Unrolled build for rust-lang#126057

rust-timer · web-flow · commit 40228785b3c1 · 2024-06-08T00:39:20.000-04:00
Rollup merge of rust-lang#126057 - Sunshine40:rustdoc-search-non-english, r=notriddle Make html rendered by rustdoc allow searching non-English identifier / alias Fix alias search result showing `undefined` description. Inspired by rust-lang/mdBook#2393 . Not sure if it's worth it adding full-text search functionality to rustdoc rendered html.
diff --git a/src/ci/docker/host-x86_64/mingw-check/Dockerfile b/src/ci/docker/host-x86_64/mingw-check/Dockerfile
@@ -61,8 +61,8 @@ ENV SCRIPT python3 ../x.py check --stage 0 --set build.optimized-compiler-builti
            /scripts/validate-toolstate.sh && \
            /scripts/validate-error-codes.sh && \
            reuse --include-submodules lint && \
-           # Runs checks to ensure that there are no ES5 issues in our JS code.
-           es-check es8 ../src/librustdoc/html/static/js/*.js && \
+           # Runs checks to ensure that there are no issues in our JS code.
+           es-check es2019 ../src/librustdoc/html/static/js/*.js && \
            eslint -c ../src/librustdoc/html/static/.eslintrc.js ../src/librustdoc/html/static/js/*.js && \
            eslint -c ../src/tools/rustdoc-js/.eslintrc.js ../src/tools/rustdoc-js/tester.js && \
            eslint -c ../src/tools/rustdoc-gui/.eslintrc.js ../src/tools/rustdoc-gui/tester.js
diff --git a/src/librustdoc/html/static/.eslintrc.js b/src/librustdoc/html/static/.eslintrc.js
@@ -5,7 +5,7 @@ module.exports = {
     },
     "extends": "eslint:recommended",
     "parserOptions": {
-        "ecmaVersion": 8,
+        "ecmaVersion": 2019,
         "sourceType": "module"
     },
     "rules": {
diff --git a/src/librustdoc/html/static/js/externs.js b/src/librustdoc/html/static/js/externs.js
@@ -41,8 +41,9 @@ let ParserState;
  *     foundElems: number,
  *     totalElems: number,
  *     literalSearch: boolean,
- *     corrections: Array<{from: string, to: integer}>,
+ *     corrections: Array<{from: string, to: integer}> | null,
  *     typeFingerprint: Uint32Array,
+ *     error: Array<string> | null,
  * }}
  */
 let ParsedQuery;
diff --git a/src/librustdoc/html/static/js/search.js b/src/librustdoc/html/static/js/search.js
@@ -89,6 +89,10 @@ const ROOT_PATH = typeof window !== "undefined" ? window.rootPath : "../";
 // of permutations we need to check.
 const UNBOXING_LIMIT = 5;
 
+// used for search query verification
+const REGEX_IDENT = /\p{ID_Start}\p{ID_Continue}*|_\p{ID_Continue}+/uy;
+const REGEX_INVALID_TYPE_FILTER = /[^a-z]/ui;
+
 // In the search display, allows to switch between tabs.
 function printTab(nb) {
     let iter = 0;
@@ -410,18 +414,21 @@ function initSearch(rawSearchIndex) {
     }
 
     /**
-     * Returns `true` if the given `c` character is valid for an ident.
+     * If the current parser position is at the beginning of an identifier,
+     * move the position to the end of it and return `true`. Otherwise, return `false`.
      *
-     * @param {string} c
+     * @param {ParserState} parserState
      *
      * @return {boolean}
      */
-    function isIdentCharacter(c) {
-        return (
-            c === "_" ||
-            (c >= "0" && c <= "9") ||
-            (c >= "a" && c <= "z") ||
-            (c >= "A" && c <= "Z"));
+    function consumeIdent(parserState) {
+        REGEX_IDENT.lastIndex = parserState.pos;
+        const match = parserState.userQuery.match(REGEX_IDENT);
+        if (match) {
+            parserState.pos += match[0].length;
+            return true;
+        }
+        return false;
     }
 
