gyojeong 0.2.0 → 0.3.0

package/apps/review-web/dist/feedback-wrapper.html

@@ -4,9 +4,9 @@
   <meta charset="utf-8">
   <meta name="viewport" content="width=device-width, initial-scale=1">
   <title>원본 문서 피드백</title>
-  <script type="module" crossorigin src="./assets/feedback-wrapper-Bhgs7aQS.js"></script>
-  <link rel="modulepreload" crossorigin href="./assets/feedback-contract-CMrinGBH.js">
-  <link rel="stylesheet" crossorigin href="./assets/feedback-wrapper-DWKFHwo5.css">
+  <script type="module" crossorigin src="./assets/feedback-wrapper-BPcSoKgo.js"></script>
+  <link rel="modulepreload" crossorigin href="./assets/feedback-contract-_524rCEj.js">
+  <link rel="stylesheet" crossorigin href="./assets/feedback-wrapper-gbapxzQa.css">
 </head>
 <body>
   <iframe id="source" title="검수할 원본 페이지"></iframe>

package/apps/review-web/dist/index.html

@@ -4,9 +4,9 @@
     <meta charset="utf-8">
     <meta name="viewport" content="width=device-width, initial-scale=1">
     <title>번역 검토</title>
-    <script type="module" crossorigin src="./assets/index-CY49dpth.js"></script>
-    <link rel="modulepreload" crossorigin href="./assets/feedback-contract-CMrinGBH.js">
-    <link rel="stylesheet" crossorigin href="./assets/index-DOaH7bry.css">
+    <script type="module" crossorigin src="./assets/index-DlzIyqhR.js"></script>
+    <link rel="modulepreload" crossorigin href="./assets/feedback-contract-_524rCEj.js">
+    <link rel="stylesheet" crossorigin href="./assets/index-CxnczdwI.css">
   </head>
   <body>
     <main id="app" aria-live="polite"></main>

package/bin/gyojeong.mjs

@@ -16,6 +16,7 @@ const COMMANDS = {
   url: ['review-cli.mjs', 'Print a review link, starting the server if it stopped', true],
   list: ['review-cli.mjs', 'List open reviews (older than 14 days are marked stale)', true],
   discard: ['review-cli.mjs', 'Delete a review without finishing it in the browser', true],
+  rules: ['review-cli.mjs', 'Sync/list rules; --compact true returns the writing rulebook only', true],
   serve: ['review-server.mjs', 'Run the review server in the foreground (default http://127.0.0.1:8377)'],
 };
 
@@ -32,7 +33,7 @@ if (!command || command === '--help' || command === '-h' || !COMMANDS[command])
   const out = [`gyojeong ${packageJson.version}`, '', 'Usage: gyojeong <command> [options]', '', ...lines].join('\n');
   if (command && !COMMANDS[command] && command !== '--help' && command !== '-h') {
     console.error(`Unknown command: ${command}\n\n${out}`);
-    process.exit(2);
+    process.exit(64);
   }
   console.log(out);
   process.exit(0);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "gyojeong",
-  "version": "0.2.0",
+  "version": "0.3.0",
   "description": "Korean review and verification layer: deterministic checks plus a local sentence-level review server.",
   "license": "Apache-2.0",
   "type": "module",

