Skip to content

feat: bump public types#9232

Merged
davidkaplanbitgo merged 1 commit into
masterfrom
WCN-1175.bump-sdk
Jul 10, 2026
Merged

feat: bump public types#9232
davidkaplanbitgo merged 1 commit into
masterfrom
WCN-1175.bump-sdk

Conversation

@davidkaplanbitgo

Copy link
Copy Markdown
Contributor

WCN-1175

TICKET: WCN-1175

@linear-code

linear-code Bot commented Jul 10, 2026

Copy link
Copy Markdown
Contributor

WCN-1175

@davidkaplanbitgo davidkaplanbitgo requested a deployment to breaking-changes-override July 10, 2026 17:21 — with GitHub Actions Waiting
@davidkaplanbitgo davidkaplanbitgo marked this pull request as ready for review July 10, 2026 19:33
@davidkaplanbitgo davidkaplanbitgo requested a review from a team as a code owner July 10, 2026 19:33
@davidkaplanbitgo davidkaplanbitgo temporarily deployed to breaking-changes-override July 10, 2026 19:43 — with GitHub Actions Inactive
@vmccarty

Copy link
Copy Markdown
Contributor

Breaking change linter override approved. Re:

None of the 98 findings is the removal of a required, client-facing, in-production response field. So the premise ("each was intentionally broken elsewhere and got a DevEx override") doesn't map 1:1 — most aren't breaks at all. But every real change traces to a deliberate, reviewed, released public-types PR with a Linear ticket, and several were explicitly coordinated with the owning service:

Irys revert → #396 (SC-5765/6/7), owner-driven
DeFi clientIdempotencyKey/restructure → #389 (CGD-1400), #414 (CGD-2007, body: "Unblocks CGD-1821 (wallet-platform) and CGD-1822 (BitGoJS)")
Canton choiceArgument optional → v6.26.0
goStaking optionality → v6.29.0
The reason there's no override "in the public-types repo" is exactly the gap you identified: public-types has no breaking-change CI, and the consuming service (wallet-platform, which owns signtxtss/sweep) drove these via tickets rather than through a spec-diff gate.

Recommendation
Both PRs (9232 → 6.41.0; 9211 → ^6.39.0, same range, subset of the same findings) are safe to override and approve on client-impact grounds — the breaking-change gate is misfiring on oneOf reindexing, not detecting real contract breaks. I'd note in the DevEx override request: "All findings are positional oneOf artifacts in the signtxtss/sweep intent union from additive public-types changes 6.22→6.41 plus the intentional Irys revert (#396); verified surviving fields still present; no required client-facing field removed."

Two things I can do next if useful:

Pull the literal override/approval paper trail from wallet-platform (owner of these endpoints) to attach as corroboration — I'd need to add that repo to the session.
For the CI/pipeline fix: the concrete bug is the positional-index diff on large unions. The durable fix is a name/discriminator-aware oneOf comparison (match variants by intentType literal, not array position) in BitGo/static-analysis's breaking-changes rule — I can scope that out.
Want me to grab the wallet-platform override records, or start on the static-analysis fix?

@davidkaplanbitgo davidkaplanbitgo merged commit 35edadb into master Jul 10, 2026
23 of 24 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants