pull request states to union. Empty/omitted reads open PRs; `all` subsumes every other member. */\n\tstates?: PullRequestStateFilter[];\n\t/** ISO date (`YYYY-MM-DD`). The most effective narrowing criterion by far on a large scope. */\n\tupdatedAfter?: string;\n\t/** ISO date (`YYYY-MM-DD`). */\n\tcreatedAfter?: string;\n\t/** Includes PRs from archived repositories. They are excluded by default. */\n\tincludeArchived?: boolean;\n\t/**\n\t * Narrows on draft state: `true` returns only drafts, `false` only ready-for-review PRs. Omitted places no\n\t * constraint. A boolean rather than a truthy flag because `false` is a distinct request, not the absence of one.\n\t */\n\tdraft?: boolean;\n\t/**\n\t * How to order the results. Omitted means `updated:desc`, which is what this search has always served.\n\t *\n\t * Validated all-or-nothing against {@link PullRequestSearchCapabilities.sorts} like every other criterion: a\n\t * key the provider can't express server-side refuses the WHOLE read rather than falling back to the default,\n\t * because combined with the provider's result ceiling another order returns a different subset than was asked\n\t * for and the paging that comes with it describes that other subset. Do not change it mid-pagination — a\n\t * cursor carries the sort it was produced under; drop the cursor instead of threading it under a new key.\n\t */\n\tsort?: PullRequestSorting;\n}\n\n/**\n * Which {@link PullRequestSearchCriteria} fields and manager-level scopes a provider can express server-side.\n * An empty `relationships` means the provider has no filtered PR search at all.\n */\nexport interface PullRequestSearchCapabilities {\n\t/** Current-user relationship facets the provider can union. Empty means the search itself is unsupported. */\n\trelationships: PullRequestFilter[];\n\t/** Individual state values the provider can union in one logical search. */\n\tstates: PullRequestStateFilter[];\n\ttext: boolean;\n\tupdatedAfter: boolean;\n\tcreatedAfter: boolean;\n\tincludeArchived: boolean;\n\t/** Whether the provider can constrain the search by draft state server-side. */\n\tdraft: boolean;\n\t/** Whether the manager's repository-descriptor scope is supported. */\n\trepositoryScope: boolean;\n\t/** Whether the manager's organization scope is supported. */\n\torganizationScope: boolean;\n\t/**\n\t * Sort keys the search can express server-side. Always contains at least `updated:desc` — the historical\n\t * default — when the search exists at all, so it is never empty for a usable surface; a provider without a\n\t * filtered PR search reports an empty `relationships`, which is already the signal there is no surface to order.\n\t */\n\tsorts: PullRequestSorting[];\n}\n\nexport interface PullRequestRef {\n\towner: string;\n\trepo: string;\n\tbranch: string;\n\tsha: string;\n\texists: boolean;\n\t/**\n\t * Web URL of the ref's repository. `undefined` when the provider cannot resolve it — Azure DevOps embeds an\n\t * abbreviated repository reference in a pull request, so a fork's URL may need a lookup that can fail.\n\t */\n\turl: string | undefined;\n\t/** HTTPS clone URL of the ref's repository, when the provider exposes it. */\n\tcloneHttps?: string;\n\t/** SSH clone URL of the ref's repository, when the provider exposes it. */\n\tcloneSsh?: string;\n\t/** Best-effort flag: whether the ref's repository is a fork. `undefined` when the provider can't tell. */\n\tisFork?: boolean;\n}\n\nexport interface PullRequestRefs {\n\tbase: PullRequestRef & { url: string };\n\thead: PullRequestRef;\n\tisCrossRepository: boolean;\n}\n\n/**\n * A pull request's membership in a stack — an ordered chain of dependent pull requests where each\n * targets the branch of the one below it.\n *\n * `baseRef` is the stack's ultimate target (its trunk) and is NOT the same as `refs.base`, which for\n * anything above the bottom layer names the layer below. Use `refs.base` to diff a single layer; use\n * `baseRef` to answer where the work ultimately lands.\n */\nexport interface PullRequestStackInfo {\n\tid: string;\n\t/** Identifies the stack within its repository. */\n\tnumber: number;\n\t/** Total pull requests in the stack. */\n\tsize: number;\n\t/** This pull request's layer, 1-based, where 1 is closest to `baseRef`. */\n\tposition: number;\n\t/** The branch the bottom of the stack targets. */\n\tbaseRef: string;\n}\n\nexport interface PullRequestReviewer {\n\tisCodeOwner?: boolean;\n\treviewer: PullRequestMember;\n\tstate: PullRequestReviewState;\n\t/**\n\t * The head commit oid this review was submitted against (GitHub). Lets a consumer tell \"the PR moved past\n\t * my review\" (oid !== the PR's current head) from a review still at the tip — without a per-review\n\t * timestamp. Only populated where the full GitHub/GHE projection is selected: the filtered search read, and\n\t * an account-wide sweep that opts in with `includeReviews`. Undefined elsewhere, including reviews the\n\t * repo-scoped read carries natively.\n\t */\n\tcommitOid?: string;\n}\n\nexport type PullRequestRepositoryIdentityDescriptor = RequireSomeWithProps<\n\tRequireSome<RepositoryIdentityDescriptor<string>, 'provider'>,\n\t'provider',\n\t'id' | 'domain' | 'repoDomain' | 'repoName'\n> &\n\tRequireSomeWithProps<RequireSome<RepositoryIdentityDescriptor<string>, 'remote'>, 'remote', 'domain'>;\n"],"mappings":";;;;;;AA+CO,IAAM,cAAA,eAAN,MAAM,YAAwC;CAInC;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CAGA;CACA;CACA;CAEA;CACA;CACA;CAtCjB,OAAgB;CAEhB,YACC,UACA,QACA,IACA,QACA,OACA,KACA,YACA,OACA,aACA,aACA,YACA,YACA,gBACA,iBACA,MACA,SACA,WACA,WACA,eACA,eACA,gBACA,gBACA,eACA,WACA,wBACA,SACA,SAGA,aACA,OACA,cAEA,MACA,QACA,cACC;EApCe,KAAA,WAAA;EACA,KAAA,SAAA;EACA,KAAA,KAAA;EACA,KAAA,SAAA;EACA,KAAA,QAAA;EACA,KAAA,MAAA;EACA,KAAA,aAAA;EACA,KAAA,QAAA;EACA,KAAA,cAAA;EACA,KAAA,cAAA;EACA,KAAA,aAAA;EACA,KAAA,aAAA;EACA,KAAA,iBAAA;EACA,KAAA,kBAAA;EACA,KAAA,OAAA;EACA,KAAA,UAAA;EACA,KAAA,YAAA;EACA,KAAA,YAAA;EACA,KAAA,gBAAA;EACA,KAAA,gBAAA;EACA,KAAA,iBAAA;EACA,KAAA,iBAAA;EACA,KAAA,gBAAA;EACA,KAAA,YAAA;EACA,KAAA,yBAAA;EACA,KAAA,UAAA;EACA,KAAA,UAAA;EAGA,KAAA,cAAA;EACA,KAAA,QAAA;EACA,KAAA,eAAA;EAEA,KAAA,OAAA;EACA,KAAA,SAAA;EACA,KAAA,eAAA;CACd;CAEH,IAAI,SAAkB;EACrB,OAAO,KAAK,UAAU;CACvB;CAEA,OAAO,GAAG,IAAgC;EACzC,OAAO,cAAA;CACR;CAEA,OAAO,WAAW,IAAsB,QAAgC;EACvE,OAAO,WAAW,GAAG,cAAc,GAAG,cAAc,GAAG,aAAa,UAAU,qBAAqB;CACpG;CAEA,OAAO,kBAAkB,IAA8B;EACtD,OAAO,QAAQ,GAAG,cAAc,GAAG,cAAc,GAAG,WAAW;CAChE;CAEA,OAAO,oBACN,IACA,YACS;EACT,OAAO,WAAW,cAAc,aAAA,aACjB,WAAW,IAAI,WAAW,UAAU,IAAA,aACpC,kBAAkB,EAAE;CACpC;AACD;AApEC,cAAA,eAAA,WAAA,CAAA,UAAS,MAAK,EAAE,EAAE,GAClB,YAAA,GAAA,WAAA;AAqED,IAAkB,4BAAX,yBAAA,2BAAA;CACN,0BAAA,cAAA;CACA,0BAAA,sBAAA;CACA,0BAAA,oBAAA;;AACD,EAAA,CAAA,CAAA;AAEA,IAAkB,4BAAX,yBAAA,2BAAA;CACN,0BAAA,aAAA;CACA,0BAAA,eAAA;CACA,0BAAA,iBAAA;CACA,0BAAA,mBAAA;CACA,0BAAA,qBAAA;;AACD,EAAA,CAAA,CAAA;AAEA,IAAkB,oCAAX,yBAAA,mCAAA;CACN,kCAAA,aAAA;CACA,kCAAA,aAAA;CACA,kCAAA,YAAA;;AACD,EAAA,CAAA,CAAA;AAEA,IAAkB,yBAAX,yBAAA,wBAAA;CACN,uBAAA,WAAA;CACA,uBAAA,YAAA;CACA,uBAAA,YAAA;;AACD,EAAA,CAAA,CAAA;AAEA,IAAkB,yBAAX,yBAAA,wBAAA;CACN,uBAAA,cAAA;CACA,uBAAA,sBAAA;CACA,uBAAA,eAAA;CACA,uBAAA,eAAA;CACA,uBAAA,aAAA;CACA,uBAAA,qBAAA;;AACD,EAAA,CAAA,CAAA;;AAyBA,IAAY,oBAAL,yBAAA,mBAAA;CACN,kBAAA,YAAA;CACA,kBAAA,cAAA;CACA,kBAAA,qBAAA;;;;;;;;CAQA,kBAAA,cAAA;CACA,kBAAA,aAAA;;AACD,EAAA,CAAA,CAAA;;;;;;;AAiCA,MAAa,yBAA6C"}
