hotfix(codegen): resolve conflict markers from #1030 merge (unblocks every open PR)

[proggeramlug]

What this fixes

The merge of #1030 over #1008 (commit 9a9a233c) left 4 unresolved git conflict markers in main:

• crates/perry-codegen/src/lower_call.rs (1 marker block at line 1875)
• crates/perry-codegen/src/type_analysis.rs (3 marker blocks at lines 1034, 1063, plus around the static-method check)

$ git show origin/main:crates/perry-codegen/src/lower_call.rs | grep -nE '<<<|>>>'
1875:<<<<<<< HEAD
1891:>>>>>>> 92d5eab9 (fix: recognize global Promise static calls)

main does not compile cleanly on a fresh checkout. Every PR that rebases onto current main inherits the markers and hits cargo-test + lint failures.

Why a tiny standalone PR

#1028, #1031, #1032 each include the same fix as a drive-by, but they're all blocked on their own CI signal. A 37-line removal is the fastest possible review-and-merge path; once this lands, the feature PRs rebase cleanly.

Resolution choice

Each conflict pairs the legacy is_global_builtin_named matcher with the new is_global_constructor_expr helper introduced in #1008. I kept the post-#973 is_global_constructor_expr side — same resolution shipping in the three feature PRs above; matches the in-flight maintainer fix per the source comments.

Verification

• cargo check -p perry-codegen --release — green
• grep -nE '<<<<<<<|>>>>>>>' crates/perry-codegen/src/*.rs — no remaining markers

No behavioral changes vs. the maintainer's intended post-#1008 state.

[proggeramlug]

Superseded by #1038 (rebase of #790) which incidentally cleaned the same conflict markers in lower_call.rs and type_analysis.rs while resolving its own rebase against current main. Verified: git show origin/main:crates/perry-codegen/src/lower_call.rs | grep -c '<<<<<<<' returns 0 now.

Thanks for catching this @proggeramlug — without #1036 surfacing the issue the loop would have kept hitting the same conflict on every PR rebase.