Refactor file parsing

oxc-project · May 5, 2024 · 0dedfc3 · 0dedfc3
1 parent 05ca458
commit 0dedfc3
Showing 1 changed file with 79 additions and 56 deletions.
diff --git a/crates/oxc_traverse/scripts/lib/parse.mjs b/crates/oxc_traverse/scripts/lib/parse.mjs
@@ -3,69 +3,92 @@ import {join as pathJoin} from 'path';
 import {fileURLToPath} from 'url';
 import assert from 'assert';
 
+const FILENAMES = ['js.rs', 'jsx.rs', 'literal.rs', 'ts.rs'];
+
+/**
+ * Parse type defs from Rust files.
+ */
 export default async function getTypesFromCode() {
     const codeDirPath = pathJoin(fileURLToPath(import.meta.url), '../../../../oxc_ast/src/ast/');
-    const filenames = ['js.rs', 'jsx.rs', 'literal.rs', 'ts.rs'];
 
-    // Parse type defs from Rust files
     const types = Object.create(null);
-    for (const filename of filenames) {
-        const code = await readFile(`${codeDirPath}${filename}`, 'utf8'),
-            lines = code.split(/\r?\n/);
-        for (let i = 0; i < lines.length; i++) {
-            if (lines[i] === '#[visited_node]') {
-                let match;
-                while (true) {
-                    match = lines[++i].match(/^pub (enum|struct) (.+?)(<'a>)? \{/);
-                    if (match) break;
-                }
-                const [, kind, name, lifetimeStr] = match,
-                    hasLifetime = !!lifetimeStr;
-                const itemLines = [];
-                while (true) {
-                    const line = lines[++i].replace(/\/\/.*$/, '').replace(/\s+/g, ' ').trim();
-                    if (line === '}') break;
-                    if (line !== '') itemLines.push(line);
-                }
+    for (const filename of FILENAMES) {
+        const code = await readFile(`${codeDirPath}${filename}`, 'utf8');
+        parseFile(code, filename, types);
+    }
+    return types;
+}
+
+function parseFile(code, filename, types) {
+    const lines = code.split(/\r?\n/);
+    for (let lineIndex = 0; lineIndex < lines.length; lineIndex++) {
+        if (lines[lineIndex] !== '#[visited_node]') continue;
+
+        let match;
+        while (true) {
+            match = lines[++lineIndex].match(/^pub (enum|struct) (.+?)(<'a>)? \{/);
+            if (match) break;
+        }
+        const [, kind, name, lifetimeStr] = match,
+            hasLifetime = !!lifetimeStr,
+            startLineIndex = lineIndex;
+
+        const itemLines = [];
+        while (true) {
+            const line = lines[++lineIndex].replace(/\/\/.*$/, '').replace(/\s+/g, ' ').trim();
+            if (line === '}') break;
+            if (line !== '') itemLines.push(line);
+        }
 
-                if (kind === 'enum') {
-                    const variants = [],
-                        inherits = [];
-                    for (const line of itemLines) {
-                        const match = line.match(/^(.+?)\((.+?)\)(?: ?= ?(\d+))?,$/);
-                        if (match) {
-                            let [, name, type, discriminant] = match;
-                            type = type.replace(/<'a>/g, '').replace(/<'a,\s*/g, '<');
-                            discriminant = discriminant ? +discriminant : null;
-                            variants.push({name, type, discriminant});
-                        } else {
-                            const match2 = line.match(/^@inherit ([A-Za-z]+)$/);
-                            assert(match2, `Cannot parse line ${i} in '${filename}' as enum variant: '${line}'`);
-                            inherits.push(match2[1]);
-                        }
-                    }
-                    types[name] = {kind: 'enum', name, hasLifetime, variants, inherits};
-                } else {
-                    const fields = [];
-                    for (let i = 0; i < itemLines.length; i++) {
-                        const line = itemLines[i];
-                        if (line.startsWith('#[')) {
-                            while (!itemLines[i].endsWith(']')) {
-                                i++;
-                            }
-                            continue;
-                        }
+        if (kind === 'enum') {
+            types[name] = parseEnum(name, hasLifetime, itemLines, filename, startLineIndex);
+        } else {
+            types[name] = parseStruct(name, hasLifetime, itemLines, filename, startLineIndex);
+        }
+    }
+}
 
-                        const match = line.match(/^pub ((?:r#)?([a-z_]+)): (.+),(?: ?\/\/.+)?$/);
-                        assert(match, `Cannot parse line ${i} in '${filename}' as struct field: '${line}'`);
-                        const [, rawName, name, rawType] = match,
-                            type = rawType.replace(/<'a>/g, '').replace(/<'a, ?/g, '<');
-                        fields.push({name, type, rawName, rawType});
-                    }
-                    types[name] = {kind: 'struct', name, hasLifetime, fields};
-                }
+function parseEnum(name, hasLifetime, lines, filename, startLineIndex) {
+    const variants = [],
+        inherits = [];
+    for (const [lineIndex, line] of lines.entries()) {
+        const match = line.match(/^(.+?)\((.+?)\)(?: ?= ?(\d+))?,$/);
+        if (match) {
+            const [, name, rawType, discriminantStr] = match,
+                type = rawType.replace(/<'a>/g, '').replace(/<'a,\s*/g, '<'),
+                discriminant = discriminantStr ? +discriminantStr : null;
+            variants.push({name, type, rawType, discriminant});
+        } else {
+            const match2 = line.match(/^@inherit ([A-Za-z]+)$/);
+            assert(
+                match2,
+                `Cannot parse line ${startLineIndex + lineIndex} in '${filename}' as enum variant: '${line}'`
+            );
+            inherits.push(match2[1]);
+        }
+    }
+    return {kind: 'enum', name, hasLifetime, variants, inherits};
+}
+
+function parseStruct(name, hasLifetime, lines, filename, startLineIndex) {
+    const fields = [];
+    for (let i = 0; i < lines.length; i++) {
+        const line = lines[i];
+        if (line.startsWith('#[')) {
+            while (!lines[i].endsWith(']')) {
+                i++;
             }
+            continue;
         }
+
+        const match = line.match(/^pub ((?:r#)?([a-z_]+)): (.+),$/);
+        assert(
+            match,
+            `Cannot parse line ${startLineIndex + i} in '${filename}' as struct field: '${line}'`
+        );
+        const [, rawName, name, rawType] = match,
+            type = rawType.replace(/<'a>/g, '').replace(/<'a, ?/g, '<');
+        fields.push({name, type, rawName, rawType});
     }
-    return types;
+    return {kind: 'struct', name, hasLifetime, fields};
 }