fix(specs): drop 80 placeholder tests from the coder pipeline spec (Refs #5472) - #5613
Merged
Merged
Conversation
…efs #5472) specs/igla/coder/pipeline.t27 carried 80 tests whose body is only `assert true` / `{ /* verify baseline */ }`; gf16_matmul_top.t27 two comment-only tests. They run nothing. Removed, pipeline resealed with the t27c built from this tree, and the ratchets follow: corpus ledger drops gf16_matmul_top (134/134, CLEAN), assertionless baseline 3905 -> 3761, published test blocks 14330 -> 14314 (master measured 14394). Split out of #5473 at the reviewer's request. Refs #5472 Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
dmitrii-f-t27
added a commit
that referenced
this pull request
Oct 2, 2026
…#5472) At the reviewer's request (changes requested on 9eab05d) the PR is split. Restored to master: the 80 placeholder tests of igla/coder/pipeline and gf16_matmul_top with the ratchets that follow them (now #5613), the two seals of specs that had none (verilog_ternary_mac_top, vivado_gf16_matmul_top), the akashic-log claim lines, and the corpus re-take and test-block pin, which move with every port PR and are redone right before merge. What remains: the two contributor-key specs, their controls, their two seals and the docs/now entry. Both specs pass 17 tests and `seal --verify`. Refs #5472 Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Contributor
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
dmitrii-f-t27
enabled auto-merge (squash)
October 2, 2026 16:58
This was referenced Oct 2, 2026
dmitrii-f-t27
added a commit
that referenced
this pull request
Oct 2, 2026
* Specify contributor key management boundaries Closes #5472 phi^2 + 1/phi^2 = 3 | TRINITY * Preserve negative controls for contributor key contracts Refs #5472 phi^2 + 1/phi^2 = 3 | TRINITY * Record contributor contract evidence and refresh corpus count Refs #5472 phi^2 + 1/phi^2 = 3 | TRINITY * fix(ci): distinguish commands and function bodies from prose and declarations Remove 149 vacuous test declarations without changing runtime source. Lower the assertionless ceiling and regenerate only the three changed seals. Add failing-before controls for both scanners; keep genuine CI debt visible. Refs #5472, #3141 phi^2 + 1/phi^2 = 3 | TRINITY * refactor(hslm): share empty forward output (#5472) Refs #5472 * chore(gates): re-pin corpus ledger and test-block count for this PR (#5472) gf16_matmul_top.t27 no longer discards anything after the merge, so its parse-no-discard entry in the corpus ledger is dropped and the cap goes from 133 to 132. The "test blocks" figure moves 14330 -> 14267: the two contributor-key specs add 17 tests and igla/coder/pipeline loses 80 placeholder tests whose only body was `assert true`. Refs #5472 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(igla): re-take the contributor-key corpus figures on the merged tree (#5472) After merging master d04bf14 the corpus is 1148 specs, not the 1136 this PR first recorded at ce4e092, so the re-take guard in the untrusted-input gate refused the block. `t27c impl-status --specs-dir specs` at b833780: 738 fully implemented, 350 with no functions, 16 partly written, 17 unwritten, 27 that do not parse; 6734 functions, 187 with no body. Refs #5472 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(seals): reseal gf16_matmul_top after the merge (#5472) The merge with master d04bf14 kept both sides' deletions in specs/port/fpga/vivado/gf16_matmul_top.t27, so the spec text changed and Seal Coverage reported the seal stale. Resealed with t27c built from this tree: only spec_hash moves, all four gen_hashes are unchanged. Refs #5472 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(gates): re-take the corpus on the merged tree, re-pin test blocks (Refs #5472) After merging master 6111696 the corpus is 1148 specs (1146 on master plus the two contributor-key specs). The re-take block is anchored at the merge commit 70ed745, with the master block (4f65684, 1146) kept above it. The test-blocks figure moves to 14271: master itself measured 14334 against a pin of 14330, and this PR adds 17 and drops 80 placeholders. Refs #5472 Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> * chore(scope): keep only the contributor key contracts in this PR (Refs #5472) At the reviewer's request (changes requested on 9eab05d) the PR is split. Restored to master: the 80 placeholder tests of igla/coder/pipeline and gf16_matmul_top with the ratchets that follow them (now #5613), the two seals of specs that had none (verilog_ternary_mac_top, vivado_gf16_matmul_top), the akashic-log claim lines, and the corpus re-take and test-block pin, which move with every port PR and are redone right before merge. What remains: the two contributor-key specs, their controls, their two seals and the docs/now entry. Both specs pass 17 tests and `seal --verify`. Refs #5472 Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <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
1 participant
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.
Refs #5472. Split out of #5473 at the reviewer's request: this is (b)+(c) of the review, as one fresh PR from master.
What
specs/igla/coder/pipeline.t27: 80 tests whose whole body isassert true/{ /* verify baseline */ }are removed. 196 real test blocks remain.specs/port/fpga/vivado/gf16_matmul_top.t27: two comment-only tests removed.coder_igla-coder-pipeline.jsonresealed with the t27c built from this tree (seal --verify: all hashes MATCH).gf16_matmul_top.t27has no seal on master and none is added.Ratchets, in the same commit
gf16_matmul_topno longer discards;t27c suite --ratchet --corpus-only→RATCHET: CLEAN, 134 / 134check_assertionless_spec_tests.py --blesspublished_figures.py --check(master measured 14394 at 4c597ec, 64 above its pin; −80 here)The test-block count moves with every port PR that lands, as the reviewer says, so it must be re-measured right before merge. Measured on
4c597ec7.Measured locally
dupe scan ok;
check_seal_currency.pyreports the same 30 stale seals as master (#5580, none of them from this PR); published figures and assertionless checks pass.Not in this PR
The Untrusted Input Gate is red on master itself (1151 specs, the re-take says 1146), because other PRs added specs. This PR adds none.
🤖 Generated with Claude Code