fix(parser): substitute macro parameters behind element selectors

Assisted-by: GLM 5.3 Flash
2026-09-20 22:02:19 +02:00
parent 522e6f2ae8
commit a7744c24bd
2 changed files with 210 additions and 2 deletions
@@ -536,6 +536,127 @@ func TestConstantExpressionFoldsWithoutExpand(t *testing.T) {
 	}
 }

+func TestParameterWithSelectorSubstitutes(t *testing.T) {
+	// The lexer folds A.S4 into one identifier token, so a parameter used
+	// with an element or lane selector never matched the whole-token
+	// substitution; the toolchain's lexer splits on the period and its
+	// substitution sees the name alone.  Several parameters carry selectors
+	// in one body here, which is the chacha8_arm64.s QR shape in miniature.
+	_, got := expand(t, `
+#define QR(A, B, C, D) VADD A.S4, B.S4, C.S4; VEOR D.B16, A.B16, D.B16
+TEXT ·f(SB), NOSPLIT, $0
+	QR(V0, V1, V2, V3)
+	RET
+`)
+	wantLines(t, got,
+		"VADD V0.S4, V1.S4, V2.S4",
+		"VEOR V3.B16, V0.B16, V3.B16",
+		"RET",
+	)
+}
+
+func TestSelectorWithCompoundArgumentPastesTextually(t *testing.T) {
+	// An argument that is itself one compound identifier pastes verbatim:
+	// A.S4 over V0.B16 spells V0.B16.S4, the operand the toolchain's
+	// split-then-substitute leaves behind.
+	_, got := expand(t, `
+#define M(A) VADD A.S4, A.S4, A.S4
+TEXT ·f(SB), NOSPLIT, $0
+	M(V0.B16)
+	RET
+`)
+	wantLines(t, got, "VADD V0.B16.S4, V0.B16.S4, V0.B16.S4", "RET")
+}
+
+func TestSelectorAlongsideBareParameter(t *testing.T) {
+	// A body may use the parameter bare and suffixed, and the argument may
+	// itself end in a selector; neither disturbs the other.
+	_, got := expand(t, `
+#define M(A) VADD A, A.S4, A
+TEXT ·f(SB), NOSPLIT, $0
+	M(V0)
+	M(V1.B16)
+	RET
+`)
+	wantLines(t, got,
+		"VADD V0, V0.S4, V0",
+		"VADD V1.B16, V1.B16.S4, V1.B16",
+		"RET",
+	)
+}
+
+func TestSelectorKeepsNonParameterPrefixes(t *testing.T) {
+	// The prefix before the period must be the whole parameter name:
+	// AB.S4 never reaches a parameter A.
+	_, got := expand(t, `
+#define M(A) VADD AB.S4, A.S4, AB.S4
+TEXT ·f(SB), NOSPLIT, $0
+	M(V0)
+	RET
+`)
+	wantLines(t, got, "VADD AB.S4, V0.S4, AB.S4", "RET")
+}
+
+func TestSelectorExpandsMacroValuedArgument(t *testing.T) {
+	// gcm_arm64.s invokes mulRound(B1) where B1 is itself an object macro:
+	// the paste stays rescannable, so B1.D1 still expands to V1.D1 the way
+	// the toolchain's rescan of substituted tokens does.
+	_, got := expand(t, `
+#define B1 V1
+#define mulRound(X) VPMULL X.D1, T1.D1, T3.Q1
+TEXT ·f(SB), NOSPLIT, $0
+	mulRound(B1)
+	RET
+`)
+	wantLines(t, got, "VPMULL V1.D1, T1.D1, T3.Q1", "RET")
+}
+
+func TestObjectMacroBehindSelectorExpands(t *testing.T) {
+	// Ordinary code writes ACC0.B16 where ACC0 is an object macro; the
+	// toolchain expands the alias because its lexer reads the selector as
+	// its own token, and the lookup here must reach the macro through the
+	// compound spelling the same way.
+	_, got := expand(t, `
+#define ACC0 V8
+TEXT ·f(SB), NOSPLIT, $0
+	VEOR ACC0.B16, ACC0.B16, ACC0.B16
+	RET
+`)
+	wantLines(t, got, "VEOR V8.B16, V8.B16, V8.B16", "RET")
+}
+
+func TestChacha8QRMacroExpands(t *testing.T) {
+	// The real QR round of chacha8_arm64.s end to end: every parameter
+	// carries a selector somewhere, and the round is sixteen instructions.
+	_, got := expand(t, `
+#define QR(A, B, C, D) \
+	VADD A.S4, B.S4, A.S4; VEOR D.B16, A.B16, D.B16; VREV32 D.H8, D.H8; \
+	VADD C.S4, D.S4, C.S4; VEOR B.B16, C.B16, V30.B16; VSHL $12, V30.S4, B.S4; VSRI $20, V30.S4, B.S4; \
+	VADD A.S4, B.S4, A.S4; VEOR D.B16, A.B16, D.B16; VTBL V31.B16, [D.B16], D.B16; \
+	VADD C.S4, D.S4, C.S4; VEOR B.B16, C.B16, V30.B16; VSHL $7, V30.S4, B.S4; VSRI $25, V30.S4, B.S4
+TEXT ·f(SB), NOSPLIT, $0
+	QR(V0, V1, V2, V3)
+	RET
+`)
+	wantLines(t, got,
+		"VADD V0.S4, V1.S4, V0.S4",
+		"VEOR V3.B16, V0.B16, V3.B16",
+		"VREV32 V3.H8, V3.H8",
+		"VADD V2.S4, V3.S4, V2.S4",
+		"VEOR V1.B16, V2.B16, V30.B16",
+		"VSHL $12, V30.S4, V1.S4",
+		"VSRI $20, V30.S4, V1.S4",
+		"VADD V0.S4, V1.S4, V0.S4",
+		"VEOR V3.B16, V0.B16, V3.B16",
+		"VTBL V31.B16, [V3.B16], V3.B16",
+		"VADD V2.S4, V3.S4, V2.S4",
+		"VEOR V1.B16, V2.B16, V30.B16",
+		"VSHL $7, V30.S4, V1.S4",
+		"VSRI $25, V30.S4, V1.S4",
+		"RET",
+	)
+}
+
 func TestNotAnExpressionFallsBack(t *testing.T) {
 	// Symbol immediates and floats must keep their ordinary parse.
 	f, errs := ParseWithOptions("t_amd64.s", "TEXT ·f(SB), NOSPLIT, $0\n\tMOVQ $1.5, AX\n\tMOVQ $·sym(SB), AX\n\tRET\n", Options{})