[colbymchenry]

Summary

A function that writes a macro before a block, or with no ; after it, counts as using the macro again. This closes regression 4 of the comparison with 1.6.2 (TEST_CONTEXT("…") { on redis) and rocksdb's XXH_ASSERT references.

First of a stack of four. The base of this PR is main.

Root cause

cbd9817 (#1736) taught the statement-macro blank to take names in capitals, so that ATOMIC_BLOCK(x) { stops reading as a K&R function. The blank takes the whole call, name included. The function no longer named the macro, so its value reference went too: 10 on redis, 16 on rocksdb.

The fix

In a statement of a function body the name stays, as name;, in the bytes the call took. Anywhere else a ; would be the error the blank avoids, so the call goes whole as before:

• in an initializer, a struct or enum body, another macro's arguments, or the rest of an expression;
• as the body of an if, for, else or do, where it would part the block from its else or while.

cStatementLines reads which lines those are from the file's code, once, and only for a file that has such a macro. The change is length-preserving and in the preParse, so the kernel gets the same bytes and needs no change.

The arguments stay blanked. A macro's can be types, which a call would not parse, and no call or reference lost against 1.6.2 on redis or rocksdb was in one. Lowercase iterator macros take the same path, so nk_foreach in glfw's nuklear.h counts as used too.

Validation

Windows 11, main 4499d00.

Tests. 7 of the 9 new tests fail on main and all pass. 11 mutations each fail a test. Three older tests that said the name goes now say it stays (extraction.test.ts, cpp-raw-string-preparse-1505.test.ts).

A/B against main.

Repo	Language	Result
redis	C	the 10 back, nothing lost or moved by this commit
rocksdb	C++ and C	15 of the 16 back on the same functions, 2 phantom nodes with 4 calls back, 20 spilled locals gone
glfw	C	+1, nk_convert really uses nk_foreach
jq, betaflight	C	nothing from this commit
leveldb, gin, flask	controls	identical

The 16th XXH_ASSERT sat on XXH3_ACCUMULATE_TEMPLATE(scalar), a K&R misreading on both arms whose body is XXH3_scalarScrambleRound, which neither arm has as a node. REG read that entry as neither arm's.

Parse errors. Over every C file of redis, jq, glfw, betaflight and linux-sub, 10,220 files: 0 worse, 0 better, error counts identical.

Kernel parity, 0 diffs. redis 639/796 (the deferrals main has), jq 66/81, glfw 69/97, rocksdb db + util 381/463, betaflight src 1177/1465.

Cost. Whole-index CPU, median of 5 interleaved rounds, base and stack built the same way: redis 23.70 to 23.84 s (+0.6%), wall 7.98 to 7.89 s. cStatementLines is 42 ms of a redis index in a profile. C preParse alone: linux-sub 6.1 to 6.6 s, redis 0.94 to 0.99 s.

Suite. 857 files passed and 6 skipped, 10,651 tests passed, 2 expected fail, 87 skipped, 4 workers, with the kernel, on the stack one clean rebase back (719ea417 on main 3bd9949). After the rebase onto 087de4c, whose hunks are apart from these: 138 files and 2,224 tests of the C, C++, Objective-C, Python and extraction suites pass on the tip.

🤖 Generated with Claude Code