fix(cgen): fix address-of and as_cast pointer codegen

[guweigang]

Summary

This PR fixes a C code generation issue where combining the address-of (&) operator and as_cast expressions (specifically during smartcasts) results in incorrect pointer casting and dereferencing syntax in the generated C output.

Details of Changes

In vlib/v/gen/c/cgen.v:

1. Gen.expr:
   • Enhanced the address-of (&) operator codegen path. When the right-hand side is an AsCast expression, we check if it evaluates to a pointer casting using the new as_cast_will_use_ptr() helper. If so, we avoid writing incorrect prefix operators (like double address-of or invalid dereference syntax).
2. Gen.write_as_cast_call_start:
   • Added a check for g.is_amp (address-of context) to ensure the compiler generates (Type*) pointer casts rather than standard value casts when smartcasting is performed in an address-of context.
3. Gen.as_cast_will_use_ptr:
   • Introduced this helper method to accurately determine if an AsCast node will result in a pointer-based representation under the hood (e.g. casting interfaces or sum types with differing base types).

Checklist

• I have formatted my V files with v fmt -w.
• All existing tests pass.

Tests Run

Rebuilt the V compiler and verified that all compiler and cgen tests pass:

./v -g -keepc -o ./vnew cmd/v
./vnew -silent test vlib/v/gen/c/

All tests passed successfully on local machine.

[chatgpt-codex-connector]

Codex usage limits have been reached for code reviews. Please check with the admins of this repo to increase the limits by adding credits.

[medvednikov]

@codex review

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 6045f791fb

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".

[guweigang]

Addressed the latest review feedback in 528f9ed:

• keep address-of semantics for temporary-backed &(call() as Foo) casts
• scope pointer-mode codegen to the direct as cast so selectors such as &((sum as Foo).field) remain valid
• add regression coverage for interface casts, sumtype selectors, and call-backed casts

Validation:

• macOS: casts tests, 36/36 passed
• Ubuntu ARM64 VM with VFLAGS="-gc none -cc gcc": 35 passed, 1 JS-only test skipped
• macOS compiler error and C output suites passed

[medvednikov]

@codex review

[medvednikov]

@codex review

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 528f9edd0b

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".

[chatgpt-codex-connector]

Hoist direct address casts out of statement expressions

When &(expr as T) reaches this temporary-evaluation branch with is_direct_amp set, the emitted value becomes a T* pointing into the branch-local ${tmp_var} declared inside the GCC/Clang statement expression. That temporary dies at the end of the expression, so cases not covered by the direct-call special case above, such as p := &((get_wrapper().sum_field as Foo)) or an if/match operand, leave p dangling while still compiling. The MSVC path already hoists to the enclosing statement; direct address casts that need a temp need the same lifetime on the statement-expression path.

Useful? React with 👍 / 👎.