package/scripts/check.mjs

@@ -79,10 +79,107 @@ function trimUrl(url) {
 function maskCode(text) {
   return text.replace(CODE_SPAN_RE, " ");
 }
-function stripComments(text) {
+var HASH_COMMENT_LANGUAGES = /* @__PURE__ */ new Set([
+  "bash",
+  "dockerfile",
+  "make",
+  "makefile",
+  "perl",
+  "php",
+  "pl",
+  "powershell",
+  "ps1",
+  "py",
+  "python",
+  "r",
+  "rb",
+  "ruby",
+  "sh",
+  "shell",
+  "toml",
+  "yaml",
+  "yml",
+  "zsh"
+]);
+var SHELL_COMMENT_LANGUAGES = /* @__PURE__ */ new Set(["bash", "sh", "shell", "zsh"]);
+var WHITESPACE_HASH_COMMENT_LANGUAGES = /* @__PURE__ */ new Set(["yaml", "yml"]);
+var ANYWHERE_HASH_COMMENT_LANGUAGES = /* @__PURE__ */ new Set([
+  "perl",
+  "php",
+  "pl",
+  "powershell",
+  "ps1",
+  "rb",
+  "ruby"
+]);
+var MAKE_LANGUAGES = /* @__PURE__ */ new Set(["make", "makefile"]);
+var PHP_ATTRIBUTE_LANGUAGES = /* @__PURE__ */ new Set(["php"]);
+var LINE_HASH_COMMENT_LANGUAGES = /* @__PURE__ */ new Set(["dockerfile"]);
+var SLASH_COMMENT_LANGUAGES = /* @__PURE__ */ new Set([
+  "c",
+  "c++",
+  "cc",
+  "cpp",
+  "cs",
+  "csharp",
+  "cxx",
+  "go",
+  "h",
+  "hpp",
+  "java",
+  "javascript",
+  "js",
+  "jsonc",
+  "jsx",
+  "kotlin",
+  "kt",
+  "less",
+  "php",
+  "rs",
+  "rust",
+  "scala",
+  "scss",
+  "swift",
+  "ts",
+  "tsx",
+  "typescript"
+]);
+var C_STYLE_BLOCK_COMMENT_LANGUAGES = /* @__PURE__ */ new Set([...SLASH_COMMENT_LANGUAGES, "css"]);
+var HTML_COMMENT_LANGUAGES = /* @__PURE__ */ new Set(["html", "markdown", "md", "svelte", "svg", "vue", "xml"]);
+var PAIRED_DELIMITERS = /* @__PURE__ */ new Map([["(", ")"], ["{", "}"], ["[", "]"], ["<", ">"]]);
+var delimitedLiteralStart = (rest, previous, language) => {
+  if (language === "ruby" || language === "rb") {
+    return rest.match(/^%(?:[qQrwWixI])?([^\w\s])/u);
+  }
+  if ((language === "perl" || language === "pl") && !/[\w$@%]/u.test(previous)) {
+    return rest.match(/^(?:qq|qw|qx|qr|q)([^\w\s])/u);
+  }
+  return null;
+};
+var slashLiteralStart = (text, index, language) => {
+  if (!["perl", "pl", "rb", "ruby"].includes(language) || text[index] !== "/") return false;
+  const beforeText = text.slice(0, index);
+  const before = beforeText.match(/\S\s*$/u)?.[0].trim() ?? "";
+  if (before === "" || /[=(:,!\[\{;~]/u.test(before.at(-1) ?? "")) return true;
+  const word = beforeText.match(/([A-Za-z]+)\s*$/u)?.[1]?.toLowerCase();
+  if (["rb", "ruby"].includes(language)) {
+    return ["and", "if", "not", "or", "return", "then", "unless", "when", "while"].includes(word ?? "");
+  }
+  return ["m", "qr", "s"].includes(word ?? "");
+};
+function stripComments(text, language = "") {
+  const normalizedLanguage = language.toLowerCase();
+  const hashComments = HASH_COMMENT_LANGUAGES.has(normalizedLanguage);
+  const slashComments = SLASH_COMMENT_LANGUAGES.has(normalizedLanguage);
+  const cStyleBlockComments = C_STYLE_BLOCK_COMMENT_LANGUAGES.has(normalizedLanguage);
+  const htmlComments = HTML_COMMENT_LANGUAGES.has(normalizedLanguage);
+  const knownLanguage = hashComments || slashComments || cStyleBlockComments || htmlComments;
   let out = "";
