fix: Make string truncation unicode aware (#810)

* fix: Make string truncation unicode aware

---------

Co-authored-by: Burak Yigit Kaya <[email protected]>

.github/workflows/ci.yml

@@ -56,12 +56,12 @@ jobs:
 
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
 
       - name: Setup node
-        uses: actions/setup-node@v3
+        uses: actions/setup-node@v4
         with:
-          node-version: '14'
+          node-version: '18'
 
       - name: yarn install
         run: yarn install --immutable
@@ -71,7 +71,7 @@ jobs:
       - name: test
         run: yarn test:ci --coverage
       - name: Codecov
-        uses: codecov/codecov-action@v3
+        uses: codecov/codecov-action@v4
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
       - name: eslint
@@ -112,9 +112,9 @@ jobs:
           fetch-depth: 0
 
       - name: Setup node
-        uses: actions/setup-node@v3
+        uses: actions/setup-node@v4
         with:
-          node-version: '14'
+          node-version: '18'
 
       - name: yarn install
         run: yarn install --immutable

jest.config.ts

@@ -1,5 +1,11 @@
 // eslint-disable-next-line
 module.exports = {
+  // TODO: Remove the snapshotFormat override below and regen snapshots
+  //       This is a remnant from the v29 upgrade
+  snapshotFormat: {
+    escapeString: true,
+    printBasicPrototype: true,
+  },
   collectCoverageFrom: [
     '**/*.ts',
     '!test/**',

package.json

@@ -71,11 +71,11 @@
     "pg": "^8.5.1",
     "source-map-support": "^0.5.21",
     "tar": "^6.2.1",
-    "typescript": "^4.5.0"
+    "typescript": "^4.7.0"
   },
   "devDependencies": {
     "@sentry/cli": "^2.20.1",
-    "@types/jest": "^27.0.3",
+    "@types/jest": "^29.0.0",
     "@typescript-eslint/eslint-plugin": "^5.4.0",
     "@typescript-eslint/parser": "^5.4.0",
     "dotenv": "^16.0.3",
@@ -84,17 +84,17 @@
     "eslint-plugin-prettier": "^4.0.0",
     "eslint-plugin-simple-import-sort": "^7.0.0",
     "husky": ">=4.3.0",
-    "jest": "27.3.1",
-    "jest-circus": "^27.3.1",
+    "jest": "^29.6.0",
+    "jest-circus": "^29.6.0",
     "lint-staged": ">=10.5.4",
     "lodash.clonedeep": "^4.5.0",
     "lodash.merge": "^4.6.2",
     "markdownlint-cli": "^0.27.1",
     "nodemon": "^2.0.7",
     "pino-pretty": "^4.5.0",
     "prettier": "^2.4.1",
-    "ts-jest": "^27.0.7",
-    "ts-node": "^10.4.0"
+    "ts-jest": "^29.1.5",
+    "ts-node": "^10.8.1"
   },
   "resolutions": {
     "set-value": ">=4.0.1"

src/webhooks/pubsub/slackScores.test.ts

@@ -429,7 +429,7 @@ describe('slackScores tests', function () {
 ├────────────────────────────────────────────────────────────────────┤
 | <https://github.com/routing-repo/discussions/001|Discussion 1>                                      |              3 |
 | <https://github.com/routing-repo/discussions/002|Discussion 2>                                      |              2 |
-| <https://github.com/test-ttt-simple/discussions/003|Overflowing Discussion Title blahblahblahblahb...> |              1 |
+| <https://github.com/test-ttt-simple/discussions/003|Overflowing Discussion Title blahblahblahblahbla…> |              1 |
 └────────────────────────────────────────────────────────────────────┘
 ┌────────────────────────────────────────────────────────────────────┐
 | Most Active Issues this Week                      │     # comments |

