fix(issues): decode @mention entities without lockfile or new deps

- Drop entities package (CI blocks pnpm-lock.yaml on PRs; reset lockfile to master)
- Restore numeric + allowlisted named entity decoding in issues.ts
- Split Greptile mid-token & case into its own test with review comment

Made-with: Cursor
This commit is contained in:
amit221
2026-03-24 15:22:21 +02:00
parent 53f0988006
commit 2735ef1f4a
4 changed files with 7654 additions and 7542 deletions

15149
pnpm-lock.yaml generated

File diff suppressed because it is too large Load Diff

View File

@@ -64,7 +64,6 @@
"dotenv": "^17.0.1",
"drizzle-orm": "^0.38.4",
"embedded-postgres": "^18.1.0-beta.16",
"entities": "^8.0.0",
"express": "^5.1.0",
"hermes-paperclip-adapter": "0.1.1",
"jsdom": "^28.1.0",

View File

@@ -14,12 +14,16 @@ describe("normalizeAgentMentionToken", () => {
expect(normalizeAgentMentionToken("Baba ")).toBe("Baba");
});
it("decodes named entities mid-token so agent names can include &", () => {
// Greptile: entity mid-token (not only trailing) — must decode & to &, not delete the sequence.
it("decodes & in the middle of a mention token", () => {
expect(normalizeAgentMentionToken("Ba&ba")).toBe("Ba&ba");
});
it("decodes & so agent names with ampersands still match", () => {
expect(normalizeAgentMentionToken("M&M")).toBe("M&M");
});
it("decodes named entities mid-token (e.g. copyright) for full HTML named coverage", () => {
it("decodes additional named entities used in rich text (e.g. ©)", () => {
expect(normalizeAgentMentionToken("Agent©Name")).toBe("Agent©Name");
});

View File

@@ -21,7 +21,6 @@ import {
projects,
} from "@paperclipai/db";
import { extractAgentMentionIds, extractProjectMentionIds } from "@paperclipai/shared";
import { decodeHTMLStrict } from "entities";
import { conflict, notFound, unprocessable } from "../errors.js";
import {
defaultIssueExecutionWorkspaceSettingsForProject,
@@ -219,12 +218,39 @@ function unreadForUserCondition(companyId: string, userId: string) {
`;
}
/**
* Decodes HTML character references in a raw @mention capture (WHATWG HTML, strict semicolon form)
* so rich-text / UI-encoded bodies still match agent names.
*/
/** Named entities commonly emitted in saved issue bodies; unknown `&name;` sequences are left unchanged. */
const WELL_KNOWN_NAMED_HTML_ENTITIES: Readonly<Record<string, string>> = {
amp: "&",
apos: "'",
copy: "\u00A9",
gt: ">",
lt: "<",
nbsp: "\u00A0",
quot: '"',
ensp: "\u2002",
emsp: "\u2003",
thinsp: "\u2009",
};
function decodeNumericHtmlEntity(digits: string, radix: 16 | 10): string | null {
const n = Number.parseInt(digits, radix);
if (Number.isNaN(n) || n < 0 || n > 0x10ffff) return null;
try {
return String.fromCodePoint(n);
} catch {
return null;
}
}
/** Decodes HTML character references in a raw @mention capture so UI-encoded bodies match agent names. */
export function normalizeAgentMentionToken(raw: string): string {
return decodeHTMLStrict(raw).trim();
let s = raw.replace(/&#x([0-9a-fA-F]+);/gi, (full, hex: string) => decodeNumericHtmlEntity(hex, 16) ?? full);
s = s.replace(/&#([0-9]+);/g, (full, dec: string) => decodeNumericHtmlEntity(dec, 10) ?? full);
s = s.replace(/&([a-z][a-z0-9]*);/gi, (full, name: string) => {
const decoded = WELL_KNOWN_NAMED_HTML_ENTITIES[name.toLowerCase()];
return decoded !== undefined ? decoded : full;
});
return s.trim();
}
export function deriveIssueUserContext(