fix(cli): preserve transitive audience-filtered type references

[kafkas]

Description

Linear ticket: Closes FER-10708

Fixes an audience-filtering issue where nested object types could render as any in docs when an OpenAPI schema reached nested inline object types through $ref chains.

The audience-filtered IR graph was only including one level of referenced type descendants. This meant direct $ref targets survived filtering, but their own nested inline child types could be dropped from the filtered IR.

Changes Made

• Updated IrGraph.addReferencedTypes to walk the full transitive type-dependency graph instead of adding only direct descendants.
• Added a V3 OpenAPI regression fixture covering nested object types reached through $ref chains and arrays.
• Added regression coverage that verifies audience-filtered FDR output preserves all non-internal transitive type definitions.
• Added a CLI changelog entry.

Testing

• Unit tests added/updated
• Manual testing completed

[claude]

Claude Code Review

This repository is configured for manual code reviews. Comment @claude review to trigger a review and subscribe this PR to future pushes, or @claude review once for a one-time review.

_{Tip: disable this comment in your organization's Code Review settings.}

[github-actions]

Docs Generation Benchmark Results

Comparing PR branch against median of 5 nightly run(s) on main (latest: 2026-05-22T05:19:54Z).

Fixture	main	PR	Delta
docs	211.4s (n=5)	201.4s (35 versions)	-10.0s (-4.7%)

Docs generation runs fern generate --docs --preview end-to-end against the benchmark fixture with 35 API versions (each version: markdown processing + OpenAPI-to-IR + FDR upload).
Delta is computed against the nightly baseline on main.
Baseline from nightly run(s) on main (latest: 2026-05-22T05:19:54Z). Trigger benchmark-baseline to refresh.
Last updated: 2026-05-22 23:32 UTC

[github-actions]

SDK Generation Benchmark Results

Comparing PR branch against median of 5 nightly run(s) on main (latest: 2026-05-22T05:19:54Z).

Full benchmark table (click to expand)

Generator	Spec	main (generator)	main (E2E)	PR (generator)	Delta
csharp-sdk	square	73s (n=5)	109s (n=5)	58s	-15s (-20.5%)
go-sdk	square	136s (n=5)	294s (n=5)	132s	-4s (-2.9%)
java-sdk	square	209s (n=5)	286s (n=5)	182s	-27s (-12.9%)
php-sdk	square	62s (n=5)	80s (n=5)	51s	-11s (-17.7%)
python-sdk	square	128s (n=5)	229s (n=5)	125s	-3s (-2.3%)
ruby-sdk-v2	square	92s (n=5)	118s (n=5)	83s	-9s (-9.8%)
rust-sdk	square	170s (n=5)	164s (n=5)	164s	-6s (-3.5%)
swift-sdk	square	63s (n=5)	746s (n=5)	49s	-14s (-22.2%)
ts-sdk	square	234s (n=5)	240s (n=5)	219s	-15s (-6.4%)

main (generator): generator-only time via --skip-scripts (includes Docker image build, container startup, IR parsing, and code generation — this is the same Docker-based flow customers use via fern generate). main (E2E): full customer-observable time including build/test scripts (nightly baseline, informational). Delta is computed against generator-only baseline.
⚠️ = generation exited with a non-zero exit code (timing may not reflect a successful run).
Baseline from nightly runs on main (latest: 2026-05-22T05:19:54Z). Trigger benchmark-baseline to refresh.
Last updated: 2026-05-22 23:35 UTC