fix: inconsistencies

Signed-off-by: Umberto Sgueglia <usgueglia@contractor.linuxfoundation.org>

Author: ulemons

services/apps/packages_worker/src/npm/backfillRepositoryUrl.ts

@@ -1,4 +1,3 @@
-import { deleteMavenPackageRepoLinks } from '@crowd/data-access-layer'
 import {
   NpmRepoUrlRow,
   getOrCreateRepoByUrl,
@@ -103,7 +102,8 @@ export async function backfillNpmRepositoryUrls(
 
     const updates: { id: string; repositoryUrl: string | null }[] = []
     // Rows that had a link and now change to a different one — their stale 'declared' link is pruned.
-    const pruneTargets: number[] = []
+    // Kept as strings (packages.id is bigint) — Number() coercion would lose precision above 2^53.
+    const pruneTargets: string[] = []
     // Rows that gained/changed a canonical URL — their repo link is (re)written after the update.
     const linkTargets: { id: string; repositoryUrl: string; host: string }[] = []
     for (const row of rows) {
@@ -120,7 +120,7 @@ export async function backfillNpmRepositoryUrls(
       else totals.rewritten++
       updates.push({ id: row.id, repositoryUrl: desired })
       if (row.repositoryUrl !== null && desired !== row.repositoryUrl) {
-        pruneTargets.push(Number(row.id))
+        pruneTargets.push(row.id)
       }
       if (canonical)
         linkTargets.push({ id: row.id, repositoryUrl: canonical.url, host: canonical.host })
@@ -135,7 +135,13 @@ export async function backfillNpmRepositoryUrls(
       await withDeadlockRetry(() =>
         qx.tx(async (t: QueryExecutor) => {
           await updateNpmRepositoryUrls(t, updates)
-          if (pruneTargets.length > 0) await deleteMavenPackageRepoLinks(t, pruneTargets)
+          if (pruneTargets.length > 0) {
+            await t.result(
+              `DELETE FROM package_repos
+               WHERE package_id = ANY($(packageIds)::bigint[]) AND source = 'declared'`,
+              { packageIds: pruneTargets },
+            )
+          }
           for (const target of linkTargets) {
             const { id: repoId } = await getOrCreateRepoByUrl(t, target.repositoryUrl, target.host)
             await upsertPackageRepo(t, target.id, repoId, 'declared', 0.8)

services/apps/packages_worker/src/utils/__tests__/canonicalizeRepoUrl.test.ts

@@ -28,6 +28,7 @@ describe('canonicalizeRepoUrl', () => {
       'https://github.com/1inch/limit-order-protocol-utils',
       'github',
     ],
+    ['ssh://git@github.com:2222/foo/bar.git', 'https://github.com/foo/bar', 'github'],
   ])('canonicalizes %s', (input, expectedUrl, expectedHost) => {
     expect(canonicalizeRepoUrl(input)).toEqual({ url: expectedUrl, host: expectedHost })
   })

services/apps/packages_worker/src/utils/canonicalizeRepoUrl.ts

@@ -56,9 +56,12 @@ export function canonicalizeRepoUrl(raw: string): CanonicalRepo | null {
 
   // ssh:// with an scp-style `host:path` (colon instead of slash) is not valid URL
   // syntax — the part after `:` looks like a port to the URL parser and throws.
-  // Rewrite it before the generic ssh://git@host/path case below.
+  // Rewrite it before the generic ssh://git@host/path case below. A numeric-only
+  // segment before the next `/` is a real port (e.g. `ssh://git@host:2222/owner/repo`),
+  // not an scp-style owner — leave those for the generic case, which URL parses fine.
   const sshScp = s.match(/^ssh:\/\/git@([^/:]+):(.+)$/)
-  if (sshScp) {
+  const sshScpIsPort = sshScp ? /^\d+$/.test(sshScp[2].split('/')[0]) : false
+  if (sshScp && !sshScpIsPort) {
     s = `https://${sshScp[1]}/${sshScp[2]}`
   } else {
     s = s.replace(/^ssh:\/\/git@([^/]+)\//, 'https://$1/')