Repository navigation
docs(t27o): golden functions -- the hand lines in converted fns of partly converting files, and the 146 fns that move without changing a caller (Closes #11159) - #11160
Merged
Conversation
…rtly converting files, and the 146 fns that move without changing a caller (Closes #11159) Analysis only, step 4 of #9000 for the files that do not convert whole: no Rust is deleted and no spec is added. 223 partly converting files (t27 at 0e00974, trios main 78671ba794, trinity, trinity-fpga, 999-multibots-telegraf) hold 817 converted fns with 6,555 hand lines outside #[cfg(test)]. 146 fns (1,375 lines, 0.32% of the swept lines) are MOVABLE: a free fn gen-rust writes with the same name and types, whose callers need no change. 54 fns (406 lines) are MOVABLE-WITH-CALLER-CHANGE, 617 (4,774 lines) are NOT-NOW: 477 are methods or name a hand-file type, 68 have gen-rust output that does not compile, 38 are trait impl methods. No call-connected group reaches 200 lines (the biggest is 93). The first bundle (the MOVABLE fns of one file, one mount) that deletes at least 200 lines is rings/ring-092-rust/src/lib.rs: 7 fns, 201 lines, +5 mount lines; its 28 crate tests pass with the mount, and 8,560,045 differential comparisons of the hand fns against the generated ones had 0 disagreements. In the crates the t27 workspace builds the largest bundle is 45 lines (cli/tri/src/fpga.rs). 44 of the 187 ranked rows are marked valid-output: unchecked (t27c test-report was stopped when the shared Mac ran low on disk); the others carry their test-report result and a rustc compile of the gen-rust output. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
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
2 participants
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.
What this is
Step 4 of #9000 for the files that do not convert whole (analysis only: no Rust is deleted, no spec is added).
docs/now/2026-10-20-golden-functions.mdmeasures the hand-written Rust lines that sit in functions t27o already converts, inside files that only partly convert, and which of those functions could be replaced by a generated function mounted into the same file with#[path]so that the lines are deleted and no caller changes. The whole-file version is #11149 (docs/now/2026-10-20-golden-candidates.md).Closes #11159
The same table (answer, scale, reasons, first bundle, top bundles, top 120 of the 187 ranked rows) is posted on #11159: #11159 (comment)
Answer
#[cfg(test)](44,864#[cfg(test)]lines sit in the same files). That is 1.51% of the 435,491 lines lane A4 swept, against 0.69% for the whole files.cli/tri/src/fpga.rswith 45 lines. trios workspace crates: 421 lines in 49 fns.exp_f64+apply_softmax, 93). The first bundle (all MOVABLE fns of one file, one mount) that deletes at least 200 lines isrings/ring-092-rust/src/lib.rs: 7 fns, 201 hand lines, +5 mount lines (net -196). Mounted in a scratch copy of the crate, its 28 tests pass as before; 8,560,045 differential comparisons of the hand fns against the generated ones show 0 disagreements. Caveat:specs/nn/attention.t27already declares four of the fns with another argument order, so a port has to reconcile with it first.Files
docs/now/2026-10-20-golden-functions.mdHand-written lines: +602 of prose in a
docs/nowfile (not Rust); no.rsor.t27file is touched.Valid output
t27c test-reportof the converted spec of each file (one spec per file, a freshZIG_LOCAL_CACHE_DIRdeleted after each run): 178 of the 223 specs ran, 100 hold tests (377 tests, 377 pass, 0 FAIL, 57 vacuous passes) and 78 hold none.valid-output: unchecked.rustc --crate-type lib(refused stubs removed, repeated after dropping each failing item; the ring crates also under#![no_std]).python3 -I tools/check_devhome_paths.py: OK, 29 files, 37 occurrences, all pinned; the document has no/Userspath and is ASCII.Not run
tri t27o --check(oracle replay) on any file; the only behavioural comparison is the ring-092 differential run above.cargo build,cargo checkorcargo test; callers come from a token index of every.rsfile of the five repositories plusgit grepby hand for every MOVABLE row with a hit elsewhere. BrowserOS, ghashtag.github.io and trios-railway were not read.dupe_scanover any spec but the ring-092 group spec (scratch).🤖 Generated with Claude Code