Backport SW API changes without service worker #2137
Conversation
corrideat
force-pushed
the
feat/sw-support
branch
from
ae6c78a to
eedb835
Compare
corrideat
force-pushed
the
feat/sw-support
branch
2 times, most recently
from
b1b751c to
5fbb5c2
Compare
Passing run #2603 ↗︎
Details:
Review all test suite changes for PR #2137 ↗︎ |
|||||||||||||||
corrideat
force-pushed
the
feat/sw-support
branch
3 times, most recently
from
f69ca19 to
a0a48d4
Compare
| @@ -27,7 +27,7 @@ export default ({ | |||
| methods: { | |||
| submit () { | |||
| this.$refs.modal.close() | |||
| this.$router.push({ query: this.$route.query, path: this.$route.query.next ?? '/dashboard' }).catch(() => {}) | |||
| this.$router.push({ query: this.$route.query, path: this.$route.query.next ?? '/' }).catch(() => {}) | |||
There was a problem hiding this comment.
I think both should work, but / always has the same number of redirects. That's why I changed it here.
corrideat
force-pushed
the
feat/sw-support
branch
2 times, most recently
from
0909f11 to
9679412
Compare
corrideat
force-pushed
the
feat/sw-support
branch
from
9679412 to
c4fd7a1
Compare
| sbp('okTurtles.events/emit', DELETED_CHATROOM, { groupContractID: contractID, chatRoomID: data.chatRoomID }) | ||
| const { identityContractID } = sbp('state/vuex/state').loggedIn |
There was a problem hiding this comment.
The DELETED_CHATROOM listener is defined to also take an identityContractID:
{ identityContractID, groupContractID, chatRoomID }So should this include identityContractID: innerSigningContractID here?
There was a problem hiding this comment.
It's optional, but, yes, I think it could be added.
There was a problem hiding this comment.
It should not be used when deleting a chatroom. If it's set, it'll skip the handler unless the user deleting it is the same as the currently logged in user.
There was a problem hiding this comment.
Then could you please add a comment saying that the identityContractID is intentionally left out?
|
So the latest Cypress failure is related to the one of the changes made today, and it could be related to the changes with removing the event queue for login (we could try simply adding a However the tests did pass with Or something else. |
This had to do with |
|
|
||
| return message | ||
| return message.contractID() |
There was a problem hiding this comment.
To avoid having to copy GIMessage across boundaries (a potentially expensive operation)
| 'gi.actions/identity/login': (...params) => { | ||
| return sbp('okTurtles.eventQueue/queueEvent', 'ACTIONS-LOGIN', ['gi.actions/identity/_private/login', ...params]) | ||
| }, | ||
| 'gi.actions/identity/logout': (...params) => { | ||
| return sbp('okTurtles.eventQueue/queueEvent', 'ACTIONS-LOGIN', ['gi.actions/identity/_private/logout', ...params]) | ||
| }, |
There was a problem hiding this comment.
Can we replace these with anonymous functions pushed onto the event queue?
There was a problem hiding this comment.
I think it's better this way for logging, as otherwise you don't know when the anonymous function is executed
corrideat
force-pushed
the
feat/sw-support
branch
from
1a57715 to
8653614
Compare
No description provided.