cl: llgoSupportAliasAsRecv = true; cl compileFuncOrMethod: support method with callback

[xushiwei]

• link directive bug: can't use alias type as recv, callback as method param xgo-dev/llgo#2677

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.04%. Comparing base (1947404) to head (37ad384).
⚠️ Report is 23 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #880      +/-   ##
==========================================
- Coverage   88.36%   88.04%   -0.33%     
==========================================
  Files          22       22              
  Lines        1857     1857              
==========================================
- Hits         1641     1635       -6     
- Misses        216      222       +6     

Flag	Coverage Δ
llgo-tests	88.04% <100.00%> (-0.33%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[fennoai]

Review: enable methods for callback funcs & alias receivers

The functional change in cl/func.go is small and coherent, and the regenerated tool/_testc/clang-c-22.1.8/* fixtures consistently reflect the new output (value receivers like VirtualFileOverlay replacing pointer-receiver impl types, with matching llgo:link directives).

What the change does

• Drops the feats&featHasCallback == 0 guard (and its TODO) so C global functions with callback parameters can now be converted to methods.
• Flips llgoSupportAliasAsRecv from false to true, enabling alias types as method receivers.

Verified as safe/non-issues

• featHasCallback is still detected and used in cl/type.go and cl/typedef.go; removing its use as a gate here orphans no symbol.
• The real safety gate — the featIgnored early-return — still runs first, so no unsupported type slips through.
• No performance regression: feats is fully computed before the branch regardless, and the newly-taken alias branch is strictly cheaper than the old one.
• No security/trust-boundary impact: this is a compile-time generator over developer-supplied headers; type identity and linkname targets are unchanged.

Main finding — see the inline comment: flipping llgoSupportAliasAsRecv to a const true leaves the else branch in tryToMethod (lines ~228-242) statically unreachable dead code. That dead branch is also the only in-code writer of ctx.typeAbbr (lines 238-239), so the previous auto-registration of alias-type abbreviations no longer happens at runtime (ctx.typeAbbr is still read at cl/ctx.go:337 and populated from config at cl/compile.go:217). Please confirm dropping that auto-population is intentional.