Skip to content

Commit bfbfa4e

Browse files
committed
Merge remote-tracking branch 'origin/dev' into codex/absorb-baseurl-override
# Conflicts: # devlog/_plan/260820_bug_pr_backlog_consolidation/070_execution_log.md
2 parents d2493a1 + 7fc5084 commit bfbfa4e

83 files changed

Lines changed: 5986 additions & 303 deletions

File tree

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

‎devlog/_plan/260820_bug_pr_backlog_consolidation/070_execution_log.md‎

Lines changed: 113 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -178,3 +178,116 @@ Two gaps named in the PR rather than carried silently: Direct `/v1/models` can s
178178
Pool-only grant (advertisement only; dispatch still checks the caller credential), and
179179
same-account gated-400 retry stays Pool-only.
180180

181+
182+
## wp12 — PR #2053 (@Ingwannu), the C4 OAuth hold
183+
184+
**PR #2149**, branch `codex/absorb-oauth-superseded-commit`, base `dev`.
185+
186+
Applied unchanged, rebased from 145 commits behind. The persist-boundary placement is the whole
187+
design: `assertBeforePersist` runs inside the file lock, after `fn(store)` and before
188+
`persist()`, so a superseded flow's in-memory mutation is discarded rather than written.
189+
Ownership is identity-checked against the flow's own `AbortController`, not a timestamp.
190+
191+
**This was a wp1 HOLD and it is resolved by shipping, not by absorbing quietly.** The PR states
192+
plainly that MAINTAINERS.md mandates security review and asks that it not be merged on my
193+
verification alone, and it names three residuals rather than letting the original claim stand:
194+
195+
1. the description claimed reauth coverage; the diff wires the hook but adds no reauth test
196+
2. `OAuthLoginSupersededError` is not in the public allowlist, so it projects to the generic string
197+
3. a never-finishing Kiro rollback blocks all replacements, by design
198+
199+
Evidence: reverting `src/oauth/` fails 2 tests including the cancel-then-replace round trip;
200+
full suite 13536 pass / 0 fail.
201+
202+
203+
## wp14 — PR #2075 (@olddonkey): the second false negative
204+
205+
**PR #2151**, branch `codex/absorb-fastwire-native-chat`, base `dev`. Closes #1886.
206+
207+
Same class of error as wp13, different cause. I saw `CONFLICTING` and treated it as a reason not
208+
to read the diff. The rescore put it at **67**: native `/v1/chat/completions` decided
209+
`service_tier` from `chatServiceTier` alone, so a `supportsServiceTier: false` declaration was
210+
fail-open. The conflict was why it could not MERGE, not why it should score LOW — and resolving
211+
it took one import line.
212+
213+
Rebase, stated exactly: `src/adapters/openai-chat.ts` conflicted because `dev` added
214+
`AdapterTierMetadata` while the PR adds `decideTier` and `ResolvedFastPolicy`. Both kept.
215+
Everything else clean. Typecheck is what confirms the resolution.
216+
217+
Evidence: reverting `src/` fails the characterization test the author had flipped from
218+
documented-known-bug to passing assertion. Full suite 13552 pass / 0 fail.
219+
220+
## Two scoring lessons, recorded together
221+
222+
wp13 and wp14 were both my errors, from two different shortcuts:
223+
224+
1. **Scoring from titles** — "preserve and replay thought signatures" reads like bookkeeping and
225+
was a core provider 400.
226+
2. **Reading merge state as value** — `CONFLICTING` says a patch cannot land today; it says
227+
nothing about whether the defect matters.
228+
229+
Both produce false negatives that are indistinguishable from correct low scores without opening
230+
the diff. The rubric was fine; the inputs I fed it were not.
231+
232+
233+
# Campaign close (final)
234+
235+
## 13 PRs open, all green, all MERGEABLE
236+
237+
| PR | Fixes | Credit | Base |
238+
|---|---|---|---|
239+
| #2137 | issue #2132 | new work | dev |
240+
| #2138 | issue #2092 | @lilinxiong | dev |
241+
| #2140 | #2100 + #2077 | @ntdatt812 | dev |
242+
| #2141 | issue #2047 | @Ingwannu | dev |
243+
| #2142 | #2131 | @bet4it | dev |
244+
| #2144 | #2105 | @lilinxiong | dev |
245+
| #2145 | issue #1950 | @Ingwannu | dev |
246+
| #2146 | issue #2097 | @Ingwannu | **#2137 branch (stacked)** |
247+
| #2147 | issue #1886 | @olddonkey | dev |
248+
| #2148 | #2109 + #2110 | @drakonkat | dev |
249+
| #2149 | #2053 | @Ingwannu | dev |
250+
| #2150 | issue #2125 | @agentHits | dev |
251+
| #2151 | issue #1886 | @olddonkey | dev |
252+
253+
Plus #2134, which opened this session.
254+
255+
## 16 PRs closed with attribution
256+
257+
#2102, #2099, #2091, #2029, #2063, #2100, #2077, #2056, #2062, #2131, #2105, #2040, #2101,
258+
#2104, #2109, #2110, #2053, #2127, #2075.
259+
260+
Every one carries a comment naming its replacement, what was carried over, and what was
261+
deliberately not. Where a contributor's own assertion had to be replaced — #2141's scorer case,
262+
#2142's `msg_ocx_0` case — the replacement is disclosed in both the comment and the PR body.
263+
264+
## 6 remain, independently verified below threshold
265+
266+
#2115 (58), #2082 (46), #2067 (38), #2054 (58), #2032 (37), #2027 (51).
267+
268+
These are not omissions. A rescore lane read every diff and scored them against the same rubric;
269+
it found exactly two false negatives in my original triage (#2127 at 83, #2075 at 67) and both
270+
were absorbed as wp13 and wp14. The remaining six are genuinely below the line, and four of them
271+
are additionally blocked (draft, CONFLICTING, or CHANGES_REQUESTED).
272+
273+
Two of them carry real bugs attached to unabsorbable patches: #2054's Cursor context collapse
274+
(#1527) and #2027's Go quota gating (#1924). The right move for both is a clean reimplementation
275+
on `dev`, not absorbing a 19-file conflicting draft. That is stated rather than silently skipped.
276+
277+
## Corrections made on top of contributor work
278+
279+
Eight PRs shipped with fixes the originals were missing, each pinned by a test verified to fail
280+
against the contributor's own source:
281+
282+
- #2141 short-only scorer returning 0 instead of UNKNOWN
283+
- #2142 duplicate `msg_ocx_0` from a collapsed index
284+
- #2145 history-only arming and non-atomic SSE overflow
285+
- #2146 selector compact bypassing the wire rewrite, Direct callers evicting catalog cache
286+
- #2148 `allowPrivateNetwork` bypassing the HTTPS gate for public hosts
287+
- #2138 `gpt-5.60` near-miss match
288+
289+
## Not merged
290+
291+
DEV-STACK-04 and DEV-GIT-PUSH-01 both put merge authorization with the user. #2137 must land
292+
before #2146.
293+

0 commit comments

Comments
 (0)