Say work item where the runner, repos, deployments and kit said issue
- G1T-Session:
- ses_01m566357hejh9yeqjbetx23f3
9 files+10−100/9 viewed
| 235 | 235 | "metadata": { "workspaceId": "wsp_1" }, | |
| 236 | 236 | "labels": { "goodFirstIssue": 3 }, | |
| 237 | 237 | "byModel": { "claudeOpus": 1.5 }, | |
| 238 | − | "data": { "issueId": "iss_1", "outputs": { "artifactId": "a" } }, | |
| 238 | + | "data": { "itemId": "iss_1", "outputs": { "artifactId": "a" } }, | |
| 239 | 239 | })); | |
| 240 | 240 | assert_eq!( | |
| 241 | 241 | sent, | |
| ⋯ | |||
| 252 | 252 | "metadata": { "workspaceId": "wsp_1" }, | |
| 253 | 253 | "labels": { "goodFirstIssue": 3 }, | |
| 254 | 254 | "by_model": { "claudeOpus": 1.5 }, | |
| 255 | − | "data": { "issue_id": "iss_1", "outputs": { "artifactId": "a" } }, | |
| 255 | + | "data": { "item_id": "iss_1", "outputs": { "artifactId": "a" } }, | |
| 256 | 256 | }) | |
| 257 | 257 | ); | |
| 258 | 258 | assert!(camel_case_keys(&sent).is_empty()); | |
| 1 | 1 | //! Runs commands against one commit and reports how each went: the merge | |
| 2 | 2 | //! queue's runner of commands, and `MODE=checks`, which ran commands written | |
| 3 | − | //! on issues before a pull request's checks were its workflows. g1t no | |
| 3 | + | //! on work items before a pull request's checks were its workflows. g1t no | |
| 4 | 4 | //! longer starts that mode; it stays for a sandbox already under way. | |
| 5 | 5 | //! | |
| 6 | 6 | //! The sandbox holds nothing but that commit, so a passing result says |
| 6 | 6 | //! `mergecheck` finds out whether a pull request merges cleanly, `deploy` | |
| 7 | 7 | //! builds a commit for g1t.page, `bump` makes a dependency update (one | |
| 8 | 8 | //! package raised in its lockfiles, pushed as g1t), `backup` cuts a | |
| 9 | − | //! repository's nightly backup bundle, and `checks` runs an issue's | |
| 9 | + | //! repository's nightly backup bundle, and `checks` runs a work item's | |
| 10 | 10 | //! commands, from before checks were workflows. See the modules of those | |
| 11 | 11 | //! names. It talks to g1t only through the public API and git. | |
| 12 | 12 | //! |
| 18 | 18 | //! branch, commit}]`, the entry being tested last. | |
| 19 | 19 | //! - `CHECKS`: commands to run on the state, as a JSON array. g1t sends | |
| 20 | 20 | //! none now: the state is checked by its `merge_group` workflows. | |
| 21 | − | //! - `CONTRACT_CHECKS`: the checks of issues already completed, which the | |
| 21 | + | //! - `CONTRACT_CHECKS`: the checks of work items already completed, which the | |
| 22 | 22 | //! default branch has to keep passing. One that fails is run again on the | |
| 23 | 23 | //! base alone; if it fails there too, it was broken already and is not | |
| 24 | 24 | //! held against the stack. |
| 2079 | 2079 | }); | |
| 2080 | 2080 | if (behind && attempts < RENAME_WAITS) throw new Error(`projects has not seen ${current} moved yet`); | |
| 2081 | 2081 | ||
| 2082 | − | // Its history goes with it, as the repository's issues do. | |
| 2082 | + | // Its history goes with it, as its pull requests do. | |
| 2083 | 2083 | const statements: D1PreparedStatement[] = [ | |
| 2084 | 2084 | this.db.prepare("UPDATE deployments SET repo = ? WHERE repo_id = ?").bind(current, move.repoId), | |
| 2085 | 2085 | ]; |
| 1 | 1 | //! Pull requests' working copies after their pull request is done. | |
| 2 | 2 | //! | |
| 3 | − | //! Every pull request made from an issue works in its own copy of the | |
| 3 | + | //! Every pull request made for a work item works in its own copy of the | |
| 4 | 4 | //! repository in the git store (`pulls--<pull id>`, `fork_for_pull`). | |
| 5 | 5 | //! Cloudflare does not say whether a fork shares objects with its source, | |
| 6 | 6 | //! and storage is billed and capped per account (1 TB), so copies are not |
| 1 | 1 | //! Forks people make into their workspaces: a fork network. | |
| 2 | 2 | //! | |
| 3 | 3 | //! A fork is an ordinary repository of the workspace it goes in, with its | |
| 4 | − | //! own access, settings, rules, issues and pull requests, that remembers the | |
| 4 | + | //! own access, settings, rules and pull requests, that remembers the | |
| 5 | 5 | //! repository it came from (`parent_id`). That is all it shares: it can be | |
| 6 | 6 | //! synced from its parent and can open pull requests into it. It is not a | |
| 7 | 7 | //! pull request's working copy (`fork_of`, namespace `pulls`, forks.rs), |
| 1 | 1 | //! `pushed`: what a push to a branch brought into it. Work asks after each | |
| 2 | 2 | //! push, to mark merged the pull requests whose heads it brought in, as a | |
| 3 | − | //! merge from the site would, and to close the issues its commits' messages | |
| 3 | + | //! merge from the site would, and to land the work items its commits' messages | |
| 4 | 4 | //! name after a closing keyword (work's pushed.rs). | |
| 5 | 5 | //! | |
| 6 | 6 | //! One walk from the new head, every parent followed, stopping at what the |
| 115 | 115 | } | |
| 116 | 116 | ||
| 117 | 117 | /** | |
| 118 | − | * What a sandbox is doing. Runs of an issue's commands (`checks`) are from | |
| 118 | + | * What a sandbox is doing. Runs of a work item's commands (`checks`) are from | |
| 119 | 119 | * before checks were workflows; a sandbox that was started for work g1t no | |
| 120 | 120 | * longer does is told apart by its kind and only cleaned up. | |
| 121 | 121 | */ |
Comments
Sign in to comment.