     /**
@@ -618,70 +625,62 @@ function initSearch(rawSearchIndex) {
      * @return {integer}
      */
     function getIdentEndPosition(parserState) {
-        const start = parserState.pos;
+        let afterIdent = consumeIdent(parserState);
         let end = parserState.pos;
-        let foundExclamation = -1;
+        let macroExclamation = -1;
         while (parserState.pos < parserState.length) {
             const c = parserState.userQuery[parserState.pos];
-            if (!isIdentCharacter(c)) {
-                if (c === "!") {
-                    if (foundExclamation !== -1) {
-                        throw ["Cannot have more than one ", "!", " in an ident"];
-                    } else if (parserState.pos + 1 < parserState.length &&
-                        isIdentCharacter(parserState.userQuery[parserState.pos + 1])
-                    ) {
+            if (c === "!") {
+                if (macroExclamation !== -1) {
+                    throw ["Cannot have more than one ", "!", " in an ident"];
+                } else if (parserState.pos + 1 < parserState.length) {
+                    const pos = parserState.pos;
+                    parserState.pos++;
+                    const beforeIdent = consumeIdent(parserState);
+                    parserState.pos = pos;
+                    if (beforeIdent) {
                         throw ["Unexpected ", "!", ": it can only be at the end of an ident"];
                     }
-                    foundExclamation = parserState.pos;
-                } else if (isPathSeparator(c)) {
-                    if (c === ":") {
-                        if (!isPathStart(parserState)) {
+                }
+                if (afterIdent) macroExclamation = parserState.pos;
+            } else if (isPathSeparator(c)) {
+                if (c === ":") {
+                    if (!isPathStart(parserState)) {
+                        break;
+                    }
+                    // Skip current ":".
+                    parserState.pos += 1;
+                } else {
+                    while (parserState.pos + 1 < parserState.length) {
+                        const next_c = parserState.userQuery[parserState.pos + 1];
+                        if (next_c !== " ") {
                             break;
                         }
-                        // Skip current ":".
                         parserState.pos += 1;
-                    } else {
-                        while (parserState.pos + 1 < parserState.length) {
-                            const next_c = parserState.userQuery[parserState.pos + 1];
-                            if (next_c !== " ") {
-                                break;
-                            }
-                            parserState.pos += 1;
-                        }
                     }
-                    if (foundExclamation !== -1) {
-                        if (foundExclamation !== start &&
-                            isIdentCharacter(parserState.userQuery[foundExclamation - 1])
-                        ) {
-                            throw ["Cannot have associated items in macros"];
-                        } else {
-                            // while the never type has no associated macros, we still
-                            // can parse a path like that
-                            foundExclamation = -1;
-                        }
-                    }
-                } else if (
-                    c === "[" ||
-                    c === "(" ||
-                    isEndCharacter(c) ||
-                    isSpecialStartCharacter(c) ||
-                    isSeparatorCharacter(c)
-                ) {
-                    break;
-                } else if (parserState.pos > 0) {
-                    throw ["Unexpected ", c, " after ", parserState.userQuery[parserState.pos - 1]];
-                } else {
-                    throw ["Unexpected ", c];
                 }
+                if (macroExclamation !== -1) {
+                    throw ["Cannot have associated items in macros"];
+                }
+            } else if (
+                c === "[" ||
+                c === "(" ||
+                isEndCharacter(c) ||
+                isSpecialStartCharacter(c) ||
+                isSeparatorCharacter(c)
+            ) {
+                break;
+            } else if (parserState.pos > 0) {
+                throw ["Unexpected ", c, " after ", parserState.userQuery[parserState.pos - 1],
+                    " (not a valid identifier)"];
+            } else {
+                throw ["Unexpected ", c, " (not a valid identifier)"];
             }
             parserState.pos += 1;
+            afterIdent = consumeIdent(parserState);
             end = parserState.pos;
         }
-        // if start == end - 1, we got the never type
-        if (foundExclamation !== -1 &&
-            foundExclamation !== start &&
-            isIdentCharacter(parserState.userQuery[foundExclamation - 1])
-        ) {
+        if (macroExclamation !== -1) {
             if (parserState.typeFilter === null) {
                 parserState.typeFilter = "macro";
             } else if (parserState.typeFilter !== "macro") {
@@ -693,7 +692,7 @@ function initSearch(rawSearchIndex) {
                     " both specified",
                 ];
             }
-            end = foundExclamation;
+            end = macroExclamation;
         }
         return end;
     }
@@ -1071,16 +1070,15 @@ function initSearch(rawSearchIndex) {
     function checkExtraTypeFilterCharacters(start, parserState) {
         const query = parserState.userQuery.slice(start, parserState.pos).trim();
 
-        for (const c in query) {
-            if (!isIdentCharacter(query[c])) {
-                throw [
-                    "Unexpected ",
-                    query[c],
-                    " in type filter (before ",
-                    ":",
-                    ")",
-                ];
-            }
+        const match = query.match(REGEX_INVALID_TYPE_FILTER);
+        if (match) {
+            throw [
+                "Unexpected ",
+                match[0],
+                " in type filter (before ",
+                ":",
+                ")",
+            ];
         }
     }
 
@@ -2127,7 +2125,7 @@ function initSearch(rawSearchIndex) {
             };
         }
 
-        function handleAliases(ret, query, filterCrates, currentCrate) {
+        async function handleAliases(ret, query, filterCrates, currentCrate) {
             const lowerQuery = query.toLowerCase();
             // We separate aliases and crate aliases because we want to have current crate
             // aliases to be before the others in the displayed results.
@@ -2163,6 +2161,15 @@ function initSearch(rawSearchIndex) {
             crateAliases.sort(sortFunc);
             aliases.sort(sortFunc);
 