   let quote = null;
+  let quoteOpen = null;
+  let quoteDepth = 0;
   let block = null;
+  let lineStart = 0;
   for (let i = 0; i < text.length; i += 1) {
     const rest = text.slice(i);
     if (block) {
@@ -90,48 +187,92 @@ function stripComments(text) {
         i += block.length - 1;
         block = null;
       }
+      if (text[i] === "\n") {
+        lineStart = i + 1;
+      }
       continue;
     }
     const char = text[i];
+    const atLineStart = /^\s*$/.test(text.slice(lineStart, i));
+    const genericSlashComment = !knownLanguage && atLineStart && rest.startsWith("//");
+    const genericHashComment = !knownLanguage && atLineStart && char === "#" && !rest.startsWith("#!") && !/^#[\w-]+\s*\{/u.test(rest);
+    const genericCStyleBlockComment = !knownLanguage && atLineStart && rest.startsWith("/*");
+    const genericHtmlComment = !knownLanguage && atLineStart && rest.startsWith("<!--");
+    const previous = text[i - 1] ?? "";
+    const makeRecipe = MAKE_LANGUAGES.has(normalizedLanguage) && text[lineStart] === "	";
+    const shellHashComment = (SHELL_COMMENT_LANGUAGES.has(normalizedLanguage) || makeRecipe) && char === "#" && (i === 0 || /[\s;&|()<>]/u.test(previous));
+    const whitespaceHashComment = WHITESPACE_HASH_COMMENT_LANGUAGES.has(normalizedLanguage) && char === "#" && (i === 0 || /\s/u.test(previous)) && !(PHP_ATTRIBUTE_LANGUAGES.has(normalizedLanguage) && rest.startsWith("#["));
+    const lineHashComment = LINE_HASH_COMMENT_LANGUAGES.has(normalizedLanguage) && char === "#" && atLineStart;
+    const anywhereHashComment = ANYWHERE_HASH_COMMENT_LANGUAGES.has(normalizedLanguage) && char === "#" && !(PHP_ATTRIBUTE_LANGUAGES.has(normalizedLanguage) && rest.startsWith("#["));
+    const makeHashComment = MAKE_LANGUAGES.has(normalizedLanguage) && char === "#" && (makeRecipe ? shellHashComment : true);
+    const languageHashComment = hashComments && char === "#" && (SHELL_COMMENT_LANGUAGES.has(normalizedLanguage) || makeRecipe ? shellHashComment : WHITESPACE_HASH_COMMENT_LANGUAGES.has(normalizedLanguage) ? whitespaceHashComment : LINE_HASH_COMMENT_LANGUAGES.has(normalizedLanguage) ? lineHashComment : MAKE_LANGUAGES.has(normalizedLanguage) ? makeHashComment : anywhereHashComment);
     if (quote) {
       out += char;
       if (char === "\\") {
         out += text[i + 1] ?? "";
         i += 1;
+      } else if (quoteOpen && char === quoteOpen) {
+        quoteDepth += 1;
       } else if (char === quote) {
-        quote = null;
+        if (quoteDepth > 0) {
+          quoteDepth -= 1;
+        } else {
+          quote = null;
+          quoteOpen = null;
+        }
       }
       continue;
     }
+    const delimitedLiteral = delimitedLiteralStart(rest, previous, normalizedLanguage);
+    if (delimitedLiteral) {
+      const opener = delimitedLiteral[1];
+      const token = delimitedLiteral[0];
+      quoteOpen = PAIRED_DELIMITERS.has(opener) ? opener : null;
+      quote = PAIRED_DELIMITERS.get(opener) ?? opener;
+      quoteDepth = 0;
+      out += token;
+      i += token.length - 1;
+      continue;
+    }