src/webhooks/pubsub/slackScores.ts

@@ -226,19 +226,24 @@ export const sendGitHubActivityMetrics = async (
       },
     },
   ];
+
+  const segmentize = (text: string): string[] =>
+    Array.from(new Intl.Segmenter('en').segment(text)).map((s) => s.segment);
+
   const addSpaces = (entry: string, column: string) => {
+    const entryLength = segmentize(entry).length;
     if (column === 'discussion') {
       return (
         entry +
         ' '.repeat(
-          DISCUSSION_COLUMN_WIDTH + LESS_THAN_SIGN_LENGTH - entry.length
+          DISCUSSION_COLUMN_WIDTH + LESS_THAN_SIGN_LENGTH - entryLength
         )
       );
     } else if (column === 'firstColumnItem') {
-      return entry + ' '.repeat(DISCUSSION_COLUMN_WIDTH - entry.length);
+      return entry + ' '.repeat(DISCUSSION_COLUMN_WIDTH - entryLength);
     }
     return (
-      ' '.repeat(NUM_COMMENTS_COLUMN_WIDTH - entry.length - SPACE_LENGTH) +
+      ' '.repeat(NUM_COMMENTS_COLUMN_WIDTH - entryLength - SPACE_LENGTH) +
       entry +
       ' '
     );
@@ -266,11 +271,15 @@ export const sendGitHubActivityMetrics = async (
         );
         scoreBoard += `| ${usernameText}| ${numCommentsText}|\n`;
       } else {
+        // This is the correct char-by-char representation of a string that may have unicode chars
+        const segmentizedTitle = segmentize(itemInfo.title);
         // Append less than sign here, because trailing spaces are ignored for text within the <> blocks for a hyperlink
         const truncatedDiscussionTitle =
-          itemInfo.title.length <= DISCUSSION_COLUMN_WIDTH
+          segmentizedTitle.length <= DISCUSSION_COLUMN_WIDTH
             ? itemInfo.title + '>'
-            : `${itemInfo.title.slice(0, DISCUSSION_COLUMN_WIDTH - 4)}...> `;
+            : `${segmentizedTitle
+                .slice(0, DISCUSSION_COLUMN_WIDTH - 2)
+                .join('')}…> `;
         const truncatedDiscussionTitleWithSpaces = addSpaces(
           // Remove all instances of ` char from string
           truncatedDiscussionTitle.replace(/`/g, ''),

src/webhooks/pubsub/stalebot.ts

@@ -158,15 +158,16 @@ export const triggerStaleBot = async (org: GitHubOrg, now: moment.Moment) => {
           (pullRequest) =>
             !pullRequest.labels.some(
               (label) =>
-                label === WORK_IN_PROGRESS_LABEL ||
+                (label as string) === WORK_IN_PROGRESS_LABEL ||
                 label.name === WORK_IN_PROGRESS_LABEL
             )
         );
         // Unfortunately, octokit doesn't allow us to filter by labels when
         // sending a GET request for pull requests, so we need to do this manually.
         const stalePullRequests = pullRequestsToCheck.filter((pullRequest) =>
           pullRequest.labels.some(
-            (label) => label === STALE_LABEL || label.name === STALE_LABEL
+            (label) =>
+              (label as string) === STALE_LABEL || label.name === STALE_LABEL
           )
         );
         const activePullRequests = pullRequestsToCheck.filter(

test/jest.tsconfig.json

@@ -1,6 +1,8 @@
 {
   "extends": "../tsconfig.json",
   "compilerOptions": {
-    "types": ["jest"]
+    "types": [
+      "jest"
+    ]
   }
 }

tsconfig.json

@@ -1,11 +1,11 @@
 {
   "compilerOptions": {
     "allowJs": false,
-    "lib": ["ES2020"],
+    "lib": ["ES2022"],
     "module": "CommonJS",
     "moduleResolution": "node",
     "allowSyntheticDefaultImports": true,
-    "target": "ES2020",
+    "target": "ES2022",
     "noImplicitReturns": true,
     "noFallthroughCasesInSwitch": true,
     "noUnusedLocals": true,
@@ -19,7 +19,9 @@
     "resolveJsonModule": true,
     "baseUrl": "./",
     "paths": {
-      "_comment": ["⚠️  Keep this in sync with jest.config.ts and package.json. ⚠️ "],
+      "_comment": [
+        "⚠️  Keep this in sync with jest.config.ts and package.json. ⚠️ "
+      ],
       "@/*": ["src/*"],
       "@api/*": ["src/api/*"],
       "@test/*": ["test/*"],