The test site used in the android `view` tests, https://www.example.com/ doesn't always resolve to a host
Categories
(Testing :: mozperftest, task, P1)
Tracking
(firefox88 fixed)
Tracking | Status | |
---|---|---|
firefox88 | --- | fixed |
People
(Reporter: acreskey, Assigned: mcomella)
References
(Depends on 1 open bug)
Details
Attachments
(1 file)
The view
test loads a website (example.com) via applink.
However, over the past 2 days this site is not resolving correctly.
The bug is to select and change the test site.
It should be one that we don't otherwise pre-connect to during browser launch.
Reporter | ||
Updated•4 years ago
|
Assignee | ||
Comment 1•3 years ago
|
||
Let's host it on our mozilla-mobile github.io page.
Assignee | ||
Comment 2•3 years ago
|
||
The test page is here – https://mozilla-mobile.github.io/perf-tools/mozperftest-test-page.html – and the source is here – https://github.com/mozilla-mobile/perf-tools/blob/master/docs/mozperftest-test-page.html
Assignee | ||
Comment 3•3 years ago
|
||
Updated•3 years ago
|
Assignee | ||
Comment 4•3 years ago
|
||
Assignee | ||
Comment 5•3 years ago
|
||
We're going to wait until mozilla-mobile/fenix#17447 is done before landing to minimize the number of variables.
Updated•3 years ago
|
Comment 6•3 years ago
|
||
There's a r+ patch which didn't land and no activity in this bug for 2 weeks.
:mcomella, could you have a look please?
For more information, please visit auto_nag documentation.
Assignee | ||
Comment 7•3 years ago
|
||
acreskey, two questions for you:
-
Do you think this is safe to land now? We haven't seen the automated improvement alert yet but at this point, it doesn't seem like we're going to see it so I think it should be safe.
-
Locally, I notice a difference of behavior when opening example.com and the test site: example.com will open a new tab while the test site will re-use the same tab (filed here). I don't think this should affect our results because we use conditioned profiles so it should be safe to proceed. Do you think this change is safe to make?
Reporter | ||
Comment 8•3 years ago
|
||
(In reply to Michael Comella (:mcomella) [needinfo or I won't see it] from comment #7)
acreskey, two questions for you:
- Do you think this is safe to land now? We haven't seen the automated improvement alert yet but at this point, it doesn't seem like we're going to see it so I think it should be safe.
I think that we've given the performance alerts sufficient time to detect the improvement.
We can discuss this at the next Bleeding Edge or Fenix Sheriffing meeting.
- Locally, I notice a difference of behavior when opening example.com and the test site: example.com will open a new tab while the test site will re-use the same tab (filed here). I don't think this should affect our results because we use conditioned profiles so it should be safe to proceed. Do you think this change is safe to make?
We are not actually using the conditioned profiles for the VIEW test because they were introducing other issues (slowdowns due to addon's not matching the current binary).
But I don't think 2) is a problem in our VIEW test case because we're just looking at the first load, right?
Assignee | ||
Comment 9•3 years ago
|
||
But I don't think 2) is a problem in our VIEW test case because we're just looking at the first load, right?
To clarify, it's not a problem because we use fresh profiles for each page load so we should never run into the case where there's an existing tab and we're selecting it. Let's land this.
Updated•3 years ago
|
Assignee | ||
Comment 10•3 years ago
|
||
I requested landing. I'll keep the NI just in case something goes wrong during landing.
Comment 11•3 years ago
|
||
Pushed by mcomella@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/9c9a41668d95 change VIEW perftest page to our github.io. r=acreskey,sparky
Comment 12•3 years ago
|
||
bugherder |
Assignee | ||
Updated•3 years ago
|
Description
•