     if (char === '"' || char === "'" || char === "`") {
       quote = char;
       out += char;
       continue;
     }
-    if (rest.startsWith("/*")) {
+    if (slashLiteralStart(text, i, normalizedLanguage)) {
+      quote = "/";
+      out += char;
+      continue;
+    }
+    if (cStyleBlockComments && rest.startsWith("/*") || genericCStyleBlockComment) {
       block = "*/";
       continue;
     }
-    if (rest.startsWith("<!--")) {
+    if (htmlComments && rest.startsWith("<!--") || genericHtmlComment) {
       block = "-->";
       continue;
     }
-    if (rest.startsWith("//") || char === "#") {
+    if (slashComments && rest.startsWith("//") || languageHashComment || genericSlashComment || genericHashComment) {
       const end = text.indexOf("\n", i);
       if (end === -1) break;
       i = end - 1;
+      lineStart = end;
       continue;
     }
     out += char;
+    if (char === "\n") {
+      lineStart = i + 1;
+    }
   }
   return out;
 }
 function normalizeCodeBlock(text) {
   const raw = text.split("\n");
   const open = raw[0] ?? "";
+  const language = open.match(/^\s*(?:`{3,}|~{3,})\s*([^\s{]+)/)?.[1] ?? "";
   const closes = raw.length > 1 && /^\s*(`{3,}|~{3,})\s*$/.test(raw[raw.length - 1] ?? "");
   const body = raw.slice(1, closes ? -1 : void 0).join("\n");
-  const lines = [open, ...stripComments(body).split("\n")];
+  const lines = [open, ...stripComments(body, language).split("\n")];
   const indent = lines[0].match(/^\s*/)[0].length;
   return lines.map((line) => line.slice(Math.min(indent, line.match(/^\s*/)[0].length))).map((line) => line.replace(/\s+$/, "")).filter((line) => line.trim() !== "").join("\n");
 }
@@ -246,25 +387,18 @@ var ENGLISH_NUMBER_WORDS = {
   triple: 3,
   both: 2,
   dozen: 12,
-  half: 50,
-  hex: 16,
-  hexadecimal: 16,
-  octal: 8,
-  binary: 2,
-  decimal: 10,
-  january: 1,
-  february: 2,
-  march: 3,
-  april: 4,
-  may: 5,
-  june: 6,
-  july: 7,
-  august: 8,
-  september: 9,
-  october: 10,
-  november: 11,
-  december: 12
+  half: 50
 };
+var MONTHS = { january: 1, february: 2, march: 3, april: 4, may: 5, june: 6, july: 7, august: 8, september: 9, october: 10, november: 11, december: 12 };
+var MONTH_DATE_RE = new RegExp(`\\b(${Object.keys(MONTHS).join("|")})\\s+\\d|\\d\\s+(${Object.keys(MONTHS).join("|")})\\b`, "g");
+var SECOND_UNIT_RE = /\b(?:\d+(?:\.\d+)?|a|an|one|per|each|every)\s+second\b|\bseconds\b/g;
+var KOREAN_BASE_RE = /(\d+)\s*진/g;
+var ENGLISH_BASES = { hex: 16, hexadecimal: 16, octal: 8, binary: 2, decimal: 10 };
+var ENGLISH_BASE_RE = new RegExp(`\\b(${Object.keys(ENGLISH_BASES).join("|")})\\b`, "g");
+function sourceBases(prose, language) {
+  if (language === "en") return new Set([...prose.toLowerCase().matchAll(ENGLISH_BASE_RE)].map((m) => String(ENGLISH_BASES[m[1]])));
+  return new Set([...prose.matchAll(KOREAN_BASE_RE)].map((m) => normalizeNumber(m[1])));
+}
 var KOREAN_NUMBER_WORDS = { \uD55C: 1, \uD558\uB098: 1, \uB450: 2, \uB458: 2, \uC138: 3, \uC14B: 3, \uB124: 4, \uB137: 4, \uB2E4\uC12F: 5, \uC5EC\uC12F: 6, \uC77C\uACF1: 7, \uC5EC\uB35F: 8, \uC544\uD649: 9, \uC5F4: 10 };
