Repository navigation
Retire the six DeclaredCallableIdentity admission rows by their trigger (#9665 merged; they refused every PR) - #9705
Merged
Conversation
…dissolved on #9665 merging Every pull_request build now has ecdeb49 (or later) as its base, so base and merge head both carry the hoist, no run can produce those six deltas, and the rows report stale -- refusing every open PR (measured on #9689 @ bfd9524: 0 unadjudicated, 6 stale). Removed by their own dissolve-on trigger; the roster is empty and empty is not permissive. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01GXfYKNQTD3VfYyQcnJpxNU
gunbai-bot
Bot
changed the title
gunbai-bot Bot
pushed a commit
that referenced
this pull request
Aug 30, 2026
…red record and this branch's two live rows Both sides removed #9665's six stale admission rows — this branch in its previous merge commit, main in #9705 with the measured receipt. The resolution keeps main's richer shrink record verbatim and re-seats this branch's two RequiredCiLane rows, which still match live deltas until #9698 merges. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01GpYEk9gSYmMkwL1489D79U
This was referenced Aug 30, 2026
gunbai-bot Bot
pushed a commit
that referenced
this pull request
Aug 30, 2026
…merging -- this refuses EVERY pull request, not just this one Not this lane's subject, taken because it stands between every branch and a green. #9698 (fa03a01) moved BuildLane and WitnessesLane from gunbc.required_ci_host_verdict_census to the new gunbc.required_ci_phase_roster. The two TransitionAdmission rows that adjudicated that move were authored with the trigger "they go STALE the moment #9698 merges and MUST be removed then" -- their own words -- and #9698 is merged, so no pull_request build can produce the deltas they name and the phase refuses on the stale rows alone. Measured, run 33292088803: FAILED PHASE namespace-wave-admission (0 unadjudicated delta(s), 2 stale admission(s)), naming both rows. This branch's own namespace delta -- the gunbc.repo_self_build -> gunbc.stage0_crate_partition_generated import the roster fix adds -- was adjudicated ExplicitlyEvaluatedZeroDelta and passed, so the refusal is entirely the two stale rows. Emptying the array to &[] is the proven form: the fourth shrink (#9705) did exactly that and shipped green. That shrink was a standalone PR and that remains the normal shape; it rides here only because main was already red. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01RTvKD3J3pnUzXmUascawfQ
briansrls
pushed a commit
that referenced
this pull request
Aug 30, 2026
…, and the required build lane gains the partition crates as the wall (#9711) * Stage0 partition roster: the three modules #9689 added, the phantom row, and the required gate that can see them The authority edits only; generated artifacts follow in the next commit. - v2.workflow.rust_crate_partition: std_target_representation into the std-core unit, extdeps_languages_rust_representation into the extdeps-languages unit, and gunbc_rust_source_type_bindings into the v1-infer unit through its own named list, because its layer is neither the v1_compiler pipeline nor the extdeps one and the grouping is physical rather than nominal. Placement is forced by each module's own crate:: references, not chosen. - The same authority drops std_lens_verdict, which named a file stage0 lib.rs deliberately does not declare (has_pub_mod false in the crate-layout registration), so the partition crates were compiling a module the monolith does not. The FILE's disposition is deliberately not settled here. - gunbc.repo_self_build gains repo_self_build_partition_crates_command, whose package list is read from the partition roster rather than restated, and the build lane runs it beside the all-bins build. That is the executing wall for the E0432 class, with rustc as the oracle. - One witness for the class a hermetic join CAN reach, with its discriminating RED authored as a fixture because the live corpus cannot express the refused state once the phantom row is gone. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01RTvKD3J3pnUzXmUascawfQ * Move the roster witness inside the required gate's prefix; convert the v1-infer count literal to an identity join Run 33290382542 measured both of these rather than my guessing at them. - The new witness sat at `test.claim.stage0_partition_roster_layout_join_witness`, which matches none of v2.workflow.required_floor required_gate_prefixes, so the floor discovered it, classified it DeclinedOutsideRequiredGate and never ran it -- while it was RED on the live corpus. A wall that reds and is declined is an inert lens, so the module path is part of the wall and the file now lives at v2.test.claim.*, which the `v2.test.