Repository navigation
feat(queen): a compiler defect names its stage, and the compiler is one file - #4767
Merged
Merged
Conversation
…ne file `bootstrap/src/compiler.rs` is 1.79 MB and holds the lexer, the parser, the typechecker and every generator. An issue reporting that `t27c parse` accepts a body of zero statements, or that `gen-c` emits `0 bad;`, is about that file - and says so by naming the STAGE rather than the path, which is how a person writes it and why every path rule here read it as naming nothing. "gen-c: `null` has no C spelling", "The lexer turns 0o777 into 0", "The emitter dropped every `x as T` cast: 645 in 67 specs" - 44 of the 286 issues still without a boundary were this. They will conflict with one another, because they really do all touch one file. The Queen runs them one at a time; that is correct, and it is not what "unreachable" meant. A run log (`Wave Loop N`) and a proof note (`formal: ... Prop. N`) are not compiler defects, and --self-test pins that they are not. Closes #4766 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
Contributor
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Closes #4766
bootstrap/src/compiler.rsis 1.79 MB and holds the lexer, the parser, the typechecker and every generator. An issue reporting thatt27c parseaccepts a body of zero statements, or thatgen-cemits0 bad;, is about that file — and says so by naming the stage, which is how a person writes it and why every path rule here read it as naming nothing:nullhas no C spelling — 68 errors"x as Tcast: 645 in 67 specs"44 of the 286 were this; the run wrote 59 once the other rules applied alongside.
They will conflict with one another, because they really do all touch one file. The Queen runs them one at a time — correct, and not what "unreachable" meant.
The board
What the remaining 237 are — counted, not guessed
Wave Loop N — …run logsformal: … (Prop. N)proof notesA boundary answers which files does this work own. A report of what happened owns none, and #3225 — "fourteen specs parse, typecheck and emit a field with no type" — names fourteen specs while the fix is in neither them nor any file it mentions. Inventing a boundary there is the one move that makes things worse: it reserves files the work does not own and blocks whoever really owns them.
🤖 Generated with Claude Code