+{"version":3,"file":"pullRequest.js","names":[],"sources":["../../../../git/src/models/pullRequest.ts"],"sourcesContent":["import { formatDate, fromNow } from '@gitlens/utils/date.js';\nimport { loggable } from '@gitlens/utils/decorators/log.js';\nimport { serializable } from '@gitlens/utils/decorators/serializable.js';\nimport type { RequireSome, RequireSomeWithProps } from '@gitlens/utils/types.js';\nimport type { IssueProject, IssueRepository } from './issue.js';\nimport type { IssueOrPullRequest, IssueOrPullRequestState as PullRequestState } from './issueOrPullRequest.js';\nimport type { ProviderReference } from './remoteProvider.js';\nimport type { RepositoryIdentityDescriptor } from './repositoryIdentities.js';\n\nexport type { PullRequestState };\n\n/**\n * The read that produced a pull request row, as a key for `getPullRequestFieldPresence` (`@gitlens/integrations`). It\n * names a READ, not a field set: the same tag can carry different fields on different providers.\n */\nexport type PullRequestProjection =\n\t/** A single-item read: by number, for a branch, or for a commit. */\n\t| 'point'\n\t/** The filtered search with the full projection, and the host's \"my pull requests\" search. */\n\t| 'search'\n\t/** The filtered search with `summary`. */\n\t| 'search-summary'\n\t/** The free-text search (`searchPullRequests`). */\n\t| 'text-search'\n\t/** An account-wide sweep with the full projection (`includeReviews`). */\n\t| 'account'\n\t/** An account-wide list or sweep with the lightweight projection. */\n\t| 'account-summary'\n\t/** A repository-scoped list or sweep. */\n\t| 'repos'\n\t/** A repository-scoped list with `summary`. */\n\t| 'repos-summary'\n\t/** The coordinate and branch batch reads. */\n\t| 'batch';\n\nexport interface PullRequestShape extends IssueOrPullRequest {\n\treadonly author: PullRequestMember;\n\t/**\n\t * Canonical base-repository identity returned by the provider. Some providers\n\t * scope pull-request ids to a repository, so consumers must not have to infer\n\t * identity from optional refs or URLs.\n\t */\n\treadonly repository?: IssueRepository;\n\treadonly mergedDate?: Date;\n\treadonly refs?: PullRequestRefs;\n\treadonly isDraft?: boolean;\n\treadonly additions?: number;\n\treadonly deletions?: number;\n\treadonly filesChanged?: number;\n\treadonly body?: string;\n\treadonly mergeableState?: PullRequestMergeableState;\n\treadonly reviewDecision?: PullRequestReviewDecision;\n\treadonly reviewRequests?: PullRequestReviewer[];\n\t/**\n\t * Reviews already submitted, as opposed to the still-pending {@link reviewRequests}. Absent means the read\n\t * didn't fetch reviews; empty means nobody has reviewed — but only where the read actually fetched them, which\n\t * is provider- and read-dependent: ask `getPullRequestFieldPresence` (`@gitlens/integrations`) for the row's\n\t * `reviews` group rather than trusting an empty array.\n\t * {@link PullRequestReviewer.commitOid} is the narrower promise: only the full GitHub/GHE projection has it.\n\t */\n\treadonly latestReviews?: PullRequestReviewer[];\n\treadonly assignees?: PullRequestMember[];\n\treadonly project?: IssueProject;\n\treadonly stack?: PullRequestStackInfo;\n\treadonly number?: number;\n\treadonly authoredByMe?: boolean;\n\t/**\n\t * The current account as this pull request's members identify it: the identity {@link authoredByMe} was matched\n\t * against, present exactly when that was. A consumer matching {@link reviewRequests} or {@link latestReviews}\n\t * entries to the current user compares against this, not the account `getCurrentAccount` returns: Azure DevOps\n\t * Server gives one person a different id in each collection than at the server level, and the members carry the\n\t * collection's. Core matches `username` too only on GitHub and GitHub Enterprise, whose rows key people by login.\n\t */\n\treadonly viewer?: { readonly id: string; readonly username?: string };\n\t/** The read that produced this row. Absent means unknown: a legacy row, or a provider or read not yet tagged. */\n\treadonly projection?: PullRequestProjection;\n}\n\n@loggable(i => i.id)\n@serializable\nexport class PullRequest implements PullRequestShape {\n\treadonly type = 'pullrequest';\n\n\tconstructor(\n\t\tpublic readonly provider: ProviderReference,\n\t\tpublic readonly author: PullRequestMember,\n\t\tpublic readonly id: string,\n\t\tpublic readonly nodeId: string | undefined,\n\t\tpublic readonly title: string,\n\t\tpublic readonly url: string,\n\t\tpublic readonly repository: IssueRepository,\n\t\tpublic readonly state: PullRequestState,\n\t\tpublic readonly createdDate: Date,\n\t\tpublic readonly updatedDate: Date,\n\t\tpublic readonly closedDate?: Date,\n\t\tpublic readonly mergedDate?: Date,\n\t\tpublic readonly mergeableState?: PullRequestMergeableState,\n\t\tpublic readonly viewerCanUpdate?: boolean,\n\t\tpublic readonly refs?: PullRequestRefs,\n\t\tpublic readonly isDraft?: boolean,\n\t\tpublic readonly additions?: number,\n\t\tpublic readonly deletions?: number,\n\t\tpublic readonly commentsCount?: number,\n\t\tpublic readonly thumbsUpCount?: number,\n\t\tpublic readonly reviewDecision?: PullRequestReviewDecision,\n\t\tpublic readonly reviewRequests?: PullRequestReviewer[],\n\t\tpublic readonly latestReviews?: PullRequestReviewer[],\n\t\tpublic readonly assignees?: PullRequestMember[],\n\t\tpublic readonly statusCheckRollupState?: PullRequestStatusCheckRollupState,\n\t\tpublic readonly project?: IssueProject,\n\t\tpublic readonly version?: number,\n\t\t/** Commits on the head that aren't on the base. Appended rather than slotted with the other counts:\n\t\t *  this constructor is positional and every mapper passes through it. */\n\t\tpublic readonly commitCount?: number,\n\t\tpublic readonly stack?: PullRequestStackInfo,\n\t\tpublic readonly filesChanged?: number,\n\t\t/** The pull request's description, as markdown. */\n\t\tpublic readonly body?: string,\n\t\tpublic readonly number?: number,\n\t\tpublic readonly authoredByMe?: boolean,\n\t\tpublic readonly projection?: PullRequestProjection,\n\t\tpublic readonly viewer?: { readonly id: string; readonly username?: string },\n\t) {}\n\n\tget closed(): boolean {\n\t\treturn this.state === 'closed';\n\t}\n\n\tstatic is(pr: unknown): pr is PullRequest {\n\t\treturn pr instanceof PullRequest;\n\t}\n\n\tstatic formatDate(pr: PullRequestShape, format?: string | null): string {\n\t\treturn formatDate(pr.mergedDate ?? pr.closedDate ?? pr.updatedDate, format ?? 