Skip to content

fix(tools): the elaboration ratchet counted a phantom error per module - #2435

Merged
gHashTag merged 1 commit into
masterfrom
fix/elab-honest-count
Aug 22, 2026
Merged

gHashTag merged 1 commit into
masterfrom
fix/elab-honest-count

Conversation

@gHashTag

Copy link
Copy Markdown
Collaborator

Refs #2325.

My own gate counted iverilog's summary line (N error(s) during elaboration.) as an error — one phantom per failing module. Proven by the shape of the correction, not argued: every module drops by exactly 1, 25 do so, and 186 − 161 = 25 = the number of modules that fail to elaborate.

The docstring also carried a false claim — "the remainder is two named design decisions" — measured over the unbound-identifier errors only and written as though it covered all 161. The full classification is now in the file, and it names a fourth thing that is nobody's decision: 4 malformed statements from a keyword-named parameter escaped at its declaration and emitted bare at its use.

🤖 Generated with Claude Code

iverilog ends a failing file with "N error(s) during elaboration." -- a
total, not an error. The counter matched any line containing " error",
so every failing module contributed one error that does not exist.

Proof rather than inference: after excluding the summary line, every
module drops by exactly 1, 25 modules do so, and 186 - 161 = 25 = the
number of modules that fail to elaborate. Nothing else moved.

The published 573 -> 186 was inflated at both ends by this. Honest on
this instrument: 161 real errors across 25 of 32 modules.

The docstring also stated a falsehood: "the remainder is two named
design decisions". That was measured over unbound-identifier errors
only, where it is exact, and written as if it described all of them.
The full classification now lives in the docstring -- including 4
malformed statements, which are an emitter defect (a keyword-named
parameter escaped at its declaration and emitted bare at its use) and
not a decision anyone owns.

Refs #2325.
@gHashTag
gHashTag enabled auto-merge (squash) August 22, 2026 15:45
@github-actions

Copy link
Copy Markdown
Contributor

📓 NotebookLM Notebook linked to this PR

This notebook contains session context, decisions, and artifacts for this work.

@github-actions

Copy link
Copy Markdown
Contributor

PR Dashboard

Generated at: 2026-08-22 15:46:04 UTC

Summary

Status Count
Total Open PRs 5
PRs with Failing Checks 2
PRs with All Checks Green 3
READY 2
FAILING 2
PENDING 0

Seal Status

  • ⚠️ STALE -- sha256(compiler.rs)=42d2b54210da != manifest seal=87e5cbd3ad94.
    The committed NMSE numbers were certified against an older compiler.rs.
    Run scripts/reseal-check.sh locally for the two-step reseal command (advisory; not a merge gate).

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