Repository navigation
LanguageModel.canonical_symbols may be a decorative field: resolve accepts a tree whose root atom was removed from the prelude set (gunbc#8505 probe A class 4) - #8542
Merged
Conversation
build_program_namespace (and the shared name_resolve admission path via namespace_bindings_augment_from_module_atoms) deep-harvested every Atom node anywhere in the resolved tree via harvest_direct_atom_binding and self-bound it, before any scope/symbol_index/canonical_symbols check ran. Since resolving an atom requires it to appear as an Atom node in the tree being resolved, this pre-empted the fallback chain for every reference, making LanguageModel.canonical_symbols (and cross-module symbol_index lookups within a single tree) decorative for the resolve() stage used by v2.compiler.compile's compile-frontend. Narrow both call sites to harvest_conj_named_bindings (direct top-level declarations only); param/bind-target scoping already runs its own narrowly-scoped harvest_direct_atom_binding at traversal time and is untouched. Adds a discriminating regression: a root atom absent from both declarations and canonical_symbols must now be rejected. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
The first version routed through normalize(), which desugars the tested atom before resolve() sees it -- the same "mis-aimed mutation" trap flagged in gunbc#8505 probe A, this time in the test itself rather than the original probe. Rebuilt it as a hand-built Conj-rooted tree admitted directly via admit_normalized_tree, isolating the module-root self-bind defect: a legitimately declared root whose declaration target is an undeclared, non-canonical atom reference. Verified by execution against the fixed namespace_bindings_augment_from_module_atoms. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
# Conflicts: # src/v2/test/claim/claim_pipeline/resolve.dag
gunbai-bot Bot
pushed a commit
that referenced
this pull request
Aug 19, 2026
Six commits, 11 files. Two .dag conflicts and one Rust API change. BOTH .dag CONFLICTS RESOLVED IN MAIN'S FAVOUR, on the established rule: main is the authority for CONTENT, this branch for SPELLING. Checked rather than assumed which side was newer -- 03_resolve.dag looked backwards at first glance because the BRANCH side had MORE code. It was main that deliberately deleted it: #8542 collapses namespace_bindings_augment_from_module_atoms to harvest_conj_named_bindings and removes add_module_named_exports, so resolve accepts a tree whose root atom left the prelude set. Taking the branch's larger side would have reverted that. Re-applied the cut to both files afterwards: 339 names qualified (65 record pattern heads, 20 payload-free heads, 6 string occurrences left alone), then 33 re-exporter repoints. THE RUST CHANGE IS A REAL IMPROVEMENT AND IS ADOPTED, NOT PAPERED OVER. #8547 made parse_module_binding return Result<ModuleBindingOutcome, ModuleBindingRefusal> instead of an Option that panicked on a header-bearing parse failure. A branch call site still used the old `?.ok_or_else(...)` shape. It now matches both arms, and keeps them DISTINCT rather than collapsing them: a parse REFUSAL reports the parser's own located diagnostic; ModuleBindingUnclassified reports "missing module declaration". Those are different facts -- a file that failed to parse is not a file that parsed fine and had no module header -- and folding the first into the second would have thrown away exactly the typed refusal #8547 landed to create. Receipts: v1_src_dag_parse 46 file(s) parse-clean, exit 0; the specimen witness w_the_three_blocker_kinds_are_distinct still returns `true`. 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
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.
Auto-opened by session-dashboard for session
royal-deer-610.Pushing to
session/royal-deer-610advances this PR.Worker attestation
Before flipping this PR to ready for review, confirm each item:
npm test,cargo test) and the result.Closes #Ndirective.Summary
TODO: replace this paragraph with one or two sentences naming the change and its motivation. Reviewers read this first.
Test plan