'MMMM Do, YYYY h:mma');\n\t}\n\n\tstatic formatDateFromNow(pr: PullRequestShape): string {\n\t\treturn fromNow(pr.mergedDate ?? pr.closedDate ?? pr.updatedDate);\n\t}\n\n\tstatic formatDateWithStyle(\n\t\tpr: PullRequestShape,\n\t\tformatting: { dateStyle: string; dateFormat: string | null },\n\t): string {\n\t\treturn formatting.dateStyle === 'absolute'\n\t\t\t? PullRequest.formatDate(pr, formatting.dateFormat)\n\t\t\t: PullRequest.formatDateFromNow(pr);\n\t}\n}\n\nexport const enum PullRequestReviewDecision {\n\tApproved = 'Approved',\n\tChangesRequested = 'ChangesRequested',\n\tReviewRequired = 'ReviewRequired',\n}\n\nexport const enum PullRequestMergeableState {\n\tUnknown = 'Unknown',\n\tMergeable = 'Mergeable',\n\tConflicting = 'Conflicting',\n\tFailingChecks = 'FailingChecks',\n\tBlockedByPolicy = 'BlockedByPolicy',\n}\n\nexport const enum PullRequestStatusCheckRollupState {\n\tSuccess = 'success',\n\tPending = 'pending',\n\tFailed = 'failed',\n}\n\nexport const enum PullRequestMergeMethod {\n\tMerge = 'merge',\n\tSquash = 'squash',\n\tRebase = 'rebase',\n}\n\nexport const enum PullRequestReviewState {\n\tApproved = 'APPROVED',\n\tChangesRequested = 'CHANGES_REQUESTED',\n\tCommented = 'COMMENTED',\n\tDismissed = 'DISMISSED',\n\tPending = 'PENDING',\n\tReviewRequested = 'REVIEW_REQUESTED',\n}\n\nexport interface PullRequestComparisonRefs {\n\trepoPath: string;\n\tbase: { ref: string; label: string };\n\thead: { ref: string; label: string };\n}\n\nexport interface PullRequestMember {\n\tid: string;\n\t/**\n\t * Absent when the provider exposes no display name for the member — same contract as `IssueMember.name`: a\n\t * fallback invented here can't be told apart from a real name downstream, so each consumer picks its own\n\t * presentation.\n\t */\n\tname?: string;\n\t/** The provider's handle for this person, when it has one — GitHub's login, Azure's `uniqueName` (a UPN,\n\t *  so an email), Bitbucket's mutable `nickname`. Display/labelling only: it is neither guaranteed present\n\t *  (GitLab's native mapper has none) nor a stable identity, so never key a match off it. */\n\tusername?: string;\n\tavatarUrl?: string;\n\turl?: string;\n}\n\n/** Provider-neutral relationship facets shared by PR listings and filtered PR search. */\nexport enum PullRequestFilter {\n\tAuthor = 'author',\n\tAssignee = 'assignee',\n\tReviewRequested = 'review-requested',\n\t/**\n\t * PRs the user has already reviewed (GitHub `reviewed-by:@me`), as opposed to `ReviewRequested` (a\n\t * still-pending request). Distinguishes \"waiting on the author after my review\" from \"waiting for my\n\t * review\". Not expressible on the repo-scoped listing — no provider SDK exposes a reviewed-by axis there —\n\t * so it is advertised for the account-wide read and for the filtered search (which can still bound itself\n\t * to repositories or an org), never under `supportedPullRequestFilters`.\n\t */\n\tReviewed = 'reviewed-by',\n\tMention = 'mention',\n}\n\n/** Selects which pull request states a read should include. `all` covers open + closed + merged. */\nexport type PullRequestStateFilter = 'open' | 'closed' | 'merged' | 'all';\n\n/**\n * A field a pull-request read can be ordered by, provider-neutral.\n *\n * Only two, and deliberately so — unlike {@link IssueSortField} this is the INTERSECTION of what GitHub's PR\n * search can order by server-side and what a {@link PullRequestShape} carries, because the filtered PR search\n * always merges its relationship × state facets in the facade and so must re-order the union itself (see\n * `getPullRequestComparator`). `created` and `updated` are both. A field GitHub cannot order PRs by, or one a\n * merged page can't reproduce, is left out rather than advertised and then approximated: relevance is unstable\n * and has no comparable value on a merge, and priority is not a pull-request concept on any provider that has PRs.\n */\nexport type PullRequestSortField =\n\t/** Creation date. */\n\t| 'created'\n\t/** Last activity. */\n\t| 'updated';\n\n/**\n * How a pull-request read is ordered, as `field:direction` — the same serializable shape {@link IssueSorting}\n * and `BranchSorting` use, so a consumer can persist it, compare it, or bind it straight to a setting or dropdown.\n */\nexport type PullRequestSorting = `${PullRequestSortField}:asc` | `${PullRequestSortField}:desc`;\n\n/**\n * The order the filtered pull-request search applies when the caller asks for none: most recently updated first,\n * which is what it served before ordering was an option. Named once rather than per layer — the criteria model\n * documents it, the GitHub query emits it, and the result-ceiling warning quotes it — so the three cannot\n * disagree about what \"the default\" is.\n */\nexport const defaultPullRequestSort: PullRequestSorting = 'updated:desc';\n\n/**\n * What the filtered pull-request search narrows on.\n *\n * Structured rather than a provider query string so callers can send free text without giving it a qualifier\n * channel, and can check {@link PullRequestSearchCapabilities} before issuing a read. Every requested criterion\n * is validated all-or-nothing before the provider runs; unsupported criteria never fall through to a wider\n * result set whose paging would no longer describe the returned items.\n *\n * Relationships are OR-ed, as are states. Omitting `relationships` removes the current-user constraint, which is\n * only safe when the manager call supplies a repository or organization scope. Omitting `states` reads open PRs.\n */\nexport interface PullRequestSearchCriteria {\n\t/**\n\t * Free text matched by the provider's own relevance rules. Tokens that look like provider qualifiers are\n\t * removed before the query is built, so text cannot change the search's scope or state.\n\t */\n\ttext?: string;\n\t/** Current-user relationship facets to union. Empty/omitted means no relationship constraint. */\n\trelationships?: PullRequestFilter[];\n\t/** Pull request states to union. Empty/omitted reads open PRs; `all` subsumes every other member. */\n\tstates?: PullRequestStateFilter[];\n\t/** ISO date (`YYYY-MM-DD`). The most effective narrowing criterion by far on a large scope. */\n\tupdatedAfter?: string;\n\t/** ISO date (`YYYY-MM-DD`). */\n\tcreatedAfter?: string;\n\t/** Includes PRs from archived repositories. They are excluded by default. */\n\tincludeArchived?: boolean;\n\t/**\n\t * Narrows on draft state: `true` returns only drafts, `false` only ready-for-review PRs. Omitted places no\n\t * constraint. A boolean rather than a truthy flag because `false` is a distinct request, not the absence of one.\n\t */\n\tdraft?: boolean;\n\t/**\n\t * How to order the results. Omitted means `updated:desc`, which is what this search has always served.\n\t *\n\t * Validated all-or-nothing against {@link PullRequestSearchCapabilities.sorts} like every other criterion: a\n\t * key the provider can't express server-side refuses the WHOLE read rather than falling back to the default,\n\t * because combined with the provider's result ceiling another order returns a different subset than was asked\n\t * for and the paging that comes with it describes that other subset. Do not change it mid-pagination — a\n\t * cursor carries the sort it was produced under; drop the cursor instead of threading it under a new key.\n\t */\n\tsort?: PullRequestSorting;\n}\n\n/**\n * Which {@link PullRequestSearchCriteria} fields and manager-level scopes a provider can express server-side.\n * An empty `relationships` means the provider has no filtered PR search at all.