` prefix admits. - witness_v1_infer_unit_members_holds was the run's single unexpected failure: `(unit.members |> count) == 9`, stale the moment the roster gained a module. It is now an identity join between the declared roster and what came back through the assignment map and partition_fold, so it cannot be repaired by retyping a number. Its blind spot -- an error in the roster itself -- is stated in place and belongs to the build lane, where rustc is the oracle. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01RTvKD3J3pnUzXmUascawfQ * Sixth admission shrink: the two RequiredCiLane rows dissolved on #9698 merging -- this refuses EVERY pull request, not just this one Not this lane's subject, taken because it stands between every branch and a green. #9698 (fa03a01) moved BuildLane and WitnessesLane from gunbc.required_ci_host_verdict_census to the new gunbc.required_ci_phase_roster. The two TransitionAdmission rows that adjudicated that move were authored with the trigger "they go STALE the moment #9698 merges and MUST be removed then" -- their own words -- and #9698 is merged, so no pull_request build can produce the deltas they name and the phase refuses on the stale rows alone. Measured, run 33292088803: FAILED PHASE namespace-wave-admission (0 unadjudicated delta(s), 2 stale admission(s)), naming both rows. This branch's own namespace delta -- the gunbc.repo_self_build -> gunbc.stage0_crate_partition_generated import the roster fix adds -- was adjudicated ExplicitlyEvaluatedZeroDelta and passed, so the refusal is entirely the two stale rows. Emptying the array to &[] is the proven form: the fourth shrink (#9705) did exactly that and shipped green. That shrink was a standalone PR and that remains the normal shape; it rides here only because main was already red. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01RTvKD3J3pnUzXmUascawfQ * Regenerated: the partition roster, its stage0 mirror, the three crate lib.rs files, and the build lane that now compiles them Produced by the generators, not by hand. main_wet (dag/gunbc/instruments/ generated_artifact_gate.dag) rewrote exactly two artifacts -- the generated partition .dag and .github/workflows/witnesses.yml -- and nothing else in the corpus moved. --required-regen then reported one drifted mirror, gunbc_stage0_crate_partition_generated.rs, and --emit-partition-crates rendered 14 files and wrote 3: the std-core, extdeps-languages and v1-infer lib.rs, which are precisely the three crates the roster change touches. The emitted build step, read off the emitted bytes rather than assumed: run: | cargo build --release -p v1-compiler --bins cargo build --release -p v1-stage0-runtime -p v1-stage0-std-core ... Both commands under one literal block scalar, the package list derived from the partition roster rather than restated, so a crate added to the partition joins the gate with no edit at the gate. REGEN NOTE FOR THE NEXT AUTHOR: main_wet is OOM-killed on a BuildBuddy runner under the default budget (SIGKILL, rc=137). /sys/fs/cgroup/memory.max and memory.high are unreadable there, so memory_governor read_host_budget_bytes falls through to MemAvailable -- which reports the HOST's memory, not the slot's -- and caps at the 15 GiB declared slot line, which can exceed what the slot actually has. Setting GUNBC_MEMORY_BUDGET_BYTES explicitly (10 GiB used here, peak RSS 7.7 GiB) makes it complete. The kill is silent through a `cmd | grep | tail` pipeline, because the pipeline reports the last stage's status. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01RTvKD3J3pnUzXmUascawfQ --------- Co-authored-by: gunbc-ci-auto-heal <gunbc-ci-auto-heal@users.noreply.github.com> Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The six
DeclaredCallableIdentity hoist to v1.std.core 2026-08-29rows insrc/v1/stage0/src/namespace_wave_admission.rs(hand-maintained roster, outside the regen population) were authored for #9665 with dissolve-on "this PR merging". #9665 merged as ecdeb49, so everypull_requestbuild now has a base that already carries the hoist: base and merge head agree, no run can produce those six deltas, andstale_admissionsreports all six ->namespace-wave-admission FAILED PHASEon every open PR (measured on #9689 @ bfd9524:0 unadjudicated delta(s), 6 stale admission(s); same on #9676 @ 680cc49). Main's own push build at ecdeb49 is green because its base is pre-#9665 -- the block is PR-only but universal, the fourth time this roster has reproduced the shape its own doc predicts. The rows are deleted by their trigger, the roster is&[]again, and empty is not permissive: a real namespace delta still refuses as UNADJUDICATED until its author adds a row. The fourth-shrink note sits beside the previous three.test:
cargo test -p v1-compiler --test namespace_wave_admission33/33 (remote, BuildBuddy 11f384e2). Discriminating check: this PR's own witnesses lane reports0 stale admission(s)where every other open PR reports 6.Part of #9664.
🤖 Generated with Claude Code
https://claude.ai/code/session_01GXfYKNQTD3VfYyQcnJpxNU