fix: ignore the gitignored files in the scan to improve performance

cspell-ignore-words.txt

@@ -4,4 +4,7 @@ cacheable
 crap
 flagger
 flaggy
+globify
+Posixified
+posixify
 shit

packages/cspell-gitignore/package.json

@@ -50,7 +50,8 @@
   },
   "dependencies": {
     "cspell-glob": "workspace:*",
-    "find-up": "^5.0.0"
+    "find-up": "^5.0.0",
+    "globify-gitignore": "^0.2.1"
   },
   "devDependencies": {
     "@types/node": "^18.11.18",

packages/cspell-gitignore/src/GitIgnore.ts

@@ -1,3 +1,4 @@
+import { globifyGitIgnoreFile, posixifyPathNormalized } from 'globify-gitignore';
 import * as path from 'path';
 import { contains } from '.';
 import { GitIgnoreHierarchy, IsIgnoredExResult, loadGitIgnore } from './GitIgnoreFile';
@@ -51,6 +52,26 @@ export class GitIgnore {
         this.resolvedGitIgnoreHierarchies.set(directory, found);
         return find;
     }
+
+    async getGlobs(root: string) {
+        const rootPosixified = posixifyPathNormalized(root);
+        const globs = await globifyGitIgnoreFile(rootPosixified);
+
+        // globify-gitignore is compatible with fast-glob, but to make it work with glob,
+        // the patterns need to be separated and normalized.
+        const rootPosixifiedSlashed = `${rootPosixified}/`;
+        const ignored = [];
+        const included = [];
+        for (const g of globs) {
+            if (g.startsWith('!')) {
+                ignored.push(g.slice(1).replace(rootPosixifiedSlashed, ''));
+            } else {
+                included.push(g.replace(rootPosixifiedSlashed, ''));
+            }
+        }
+        return [ignored, included];
+    }
+
     filterOutIgnored(files: string[]): Promise<string[]>;
     filterOutIgnored(files: Iterable<string>): Promise<string[]>;
     filterOutIgnored(files: AsyncIterable<string>): AsyncIterable<string>;

packages/cspell/src/lint/lint.ts

@@ -380,6 +380,24 @@ async function determineFilesToCheck(
             globOptions.dot = enableGlobDot;
         }
 
+        if (gitIgnore) {
+            // Add the gitignore file in the root to the globs to improve performance.
+            const [gitIgnoredGlobs, gitIncludedGlobs] = await gitIgnore.getGlobs(root);
+
+            allGlobs.push(...gitIncludedGlobs);
+            fileGlobs.push(...gitIncludedGlobs);
+
+            let globOptionsIgnore: string[] = [];
+            if (Array.isArray(globOptions.ignore)) {
+                globOptionsIgnore = globOptions.ignore;
+            } else if (typeof globOptions.ignore === 'string') {
+                globOptionsIgnore = [globOptions.ignore];
+            }
+
+            globOptionsIgnore.push(...gitIgnoredGlobs);
+            globOptions.ignore = globOptionsIgnore;
+        }
+
         const filterFiles = opFilter(filterFilesFn(globMatcher));
         const foundFiles = await (hasFileLists
             ? useFileLists(fileLists, allGlobs, root, enableGlobDot)