Ajax: Fix the invalid script test in Android 2.3 #1573
Conversation
|
We rely in 2.x on setting I'm not sure if we're doing anything substantially different in 1.x, though. We set |
Android 2.3 doesn't support async script tags so if the injected script tag refers to a throwing script, the error is thrown synchronously making the window.onerror handler to not fire. Catch the error instead.
|
👍 |
|
Hm, actually i don't see this test failure in android 2.3, tested on browserstack samsung galaxy note. And this is the same domain right? Shouldn't this prevent script injection? |
|
It does look like when we switch to all-script-tag execution this won't be an issue. I think the comment in the PR isn't quite right since if it's a local request and going through XHR. Maybe |
|
@dmethvin did you check this test in android? Do you see a failure? |
It shows up in TestSwarm & locally for me in the Android emulator.
Right, it might be related to |
|
@mzgol testswarm? you mean on http://swarm.jquery.org/project/jquery?
It doesn't for me, could you check on the browserstack and make a screenshot? |
No; Androids are tested at http://swarm.jquery.org/project/jqueryweekly. On TestSwarm it just hangs always at this test; see http://swarm.jquery.org/result/1810062 & http://swarm.jquery.org/result/1816812. I don't have a screenshot at hand, I can shoot one later. |
|
How about http://swarm.jquery.org/result/1822771 or http://swarm.jquery.org/result/1816951, could it be server related problem? |
These are runs from |
|
Now this explains the -
Screenshot is no longer required, thank you @mzgol |
|
BTW, this might show one of the issues we have with TestSwarm stability. It should just error on this test in the same way as the test locally does and not timeout the whole TestSwarm job instead. Maybe sth similar is causing non-deterministic issues with other tests? |
Unfortunately, the failures didn't disappear, see: http://swarm.jquery.org/job/3169 (I force-run the test to check if it helped). |
|
Maybe Android 2.3 doesn't fire |
Yeah, this PR will most likely be wrong after rebasing; I'll check what can be done. |
|
@mzgol I think this should close, but will leave for your confirmation. |
|
The problem still exists and will need a workaround (see http://swarm.jquery.org/result/2048250) but it'll now have to be different anyway so I'm closing this PR. |
Android 2.3 doesn't fire the window.onerror handler, just accept the reality there and skip the test. Refs jquerygh-1573
Android 2.3 doesn't fire the window.onerror handler, just accept the reality there and skip the test. Refs jquerygh-1573 Fixes jquerygh-2457
Android 2.3 doesn't fire the window.onerror handler, just accept the reality there and skip the test. Refs jquerygh-1573 Fixes jquerygh-2457
Android 2.3 doesn't fire the window.onerror handler, just accept the reality there and skip the test. Refs jquerygh-1573 Fixes jquerygh-1786
Android 2.3 doesn't fire the window.onerror handler, just accept the reality there and skip the test. Refs jquerygh-1573 Refs jquerygh-1786 Refs jquery/jquery.com#108
Android 2.3 doesn't fire the window.onerror handler, just accept the reality there and skip the test. Refs gh-1573 Refs gh-1786 Refs jquery/jquery.com#108 Closes gh-2458
Android 2.3 doesn't fire the window.onerror handler, just accept the reality there and skip the test. (cherry-picked from 6044fb6) Refs gh-1573 Refs gh-1786 Refs jquery/jquery.com#108 Closes gh-2458
Android 2.3 doesn't fire the window.onerror handler, just accept the reality there and skip the test. Refs jquerygh-1573 Refs jquerygh-1786 Refs jquery/jquery.com#108 Closes jquerygh-2458
4 participants
Android 2.3 doesn't support async script tags so if the injected script tag
refers to a throwing script, the error is thrown synchronously making the
window.onerror handler to not fire. Catch the error instead.