- Ya
- 14 Agosti 2026, 04:43 UTC
- Mwandishi
- kamo
- Ahadi ya
- 1a050fb
Two round trips still sat in front of a rendered inbox, and by this point they were the only thing left there — the folder tree and the first page already arrive early, so the member was watching a boot skeleton on top of mail that had already loaded. The mailbox probe waited for the provider lookup to answer before it would start. Nothing about the probe depends on that answer; only the reason reported when both fail does, and that ordering is preserved by checking the provider first when the reason is chosen. So they now run together, and an org with no provider still reports PROVIDER_NOT_CONFIGURED without waiting for a probe that could not change it. The cost is one wasted request for an org that has not configured email — an org whose members are not on this page. The second is the skeleton itself. When this tab already holds a snapshot of the member's own mail, that is evidence, not a guess, that the mailbox worked moments ago; blocking on two lookups to confirm what the tab watched be true means covering real messages with a placeholder. The snapshot is read once at mount, so the skeleton cannot vanish mid-load, and if availability does resolve to unavailable the explanatory panel still takes over. This decoders to block on the answer; it does not decide it.