\n */\nexport interface PullRequestSearchCapabilities {\n\t/** Current-user relationship facets the provider can union. Empty means the search itself is unsupported. */\n\trelationships: PullRequestFilter[];\n\t/** Individual state values the provider can union in one logical search. */\n\tstates: PullRequestStateFilter[];\n\ttext: boolean;\n\tupdatedAfter: boolean;\n\tcreatedAfter: boolean;\n\tincludeArchived: boolean;\n\t/** Whether the provider can constrain the search by draft state server-side. */\n\tdraft: boolean;\n\t/** Whether the manager's repository-descriptor scope is supported. */\n\trepositoryScope: boolean;\n\t/** Whether the manager's organization scope is supported. */\n\torganizationScope: boolean;\n\t/**\n\t * Sort keys the search can express server-side. Always contains at least `updated:desc` — the historical\n\t * default — when the search exists at all, so it is never empty for a usable surface; a provider without a\n\t * filtered PR search reports an empty `relationships`, which is already the signal there is no surface to order.\n\t */\n\tsorts: PullRequestSorting[];\n}\n\nexport interface PullRequestRef {\n\t/**\n\t * `undefined` when the ref's repository was deleted (a removed fork): GitHub's rows say so, with `exists` false.\n\t * provider-apis' converters fill `''` instead, so test for a value rather than for `undefined`.\n\t */\n\towner: string | undefined;\n\t/** `undefined` (or `''`) when the ref's repository was deleted, as {@link owner}. */\n\trepo: string | undefined;\n\tbranch: string;\n\tsha: string;\n\texists: boolean;\n\t/**\n\t * Web URL of the ref's repository. `undefined` when the provider cannot resolve it — Azure DevOps embeds an\n\t * abbreviated repository reference in a pull request, so a fork's URL may need a lookup that can fail.\n\t */\n\turl: string | undefined;\n\t/** HTTPS clone URL of the ref's repository, when the provider exposes it. */\n\tcloneHttps?: string;\n\t/** SSH clone URL of the ref's repository, when the provider exposes it. */\n\tcloneSsh?: string;\n\t/** Best-effort flag: whether the ref's repository is a fork. `undefined` when the provider can't tell. */\n\tisFork?: boolean;\n}\n\nexport interface PullRequestRefs {\n\tbase: PullRequestRef & { url: string };\n\thead: PullRequestRef;\n\tisCrossRepository: boolean;\n}\n\n/**\n * A pull request's membership in a stack — an ordered chain of dependent pull requests where each\n * targets the branch of the one below it.\n *\n * `baseRef` is the stack's ultimate target (its trunk) and is NOT the same as `refs.base`, which for\n * anything above the bottom layer names the layer below. Use `refs.base` to diff a single layer; use\n * `baseRef` to answer where the work ultimately lands.\n */\nexport interface PullRequestStackInfo {\n\tid: string;\n\t/** Identifies the stack within its repository. */\n\tnumber: number;\n\t/** Total pull requests in the stack. */\n\tsize: number;\n\t/** This pull request's layer, 1-based, where 1 is closest to `baseRef`. */\n\tposition: number;\n\t/** The branch the bottom of the stack targets. */\n\tbaseRef: string;\n}\n\nexport interface PullRequestReviewer {\n\tisCodeOwner?: boolean;\n\treviewer: PullRequestMember;\n\tstate: PullRequestReviewState;\n\t/**\n\t * The head commit oid this review was submitted against (GitHub). Lets a consumer tell \"the PR moved past\n\t * my review\" (oid !== the PR's current head) from a review still at the tip — without a per-review\n\t * timestamp. Only populated where the full GitHub/GHE projection is selected: the filtered search read, and\n\t * an account-wide sweep that opts in with `includeReviews`. Undefined elsewhere, including reviews the\n\t * repo-scoped read carries natively.\n\t */\n\tcommitOid?: string;\n\t/**\n\t * The reviewer is a group the current user is a member of (a team or another group, directly or through a group\n\t * nested in it) rather than a person: a request to it is one the user can answer as a member, not one addressed to\n\t * them by name. Only the Azure DevOps reads that resolve the user's groups set it; absent elsewhere, including for\n\t * groups a read could not tell the user belongs to.\n\t */\n\tisMyGroup?: boolean;\n}\n\nexport type PullRequestRepositoryIdentityDescriptor = RequireSomeWithProps<\n\tRequireSome<RepositoryIdentityDescriptor<string>, 'provider'>,\n\t'provider',\n\t'id' | 'domain' | 'repoDomain' | 'repoName'\n> &\n\tRequireSomeWithProps<RequireSome<RepositoryIdentityDescriptor<string>, 'remote'>, 'remote', 'domain'>;\n"],"mappings":";;;;;;AAgFO,IAAM,cAAA,eAAN,MAAM,YAAwC;CAInC;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CACA;CAGA;CACA;CACA;CAEA;CACA;CACA;CACA;CACA;CAxCjB,OAAgB;CAEhB,YACC,UACA,QACA,IACA,QACA,OACA,KACA,YACA,OACA,aACA,aACA,YACA,YACA,gBACA,iBACA,MACA,SACA,WACA,WACA,eACA,eACA,gBACA,gBACA,eACA,WACA,wBACA,SACA,SAGA,aACA,OACA,cAEA,MACA,QACA,cACA,YACA,QACC;EAtCe,KAAA,WAAA;EACA,KAAA,SAAA;EACA,KAAA,KAAA;EACA,KAAA,SAAA;EACA,KAAA,QAAA;EACA,KAAA,MAAA;EACA,KAAA,aAAA;EACA,KAAA,QAAA;EACA,KAAA,cAAA;EACA,KAAA,cAAA;EACA,KAAA,aAAA;EACA,KAAA,aAAA;EACA,KAAA,iBAAA;EACA,KAAA,kBAAA;EACA,KAAA,OAAA;EACA,KAAA,UAAA;EACA,KAAA,YAAA;EACA,KAAA,YAAA;EACA,KAAA,gBAAA;EACA,KAAA,gBAAA;EACA,KAAA,iBAAA;EACA,KAAA,iBAAA;EACA,KAAA,gBAAA;EACA,KAAA,YAAA;EACA,KAAA,yBAAA;EACA,KAAA,UAAA;EACA,KAAA,UAAA;EAGA,KAAA,cAAA;EACA,KAAA,QAAA;EACA,KAAA,eAAA;EAEA,KAAA,OAAA;EACA,KAAA,SAAA;EACA,KAAA,eAAA;EACA,KAAA,aAAA;EACA,KAAA,SAAA;CACd;CAEH,IAAI,SAAkB;EACrB,OAAO,KAAK,UAAU;CACvB;CAEA,OAAO,GAAG,IAAgC;EACzC,OAAO,cAAA;CACR;CAEA,OAAO,WAAW,IAAsB,QAAgC;EACvE,OAAO,WAAW,GAAG,cAAc,GAAG,cAAc,GAAG,aAAa,UAAU,qBAAqB;CACpG;CAEA,OAAO,kBAAkB,IAA8B;EACtD,OAAO,QAAQ,GAAG,cAAc,GAAG,cAAc,GAAG,WAAW;CAChE;CAEA,OAAO,oBACN,IACA,YACS;EACT,OAAO,WAAW,cAAc,aAAA,aACjB,WAAW,IAAI,WAAW,UAAU,IAAA,aACpC,kBAAkB,EAAE;CACpC;AACD;AAtEC,cAAA,eAAA,WAAA,CAAA,UAAS,MAAK,EAAE,EAAE,GAClB,YAAA,GAAA,WAAA;AAuED,IAAkB,4BAAX,yBAAA,2BAAA;CACN,0BAAA,cAAA;CACA,0BAAA,sBAAA;CACA,0BAAA,oBAAA;;AACD,EAAA,CAAA,CAAA;AAEA,IAAkB,4BAAX,yBAAA,2BAAA;CACN,0BAAA,aAAA;CACA,0BAAA,eAAA;CACA,0BAAA,iBAAA;CACA,0BAAA,mBAAA;CACA,0BAAA,qBAAA;;AACD,EAAA,CAAA,CAAA;AAEA,IAAkB,oCAAX,yBAAA,mCAAA;CACN,kCAAA,aAAA;CACA,kCAAA,aAAA;CACA,kCAAA,YAAA;;AACD,EAAA,CAAA,CAAA;AAEA,IAAkB,yBAAX,yBAAA,wBAAA;CACN,uBAAA,WAAA;CACA,uBAAA,YAAA;CACA,uBAAA,YAAA;;AACD,EAAA,CAAA,CAAA;AAEA,IAAkB,yBAAX,yBAAA,wBAAA;CACN,uBAAA,cAAA;CACA,uBAAA,sBAAA;CACA,uBAAA,eAAA;CACA,uBAAA,eAAA;CACA,uBAAA,aAAA;CACA,uBAAA,qBAAA;;AACD,EAAA,CAAA,CAAA;;AAyBA,IAAY,oBAAL,yBAAA,mBAAA;CACN,kBAAA,YAAA;CACA,kBAAA,cAAA;CACA,kBAAA,qBAAA;;;;;;;;CAQA,kBAAA,cAAA;CACA,kBAAA,aAAA;;AACD,EAAA,CAAA,CAAA;;;;;;;AAiCA,MAAa,yBAA6C"}

package/dist/git/utils/issue.utils.d.ts

@@ -41,6 +41,10 @@ export declare function effectiveIssueSort(sort: IssueSorting | undefined): Issu
  */
 export declare function getIssueComparator(sort: IssueSorting): ((a: IssueShape, b: IssueShape) => number) | undefined;
 export declare function getRepositoryIdentityForIssue(issue: IssueShape | Issue): IssueRepositoryIdentityDescriptor;
+/**
+ * Deliberately leaves `projection` out: this copy drops `repository.accessLevel`, so a tag would have field presence
+ * report the viewer's access as fetched when the copy no longer has it.
