Commit Graph

2 Commits

Author SHA1 Message Date
Grant Whitmer
b6a7045907 G7.4/G11.3: import tool — GitHub upstream, Windy Git downstream, CI here
All checks were successful
check / gate (push) Successful in 17s
Direction is deliberate. For the migration quarter GitHub stays upstream and
Windy Git is a PULL mirror, so it cannot diverge — worst case it is stale, not
wrong. Making Windy Git authoritative before Grant flips G11.6 would create a
two-writer problem nobody asked for. I-4's push-mirror is the steady state for
repos that originate here.

Refuses windy-pro by name: six checkouts, a build counter forked three ways, and
two sessions recording different HEADs hours apart. That gets resolved by
reading, not by importing (G11.5).

Records the Cloudflare trap: the public endpoint answers 403 error 1010 because
CF blocks urllib's user-agent as a bot signature. It reads like Gitea rejecting
the token and is not — the identical call against localhost:3080 on the host
succeeds. Bulk import belongs on the host anyway.

Verified: windy-calendar imported, 731 KB, mirroring.

Survey: 135 private repos, and GitHub Actions cannot run on ANY of them. Sampled
windy-pro, windy-mind, eternitas, windy-registry, windy-agent, WindyCloud — every
recent run is startup_failure, as recently as today.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-12 12:12:02 -04:00
Grant Whitmer
659991b2bd G0: cell substrate — invariants made executable
Strand G0 complete and VERIFIED against real Postgres, not asserted.

  - FastAPI plane, fail-closed provider seams, repair-pointer error taxonomy
  - migration 001: all 10 tables incl. repo_type NOT NULL and model_cards (I-7)
  - 17 invariant tests, ruff clean, vocabulary audit clean

Two bugs found by RUNNING it that review would not have caught:

  1. SQLAlchemy Enum persists .name, not .value — so RepoState.deleted_soft
     and CreatedVia.imported would have written labels migration 001 never
     declared, failing at runtime rather than at review. Pinned via
     values_callable.
  2. op.create_table asks each Enum to emit its own CREATE TYPE with no
     checkfirst, so the second reference raised DuplicateObject and the
     migration died halfway. Types are now created once, referenced with
     create_type=False.

Proven live, with the hostile env var set:
  - I-12: COMMIT_SHA=deadbeef... in the environment, /version reports real HEAD.
    That env pin is the documented root cause of nine sibling services
    misreporting their commit; here it is structurally ignored.
  - I-8: three unconfigured providers -> status degraded, HTTP 503, each saying
    'refusing to report healthy'. No mock, no false green.
  - G0.4: upgrade -> downgrade -> upgrade round-trip clean (10 -> 0 -> 10).

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-11 14:19:28 -04:00