Hi all
sorry for quite some emails from CI server, but finally I have been able to run Selenium tests using BrowserStack on our CI server. Many tests need to be adjusted to run there, but the core of the work is done.
On 2/17/14 9:51 AM, Michal Čihař wrote:
Hi all
sorry for quite some emails from CI server, but finally I have been able to run Selenium tests using BrowserStack on our CI server. Many tests need to be adjusted to run there, but the core of the work is done.
This is great news. Thanks for your work.
Hi
Dne Mon, 17 Feb 2014 11:58:23 -0500 Isaac Bennetch bennetch@gmail.com napsal(a):
On 2/17/14 9:51 AM, Michal Čihař wrote:
Hi all
sorry for quite some emails from CI server, but finally I have been able to run Selenium tests using BrowserStack on our CI server. Many tests need to be adjusted to run there, but the core of the work is done.
This is great news. Thanks for your work.
Now there is another work: to fix current Selenium tests :-). Many of them were not updated to current code, many have wrong assumptions about their environment.
Michal Čihař a écrit :
Hi
Dne Mon, 17 Feb 2014 11:58:23 -0500 Isaac Bennetch bennetch@gmail.com napsal(a):
On 2/17/14 9:51 AM, Michal Čihař wrote:
Hi all
sorry for quite some emails from CI server, but finally I have been able to run Selenium tests using BrowserStack on our CI server. Many tests need to be adjusted to run there, but the core of the work is done.
This is great news. Thanks for your work.
Now there is another work: to fix current Selenium tests :-). Many of them were not updated to current code, many have wrong assumptions about their environment.
Michal, should test fixing be added to http://wiki.phpmyadmin.net/pma/GSoC_2014_Ideas_List#Automated_testing ?
Hi
Dne Mon, 17 Feb 2014 15:14:28 -0500 Marc Delisle marc@infomarc.info napsal(a):
should test fixing be added to http://wiki.phpmyadmin.net/pma/GSoC_2014_Ideas_List#Automated_testing ?
I'd really like to have them fixed before GSoC, so IMHO no.
Currently many of failing tests are simply because the test server does not have configured pmadb or they are not running as root. I thinks the tests should properly detect this and mark themselves as skipped rather than fail.
Hi
Dne Mon, 17 Feb 2014 22:24:52 +0100 Michal Čihař michal@cihar.com napsal(a):
Currently many of failing tests are simply because the test server does not have configured pmadb or they are not running as root. I thinks the tests should properly detect this and mark themselves as skipped rather than fail.
As you could see from the flood of CI mails, it took some time, but the tests are mostly passing now. However there can be still some hidden timing issue, which can pop up randomly - like missing waiting for element to appear before attempting to use it. Anyway basic Selenium tests should be now working on our CI server.
There are still some work to do:
- I've marked some tests as incomplete with description what needs to be fixed in them. - For now we do not test any superuser or configuration storage depending functionality, I'll add these later, but I think this is anyway the most usual setup which user faces.
Le 2014-02-24 11:56, Michal Čihař a écrit :
Hi
Dne Mon, 17 Feb 2014 22:24:52 +0100 Michal Čihař michal@cihar.com napsal(a):
Currently many of failing tests are simply because the test server does not have configured pmadb or they are not running as root. I thinks the tests should properly detect this and mark themselves as skipped rather than fail.
As you could see from the flood of CI mails, it took some time, but the tests are mostly passing now. However there can be still some hidden timing issue, which can pop up randomly - like missing waiting for element to appear before attempting to use it. Anyway basic Selenium tests should be now working on our CI server.
There are still some work to do:
- I've marked some tests as incomplete with description what needs to be fixed in them.
- For now we do not test any superuser or configuration storage depending functionality, I'll add these later, but I think this is anyway the most usual setup which user faces.
Thanks Michal, this turned out to be more tricky than expected!
-
Isaac Bennetch -
Marc Delisle -
Michal Čihař