+ */
 export declare function serializeIssue(value: IssueShape): IssueShape;
 //#endregion
 //# sourceMappingURL=issue.utils.d.ts.map

package/dist/git/utils/issue.utils.js

@@ -93,6 +93,10 @@ function getRepositoryIdentityForIssue(issue) {
 		}
 	};
 }
+/**
+* Deliberately leaves `projection` out: this copy drops `repository.accessLevel`, so a tag would have field presence
+* report the viewer's access as fetched when the copy no longer has it.
+*/
 function serializeIssue(value) {
 	return {
 		type: value.type,

package/dist/git/utils/issue.utils.js.map

@@ -1 +1 @@
-{"version":3,"file":"issue.utils.js","names":[],"sources":["../../../../git/src/utils/issue.utils.ts"],"sourcesContent":["import type {\n\tIssue,\n\tIssueRepositoryIdentityDescriptor,\n\tIssueShape,\n\tIssueSortField,\n\tIssueSorting,\n} from '../models/issue.js';\nimport { defaultIssueSort } from '../models/issue.js';\n\n/**\n * How each sort field is read off an {@link IssueShape}, for the reads that MERGE several provider queries and so\n * have to order the union themselves instead of letting the provider do it.\n *\n * Only the fields the shape actually carries appear. `priority`, `dueDate` and `resolved` are absent because\n * `IssueOrPullRequest`/`IssueShape` doesn't model them: a provider may well order by them server-side on a\n * single-query read, but a merged one cannot — which is exactly what {@link getIssueComparator} returning\n * `undefined` lets its caller refuse, instead of concatenating unordered runs and calling them sorted.\n */\nconst issueSortValues: Partial<Record<IssueSortField, (issue: IssueShape) => number | Date | string | undefined>> = {\n\tcreated: i => i.createdDate,\n\tupdated: i => i.updatedDate,\n\tclosed: i => i.closedDate,\n\tcomments: i => i.commentsCount,\n\treactions: i => i.thumbsUpCount,\n\ttitle: i => i.title,\n};\n\n/**\n * The order a read that ALWAYS orders is being made under: the caller's key, or the facade's default.\n *\n * The one place `?? defaultIssueSort` may be written. Several things downstream must agree on the answer — the\n * emitted provider qualifier, the fingerprint sealed into a cursor, the comparator that re-sorts a merged page and\n * the ceiling warning that names the order — and re-deriving the fallback at each is how they come to disagree,\n * which is not cosmetic: a cursor sealed with a key its query wasn't built with makes the fingerprint check refuse\n * valid resumptions.\n *\n * Note this is for a surface that CAN always order. A read whose provider may not express even the default resolves\n * through `resolveIssueSort` instead, which can answer \"leave the provider's own order\".\n */\nexport function effectiveIssueSort(sort: IssueSorting | undefined): IssueSorting {\n\treturn sort ?? defaultIssueSort;\n}\n\n/**\n * A comparator for one sort key over normalized issues, or `undefined` when the field isn't derivable from an\n * {@link IssueShape} — the signal that the key is only honorable on a SINGLE-origin read, where the provider\n * already ordered the page.\n *\n * `@gitkraken/provider-apis` carries a twin of this over its own `Issue` shape (`src/issueSort.ts`), which it uses\n * for the merges it performs itself. It is not importable — the SDK does not export it, and it takes a different\n * shape — so the two are separate implementations of one rule, and the rule below is the part that must not drift.\n *\n * Missing values sort LAST in both directions. That is deliberately not the same as negating the descending\n * comparator (which would float them to the front when ascending), and it is why they're partitioned before any\n * arithmetic rather than mapped to an infinite sentinel: two missing values have to compare equal, and\n * `-Infinity - -Infinity` is `NaN`, which makes the comparator inconsistent and the resulting order\n * implementation-defined.\n *\n * Note it orders whatever it is handed. Sorting a page that was already capped does not make it the top N.\n *\n * KNOWN DIVERGENCE, `title` only: strings are compared by code unit, so `'B' < 'a'`. A single-scope read is\n * ordered by the provider's own database instead — case-insensitive and accent-aware — so the same `title:asc`\n * yields \"Banana, apple, ábaco\" from a merged page and \"ábaco, apple, Banana\" from one repository. Deliberately\n * NOT fixed with `localeCompare`: the SDK's twin comparator (which performs the merges core doesn't, notably\n * GitLab's repo-ids aggregate) compares by code unit, so switching here would make two paths of the SAME\n * multi-repo read disagree — a worse divergence than the one it closes, and it would break the parity test in\n * `issueSortCapabilities.test.ts`. Closing this properly means changing the rule on BOTH sides at once; until\n * then it is a cosmetic ordering difference on one key, not wrong data.\n */\nexport function getIssueComparator(sort: IssueSorting): ((a: IssueShape, b: IssueShape) => number) | undefined {\n\t// The union's shape guarantees both halves, so this needs no validation — only the cast tsc can't infer.\n\tconst [field, direction] = sort.split(':') as [IssueSortField, 'asc' | 'desc'];\n\tconst getValue = issueSortValues[field];\n\tif (getValue == null) return undefined;\n\n\treturn (a, b) => {\n\t\tconst left = toComparable(getValue(a));\n\t\tconst right = toComparable(getValue(b));\n\t\tif (left == null || right == null) {\n\t\t\tif (left == null && right == null) return 0;\n\t\t\treturn left == null ? 1 : -1;\n\t\t}\n\n\t\tconst ordered = left < right ? -1 : left > right ? 