e2e: read every page when looking for a resource in a list

[NekoPunch (orangeCatDeveloper)]

Fixes #1126

TestActorSnapshotLifecycle reports snapshot %q missing from ListActorSnapshots on a cluster that has been up a while, even though the snapshot exists — GetActorSnapshot succeeds on it one line earlier.

The test called ListActorSnapshots once and treated the first page as the whole list. Lists are name-ordered, so a snapshot created just now lands wherever its name sorts: on the cluster where this reproduced it was 1408th of 1518, against a page size of 1000. createActor read ListActors the same way.

Both now drain the pages through one generic helper. Nothing about the product changes; the assertion just looks at the whole list it always meant to.

CI does not hit this — every run gets a fresh cluster and the demo atespace stays well under one page — so the paging path will not be exercised there either. It bites long-lived and shared clusters, and it cost time during the #1050 investigation by looking like a real snapshot loss.

Measured on the cluster where it reproduced, and re-run locally after a rebase onto current main:

before:  15/35 runs failed
after:    0/35 runs failed
Fisher one-sided p = 4.5e-6

$ bash hack/run-e2e-kind.sh -v -args --no-color --test.run TestActorSnapshotLifecycle
--- PASS: TestActorSnapshotLifecycle (5.53s)
ok  github.com/agent-substrate/substrate/internal/e2e/suites/demo	6.184s