fix(docs): unblock markdown --append on Thai/CJK/emoji content

internal/cmd/docs_markdown.go

@@ -5,6 +5,7 @@ import (
 	"regexp"
 	"strings"
 	"unicode/utf16"
+	"unicode/utf8"
 )
 
 const (
@@ -455,18 +456,16 @@ func ParseInlineFormatting(text string) ([]TextStyle, string) {
 	return styles, strippedText
 }
 
-// nextRune returns the first rune and its byte size from a string
+// nextRune returns the first rune and its byte size from a string.
+// For a string consisting of a single multi-byte rune (e.g. Thai or other
+// non-ASCII text), the previous range-based implementation returned size 0,
+// which caused callers like ParseInlineFormatting to spin in an infinite loop.
 func nextRune(s string) (string, int) {
-	for i, r := range s {
-		if i > 0 {
-			return s[:i], i
-		}
-		if len(s) == 1 {
-			return s, 1
-		}
-		_ = r
+	if s == "" {
+		return "", 0
 	}
-	return "", 0
+	_, size := utf8.DecodeRuneInString(s)
+	return s[:size], size
 }
 
 func parseHeading(line string) (int, string) {

internal/cmd/docs_markdown_thai_test.go

@@ -0,0 +1,110 @@
+package cmd
+
+import (
+	"testing"
+	"time"
+)
+
+// withTimeout runs fn in a goroutine and fails the test if it does not return
+// within d. Used to catch infinite loops without blocking the whole test run.
+func withTimeout(t *testing.T, d time.Duration, name string, fn func()) {
+	t.Helper()
+	done := make(chan struct{})
+	go func() {
+		defer close(done)
+		fn()
+	}()
+	select {
+	case <-done:
+	case <-time.After(d):
+		t.Fatalf("%s: timed out after %s (suspected infinite loop)", name, d)
+	}
+}
+
+// TestNextRune_SingleMultiByteRune is the direct unit-level regression test for
+// the bug that caused `gog docs write --markdown --append` to hang on any
+// content ending in a non-ASCII rune. The previous range-based nextRune
+// returned size=0 for a string that contained exactly one multi-byte rune
+// (e.g. a single Thai character), and ParseInlineFormatting then advanced
+// currentByte by 0, looping forever.
+func TestNextRune_SingleMultiByteRune(t *testing.T) {
+	cases := []struct {
+		name     string
+		in       string
+		wantStr  string
+		wantSize int
+	}{
+		{name: "empty", in: "", wantStr: "", wantSize: 0},
+		{name: "single ascii", in: "a", wantStr: "a", wantSize: 1},
+		{name: "two ascii", in: "ab", wantStr: "a", wantSize: 1},
+		{name: "single thai (3 bytes)", in: "ก", wantStr: "ก", wantSize: 3},
+		{name: "single emoji (4 bytes)", in: "😀", wantStr: "😀", wantSize: 4},
+		{name: "two thai", in: "กข", wantStr: "ก", wantSize: 3},
+		{name: "thai then ascii", in: "กa", wantStr: "ก", wantSize: 3},
+		{name: "ascii then thai", in: "aก", wantStr: "a", wantSize: 1},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			gotStr, gotSize := nextRune(tc.in)
+			if gotStr != tc.wantStr || gotSize != tc.wantSize {
+				t.Fatalf("nextRune(%q) = (%q, %d), want (%q, %d)",
+					tc.in, gotStr, gotSize, tc.wantStr, tc.wantSize)
+			}
+		})
+	}
+}
+
+// TestParseInlineFormatting_Thai ensures the parser returns in finite time on
+// Thai content. With the buggy nextRune this hung forever; we cap each call at
+// 2 seconds via withTimeout to keep the test suite fast even on regression.
+func TestParseInlineFormatting_Thai(t *testing.T) {
+	inputs := []string{
+		"ก",         // single Thai rune
+		"ส่วนคำถาม", // common heading text from the bug report
+		"คำถามที่พบบ่อย",                 // FAQ heading
+		"**ตัวหนา** ปกติ *เอียง* `code`", // bold/italic/code mixed with Thai
+		"พิมพ์ภาษาไทย 😀",                 // emoji at the end (4-byte rune)
+	}
+	for _, in := range inputs {
+		in := in
+		t.Run(in, func(t *testing.T) {
+			withTimeout(t, 2*time.Second, "ParseInlineFormatting", func() {
+				styles, stripped := ParseInlineFormatting(in)
+				_ = styles
+				if stripped == "" {
+					t.Fatalf("ParseInlineFormatting(%q) returned empty stripped text", in)
+				}
+			})
+		})
+	}
+}
+
+// TestMarkdownToDocsRequests_ThaiAppend exercises the full path used by
+// `gog docs write --markdown --append <thai-md>`: parse markdown, then convert
+// to Docs API requests at a non-zero base index. Each input ends in a Thai
+// rune, which is the trigger condition for the original hang.
+func TestMarkdownToDocsRequests_ThaiAppend(t *testing.T) {
+	const sample = `## ส่วนคำถาม
+
+คำถามที่พบบ่อยของลูกค้า
+
+- ราคาเท่าไหร่
+- ส่งของเมื่อไหร่
+
+> ติดต่อสอบถามเพิ่มเติม
+`
+	withTimeout(t, 5*time.Second, "MarkdownToDocsRequests", func() {
+		elements := ParseMarkdown(sample)
+		if len(elements) == 0 {
+			t.Fatal("ParseMarkdown returned no elements for Thai sample")
+		}
+		// baseIndex = 100 mimics appending at the tail of an existing doc.
+		reqs, plain, _ := MarkdownToDocsRequests(elements, 100, "")
+		if plain == "" {
+			t.Fatal("MarkdownToDocsRequests returned empty plain text for Thai sample")
+		}
+		if len(reqs) == 0 {
+			t.Fatal("MarkdownToDocsRequests returned no requests for Thai sample")
+		}
+	})
+}