Address only the longer of two @names in a group, and not a Bot named by an email address

CHANGELOG.md

@@ -8,6 +8,13 @@ Newest first. `Unreleased` is what is on `main` and not yet tagged.
 
 ## Unreleased
 
+### `@Ops Lead` in a group addresses Ops Lead, not Ops as well
+
+In a group conversation, a reply naming `@Ops Lead` also addressed a Bot called Ops, because the
+shorter name matched at the same `@`, so both answered. An email address addressed a Bot by its
+domain: `jo@sam.com` reached a Bot called Sam. Where two names start at the same `@`, only the
+longer one is now addressed, and an `@` straight after a letter or digit is not a mention.
+
 ### A webhook with many long top-level fields no longer stops the server
 
 A trigger's event is cut to 32 KiB before it is recorded, keeping each top-level text field up to

server/src/channels/group.ts

@@ -460,20 +460,38 @@ export function groupTurnMessage(
  * The peers a reply addresses, in the order the reply names them.
  *
  * `@Name` with the whole name, case-insensitive, and not followed by more of a word, so a Bot called
- * "Ops" is not addressed by "@Opsgenie". Never the speaker itself.
+ * "Ops" is not addressed by "@Opsgenie". Not preceded by one either, so "jo@sam.com" does not
+ * address "Sam". Where two names start at the same `@`, the longer one is meant: "@Ops Lead"
+ * addresses "Ops Lead" and not "Ops". Never the speaker itself.
  */
 export function mentionedPeers(
   reply: string,
   roster: readonly GroupBot[],
   speakerId: string,
 ): GroupBot[] {
-  return roster
-    .flatMap((bot) => {
-      if (bot.id === speakerId || !bot.name.trim()) return [];
-      const name = bot.name.trim().replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
-      const at = reply.search(new RegExp(`@${name}(?![\\p{L}\\p{N}_])`, "iu"));
-      return at < 0 ? [] : [{ bot, at }];
-    })
+  // The speaker is matched too, so that its own longer name still hides a shorter peer's.
+  const hits = roster.flatMap((bot) => {
+    if (!bot.name.trim()) return [];
+    const name = bot.name.trim().replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
+    const pattern = new RegExp(
+      `(?<![\\p{L}\\p{N}_])@${name}(?![\\p{L}\\p{N}_])`,
+      "giu",
+    );
+    return Array.from(reply.matchAll(pattern), (match) => ({
+      bot,
+      at: match.index,
+      length: match[0].length,
+    }));
+  });
+  const first = new Map<string, { bot: GroupBot; at: number }>();
+  for (const hit of hits) {
+    if (hit.bot.id === speakerId || first.has(hit.bot.id)) continue;
+    const shadowed = hits.some(
+      (other) => other.at === hit.at && other.length > hit.length,
+    );
+    if (!shadowed) first.set(hit.bot.id, hit);
+  }
+  return [...first.values()]
     .sort((left, right) => left.at - right.at)
     .map(({ bot }) => bot);
 }

server/tests/group-conversations.test.ts

@@ -60,3 +60,20 @@ test("a reply addresses peers by their whole @name, never itself", () => {
     [],
   );
 });
+
+test("a longer name at the same @ is the one addressed, and an email address is not a mention", () => {
+  const roster = [
+    { id: "ada", name: "Ada" },
+    { id: "ops", name: "Ops" },
+    { id: "lead", name: "Ops Lead" },
+    { id: "sam", name: "Sam" },
+  ];
+  const ids = (reply: string, speaker = "ada") =>
+    mentionedPeers(reply, roster, speaker).map((bot) => bot.id);
+  expect(ids("@Ops Lead please look")).toEqual(["lead"]);
+  expect(ids("@Ops Lead first, then @Ops")).toEqual(["lead", "ops"]);
+  // The speaker's own name still counts as the longer one.
+  expect(ids("I, @Ops Lead, will do it", "lead")).toEqual([]);
+  expect(ids("write to jo@sam.com")).toEqual([]);
+  expect(ids("(@Sam) can you check")).toEqual(["sam"]);
+});