Rethink how send-tab falls back to the sync-based sendtab mechanism
Categories
(Firefox :: Firefox Accounts, defect, P3)
Tracking
()
Tracking | Status | |
---|---|---|
firefox79 | --- | fixed |
People
(Reporter: markh, Assigned: vladikoff)
References
Details
Attachments
(1 file)
If there's some intermittent server error sending a tab to Fenix, we fall back to the old sync command based mechanism - however, Fenix doesn't support this mechanism. Things like our VR products are probably in the same bucket.
It's probably time to consider if we should kill that fallback.
Comment 1•4 years ago
|
||
The priority flag is not set for this bug.
:markh, could you have a look please?
For more information, please visit auto_nag documentation.
Comment 2•4 years ago
|
||
Once we hit something like >90% Fennec to Fenix migration might be the right time?
Reporter | ||
Comment 3•4 years ago
|
||
(In reply to Edouard Oger [:eoger] from comment #2)
Once we hit something like >90% Fennec to Fenix migration might be the right time?
That sounds right to me! I'm marking this as P3 so it doesn't get in our way before then.
Comment 4•4 years ago
|
||
I'm trying to better understand what this change means - from a user standpoint how does the fallback mechanism work now, and what would be the user impact if we were to remove it?
Comment 5•4 years ago
|
||
Right now Fennec can only use the legacy send tab protocol that relies on Sync.
Currently Firefox for Desktop and Firefox for iOS can receive and send tabs to Fennec using the legacy protocol, but can also send and receive tabs using the new protocol (which is the preferred one).
Currently on Desktop, when we fail sending a tab using the new protocol for whatever reason, we fall back to using the legacy one.
The problem is that in the Fenix case, which only works with the new protocol, we'll report the tab was sent (through the fallback) but in reality Fenix will never get that tab.
What's proposed here is killing entirely the legacy system, and in consequence also the legacy fall back.
Reporter | ||
Comment 6•4 years ago
|
||
And to clarify, we could spend extra effort trying to prevent the above scenario, but IMO we are better off spending that effort killing the old.
In general, this would only be visible to, say Fennec users who refuse to upgrade to Fenix, or users stuck on 4 year old desktop Firefox versions - stuff would break a little for them. Up-to-date users would see zero change (except in the edge or error cases, as above)
Comment 7•4 years ago
|
||
FWIW, I think this fallback behaviour will also interact badly with rate-limiting, as described in https://github.com/mozilla/application-services/issues/3105#issuecomment-635166875
Assignee | ||
Comment 8•4 years ago
|
||
Updated•4 years ago
|
Updated•4 years ago
|
Pushed by vlad@vladikoff.com: https://hg.mozilla.org/integration/autoland/rev/03e3995a8488 Do not fallback to old SendTab if new SendTab fails. r=rfkelly,markh
Comment 10•4 years ago
|
||
bugherder |
Description
•