Conversation
"refactor the validate_subtree method using the fact that FuncDefns are always children of the module_root to turn validate_subtree into an outer loop over module children and an inner loop over the descendants of each FuncDefn using it's var_decls"
"simplify validate_subtree in that only FuncDefn's have descendants, asserting that any other node has no children"
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #2969 +/- ##
=======================================
Coverage 83.87% 83.88%
=======================================
Files 267 268 +1
Lines 52941 52949 +8
Branches 46857 46861 +4
=======================================
+ Hits 44404 44415 +11
+ Misses 6268 6266 -2
+ Partials 2269 2268 -1
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
acl-cqc
changed the title
| /// Error in a [`ExtensionOp`] serialized as an [Opaque]. | ||
| /// | ||
| /// [ExtensionOp]: crate::ops::ExtensionOp | ||
| /// [`ExtensionOp`]: crate::ops::ExtensionOp |
There was a problem hiding this comment.
is your agent enjoying fixing unrelated things?
There was a problem hiding this comment.
Haha! No, turns out the old [ExtensionOp]: ... was not actually doing anything at all and could have been removed, because the reference in the text of the comment is and was [`ExtensionOp`] - which was working instead because we use ExtensionOp at the top of the file. Agent thought it was being clever and removed the file-level import and broke the docs...which I fixed by hand
| }; | ||
| match op_type { | ||
| OpType::ExtensionOp(ext_op) => validate_ext(ext_op)?, | ||
| OpType::ExtensionOp(ext_op) => { |
There was a problem hiding this comment.
also an unrelated simplification?
There was a problem hiding this comment.
Yes but one I asked it to do specifically as a separate step when I saw that the validate_ext closure was only used once
| b_child.finish_with_outputs([b_child_in_wire])?; | ||
| let b_child_2 = b.dfg_builder(Signature::new([], [Type::UNIT]), [])?; |
There was a problem hiding this comment.
not sure whether the signature changes here are related? Was the test wrong before?
There was a problem hiding this comment.
Yeah, I should have mentioned, see inidividual commit comments. The refactor makes validate_subtree process nodes in a different order, so we start hitting UnconnectedPort errors first, rather than the desired nonlocal-edge-error. So I had to fix the ports in order to make the test return the right error...
## 🤖 New release * `hugr-model`: 0.26.1 -> 0.27.0 (✓ API compatible changes) * `hugr-core`: 0.26.1 -> 0.27.0 (✓ API compatible changes) * `hugr-llvm`: 0.26.1 -> 0.27.0 (✓ API compatible changes) * `hugr-persistent`: 0.5.1 -> 0.6.0 (✓ API compatible changes) * `hugr`: 0.26.1 -> 0.27.0 (✓ API compatible changes) * `hugr-cli`: 0.26.1 -> 0.27.0 (✓ API compatible changes) <details><summary><i><b>Changelog</b></i></summary><p> ## `hugr-model` <blockquote> ## [0.26.0](hugr-model-v0.25.7...hugr-model-v0.26.0) - 2026-03-16 ### Testing - Replace model text snapshots with roundtrip tests ([#2933](#2933)) - Add missing width arg in model-call example ([#2945](#2945)) </blockquote> ## `hugr-core` <blockquote> ## [0.27.0](hugr-core-v0.26.1...hugr-core-v0.27.0) - 2026-03-31 ### Bug Fixes - Use valid identifiers when constructing AST model ([#2973](#2973)) - Panic on constant folding with opaque consts ([#2986](#2986)) ### New Features - *(py)* Allow missing ext versions ExtensionDesc metadata ([#2979](#2979)) - [**breaking**] Deprecate JSON serialization format ([#2991](#2991)) - [**breaking**] Bump public `portgraph` dependency to `0.16.0` ([#2998](#2998)) ### Refactor - Simplify validate_subtree ([#2969](#2969)) - further cleanups to SiblingSubgraph + generalize over trait HugrConvexChecker ([#2957](#2957)) </blockquote> ## `hugr-llvm` <blockquote> ## [0.27.0](hugr-llvm-v0.26.1...hugr-llvm-v0.27.0) - 2026-03-31 ### New Features - *(hugr-llvm)* Emit rounding operation for `FloatOps::fround` extension ops ([#2949](#2949)) </blockquote> ## `hugr-persistent` <blockquote> ## [0.6.0](hugr-persistent-v0.5.1...hugr-persistent-v0.6.0) - 2026-03-31 ### New Features - [**breaking**] Deprecate JSON serialization format ([#2991](#2991)) </blockquote> ## `hugr` <blockquote> ## [0.27.0](hugr-v0.26.1...hugr-v0.27.0) - 2026-03-31 ### Bug Fixes - Use valid identifiers when constructing AST model ([#2973](#2973)) - Panic on constant folding with opaque consts ([#2986](#2986)) ### New Features - *(py)* Allow missing ext versions ExtensionDesc metadata ([#2979](#2979)) - [**breaking**] Bump public `portgraph` dependency to `0.16.0` ([#2998](#2998)) - [**breaking**] Deprecate JSON serialization format ([#2991](#2991)) ### Refactor - Simplify validate_subtree ([#2969](#2969)) - further cleanups to SiblingSubgraph + generalize over trait HugrConvexChecker ([#2957](#2957)) </blockquote> ## `hugr-cli` <blockquote> ## [0.27.0](hugr-cli-v0.26.1...hugr-cli-v0.27.0) - 2026-03-31 ### Bug Fixes - Do extension resolution on exts loaded by hugr-cli ([#2987](#2987)) ### New Features - *(py)* Allow missing ext versions ExtensionDesc metadata ([#2979](#2979)) - [**breaking**] Deprecate JSON serialization format ([#2991](#2991)) </blockquote> </p></details> --- This PR was generated with [release-plz](https://github.com/release-plz/release-plz/). --------- Co-authored-by: Agustín Borgna <[email protected]>
2 participants
...now that all FuncDefns are at top level.
closes #2346.
Most of this was done by agentic AI...it worked ok! (Maybe the naming could be better? And I wonder if it has put in too many "obvious" comments)
Also fix unconnected ports in the
non_cfg_ancestortest: the new validate_subtree checks nodes in a different order so was hitting these first.