added workflow corrections - #46
Merged
Merged
Conversation
Seth Juarez (sethjuarez)
added a commit
that referenced
this pull request
Aug 5, 2026
Picks up the Rust immediate-scalar coercion fix (typra #73). Verified against the published package, not a local tarball. Measured on this branch: cargo test --no-fail-fast 3 failures -> 2 failures tests/property_scalar_coercion_vectors 0 passed/1 failed -> 1 passed/0 failed The regeneration delta is confined to a single file, runtime/rust/prompty/src/model/core/property.rs (5 lines): the integral coercion now precedes the fractional one, and the gratuitous f32 narrowing is gone. Generated output for Go, TypeScript, Python, C#, Java and Swift is byte-identical to 0.4.21, so those runtimes cannot have regressed. Remaining Rust failures are both typra #46 (name-keyed scalar shorthand does not infer kind/default); revert-checked to confirm 0.4.22 changes nothing about them. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
Seth Juarez (sethjuarez)
added a commit
that referenced
this pull request
Aug 5, 2026
The named-collection contract was executed in Rust only. Six backends loaded and saved these shapes with nothing asserting the behaviour, so any divergence was silent. This ports the Rust reference suite to Go. Measured result: 9 of 13 vectors pass; 4 fail. All four failures are the immediate-scalar inference cases and are a pre-existing emitter defect (typra #46) that this test newly *reports* rather than introduces. No other Go suite changes state -- `go test -count=1 ./...` is otherwise green and `go vet ./model/` is clean. Go's manifestation differs from Rust's, which is the point of porting it: vector requires : kind "string", default "Seattle", example absent Go produces : kind "", example "Seattle", default absent Rust produces : kind "Seattle" Go already routes the scalar to a non-discriminator field (it honours coercionProperty, which resolves to `example`) and *still* fails, because nothing infers `kind` from the value's JSON type. That makes Go a natural experiment showing the fix needs two independent parts -- shorthand target selection AND JSON-type-to-kind inference -- not target selection alone. Verified: 9 passing roundtrip/rejection vectors plus 4 targeted failures demonstrate the harness discriminates rather than failing uniformly. Not verified: the same port for TypeScript, Python, C#, Java and Swift. Java and Swift have no runtime in this repo. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
Seth Juarez (sethjuarez)
added a commit
that referenced
this pull request
Aug 5, 2026
…d models Second backend port of the named-collection contract, after Go (ba4866e). The contract previously executed in Rust only. Measured (npm test, packages/core): before 1518 passed / 0 failed after 1528 passed / 4 failed (+10 passing, +4 failing, 14 new) The 4 failures are the immediate-scalar inference cases -- a pre-existing emitter defect (typra #46) that this test newly reports rather than introduces. No other core test changes state. TypeScript manifests like Rust and unlike Go, which is the value of having all three executable: vector requires : kind "float", default 1.5, example absent TypeScript : kind "1.5" <- scalar lands in the discriminator Rust : kind "1.5" <- same Go : kind "", example 1.5 This confirms by execution, not static audit, the reported split where Rust and TypeScript ignore `coercionProperty` while Go and C# honour it. Because Go already routes the scalar to a non-discriminator field and still fails, shorthand target selection alone cannot close this -- a JSON-type-to-kind inference step is also required. Two unrelated pre-existing failures exist elsewhere in the workspace and are NOT caused by this change; verified by removing this file and re-running those packages, which fail identically: packages/openai listModels leaves unknown models without enrichment packages/foundry listAzureModels does not set modalities Measurement note: run the TypeScript suite with `npm test`, not a bare `npx vitest run` from the workspace root. The per-package configs supply `globals`, so a root invocation reports 154 spurious file-level "describe is not defined" collection errors. Not verified: Python, C#, Java and Swift ports of this vector. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
Seth Juarez (sethjuarez)
added a commit
that referenced
this pull request
Aug 5, 2026
…dels Third backend port of the named-collection contract, after Go (ba4866e) and TypeScript (36c9619). The contract previously executed in Rust only. Measured (dotnet test, Prompty.Core.Tests): before 1051 passed / 48 failed (1099 total) after 1060 passed / 52 failed (1112 total) +13 tests, +9 passing, +4 failing. The 4 failures are the immediate-scalar inference cases -- pre-existing emitter defect typra #46, newly reported rather than introduced. The other 48 failures are unchanged and remain attributable to typra #53 (fixture generator omits required model-typed fields). With four backends now executable, the reported coercionProperty split is confirmed by execution rather than static audit. All four fail the SAME four vectors, in exactly two manifestations: vector requires kind "string", default "Seattle", example absent C# (honours it) example "Seattle" Go (honours it) kind "", example "Seattle" Rust (ignores it) kind "Seattle" TS (ignores it) kind "Seattle" Neither group passes. Backends that already route the scalar to a non-discriminator declared field still fail, because nothing infers `kind` from the value's JSON type. This is a cross-backend acceptance gate for #46: shorthand target selection alone cannot close it. Harness note for future ports: build loader input with a recursive JsonElement-to-native conversion (JsonElementToDict/JsonElementToObject, the convention already used by SpecVectorTests). Passing a JsonSerializer.Deserialize<Dictionary<string, object?>> result leaves raw JsonElement values, which the emitted loaders do not traverse -- inputs then load as empty and all 13 vectors fail misleadingly. Not verified: Python, Java and Swift ports. Java and Swift have no runtime in this repository. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
Seth Juarez (sethjuarez)
added a commit
that referenced
this pull request
Aug 5, 2026
…dels Fifth and final executable backend for the named-collection contract, after Go (ba4866e), TypeScript (36c9619) and C# (1a0c3e2). Rust was this contract's only executable home for most of this effort. Measured: .venv/Scripts/python -m pytest tests/test_named_collection_vectors.py 13 passed / 0 failed ruff check All checks passed! ruff format 1 file already formatted Python confirms the two-part #46 fix (emitter 0.4.23 plus the @entryShorthand("default") declaration in e2152ef) holds in a fifth independent backend. All five now pass the contract: rust 4 targets 0 failed go 13 vectors 0 failed ts 14 tests 0 failed csharp 13 vectors 0 failed python 13 vectors 0 failed Also corrects the record on Python's status. This effort has been carrying "Python does not build (typra #43)" as a standing assumption; it is stale. Measured full suite on this commit: .venv/Scripts/python -m pytest tests/ -q 1468 passed / 1 failed / 17 skipped / 56 deselected (11m48s) The single failure is test_spec_vectors.py::test_wire_vector [kind_to_json_type_mapping], and it is not a Python defect: ValueError: tools[0].parameters[4].items: missing required field spec/vectors/wire/wire_vectors.json declares that parameter as {"name": "an_array", "kind": "array"} with no items, so the vector requires ArrayProperty.items to be optional. It cannot be made optional -- schema/model/core/properties.tsp:80 must stay `items:` because `items?:` emits uncompilable Rust (typra #71). Verified by experiment: applying items?: and regenerating produces E0308 at property.rs:113 and :247, and the change was reverted. C# fails the identical vector: SpecVectorWireTests.OpenAI_Chat_WireFormat (name: "kind_to_json_type_mapping") with "tools.parameters.items: missing required field". So #71 blocks a canonical spec vector wherever wire_vectors.json is executed, rather than costing one isolated test. Not verified: Java and Swift have no runtime in this repository. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.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.
No description provided.