This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
It would be great if Synergy would provide another view on test run which would calculate how many test cases from all test specifications available in selected versions were executed. The UI could look like this: Include test specifications available for NetBeans versions: [ ] 7.3 [x] 7.4 [x] 8.0 [x] 8.1 Windows [x] Ubuntu [x] Mac OS X [x] Linux - Other [ ] Java Editor [x] 102/300 150/180 92/120 5/60 344/600 Git [ ] 0/50 20/100 0/50 0/0 Profiler [x] 25/70 120/140 30/70 40/140 175/280 127/370 270/320 122/190 519/880 Always the latest available test specification will be taken into account so for example although Java Editor will have three test specifications for versions 7.4, 8.0 and 8.1, only assignments for the last one 8.1 will be calculated.
Jirko I pushed first progress (mainly because I have a different pc in office and in home so that I can access work from Friday in laptop in office...). It is not directly visible, but you can access it by following link: http://services.netbeans.org/synergy/client/app/index_dev.html#/run/18/coverage (you can replace the 18 which is test run ID by ID of any test run - can be found in url). I need to verify that it correctly filters specifications and counts coverage. Then I will add total row/column and the checkbox to filter out platform/version. Note for me: Platforms need to be collected from assignments, not from API so that in case of platform rename, displayed columns with platforms will always match platforms in assignments.
Hm, looks pretty good Láďo! Once there will be the totals and opt-out checkboxes it will be exactly what I need. Two comments: 1. When I clicked the URL to TR #18, it was only showing some empty template, however several subsequent reloads finally displayed the NetBeans IDE 8.1 Full Testing Test Run Coverage title. 2. I don't understand why there are sometimes rows with all 0/0 values like for example "[API Support] Window Wizard" or "[Web Client] Cordova support" if all versions from 8.1 down to 6.5 are selected.
It might take some time to load, the "Show coverage" button gets enabled once everything is loaded and some versions are selected. The problem with 0/0 is that right now, it takes total cases from test assignments, so no assignment => 0/0. I will fix it to fetch number of cases for 0/0 specifications.
Jirko, how it should behave in following use case? 1. There is specification "Java Editor" in version 1 and version 2 . In version 1, it has 3 test cases, in version 2 it has 4 2. There are 2 assignments against version 1 for platform Windows 3. In code coverage, you would like to see version 1 and version 2 What happens right now that if there is assignment for given platform, it takes total cases from assignment, for "unused specs in platform", it retrieves number of cases of latest version from server, so the table would look like Windows [x] Ubuntu [x] Total Java Editor [x] 0/6 0/4 0/10 Is this ok? The 0/6 = 2 assignments * version1 (3 cases) and 0/4 = 1 assignment * version 2 (4 cases) Latest changes should be live in 15 minutes, so you can try it but I think there are still some cases to fix but it should work fine with only 1 version selected Thank you
And another question :) Right now, if you select e.g. version 2 in code coverage, it will show also test assignments against version 1. Should I filter them out?
(In reply to Vladimir Riha from comment #4) > Jirko, how it should behave in following use case? > > 1. There is specification "Java Editor" in version 1 and version 2 . In > version 1, it has 3 test cases, in version 2 it has 4 > 2. There are 2 assignments against version 1 for platform Windows > 3. In code coverage, you would like to see version 1 and version 2 > > What happens right now that if there is assignment for given platform, it > takes total cases from assignment, for "unused specs in platform", it > retrieves number of cases of latest version from server, so the table would > look like > > Windows [x] Ubuntu [x] Total > Java Editor [x] 0/6 0/4 0/10 > > Is this ok? The 0/6 = 2 assignments * version1 (3 cases) and 0/4 = 1 > assignment * version 2 (4 cases) Yes, this is okay. It will be a bit confusing to see different totals in one row but I assume this will happen rarely. (In reply to Vladimir Riha from comment #5) > Right now, if you select e.g. version 2 in code coverage, it will show also > test assignments against version 1. Should I filter them out? Yes, if I only want version 2 displayed I don't care if somebody incorrectly picked version 1 for its test assignment.
So hopefully fixed :) I've just fixed broken URL. Jirko please feel free to reopen or tell me if something should be changed/improved/fixed. Thank you
Excellent job Lado! I am totally happy. Thanks!