Updated the spec so that it reflects the behaviour implemented in ably-cocoa PR #1847 #181
Conversation
textile/features.textile
Outdated
| **** @(RSH3g1a)@ Transitions to @WaitingForDeregistration@. | ||
| *** @(RSH3g2)@ On event @Deregistered@: | ||
| **** @(RSH3g2a)@ Clears all local @DeviceDetails@. | ||
| **** @(RSH3g2a)@ Generates a new pair of device @id@ and @deviceSecret@ as described in "RSH3a2b":#RSH3a2b. Clears all other local @DeviceDetails@. |
There was a problem hiding this comment.
This only partially describes the ably-cocoa behaviour, doesn't it? In ably-cocoa, id and deviceSecret generation are completely decoupled from the CalledActivate event, meaning that RSH3a2b and RSH8b would also need to be changed if we wished for the spec to describe the behaviour of ably-cocoa.
Also, same point as #182 (comment) re spec point modification.
|
I'm not really familiar with the context here or the push activation state machine implemented in sdks, so going to untag myself and suggest @owenpearson instead |
SimonWoolf
requested review from
owenpearson
and removed request for
SimonWoolf
maratal
force-pushed
the
PCD2-RSH3g2a-never-nil-fix
branch
from
cd432e1 to
e99fb27
Compare
|
@maratal are you intending to update this one? |
Will take a closer look. |
maratal
force-pushed
the
PCD2-RSH3g2a-never-nil-fix
branch
from
e99fb27 to
05da9d7
Compare
@lawrence-forooghian finally fixed |
Updated the spec so that it reflects the behaviour implemented in ably/ably-cocoa#1847, where a new device ID is generated as soon as the previous ones are cleared, so that id is never null.
Originally posted by @lawrence-forooghian here
Closes #180