 var KOREAN_NUMERAL = "\uD55C|\uD558\uB098|\uB450|\uB458|\uC138|\uC14B|\uB124|\uB137|\uB2E4\uC12F|\uC5EC\uC12F|\uC77C\uACF1|\uC5EC\uB35F|\uC544\uD649|\uC5F4";
 var KOREAN_COUNTER = "\uAC1C|\uBC88|\uAC00\uC9C0|\uBA85|\uB300|\uC904|\uB2E8\uACC4|\uBC30|\uC2DC\uAC04|\uBD84|\uCD08|\uC77C|\uC8FC|\uB2EC|\uACF3|\uCABD|\uC7A5|\uCE78";
@@ -289,7 +423,9 @@ function digitNumbers(prose) {
 function justifiedNumbers(prose, language) {
   const out = /* @__PURE__ */ new Set();
   if (language === "en") {
-    for (const m of prose.toLowerCase().matchAll(/\b[a-z]+\b/g)) {
+    const lower = prose.toLowerCase();
+    for (const m of lower.matchAll(MONTH_DATE_RE)) out.add(String(MONTHS[m[1] ?? m[2]]));
+    for (const m of lower.replace(SECOND_UNIT_RE, " ").matchAll(/\b[a-z]+\b/g)) {
       if (Object.hasOwn(ENGLISH_NUMBER_WORDS, m[0])) out.add(String(ENGLISH_NUMBER_WORDS[m[0]]));
     }
     if (/\bnon-?negative\b|\bzero-based\b/i.test(prose)) out.add("0");
@@ -304,14 +440,17 @@ function justifiedNumbers(prose, language) {
 }
 function checkNumbers(sourceBlocks, draftBlocks, sourceLanguage, add) {
   const sourceProse = numberProse(sourceBlocks);
-  const draftProse = numberProse(draftBlocks);
+  const draftText = numberProse(draftBlocks);
+  const bases = sourceBases(sourceProse, sourceLanguage);
+  const draftBases = new Set([...draftText.matchAll(KOREAN_BASE_RE)].map((m) => normalizeNumber(m[1])));
+  const draftProse = draftText.replace(KOREAN_BASE_RE, " ");
   const sourceDigits = digitNumbers(sourceProse);
   const draftDigits = digitNumbers(draftProse);
   const justified = /* @__PURE__ */ new Set([...sourceDigits, ...justifiedNumbers(sourceProse, sourceLanguage)]);
   const draftWords = justifiedNumbers(draftProse, "ko");
-  const draftAll = /* @__PURE__ */ new Set([...draftDigits, ...draftWords]);
+  const draftAll = /* @__PURE__ */ new Set([...draftDigits, ...draftWords, ...draftBases]);
   const addedWords = [...draftWords].filter((n) => Number(n) >= 3);
-  for (const n of [...draftDigits, ...addedWords]) {
+  for (const n of [...draftDigits, ...addedWords, ...[...draftBases].filter((b) => !bases.has(b))]) {
     if (!justified.has(n)) add("fail", "number", `\uC6D0\uBB38\uC5D0 \uC5C6\uB294 \uC22B\uC790\uAC00 \uC0DD\uACBC\uC2B5\uB2C8\uB2E4: ${n}`, { token: n });
   }
   for (const n of sourceDigits) {
@@ -322,7 +461,7 @@ var EN_WEAK_PROHIBITION = [/\bavoid(?:s|ing)?\b/i, /\b(?:forbid|forbids|disallow
 var EN_PERMISSION = /\b(?:may|allowed|permitted|optional|optionally|free to|fine|okay|ok|need not|needn't|not required|no need|(?:do|does)n't need|(?:do|does) not need|(?:do|does)n't have to|(?:do|does) not have to|acceptable)\b/i;
 var EN = {
   prohibition: [
-    /(?:^|[.!?:]\s+|\n\s*(?:[-*+]|\d+[.)])?\s*|\*\*|>\s*)(?:do not|don't|never|avoid)\b(?!\s+(?:need|have to|require))/i,
+    /(?:^|[.!?:;]\s+|\n\s*(?:[-*+]|\d+[.)])?\s*|\*\*|>\s*)(?:do not|don't|never|avoid)\b(?!\s+(?:need|have to|require))/i,
     // "should not", and "should, however, not"
     /\b(?:must|shall|should)(?:\s+not|n't|,[^,.;!?\n]{1,25},\s+not)\b/i,
     /\b(?:is|are|isn't|aren't)\s+not\s+(?:allowed|permitted)\b/i,