Skip to content

bank-feeds: release v12.0.2 (in-house codegen) - #662

Open
codat-sdk-generator[bot] wants to merge 1 commit into
mainfrom
generate/bank-feeds
Open

bank-feeds: release v12.0.2 (in-house codegen)#662
codat-sdk-generator[bot] wants to merge 1 commit into
mainfrom
generate/bank-feeds

Conversation

@codat-sdk-generator

Copy link
Copy Markdown

Regenerated by generate-and-pr run 31394959499 (https://github.com/codat-internal/sdk-codegen/actions/runs/31394959499) from codat-internal/sdk-codegen @ b6a39d2. Version 12.0.1 -> 12.0.2, stamped in packaging and RELEASES.md. All blocking gates ran green in the linked run. Merging this PR is the release trigger.

@mcollinscodat

Copy link
Copy Markdown

Analysis
Fresh worktree at b6a39d2 (the generating commit named in the PR body, which is also current sdk-codegen main), regenerated bank-feeds Python, cloned client-sdk-python main depth-1, fetched generate/bank-feeds as pr662, then ran the same staging script CI runs. I confirmed the PR commit's parent is exactly today's main (2348452), so the clone is the right base.

Result: one diff, and it's the clock

bank-feeds/RELEASES.md

That's the whole diff against the PR branch, and it's the timestamp line only (13:59:45 in CI vs 14:21:37 in our run). Nothing differs outside bank-feeds/. The staging script printed old_version=12.0.1, version=12.0.2, matching the PR. I also re-ran the surface gate (extract_surface + diff_surface --fail-on-regression) on my tree: green, same as CI.

The version question is answered

bank-feeds/pyproject.toml on main already says 12.0.1, so a patch bump gives 12.0.2. Correct behaviour. The catch: PyPI never got 12.0.1. It's absent from the release list entirely, not yanked, so Speakeasy bumped the repo and appended a RELEASES.md entry but the publish never ran. Two consequences worth knowing:

  • PyPI will jump 12.0.0 → 12.0.2. Fine, just a skipped number.
  • bank-feeds/RELEASES.md on main already carries [PyPI v12.0.1] https://pypi.org/project/codat-bankfeeds/12.0.1, a link to a version that doesn't exist. Pre-existing, not caused by this PR.

The oracle pin is clean here: bank-feeds is pinned at 12.0.0, which is what's actually published, so there's no repeat of the payables 9.1.0-vs-9.0.2 trap.

The 24 deletions all have a home

  • 18 files are Speakeasy's own runtime (utils/*, basesdk.py, sdkconfiguration.py, types/basemodel.py), pinned as accepted absences in python_runtime_surface_absent.json.
  • models/shared/banktransactions.py goes, but BankTransactions still imports. I checked by importing: codat_bankfeeds.models.shared.BankTransactions resolves to BankTransaction through the lazy alias map.
  • Of the 5 dropped docs pages, bankaccountsmetadata and bankaccountssupplementaldata are the shared-alias decision that's committed in pipeline/compat_plan/bank-feeds.python.json, and both names still import. The other three (createbatchsourceaccount1, createbatchsourceaccountsourceaccountsresponsebody, one) are types that don't exist in the Speakeasy 12.0.0 surface either, so nothing is lost.
  • The 3 renames are docs pages tracking model renames.

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.

1 participant