1 : 0;\n\t\treturn direction === 'asc' ? ordered : -ordered;\n\t};\n}\n\nfunction toComparable(value: number | Date | string | undefined): number | string | undefined {\n\treturn value instanceof Date ? value.getTime() : value;\n}\n\nexport function getRepositoryIdentityForIssue(issue: IssueShape | Issue): IssueRepositoryIdentityDescriptor {\n\tif (issue.repository == null) throw new Error('Missing repository');\n\n\treturn {\n\t\tremote: {\n\t\t\turl: issue.repository.url,\n\t\t\tdomain: issue.provider.domain,\n\t\t},\n\t\tname: `${issue.repository.owner}/${issue.repository.repo}`,\n\t\tprovider: {\n\t\t\tid: issue.provider.id,\n\t\t\tdomain: issue.provider.domain,\n\t\t\trepoDomain: issue.repository.owner,\n\t\t\trepoName: issue.repository.repo,\n\t\t\trepoOwnerDomain: issue.repository.owner,\n\t\t},\n\t};\n}\n\nexport function serializeIssue(value: IssueShape): IssueShape {\n\tconst serialized: IssueShape = {\n\t\ttype: value.type,\n\t\tprovider: {\n\t\t\tid: value.provider.id,\n\t\t\tname: value.provider.name,\n\t\t\tdomain: value.provider.domain,\n\t\t\ticon: value.provider.icon,\n\t\t},\n\t\tid: value.id,\n\t\tnodeId: value.nodeId,\n\t\tissueType: value.issueType,\n\t\ttitle: value.title,\n\t\turl: value.url,\n\t\tcreatedDate: value.createdDate,\n\t\tupdatedDate: value.updatedDate,\n\t\tclosedDate: value.closedDate,\n\t\tclosed: value.closed,\n\t\tstate: value.state,\n\t\tproviderState:\n\t\t\tvalue.providerState == null\n\t\t\t\t? undefined\n\t\t\t\t: {\n\t\t\t\t\t\tid: value.providerState.id,\n\t\t\t\t\t\tname: value.providerState.name,\n\t\t\t\t\t\tcolor: value.providerState.color,\n\t\t\t\t\t\tcategory: value.providerState.category,\n\t\t\t\t\t},\n\t\tauthor:\n\t\t\tvalue.author == null\n\t\t\t\t? undefined\n\t\t\t\t: {\n\t\t\t\t\t\tid: value.author.id,\n\t\t\t\t\t\tname: value.author.name,\n\t\t\t\t\t\tusername: value.author.username,\n\t\t\t\t\t\tavatarUrl: value.author.avatarUrl,\n\t\t\t\t\t\turl: value.author.url,\n\t\t\t\t\t},\n\t\trepository:\n\t\t\tvalue.repository == null\n\t\t\t\t? undefined\n\t\t\t\t: {\n\t\t\t\t\t\towner: value.repository.owner,\n\t\t\t\t\t\trepo: value.repository.repo,\n\t\t\t\t\t\turl: value.repository.url,\n\t\t\t\t\t},\n\t\tproject:\n\t\t\tvalue.project == null\n\t\t\t\t? undefined\n\t\t\t\t: {\n\t\t\t\t\t\tid: value.project.id,\n\t\t\t\t\t\tname: value.project.name,\n\t\t\t\t\t\tresourceId: value.project.resourceId,\n\t\t\t\t\t\tresourceName: value.project.resourceName,\n\t\t\t\t\t},\n\t\tassignees: value.assignees.map(assignee => ({\n\t\t\tid: assignee.id,\n\t\t\tname: assignee.name,\n\t\t\tusername: assignee.username,\n\t\t\tavatarUrl: assignee.avatarUrl,\n\t\t\turl: assignee.url,\n\t\t})),\n\t\tlabels:\n\t\t\tvalue.labels == null\n\t\t\t\t? undefined\n\t\t\t\t: value.labels.map(label => ({\n\t\t\t\t\t\tcolor: label.color,\n\t\t\t\t\t\tname: label.name,\n\t\t\t\t\t})),\n\t\tcommentsCount: value.commentsCount,\n\t\tthumbsUpCount: value.thumbsUpCount,\n\t\tbody: value.body,\n\t\tbodyFormat: value.bodyFormat,\n\t\titerations:\n\t\t\tvalue.iterations == null\n\t\t\t\t? undefined\n\t\t\t\t: value.iterations.map(iteration => ({\n\t\t\t\t\t\tid: iteration.id,\n\t\t\t\t\t\tname: iteration.name,\n\t\t\t\t\t\tisActive: iteration.isActive,\n\t\t\t\t\t\tstartDate: iteration.startDate,\n\t\t\t\t\t\tendDate: iteration.endDate,\n\t\t\t\t\t})),\n\t};\n\treturn serialized;\n}\n"],"mappings":";;;;;;;;;;;AAkBA,MAAM,kBAA8G;CACnH,UAAS,MAAK,EAAE;CAChB,UAAS,MAAK,EAAE;CAChB,SAAQ,MAAK,EAAE;CACf,WAAU,MAAK,EAAE;CACjB,YAAW,MAAK,EAAE;CAClB,QAAO,MAAK,EAAE;AACf;;;;;;;;;;;;;AAcA,SAAgB,mBAAmB,MAA8C;CAChF,OAAO,QAAA;AACR;;;;;;;;;;;;;;;;;;;;;;;;;;;AA4BA,SAAgB,mBAAmB,MAA4E;CAE9G,MAAM,CAAC,OAAO,aAAa,KAAK,MAAM,GAAG;CACzC,MAAM,WAAW,gBAAgB;CACjC,IAAI,YAAY,MAAM,OAAO,KAAA;CAE7B,QAAQ,GAAG,MAAM;EAChB,MAAM,OAAO,aAAa,SAAS,CAAC,CAAC;EACrC,MAAM,QAAQ,aAAa,SAAS,CAAC,CAAC;EACtC,IAAI,QAAQ,QAAQ,SAAS,MAAM;GAClC,IAAI,QAAQ,QAAQ,SAAS,MAAM,OAAO;GAC1C,OAAO,QAAQ,OAAO,IAAI;EAC3B;EAEA,MAAM,UAAU,OAAO,QAAQ,KAAK,OAAO,QAAQ,IAAI;EACvD,OAAO,cAAc,QAAQ,UAAU,CAAC;CACzC;AACD;AAEA,SAAS,aAAa,OAAwE;CAC7F,OAAO,iBAAiB,OAAO,MAAM,QAAQ,IAAI;AAClD;AAEA,SAAgB,8BAA8B,OAA8D;CAC3G,IAAI,MAAM,cAAc,MAAM,MAAM,IAAI,MAAM,oBAAoB;CAElE,OAAO;EACN,QAAQ;GACP,KAAK,MAAM,WAAW;GACtB,QAAQ,MAAM,SAAS;EACxB;EACA,MAAM,GAAG,MAAM,WAAW,MAAM,GAAG,MAAM,WAAW;EACpD,UAAU;GACT,IAAI,MAAM,SAAS;GACnB,QAAQ,MAAM,SAAS;GACvB,YAAY,MAAM,WAAW;GAC7B,UAAU,MAAM,WAAW;GAC3B,iBAAiB,MAAM,WAAW;EACnC;CACD;AACD;AAEA,SAAgB,eAAe,OAA+B;CAoF7D,OAAO;EAlFN,MAAM,MAAM;EACZ,UAAU;GACT,IAAI,MAAM,SAAS;GACnB,MAAM,MAAM,SAAS;GACrB,QAAQ,MAAM,SAAS;GACvB,MAAM,MAAM,SAAS;EACtB;EACA,IAAI,MAAM;EACV,QAAQ,MAAM;EACd,WAAW,MAAM;EACjB,OAAO,MAAM;EACb,KAAK,MAAM;EACX,aAAa,MAAM;EACnB,aAAa,MAAM;EACnB,YAAY,MAAM;EAClB,QAAQ,MAAM;EACd,OAAO,MAAM;EACb,eACC,MAAM,iBAAiB,OACpB,KAAA,IACA;GACA,IAAI,MAAM,cAAc;GACxB,MAAM,MAAM,cAAc;GAC1B,OAAO,MAAM,cAAc;GAC3B,UAAU,MAAM,cAAc;EAC/B;EACH,QACC,MAAM,UAAU,OACb,KAAA,IACA;GACA,IAAI,MAAM,OAAO;GACjB,MAAM,MAAM,OAAO;GACnB,UAAU,MAAM,OAAO;GACvB,WAAW,MAAM,OAAO;GACxB,KAAK,MAAM,OAAO;EACnB;EACH,YACC,MAAM,cAAc,OACjB,KAAA,IACA;GACA,OAAO,MAAM,WAAW;GACxB,MAAM,MAAM,WAAW;GACvB,KAAK,MAAM,WAAW;EACvB;EACH,SACC,MAAM,WAAW,OACd,KAAA,IACA;GACA,IAAI,MAAM,QAAQ;GAClB,MAAM,MAAM,QAAQ;GACpB,YAAY,MAAM,QAAQ;GAC1B,cAAc,MAAM,QAAQ;EAC7B;EACH,WAAW,MAAM,UAAU,KAAI,cAAa;GAC3C,IAAI,SAAS;GACb,MAAM,SAAS;GACf,UAAU,SAAS;GACnB,WAAW,SAAS;GACpB,KAAK,SAAS;EACf,EAAE;EACF,QACC,MAAM,UAAU,OACb,KAAA,IACA,MAAM,OAAO,KAAI,WAAU;GAC3B,OAAO,MAAM;GACb,MAAM,MAAM;EACb,EAAE;EACL,eAAe,MAAM;EACrB,eAAe,MAAM;EACrB,MAAM,MAAM;EACZ,YAAY,MAAM;EAClB,YACC,MAAM,cAAc,OACjB,KAAA,IACA,MAAM,WAAW,KAAI,eAAc;GACnC,IAAI,UAAU;GACd,MAAM,UAAU;GAChB,UAAU,UAAU;GACpB,WAAW,UAAU;GACrB,SAAS,UAAU;EACpB,EAAE;CAEU;AACjB"}
+{"version":3,"file":"issue.utils.js","names":[],"sources":["../../../../git/src/utils/issue.utils.ts"],"sourcesContent":["import type {\n\tIssue,\n\tIssueRepositoryIdentityDescriptor,\n\tIssueShape,\n\tIssueSortField,\n\tIssueSorting,\n} from '../models/issue.js';\nimport { defaultIssueSort } from '../models/issue.js';\n\n/**\n * How each sort field is read off an {@link IssueShape}, for the reads that MERGE several provider queries and so\n * have to order the union themselves instead of letting the provider do it.\n *\n * Only the fields the shape actually carries appear. `priority`, `dueDate` and `resolved` are absent because\n * `IssueOrPullRequest`/`IssueShape` doesn't model them: a provider may well order by them server-side on a\n * single-query read, but a merged one cannot — which is exactly what {@link getIssueComparator} returning\n * `undefined` lets its caller refuse, instead of concatenating unordered runs and calling them sorted.\n */\nconst issueSortValues: Partial<Record<IssueSortField, (issue: IssueShape) => number | Date | string | undefined>> = {\n\tcreated: i => i.createdDate,\n\tupdated: i => i.updatedDate,\n\tclosed: i => i.closedDate,\n\tcomments: i => i.commentsCount,\n\treactions: i => i.thumbsUpCount,\n\ttitle: i => i.title,\n};\n\n/**\n * The order a read that ALWAYS orders is being made under: the caller's key, or the facade's default.\n *\n * The one place `?? defaultIssueSort` may be written. Several things downstream must agree on the answer — the\n * emitted provider qualifier, the fingerprint sealed into a cursor, the comparator that re-sorts a merged page and\n * the ceiling warning that names the order — and re-deriving the fallback at each is how they come to disagree,\n * which is not cosmetic: a cursor sealed with a key its query wasn't built with makes the fingerprint check refuse\n * valid resumptions.\n *\n * Note this is for a surface that CAN always order. A read whose provider may not express even the default resolves\n * through `resolveIssueSort` instead, which can answer \"leave the provider's own order\".\n */\nexport function effectiveIssueSort(sort: IssueSorting | undefined): IssueSorting {\n\treturn sort ?? defaultIssueSort;\n}\n\n/**\n * A comparator for one sort key over normalized issues, or `undefined` when the field isn't derivable from an\n * {@link IssueShape} — the signal that the key is only honorable on a SINGLE-origin read, where the provider\n * already ordered the page.\n *\n * `@gitkraken/provider-apis` carries a twin of this over its own `Issue` shape (`src/issueSort.ts`), which it uses\n * for the merges it performs itself. It is not importable — the SDK does not export it, and it takes a different\n * shape — so the two are separate implementations of one rule, and the rule below is the part that must not drift.\n *\n * Missing values sort LAST in both directions. That is deliberately not the same as negating the descending\n * comparator (which would float them to the front when ascending), and it is why they're partitioned before any\n * arithmetic rather than mapped to an infinite sentinel: two missing values have to compare equal, and\n * `-Infinity - -Infinity` is `NaN`, which makes the comparator inconsistent and the resulting order\n * implementation-defined.\n *\n * Note it orders whatever it is handed. Sorting a page that was already capped does not make it the top N.\n *\n * KNOWN DIVERGENCE, `title` only: strings are compared by code unit, so `'B' < 'a'`. A single-scope read is\n * ordered by the provider's own database instead — case-insensitive and accent-aware — so the same `title:asc`\n * yields \"Banana, apple, ábaco\" from a merged page and \"ábaco, apple, Banana\" from one repository. Deliberately\n * NOT fixed with `localeCompare`: the SDK's twin comparator (which performs the merges core doesn't, notably\n * GitLab's repo-ids aggregate) compares by code unit, so switching here would make two paths of the SAME\n * multi-repo read disagree — a worse divergence than the one it closes, and it would break the parity test in\n * `issueSortCapabilities.test.ts`. Closing this properly means changing the rule on BOTH sides at once; until\n * then it is a cosmetic ordering difference on one key, not wrong data.\n */\nexport function getIssueComparator(sort: IssueSorting): ((a: IssueShape, b: IssueShape) => number) | undefined {\n\t// The union's shape guarantees both halves, so this needs no validation — only the cast tsc can't infer.\n\tconst [field, direction] = sort.split(':') as [IssueSortField, 'asc' | 'desc'];\n\tconst getValue = issueSortValues[field];\n\tif (getValue == null) return undefined;\n\n\treturn (a, b) => {\n\t\tconst left = toComparable(getValue(a));\n\t\tconst right = toComparable(getValue(b));\n\t\tif (left == null || right == null) {\n\t\t\tif (left == null && right == null) return 0;\n\t\t\treturn left == null ? 1 : -1;\n\t\t}\n\n\t\tconst ordered = left < right ? -1 : left > right ? 1 : 0;\n\t\treturn direction === 'asc' ? ordered : -ordered;\n\t};\n}\n\nfunction toComparable(value: number | Date | string | undefined): number | string | undefined {\n\treturn value instanceof Date ? value.getTime() : value;\n}\n\nexport function getRepositoryIdentityForIssue(issue: IssueShape | Issue): IssueRepositoryIdentityDescriptor {\n\tif (issue.repository == null) throw new Error('Missing repository');\n\n\treturn {\n\t\tremote: {\n\t\t\turl: issue.repository.url,\n\t\t\tdomain: issue.provider.domain,\n\t\t},\n\t\tname: `${issue.repository.owner}/${issue.repository.repo}`,\n\t\tprovider: {\n\t\t\tid: issue.provider.id,\n\t\t\tdomain: issue.provider.domain,\n\t\t\trepoDomain: issue.repository.owner,\n\t\t\trepoName: issue.repository.repo,\n\t\t\trepoOwnerDomain: issue.repository.owner,\n\t\t},\n\t};\n}\n\n/**\n * Deliberately leaves `projection` out: this copy drops `repository.accessLevel`, so a tag would have field presence\n * report the viewer's access as fetched when the copy no longer has it.