[dinwwwh]

Requested by dinwwwh · project thread

Summary

This adds two OpenAPI Generator 7.26 server generators to two existing README lines. Each one compiles from the 3.1 input but not from the downgraded 3.0.

• Rust line. The spring server also fails to compile a required property that is a nullable format: byte string, such as Zod's z.base64().nullable(). It keeps the value as JsonNullable<byte[]>, and its equals/hashCode call Arrays.equals(JsonNullable<byte[]>, …), which matches no overload. An optional one compiles. The Rust-only clause about optional properties now says "In the Rust client".
• Java line. The spring server writes the same empty enum the Java clients write for { type: "null" } alone, in a property, an array item or a component. The python-fastapi server writes class Nul(, Enum): for such a component, which is a SyntaxError. A z.null() property or array item there is fine.

How I found it

I compile-checked OpenAPI Generator's server generators on the oRPC Zod, Valibot and ArkType shape catalogs. Each was split into one doc per shape and generated from 3.1 and from the downgraded 3.0. I also ran python-fastapi end to end through FastAPI's TestClient on 112 fuzz and shape docs.

Shape	Generator	3.1	3.0
required z.base64().nullable()	spring	compiles	no suitable method found for equals(JsonNullable<byte[]>, …)
optional z.base64().nullable(), array of it, plain z.base64()	spring	compiles	compiles
z.null() property, array item, component	spring	compiles	enum constant expected here
z.null() component (property or array item)	python-fastapi	runs	SyntaxError in class Nul(, Enum):
z.null() inline property or array item	python-fastapi	runs	runs

Other shapes were clean. Once the null-only shapes are removed and safe-integer bounds get format: int64, spring compiles every other catalog shape from 3.0. Without that format, Zod's ±2^53 bounds overflow int in both versions. python-fastapi's only other 3.0-only rejects are tuples, which 3.1 leaves untyped.

OpenAPI Generator's python client writes the same class Nul(, Enum):. I left it out because it fails to generate from the 3.1 doc at all, so the 3.1 input is no way around it there.

Test plan

• pnpm exec uncheck packages/downgrader/README.md passes.

---

Generated by Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[codspeed]

Merging this PR will not alter performance

1 warning

⚠️ 1 benchmark is dominated by system calls

System calls cannot be simulated consistently and are left out of the measure. The reported value understates the real cost. Switch to the Walltime instrument to measure them.

Learn more about system calls

See affected benchmarks

✅ 29 untouched benchmarks
⏩ 2 skipped benchmarks¹

---

_{Comparing claude/project-thread-fvev8j (9c109e3) with main (e94e73c)}

Footnotes

1. 2 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[pullfrog]

✅ No new issues found.

Reviewed changes

• packages/downgrader/README.md — the two OpenAPI Generator limitation bullets now attribute the nullable format: byte required-property compile failure to the spring server (with the optional-property caveat scoped to the Rust client), and attribute the null-only enum failure to spring and python-fastapi (the latter scoped to a component).

Both additions were checked against OpenAPI Generator v7.26.0 source and both hold: the required nullable format: byte Spring model leaves the field as JsonNullable<byte[]> while pojo.mustache still emits Arrays.equals/Arrays.hashCode (the optional-nullable flag is unset), matching open issue #17935; and the null-only enum renders an empty enum body for spring/Java and class <Name>(, Enum): for a python-fastapi component. uncheck packages/downgrader/README.md passes.

  ^{｜ View workflow run ｜ Using deepseek-v4.1-flash (free via Pullfrog for OSS) ｜ 𝕏}