Merge pull request #1 from edoardopirovano/resolve-clash

Handle multiple languages for an extension
aeisenberg · Jun 23, 2021 · a49cee0 · a49cee0
2 parents 86d4428 + 600ba0e
commit a49cee0
Show file tree

Hide file tree

Showing 4 changed files with 64 additions and 4 deletions.
diff --git a/src/languages.ts b/src/languages.ts
@@ -11,6 +11,17 @@ interface ExtensionsTypes {
 
 export interface DetectorOptions {}
 
+/**
+ * The extension map can contain multiple languages with the same extension,
+ * but we only want a single one. For the moment, these clashes are resolved
+ * by the simple heuristic below listing high-priority languages. We may want
+ * to consider smarter heuristics to correctly identify languages in cases
+ * where the extension is ambiguous. The ordering of the list matters and
+ * languages earlier on will get a higher priority when resolving clashes.
+ */
+const importantLanguages = ["javascript", "typescript", "ruby", "python", "java", "c", "c++", "c#", "rust", "scala", "perl", "go"];
+
+
 /**
  * detecte program language through file extension
  *
@@ -39,7 +50,20 @@ export class Languages {
       const languageMode = languageMap[language];
       const languageExtensions = (languageMode && languageMode.extensions) || [];
       languageExtensions.forEach((extension: string) => {
-        extensions[extension.toLowerCase()] = language.toLowerCase();
+        const lowerCaseExtension = extension.toLowerCase();
+        const lowerCaseLanguage = language.toLowerCase()
+        if (!extensions[lowerCaseExtension]) {
+          extensions[lowerCaseExtension] = lowerCaseLanguage;
+        } else {
+          const currentLanguagePriority = importantLanguages.indexOf(extensions[lowerCaseExtension]);
+          if (currentLanguagePriority === -1) {
+            extensions[lowerCaseExtension] = lowerCaseLanguage;
+          } else {
+            const otherPriority = importantLanguages.indexOf(lowerCaseLanguage);
+            if (otherPriority !== -1 && otherPriority < currentLanguagePriority)
+              extensions[lowerCaseExtension] = lowerCaseLanguage;
+          }
+        }
       });
     });
 

diff --git a/test/data2/x.cs b/test/data2/x.cs
@@ -0,0 +1,16 @@
+using System;
+
+// A comment
+/*
+ * A multi-line comment
+ */
+namespace HelloWorldApp {
+    class Geeks {
+        static void Main(string[] args) {
+            if (true) {
+                Console.WriteLine("Hello World!");
+            }
+        }
+    }
+}
+
diff --git a/test/directory.test.ts b/test/directory.test.ts
@@ -9,17 +9,24 @@ describe('Directory', () => {
       cwd: __dirname + '/data2'
     }, {
       files: [
+        'data2/x.cs',
         'data2/x.html',
         'data2/x.js',
         'data2/x.py',
         'data2/x.rb',
       ],
       info: {
-        code: 23,
-        comment: 23,
-        total: 61,
+        code: 33,
+        comment: 27,
+        total: 78,
       },
       languages: {
+        "c#": {
+          code: 10,
+          comment: 4,
+          sum: 1,
+          total: 17,
+        },
         html: {
           code: 4,
           comment: 4,

diff --git a/test/file.test.ts b/test/file.test.ts
@@ -56,6 +56,19 @@ describe('File', () => {
     });
   });
 
+  it('should calculate info for a csharp file', async () => {
+    await doTest('x.cs', {
+      languages: 'c#',
+      lines: {
+        code: 10,
+        comment: 4,
+        total: 17
+      },
+      name: 'x.cs',
+      size: 253
+    });
+  });
+
   async function doTest(fileName: string, expectedFileInfo: FileInfo) {
     const fullPath = slash(path.join(__dirname, `/data2/${fileName}`));
     const actualFileInfo = await new LocFile(fullPath).getFileInfo();