fix(installer): Codex TOML block finder preserves trailing array-of-tables siblings (#1351) (#1370)

The Codex installer's `findNextTableHeader` skipped `[[array-of-tables]]` headers instead of treating them as a block boundary, so any `[[...]]` block after `[mcp_servers.codegraph]` in ~/.codex/config.toml was silently deleted on install/upgrade/uninstall. Now treats both `[...]` and `[[...]]` as boundaries, with a small line lexer so header-shaped text inside multiline strings/arrays isn't mistaken for a boundary. Adds round-trip regression coverage (install → reinstall → uninstall) + CHANGELOG entry.

Fixes #1351. Supersedes #624.

Thanks @KtzeAbyss.

src/installer/targets/toml.ts

@@ -7,13 +7,14 @@
  *
  * Strategy: treat the file as text. Find the `[mcp_servers.codegraph]`
  * header line, splice it (and the lines that follow it until the next
- * `[...]` header or EOF) in or out. Everything outside that block is
- * preserved verbatim, byte-for-byte.
+ * `[...]` / `[[...]]` header or EOF) in or out. A small lexical scan keeps
+ * header-shaped text inside multiline values out of the boundary search.
+ * Everything outside that block is preserved verbatim, byte-for-byte.
  *
  * Limitations (acceptable for our narrow use):
- *   - Only handles top-level table headers; not array-of-tables or
- *     subtables nested inside `[mcp_servers]` itself (we always write
- *     the full dotted key `[mcp_servers.codegraph]`).
+ *   - Only writes a top-level table header. Array-of-tables and sibling
+ *     subtables are preserved as opaque blocks (we always write the full
+ *     dotted key `[mcp_servers.codegraph]`).
  *   - Doesn't validate sibling TOML — if the file is malformed
  *     elsewhere, our injection won't fix it but won't make it worse.
  *   - Quotes string values with double quotes; escapes `\` and `"`.
@@ -79,7 +80,7 @@ export function upsertTomlTable(
     };
   }
 
-  // Find the end of this block: next `[...]` header (at line start) or EOF.
+  // Find the end of this block: next table header or EOF.
   const blockEnd = findNextTableHeader(fileContent, headerIdx + headerLine.length);
   const existingBlock = fileContent.substring(headerIdx, blockEnd).replace(/\n+$/, '');
 
@@ -133,22 +134,121 @@ function findHeaderIndex(content: string, headerLine: string): number {
 }
 
 /**
- * Find the byte index of the next top-level `[...]` table header
- * (excluding array-of-tables `[[...]]`) starting from `from`, or
- * return content length when none.
+ * Find the byte index of the next `[...]` or `[[...]]` table header
+ * starting from `from`, or return content length when none.
  */
 function findNextTableHeader(content: string, from: number): number {
-  // Look for "\n[" but skip "\n[[" (array of tables).
-  let i = from;
-  while (i < content.length) {
-    const nlIdx = content.indexOf('\n[', i);
-    if (nlIdx === -1) return content.length;
-    if (content[nlIdx + 2] === '[') {
-      // [[...]] — keep searching past it.
-      i = nlIdx + 2;
-      continue;
+  const state: TomlLexState = { multilineString: null, arrayDepth: 0, inlineTableDepth: 0 };
+  let lineStart = from;
+  let isHeaderRemainder = true;
+
+  while (lineStart < content.length) {
+    const newlineIdx = content.indexOf('\n', lineStart);
+    const lineEnd = newlineIdx === -1 ? content.length : newlineIdx;
+    const line = content.slice(lineStart, lineEnd);
+
+    if (
+      !isHeaderRemainder &&
+      state.multilineString === null &&
+      state.arrayDepth === 0 &&
+      state.inlineTableDepth === 0 &&
+      isTomlTableHeader(line)
+    ) {
+      return lineStart;
     }
-    return nlIdx + 1;
+
+    scanTomlLine(line, state);
+    if (newlineIdx === -1) break;
+    lineStart = newlineIdx + 1;
+    isHeaderRemainder = false;
   }
+
   return content.length;
 }
+
+type MultilineStringDelimiter = '"""' | "'''";
+
+interface TomlLexState {
+  multilineString: MultilineStringDelimiter | null;
+  arrayDepth: number;
+  inlineTableDepth: number;
+}
+
+const TOML_KEY_PART = String.raw`(?:[A-Za-z0-9_-]+|"(?:\\.|[^"\\])*"|'[^']*')`;
+const TOML_DOTTED_KEY = String.raw`${TOML_KEY_PART}(?:[ \t]*\.[ \t]*${TOML_KEY_PART})*`;
+const TOML_TABLE = String.raw`\[[ \t]*${TOML_DOTTED_KEY}[ \t]*\]`;
+const TOML_ARRAY_TABLE = String.raw`\[\[[ \t]*${TOML_DOTTED_KEY}[ \t]*\]\]`;
+const TOML_TABLE_HEADER = new RegExp(
+  String.raw`^[ \t]*(?:${TOML_TABLE}|${TOML_ARRAY_TABLE})[ \t]*(?:#.*)?\r?$`
+);
+
+function isTomlTableHeader(line: string): boolean {
+  return TOML_TABLE_HEADER.test(line);
+}
+
+/** Track value constructs that may legally span lines so bracket-shaped string
+ * content and nested arrays cannot be mistaken for sibling table headers. */
+function scanTomlLine(line: string, state: TomlLexState): void {
+  for (let i = 0; i < line.length;) {
+    if (state.multilineString !== null) {
+      const end = findMultilineStringEnd(line, i, state.multilineString);
+      if (end === -1) return;
+      i = end + state.multilineString.length;
+      state.multilineString = null;
+      continue;
+    }
+
+    if (line[i] === '#') return;
+
+    const multiline = line.startsWith('"""', i)
+      ? '"""'
+      : line.startsWith("'''", i)
+        ? "'''"
+        : null;
+    if (multiline !== null) {
+      state.multilineString = multiline;
+      i += multiline.length;
+      continue;
+    }
+
+    const ch = line[i]!;
+    if (ch === '"' || ch === "'") {
+      i = skipSingleLineString(line, i, ch);
+      continue;
+    }
+    if (ch === '[') state.arrayDepth++;
+    else if (ch === ']' && state.arrayDepth > 0) state.arrayDepth--;
+    else if (ch === '{') state.inlineTableDepth++;
+    else if (ch === '}' && state.inlineTableDepth > 0) state.inlineTableDepth--;
+    i++;
+  }
+}
+
+function findMultilineStringEnd(
+  line: string,
+  from: number,
+  delimiter: MultilineStringDelimiter,
+): number {
+  let end = line.indexOf(delimiter, from);
+  while (delimiter === '"""' && end !== -1 && isBackslashEscaped(line, end)) {
+    end = line.indexOf(delimiter, end + 1);
+  }
+  return end;
+}
+
+function isBackslashEscaped(line: string, index: number): boolean {
+  let backslashes = 0;
+  for (let i = index - 1; i >= 0 && line[i] === '\\'; i--) backslashes++;
+  return backslashes % 2 === 1;
+}
+
+function skipSingleLineString(line: string, start: number, quote: '"' | "'"): number {
+  for (let i = start + 1; i < line.length; i++) {
+    if (quote === '"' && line[i] === '\\') {
+      i++;
+      continue;
+    }
+    if (line[i] === quote) return i + 1;
+  }
+  return line.length;
+}