Adopt basecamp-sdk v0.12.0 - #610
Conversation
Eleven Everything() aggregate methods gained a trailing *EverythingTaskFilters parameter: the nine paginated todo and card selectors, plus the two unpaginated overdue endpoints. Pass nil at every call site, so this bump is behavior-preserving — no test changes, and the account-wide listings issue exactly the requests they did before. The filters themselves are the point of the signature change, but threading real values through belongs with the flags that produce them. This commit is the mechanical half. SDK v0.12.0 is 7e2925d25078; the API provenance moves to bc3 d0edc1283b23.
github-actions
Bot
added
commands
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 4e232e9e30
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| charm.land/bubbletea/v2 v2.0.8 | ||
| charm.land/lipgloss/v2 v2.0.5 | ||
| github.com/basecamp/basecamp-sdk/go v0.11.0 | ||
| github.com/basecamp/basecamp-sdk/go v0.12.0 |
There was a problem hiding this comment.
When this bump lands independently, the repository pins v0.12.0 here while API-COVERAGE.md still identifies v0.11.0 as the pinned SDK and does not document the v0.12 command surface. That leaves the repository's SDK coverage record inaccurate and defers required commands for newly available service methods to a later change; include the corresponding command and coverage updates with this bump.
AGENTS.md reference: AGENTS.md:L118-L130
Useful? React with 👍 / 👎.
|
Closing as superseded by #611, which now targets Review was right that this is not a valid independent landing unit. At #611 carries this commit plus the surface work, so A+B is one complete SDK Correcting my own earlier claim: I described this PR as "fully green (22 |
Bumps the SDK to v0.12.0 and lands the command surface it opened, as one complete unit: splitting the bump from the coverage that documents it produced a PR that was green but violated the completeness bar (#610, closed). Adds bookmarks, drafts, notes and calendars, the Up Next verbs, and the remaining v0.12.0 operations on todos and checkins. Every new SDK service method gets a command file, catalog entry, registration and an API-COVERAGE.md row. Review found four defects that bin/ci could not see: notes set could silently discard piped content — stdin was consulted only after an argument and --file had been ruled out, so a generated body lost to a fallback file in the one command that replaces everything. checkins reminders silently ignored --project/--in/--questionnaire, inherited as persistent flags from the parent while the handler always called the account-wide endpoint. --ids-only emitted the wrong id twice over: assignments exposed the parent card rather than the prioritized step, so piping into deprioritize reported success and changed nothing (204 either way); reminders exposed no id at all, so it printed nothing. API-COVERAGE.md claimed a 400's exit code moves from 7 to 9. It does not — there is no validation mapping and clioutput defaults to ExitAPI, so a 400 still exits 7. The note-clearing gap is documented and deferred to its own change.
PR 1 of 3. Bump only — no new commands. B (
bc5-command-surface) and C(
account-wide-task-filters) stack on top.What changed
Everything()gained a parameter in v0.12.0. Eleven call sites acrosscards.goandtodos.gopassnil, preserving today's behaviour exactly.internal/version/sdk-provenance.jsonmoves to v0.12.0.vendorHash
nix/package.nixis updated tosha256-+j9bY0gS…viamake update-nix-hash.A go.mod change invalidates it and the
nix-buildjob gates on it, so the bumpcannot land without this. Recomputed and re-verified by rebuild — through the
shared classifier that #609 just extracted, which is a nice first real exercise
of it.
Rebase provenance
Rebased from
1bf961f2ontoec979fc8.git range-diffagainstbackup/sdk-0.12.0-pre-rebaseshows the only intra-commit change is thevendorHashline; everything else is byte-identical.go.mod resolved cleanly because the two moves are disjoint — main bumped goldmark
1.8.4 → 1.8.5 and left the SDK alone; this branch bumps the SDK and leaves
goldmark alone. Result is SDK v0.12.0 + goldmark v1.8.5, which is the intended
resolution. No
replacedirective.Verification
bin/ciexit 0 at this head.carried goldmark 1.8.4, main carries 1.8.5, and
internal/commands/notes.gorenders through
richtext.MarkdownToHTML— so the renderer under that pathchanged on rebase. Rendering one corpus (headings, nested lists, tables, fenced
code, footnotes, emoji, raw HTML, hard breaks, entities) under both versions
produces byte-identical output.
notes set --file→notes showreturns correct HTML for headings, emphasis, lists, links (
target/rel),blockquote and entity escaping. Test note restored to its prior content
byte-for-byte afterwards.
Backup branch:
backup/sdk-0.12.0-pre-rebaseSummary by cubic
Upgrade to
github.com/basecamp/basecamp-sdk/gov0.12.0 and updateEverything()selectors to accept the new filters param, passingnilto keep behavior unchanged. No user-facing changes; sets up future account-wide filtering.go.mod/go.sum.nix/package.nixvendorHash.internal/version/sdk-provenance.json.Written for commit 4e232e9. Summary will update on new commits.