fix(cli): surface YAML and AsyncAPI errors from docs serve

[SarthakB11]

Description

Refs #2709 (diagnostic portion).

Per @Sehat1137's recommendation on #2858 ("a smaller PR limited to that behavior, with deterministic parsing and regression tests"), this is the diagnostic-only redo: no new CLI options, no parser plugin, default yaml.safe_load preserved. The pluggable-parser proposal from #2709 is intentionally deferred per maintainer feedback.

faststream docs serve <asyncapi.yaml> previously collapsed three different failure modes into one opaque SCHEMA_NOT_SUPPORTED message:

• YAML parse errors were not surfaced. yaml.safe_load ran outside any try, so the user either got an uncaught traceback or, more often, a downstream pydantic failure on the garbage that partial parsing produced.
• AsyncAPI validation errors were dropped by contextlib.suppress(ValidationError), so there was no indication of which schema version was tried or why it failed.
• YAML scalar gotchas like protocolVersion: 3.2 (parsed as a float, rejected by AsyncAPI's string type) gave no hint that the value just needed quoting. This is the exact repro from Bug: PyYAML cannot correct parse asyncapi.yaml #2709.

Fix

Edits only _parse_and_serve in faststream/_internal/cli/docs.py:

1. Wrap yaml.safe_load in try / except yaml.YAMLError, echo the parser exception, exit 1.
2. Replace suppress(ValidationError) with an explicit collector that records (version_label, ValidationError). When neither v3.0 nor v2.6 validates, print SCHEMA_NOT_SUPPORTED followed by both per-version error blocks on stderr.
3. For each block, if any sub-error has type == "string_type" and an int | float | bool input, append a one-line hint pointing at the location and suggesting YAML quoting (e.g. '3.2').

Default parser is unchanged. _parse_and_serve's public signature is untouched. dto.py is untouched. No new modules, no parser registry, no --yaml-parser flag.

Tests

Two regression tests in tests/cli/test_asyncapi_docs.py:

• test_serve_asyncapi_yaml_unquoted_scalar_reports_hint mutates the existing yaml_asyncapi_doc fixture to protocolVersion: 3.2, asserts exit code 1 and the v3.0 / v2.6 error enumeration plus the quoting hint.
• test_serve_asyncapi_reports_yaml_parse_error feeds malformed YAML (foo: [unclosed) and asserts the parser exception is surfaced rather than swallowed.

Both follow the existing faststream_cli + generate_template + wait_for_stderr pattern in the same file.

Type of change

• Bug fix (a non-breaking change that resolves an issue)

Checklist

• My code adheres to the style guidelines of this project (just lint shows no errors)
• I have conducted a self-review of my own code
• I have made the necessary changes to the documentation
• My changes do not generate any new warnings
• I have added tests to validate the effectiveness of my fix or the functionality of my new feature
• Both new and existing unit tests pass successfully on my local environment by running just test-coverage
• I have ensured that static analysis tests are passing by running just static-analysis
• I have included code examples to illustrate the modifications

[Sehat1137]

Sorry for the long wait on this PR, and thank you for your patience.

The overall direction looks good, and the improved validation diagnostics are useful. Before we merge it, could you please make the following adjustments:

1. Show the YAML-specific quoting hint only for .yaml and .yml inputs. Currently, a JSON document containing a numeric value may also receive a misleading YAML hint.
2. Add a regression test confirming that JSON inputs do not produce the YAML-specific hint.
3. Emit the quoting hint only once, rather than potentially once per AsyncAPI version.
4. Please add a test covering the exact reproduction from Bug: PyYAML cannot correct parse asyncapi.yaml #2709, including protocolVersion: ! "3.2".

The YAML parsing error handling and surfaced Pydantic validation errors are otherwise a useful improvement. Thanks for working on this.

[SarthakB11]

All four are addressed: the hint now fires only for .yaml/.yml and prints once, a JSON test confirms it stays absent, and a new test covers the protocolVersion: ! "3.2" case from #2709.

[Sehat1137]

Can you resolve problem with tests?

[SarthakB11]

Looked into the 3 failures: test_worker_id_extra_option.py, test_run_asgi.py::test_many_workers, test_logs.py::test_run_as_asgi_mp_with_log_level. All three are in the multi-worker CLI subprocess tests, a different subsystem from this PR's diff (faststream/_internal/cli/docs.py and its own test file only). All 6 tests in that area pass locally for me on this branch. Rebased onto latest main and pushed to get a clean CI run, that should confirm it either way.

[Sehat1137]

@Sarthakb1 problem with the tests is relevant, try to rebase ur branch