Added by Xiaojing_liu over 3 years ago. Updated over 3 years ago.
Description
Using Firefox browser to see a job's details on OSD, the page is not correct.
There is javascript error: 'ReferenceError: setupResult is not defined'
There is no problem in Chrome.
The job details should show correct in Firefox
As the problem seems to be related to firefox version
Files
| Screenshot_20200910_124500.png (152 KB) Screenshot_20200910_124500.png | Xiaojing_liu, 2020-09-10 04:50 |
could this potentially be fixed by https://github.com/os-autoinst/openQA/pull/3376 ?
I read out the test URL from the screenshot, it is https://openqa.suse.de/tests/4668581
For me the page renders just fine in Firefox 78.0.2esr on openSUSE Leap 15.2 . Does the problem reproduce when checking multiple times? On other computers? What's the version of Firefox on your setup? Does the problem reproduce with other jobs or just this single one? Can you crosscheck against other openQA instances, e.g. o3 and personal instances?
okurz wrote:
could this potentially be fixed by https://github.com/os-autoinst/openQA/pull/3376 ?
I checked with this pr, it is not fixed by it.
I read out the test URL from the screenshot, it is https://openqa.suse.de/tests/4668581
For me the page renders just fine in Firefox 78.0.2esr on openSUSE Leap 15.2 . Does the problem reproduce when checking multiple times? On other computers? What's the version of Firefox on your setup? Does the problem reproduce with other jobs or just this single one? Can you crosscheck against other openQA instances, e.g. o3 and personal instances?
The problem could be reproduced every time when showing the job details. Mine and one colleague's firefox version are both '68.2.0'. The problem will happen when showing any jobs, so it does not matter the jobs. I also checked my environment, the issue also happens. I haven't checked o3, I cannot access to it right now (Because of the network in my home). But one colleague told me, o3 has the same problem with OSD.
One colleague's firefox is 78.1.0esr (64-bit), there is no problem in his test. So it's related to firefox version? do we need to support the older version?
Sorry to overwrite those settings.
Xiaojing_liu wrote:
One colleague's firefox is 78.1.0esr (64-bit), there is no problem in his test. So it's related to firefox version? do we need to support the older version?
We should support older versions but do not need to support "completely outdated versions". In this case though I would like to understand what specific change introduced the problem so that at least in the future we can avoid trying to be too innovative.
@mkittler @cdywan do you recall what JavaScript changes you might have introduced that went into OSD yesterday that could explain this problem?
#3352 was the last change to that function, although it wouldn't explain a not defined error - the signature changed, not the name of it, and this is not C++
https://bugzilla.mozilla.org/show_bug.cgi?id=1361856
This is fixed in all supported versions of Firefox. I heavily object supporting even older versions since it:
In this case though I would like to understand what specific change introduced the problem so that at least in the future we can avoid trying to be too innovative.
I think this can easily be solved by testing with ESR which (afaik) was done.
Yes, ESR was tested. I agree that in this case we do not need to change anything in our current code as all properly supported and maintained browser versions seem to be ok.
So I will close this as a Support ticket with my strong suggestion to users: Please make sure to run up-to-date secure systems