\n */\nexport function serializeIssue(value: IssueShape): IssueShape {\n\tconst serialized: IssueShape = {\n\t\ttype: value.type,\n\t\tprovider: {\n\t\t\tid: value.provider.id,\n\t\t\tname: value.provider.name,\n\t\t\tdomain: value.provider.domain,\n\t\t\ticon: value.provider.icon,\n\t\t},\n\t\tid: value.id,\n\t\tnodeId: value.nodeId,\n\t\tissueType: value.issueType,\n\t\ttitle: value.title,\n\t\turl: value.url,\n\t\tcreatedDate: value.createdDate,\n\t\tupdatedDate: value.updatedDate,\n\t\tclosedDate: value.closedDate,\n\t\tclosed: value.closed,\n\t\tstate: value.state,\n\t\tproviderState:\n\t\t\tvalue.providerState == null\n\t\t\t\t? undefined\n\t\t\t\t: {\n\t\t\t\t\t\tid: value.providerState.id,\n\t\t\t\t\t\tname: value.providerState.name,\n\t\t\t\t\t\tcolor: value.providerState.color,\n\t\t\t\t\t\tcategory: value.providerState.category,\n\t\t\t\t\t},\n\t\tauthor:\n\t\t\tvalue.author == null\n\t\t\t\t? undefined\n\t\t\t\t: {\n\t\t\t\t\t\tid: value.author.id,\n\t\t\t\t\t\tname: value.author.name,\n\t\t\t\t\t\tusername: value.author.username,\n\t\t\t\t\t\tavatarUrl: value.author.avatarUrl,\n\t\t\t\t\t\turl: value.author.url,\n\t\t\t\t\t},\n\t\trepository:\n\t\t\tvalue.repository == null\n\t\t\t\t? undefined\n\t\t\t\t: {\n\t\t\t\t\t\towner: value.repository.owner,\n\t\t\t\t\t\trepo: value.repository.repo,\n\t\t\t\t\t\turl: value.repository.url,\n\t\t\t\t\t},\n\t\tproject:\n\t\t\tvalue.project == null\n\t\t\t\t? undefined\n\t\t\t\t: {\n\t\t\t\t\t\tid: value.project.id,\n\t\t\t\t\t\tname: value.project.name,\n\t\t\t\t\t\tresourceId: value.project.resourceId,\n\t\t\t\t\t\tresourceName: value.project.resourceName,\n\t\t\t\t\t},\n\t\tassignees: value.assignees.map(assignee => ({\n\t\t\tid: assignee.id,\n\t\t\tname: assignee.name,\n\t\t\tusername: assignee.username,\n\t\t\tavatarUrl: assignee.avatarUrl,\n\t\t\turl: assignee.url,\n\t\t})),\n\t\tlabels:\n\t\t\tvalue.labels == null\n\t\t\t\t? undefined\n\t\t\t\t: value.labels.map(label => ({\n\t\t\t\t\t\tcolor: label.color,\n\t\t\t\t\t\tname: label.name,\n\t\t\t\t\t})),\n\t\tcommentsCount: value.commentsCount,\n\t\tthumbsUpCount: value.thumbsUpCount,\n\t\tbody: value.body,\n\t\tbodyFormat: value.bodyFormat,\n\t\titerations:\n\t\t\tvalue.iterations == null\n\t\t\t\t? undefined\n\t\t\t\t: value.iterations.map(iteration => ({\n\t\t\t\t\t\tid: iteration.id,\n\t\t\t\t\t\tname: iteration.name,\n\t\t\t\t\t\tisActive: iteration.isActive,\n\t\t\t\t\t\tstartDate: iteration.startDate,\n\t\t\t\t\t\tendDate: iteration.endDate,\n\t\t\t\t\t})),\n\t};\n\treturn serialized;\n}\n"],"mappings":";;;;;;;;;;;AAkBA,MAAM,kBAA8G;CACnH,UAAS,MAAK,EAAE;CAChB,UAAS,MAAK,EAAE;CAChB,SAAQ,MAAK,EAAE;CACf,WAAU,MAAK,EAAE;CACjB,YAAW,MAAK,EAAE;CAClB,QAAO,MAAK,EAAE;AACf;;;;;;;;;;;;;AAcA,SAAgB,mBAAmB,MAA8C;CAChF,OAAO,QAAA;AACR;;;;;;;;;;;;;;;;;;;;;;;;;;;AA4BA,SAAgB,mBAAmB,MAA4E;CAE9G,MAAM,CAAC,OAAO,aAAa,KAAK,MAAM,GAAG;CACzC,MAAM,WAAW,gBAAgB;CACjC,IAAI,YAAY,MAAM,OAAO,KAAA;CAE7B,QAAQ,GAAG,MAAM;EAChB,MAAM,OAAO,aAAa,SAAS,CAAC,CAAC;EACrC,MAAM,QAAQ,aAAa,SAAS,CAAC,CAAC;EACtC,IAAI,QAAQ,QAAQ,SAAS,MAAM;GAClC,IAAI,QAAQ,QAAQ,SAAS,MAAM,OAAO;GAC1C,OAAO,QAAQ,OAAO,IAAI;EAC3B;EAEA,MAAM,UAAU,OAAO,QAAQ,KAAK,OAAO,QAAQ,IAAI;EACvD,OAAO,cAAc,QAAQ,UAAU,CAAC;CACzC;AACD;AAEA,SAAS,aAAa,OAAwE;CAC7F,OAAO,iBAAiB,OAAO,MAAM,QAAQ,IAAI;AAClD;AAEA,SAAgB,8BAA8B,OAA8D;CAC3G,IAAI,MAAM,cAAc,MAAM,MAAM,IAAI,MAAM,oBAAoB;CAElE,OAAO;EACN,QAAQ;GACP,KAAK,MAAM,WAAW;GACtB,QAAQ,MAAM,SAAS;EACxB;EACA,MAAM,GAAG,MAAM,WAAW,MAAM,GAAG,MAAM,WAAW;EACpD,UAAU;GACT,IAAI,MAAM,SAAS;GACnB,QAAQ,MAAM,SAAS;GACvB,YAAY,MAAM,WAAW;GAC7B,UAAU,MAAM,WAAW;GAC3B,iBAAiB,MAAM,WAAW;EACnC;CACD;AACD;;;;;AAMA,SAAgB,eAAe,OAA+B;CAoF7D,OAAO;EAlFN,MAAM,MAAM;EACZ,UAAU;GACT,IAAI,MAAM,SAAS;GACnB,MAAM,MAAM,SAAS;GACrB,QAAQ,MAAM,SAAS;GACvB,MAAM,MAAM,SAAS;EACtB;EACA,IAAI,MAAM;EACV,QAAQ,MAAM;EACd,WAAW,MAAM;EACjB,OAAO,MAAM;EACb,KAAK,MAAM;EACX,aAAa,MAAM;EACnB,aAAa,MAAM;EACnB,YAAY,MAAM;EAClB,QAAQ,MAAM;EACd,OAAO,MAAM;EACb,eACC,MAAM,iBAAiB,OACpB,KAAA,IACA;GACA,IAAI,MAAM,cAAc;GACxB,MAAM,MAAM,cAAc;GAC1B,OAAO,MAAM,cAAc;GAC3B,UAAU,MAAM,cAAc;EAC/B;EACH,QACC,MAAM,UAAU,OACb,KAAA,IACA;GACA,IAAI,MAAM,OAAO;GACjB,MAAM,MAAM,OAAO;GACnB,UAAU,MAAM,OAAO;GACvB,WAAW,MAAM,OAAO;GACxB,KAAK,MAAM,OAAO;EACnB;EACH,YACC,MAAM,cAAc,OACjB,KAAA,IACA;GACA,OAAO,MAAM,WAAW;GACxB,MAAM,MAAM,WAAW;GACvB,KAAK,MAAM,WAAW;EACvB;EACH,SACC,MAAM,WAAW,OACd,KAAA,IACA;GACA,IAAI,MAAM,QAAQ;GAClB,MAAM,MAAM,QAAQ;GACpB,YAAY,MAAM,QAAQ;GAC1B,cAAc,MAAM,QAAQ;EAC7B;EACH,WAAW,MAAM,UAAU,KAAI,cAAa;GAC3C,IAAI,SAAS;GACb,MAAM,SAAS;GACf,UAAU,SAAS;GACnB,WAAW,SAAS;GACpB,KAAK,SAAS;EACf,EAAE;EACF,QACC,MAAM,UAAU,OACb,KAAA,IACA,MAAM,OAAO,KAAI,WAAU;GAC3B,OAAO,MAAM;GACb,MAAM,MAAM;EACb,EAAE;EACL,eAAe,MAAM;EACrB,eAAe,MAAM;EACrB,MAAM,MAAM;EACZ,YAAY,MAAM;EAClB,YACC,MAAM,cAAc,OACjB,KAAA,IACA,MAAM,WAAW,KAAI,eAAc;GACnC,IAAI,UAAU;GACd,MAAM,UAAU;GAChB,UAAU,UAAU;GACpB,WAAW,UAAU;GACrB,SAAS,UAAU;EACpB,EAAE;CAEU;AACjB"}

package/dist/git/utils/pullRequest.utils.d.ts

@@ -32,6 +32,10 @@ export declare function getPullRequestIdentityFromMaybeUrl(search: string): Pull
 export declare function getPullRequestNumberFromUrl(url: string): string | undefined;
 export declare function getRepositoryIdentityForPullRequest(pr: PullRequest, headRepo?: boolean): PullRequestRepositoryIdentityDescriptor;
 export declare function isMaybeNonSpecificPullRequestSearchUrl(search: string): boolean;
+/**
+ * Deliberately leaves `projection` out: this copy drops `filesChanged`, `repository` and the class-only check rollup,
+ * commit count and `viewerCanUpdate`, so a tag would have field presence report as fetched what the copy no longer has.
+ */
 export declare function serializePullRequest(value: PullRequest): PullRequestShape;
 //#endregion
 //# sourceMappingURL=pullRequest.utils.d.ts.map

package/dist/git/utils/pullRequest.utils.js

@@ -112,6 +112,10 @@ function getRepositoryIdentityForPullRequest(pr, headRepo = true) {
 function isMaybeNonSpecificPullRequestSearchUrl(search) {
 	return getPullRequestIdentityFromMaybeUrl(search) != null;
 }
+/**
+* Deliberately leaves `projection` out: this copy drops `filesChanged`, `repository` and the class-only check rollup,
+* commit count and `viewerCanUpdate`, so a tag would have field presence report as fetched what the copy no longer has.
+*/
 function serializePullRequest(value) {
 	return {
 		type: value.type,
@@ -176,6 +180,10 @@ function serializePullRequest(value) {
 		latestReviews: value.latestReviews,
 		assignees: value.assignees,
 		authoredByMe: value.authoredByMe,
+		viewer: value.viewer != null ? {
+			id: value.viewer.id,
+			username: value.viewer.username
+		} : void 0,
 		project: value.project ? {
 			id: value.project.id,
 			name: value.project.name,