+            const fetchDesc = alias => {
+                return searchIndexEmptyDesc.get(alias.crate).contains(alias.bitIndex) ?
+                    "" : searchState.loadDesc(alias);
+            };
+            const [crateDescs, descs] = await Promise.all([
+                Promise.all(crateAliases.map(fetchDesc)),
+                Promise.all(aliases.map(fetchDesc)),
+            ]);
+
             const pushFunc = alias => {
                 alias.alias = query;
                 const res = buildHrefAndPath(alias);
@@ -2176,7 +2183,13 @@ function initSearch(rawSearchIndex) {
                 }
             };
 
+            aliases.forEach((alias, i) => {
+                alias.desc = descs[i];
+            });
             aliases.forEach(pushFunc);
+            crateAliases.forEach((alias, i) => {
+                alias.desc = crateDescs[i];
+            });
             crateAliases.forEach(pushFunc);
         }
 
@@ -2538,7 +2551,8 @@ function initSearch(rawSearchIndex) {
             sorted_returned,
             sorted_others,
             parsedQuery);
-        handleAliases(ret, parsedQuery.original.replace(/"/g, ""), filterCrates, currentCrate);
+        await handleAliases(ret, parsedQuery.original.replace(/"/g, ""),
+            filterCrates, currentCrate);
         await Promise.all([ret.others, ret.returned, ret.in_args].map(async list => {
             const descs = await Promise.all(list.map(result => {
                 return searchIndexEmptyDesc.get(result.crate).contains(result.bitIndex) ?
diff --git a/src/tools/rustdoc-gui/.eslintrc.js b/src/tools/rustdoc-gui/.eslintrc.js
@@ -6,7 +6,7 @@ module.exports = {
     },
     "extends": "eslint:recommended",
     "parserOptions": {
-        "ecmaVersion": 2018,
+        "ecmaVersion": 2019,
         "sourceType": "module"
     },
     "rules": {
diff --git a/src/tools/rustdoc-js/.eslintrc.js b/src/tools/rustdoc-js/.eslintrc.js
@@ -6,7 +6,7 @@ module.exports = {
     },
     "extends": "eslint:recommended",
     "parserOptions": {
-        "ecmaVersion": 8,
+        "ecmaVersion": 2019,
         "sourceType": "module"
     },
     "rules": {
diff --git a/tests/rustdoc-js-std/parser-errors.js b/tests/rustdoc-js-std/parser-errors.js
@@ -24,7 +24,7 @@ const PARSED = [
         original: "-> *",
         returned: [],
         userQuery: "-> *",
-        error: "Unexpected `*` after ` `",
+        error: "Unexpected `*` after ` ` (not a valid identifier)",
     },
     {
         query: 'a<"P">',
@@ -204,16 +204,25 @@ const PARSED = [
         original: "_:",
         returned: [],
         userQuery: "_:",
-        error: "Unexpected `:` (expected path after type filter `_:`)",
+        error: "Unexpected `_` (not a valid identifier)",
     },
     {
-        query: "_:a",
+        query: "ab:",
         elems: [],
         foundElems: 0,
-        original: "_:a",
+        original: "ab:",
         returned: [],
-        userQuery: "_:a",
-        error: "Unknown type filter `_`",
+        userQuery: "ab:",
+        error: "Unexpected `:` (expected path after type filter `ab:`)",
+    },
+    {
+        query: "a:b",
+        elems: [],
+        foundElems: 0,
+        original: "a:b",
+        returned: [],
+        userQuery: "a:b",
+        error: "Unknown type filter `a`",
     },
     {
         query: "a-bb",
@@ -240,7 +249,7 @@ const PARSED = [
         original: "ab'",
         returned: [],
         userQuery: "ab'",
-        error: "Unexpected `'` after `b`",
+        error: "Unexpected `'` after `b` (not a valid identifier)",
     },
     {
         query: "a->",
diff --git a/tests/rustdoc-js/basic.js b/tests/rustdoc-js/basic.js
@@ -1,6 +1,6 @@
 const EXPECTED = {
     'query': 'Fo',
     'others': [
-        { 'path': 'basic', 'name': 'Foo' },
+        { 'path': 'basic', 'name': 'Foo', 'desc': 'Docs for Foo' },
     ],
 };
diff --git a/tests/rustdoc-js/doc-alias.js b/tests/rustdoc-js/doc-alias.js
diff --git a/tests/rustdoc-js/doc-alias.rs b/tests/rustdoc-js/doc-alias.rs
diff --git a/tests/rustdoc-js/non-english-identifier.js b/tests/rustdoc-js/non-english-identifier.js
diff --git a/tests/rustdoc-js/non-english-identifier.rs b/tests/rustdoc-js/non-english-identifier.rs
