ci(freebsd-amd64): accept 'unresolved reference to' wording in etext/end XFAIL matcher

[quaesitor-scientiam]

Summary

Run 30421217468 is a matcher false negative, not a real regression: this branch's tcc.exe now reports tcc: error: unresolved reference to 'etext'/'end' instead of the undefined symbol '...' wording is_known_etext_end() originally matched. Confirmed directly from that run's log:

hello.c: exit=1: tcc: error: unresolved reference to 'etext'
tcc: error: unresolved reference to 'end'
gc_alloc.c: exit=1: tcc: error: unresolved reference to 'etext'
tcc: error: unresolved reference to 'end'

Same known linker bug (tcc's FreeBSD backend never defines the plain BSD-style etext/end that BDWGC references directly - see the step's own root-cause comment above it), just a different tcc diagnostic wording. Fixed by accepting both wordings rather than replacing one with the other, so a differently-built tcc.exe landing here later isn't a second false negative.

Workflow-only change - no binary rebuild, no publication. Every existing guard is untouched:

• exit-code rejection (reject exit 0 and signal-terminated exits >128)
• the exclusivity check (any tcc error line that ISN'T specifically etext/end, under either wording, still fails the job for real)
• the outer aggregate-shape match (PASS crash, FAIL gc_alloc/hello (compile error))

The real upstream tcc fix (da58264) isn't shipped in any binary on this branch yet and will be validated separately on native FreeBSD - this lane is still expected to XFAIL the same way after this PR, just correctly recognized as the known case instead of hard-failing as unexpected.

Test plan

• YAML validated (yaml.safe_load) and the modified step's shell syntax-checked (bash -n).
• is_known_etext_end() extracted and run standalone against the exact captured error text from run 30421217468 - now correctly recognized as the known XFAIL.
• Negative controls verified in the same standalone test: the old undefined symbol wording still matches (backward compatible), an unrelated error text is still correctly rejected, and exit code 0 is still correctly rejected.
• Confirm this branch's CI goes green (XFAIL, not a hard failure) on the next run against this commit.

[quaesitor-scientiam]

Is ready for merge @GGRei

[GGRei]

@codex review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. You're on a roll.

Reviewed commit: 2ea352fc9c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".