feat(fmt): go-fmt-style recursive formatting and canonical blank-line layout

Assisted-by: Qwen 3.8 Max Preview

format/format.go

@@ -27,14 +27,6 @@ func Source(path, src string) string {
 		mnemLen int
 		funcID  int
 	}
-	const (
-		kBlank = iota
-		kComment
-		kPreproc
-		kDirective
-		kLabel
-		kInstr
-	)
 
 	infos := make([]info, len(lines))
 	funcID := -1
@@ -70,8 +62,8 @@ func Source(path, src string) string {
 		infos[i] = inf
 	}
 
-	// Second pass: render.
-	var b strings.Builder
+	// Second pass: render each line.
+	outs := make([]outLine, 0, len(lines))
 	inBody := false
 	for i, line := range lines {
 		inf := infos[i]
@@ -106,10 +98,87 @@ func Source(path, src string) string {
 				inBody = false
 			}
 		}
-		b.WriteString(strings.TrimRight(out, " \t"))
-		b.WriteByte('\n')
+		outs = append(outs, outLine{kind: inf.kind, text: strings.TrimRight(out, " \t")})
 	}
-	return b.String()
+	return normalizeSpacing(outs)
+}
+
+// Line classification, shared by the formatting passes.
+const (
+	kBlank = iota
+	kComment
+	kPreproc
+	kDirective
+	kLabel
+	kInstr
+)
+
+// outLine is one rendered line together with its classification.
+type outLine struct {
+	kind int
+	text string
+}
+
+// normalizeSpacing enforces the canonical blank-line layout: runs of blank
+// lines collapse to one, and a new block — a label, or a TEXT or GLOBL
+// directive — is preceded by exactly one blank line.  Comments immediately
+// above a block belong to it, so the blank line is inserted before them.  No
+// blank line is forced at the top of the file, right after a TEXT (the
+// function's first label), or between stacked labels that share an address.
+func normalizeSpacing(outs []outLine) string {
+	blockStart := func(ol outLine) bool {
+		switch ol.kind {
+		case kLabel:
+			return true
+		case kDirective:
+			// TEXT and GLOBL open a block; DATA continues a GLOBL block.
+			return strings.HasPrefix(ol.text, "TEXT") || strings.HasPrefix(ol.text, "GLOBL")
+		}
+		return false
+	}
+	insert := make([]bool, len(outs))
+	for i, ol := range outs {
+		if !blockStart(ol) {
+			continue
+		}
+		j := i
+		for j > 0 && outs[j-1].kind == kComment {
+			j--
+		}
+		if j == 0 {
+			continue // top of file
+		}
+		switch prev := outs[j-1]; {
+		case prev.kind == kBlank, prev.kind == kLabel:
+			continue // already separated, or stacked labels
+		case prev.kind == kDirective && strings.HasPrefix(prev.text, "TEXT"):
+			continue // the function's first label
+		}
+		insert[j] = true
+	}
+
+	var b strings.Builder
+	prevBlank := true // also suppresses leading blanks
+	for i, ol := range outs {
+		if insert[i] && !prevBlank {
+			b.WriteByte('\n')
+		}
+		if ol.kind == kBlank {
+			if !prevBlank {
+				b.WriteByte('\n')
+			}
+			prevBlank = true
+			continue
+		}
+		b.WriteString(ol.text)
+		b.WriteByte('\n')
+		prevBlank = false
+	}
+	out := strings.TrimRight(b.String(), "\n")
+	if out == "" {
+		return ""
+	}
+	return out + "\n"
 }
 
 // renderInstr renders an instruction line: a tab, the mnemonic padded to the

format/format_test.go

@@ -77,6 +77,64 @@ func TestDocCommentIndent(t *testing.T) {
 	}
 }
 
+// TestBlankLines checks the blank-line canonicalisation: exactly one blank
+// line before a new block (a label, or TEXT/GLOBL), runs of blanks collapsed
+// to one, and no blank forced after TEXT, between stacked labels, or at the
+// top of the file.  Leading comments belong to the block they precede.
+func TestBlankLines(t *testing.T) {
+	in := "#include \"textflag.h\"\n" +
+		"TEXT ·f(SB), NOSPLIT, $0\n" +
+		"first:\n" + // first label: no blank after TEXT
+		"XORQ AX, AX\n" +
+		"JMP next\n" + // unlabeled glue: fmt inserts a blank before next:
+		"next:\n" +
+		"stacked:\n" + // stacked labels share an address: no blank between
+		"INCQ AX\n" +
+		"\n" +
+		"\n" + // two blanks collapse to one
+		"// separated block\n" + // comment belongs to the label below
+		"later:\n" +
+		"RET\n" +
+		"// func g()\n" + // doc comment: blank goes before it
+		"TEXT ·g(SB), NOSPLIT, $0\n" +
+		"RET\n" +
+		"GLOBL ·mask(SB), RODATA, $8\n" + // blank before GLOBL…
+		"DATA ·mask+0(SB)/4, $1\n" + // …but not before DATA
+		"\n" +
+		"\n" +
+		"\n" // trailing blanks dropped
+
+	want := "#include \"textflag.h\"\n" +
+		"\n" +
+		"TEXT ·f(SB), NOSPLIT, $0\n" +
+		"first:\n" +
+		"\tXORQ AX, AX\n" +
+		"\tJMP  next\n" +
+		"\n" +
+		"next:\n" +
+		"stacked:\n" +
+		"\tINCQ AX\n" +
+		"\n" +
+		"\t// separated block\n" + // body comment before a label stays indented
+		"later:\n" +
+		"\tRET\n" +
+		"\n" +
+		"// func g()\n" +
+		"TEXT ·g(SB), NOSPLIT, $0\n" +
+		"\tRET\n" +
+		"\n" +
+		"GLOBL ·mask(SB), RODATA, $8\n" +
+		"DATA ·mask+0(SB)/4, $1\n"
+
+	got := Source("b_amd64.s", in)
+	if got != want {
+		t.Fatalf("formatting mismatch:\n--- got ---\n%q\n--- want ---\n%q", got, want)
+	}
+	if again := Source("b_amd64.s", got); again != got {
+		t.Fatalf("not idempotent:\n%q", again)
+	}
+}
+
 func TestOperandSpacing(t *testing.T) {
 	cases := map[string]string{
 		"4(SI)":           "4(SI)",