fix: fall back to file extension for usages language with custom match patterns

src/commands/usages/usages.test.ts

@@ -1,3 +1,6 @@
+import fs from 'fs'
+import os from 'os'
+import path from 'path'
 import { test } from '@oclif/test'
 import { expect } from 'vitest'
 import { BASE_URL } from '../../api/common'
@@ -135,4 +138,31 @@ describe('usages', () => {
                 expect(ctx.stdout).toMatchSnapshot()
             },
         )
+
+    const htmlOutputDir = fs.mkdtempSync(path.join(os.tmpdir(), 'dvc-usages-'))
+    const htmlOutputPath = path.join(htmlOutputDir, 'usages.json')
+    afterAll(() => fs.rmSync(htmlOutputDir, { recursive: true, force: true }))
+
+    test.stdout()
+        .command([
+            'usages',
+            '--include',
+            'test-utils/fixtures/usages/sample.html',
+            '--match-pattern',
+            'html=isFeatureEnabled\\("([^"]*)"\\)',
+            '--format',
+            'json',
+            '--output',
+            htmlOutputPath,
+        ])
+        .it(
+            'sets language from the file extension for custom match patterns',
+            () => {
+                const [usage] = JSON.parse(
+                    fs.readFileSync(htmlOutputPath, 'utf-8'),
+                )
+                expect(usage.key).toEqual('html-flag')
+                expect(usage.references[0].language).toEqual('html')
+            },
+        )
 })

src/utils/parsers/BaseParser.ts

@@ -6,7 +6,7 @@ import {
     VariableUsageMatch,
 } from './types'
 import * as usage from '../../commands/usages/types'
-import { LANGUAGE_MAP } from '../parsers'
+import { getLanguageForExtension } from '../parsers'
 
 type MatchWithRange = {
     // Variable name
@@ -395,7 +395,7 @@ export abstract class BaseParser {
                 },
                 fileName: file.name,
                 content: bufferedContent,
-                language: LANGUAGE_MAP[fileExtension],
+                language: getLanguageForExtension(fileExtension),
                 ...(match.isUnknown ? { isUnknown: true } : {}),
             })
         })

src/utils/parsers/index.ts

@@ -55,3 +55,16 @@ export const LANGUAGE_MAP: Record<string, string> = {
     php: 'php',
     dart: 'dart',
 }
+
+const SAFE_EXTENSION_PATTERN = /^[a-z0-9]{1,16}$/i
+
+/**
+ * Map a file extension to a language, falling back to the extension itself
+ * for custom match patterns. Unsafe extensions fall back to 'unknown'.
+ */
+export const getLanguageForExtension = (extension: string): string => {
+    if (Object.hasOwn(LANGUAGE_MAP, extension)) return LANGUAGE_MAP[extension]
+    return SAFE_EXTENSION_PATTERN.test(extension)
+        ? extension.toLowerCase()
+        : 'unknown'
+}

src/utils/parsers/language.test.ts

@@ -0,0 +1,23 @@
+import { expect } from 'vitest'
+import { getLanguageForExtension } from '.'
+
+describe('getLanguageForExtension', () => {
+    it('uses the mapped language for built-in extensions', () => {
+        expect(getLanguageForExtension('ts')).toEqual('typescript')
+        expect(getLanguageForExtension('kt')).toEqual('kotlin')
+    })
+
+    it('falls back to the extension for unmapped alphanumeric extensions', () => {
+        expect(getLanguageForExtension('html')).toEqual('html')
+        expect(getLanguageForExtension('Vue')).toEqual('vue')
+        expect(getLanguageForExtension('constructor')).toEqual('constructor')
+    })
+
+    it('falls back to unknown for unsafe or empty extensions', () => {
+        expect(getLanguageForExtension('<img src=x onerror=alert(1)>')).toEqual(
+            'unknown',
+        )
+        expect(getLanguageForExtension('a'.repeat(17))).toEqual('unknown')
+        expect(getLanguageForExtension('')).toEqual('unknown')
+    })
+})

test-utils/fixtures/usages/sample.html

@@ -0,0 +1,9 @@
+<html>
+    <body>
+        <script>
+            if (isFeatureEnabled("html-flag")) {
+                showBanner()
+            }
+        </script>
+    </body>
+</html>