Feature Requests
298 results found
-
Show number of lines uniquely covered by a test
Today, the test fail/pass indicators on each line show which tests are executing that line.
I would like to have an option (maybe in the Metrics window) which tells me the number of lines covered solely by one test.
My goal behind is to identify code that is only run by few system/integration tests and clearly needs more tests. Such tests might start/run half of the system just to prepare a test so that - from the coverage point of view - the code looks like to be covered but - from the testing point of view - actually isn't…
4 votes -
Force tests to run inside the same process
NCrunch spins up multiple processes to run tests. This is great most of the time.
But if a suite of tests has some expensive setup that only needs to happen once, that setup will run once for each process. Increasing the overall time it takes to run the whole test suite.
Provide a mechanism to force (suggest) NCrunch to run 2 tests in the same process. Something like a [SharedContext("Name")] attribute. Then when 2 or more tests are queued that have the same shared context, NCrunch should run those in the same process.
3 votes -
Option to disable stylecop MSBuild integration
Currently, FxCop Analysis can be switched off for Performance reasons. For the same reason it would be great to have a Setting where the StyleCop build Integration (http://www.nuget.org/packages/StyleCop.MSBuild/) can be switched off as well.
An additional benefit would be that there is no need anymore to occasionally reset the NCrunch test pipeline because of build failures (it cannot copy the StyleCop NuGet package which is at that time in use by another NCrunch process).
3 votes -
Show risk information in code margin.
Show colour indicator and number indicator. Number will show how many tests cover that method and colour will say if its enough tests to fill safe.
2 votes -
make grid nodes discoverable on the network
I think it would be great to use zeroconf or some such to provide a way to find available grid nodes on the network. Having to know the name or ip of each machine is cumbersome.
Would suggest requiring the user to assign trust to the nodes before they get used for any calculations though.
3 votes -
Risk/Progress window move to toolbar (status bar) with progress bar. Add text "xxx/yyy test pass"
Risk/Progress window move to toolbar (status bar) with progress bar. Add text "xxx/yyy test pass"
Reason: unable dock this windows in right position when clip/unclip (auto hide ) dock panels.
1 vote -
Provide an option to only enable grid node when machines has no active user sessions
I'd love to make use of the latent power of all the locked desktops in my office but don't want to risk using nodes when someone is logged in.
1 vote -
1 vote
-
Allow coverage marker colour to be defined for tests in progress
When tests execute the colour of the coverage markers darken. This is very useful for determining when affected tests have completed. However, the contrast is too subtle. It would be good if this could be changed in the 'Coverage Markers' configuration section. Either with an absolute colour or at least to have control over the percentage of darkening.
2 votes -
Remote server: Run test on all (connected) servers
I have a test which is required to be run manually to reset the environment to a initial state.
I would like the option to run the test on all servers at once. Now I have to select "run test on server X" for each server I have which is tedious.(This feature would also be good to ensure a particular test is passing on all servers).
1 vote -
Support UITestMethod in Windows Store apps
Posted in response to: forum.ncrunch.net/yafpostsm5807UITestMethod-not-recognized.aspx.
Currently NCrunch ignores test methods with UITestMethod attribute which forces the test to run on the UI thread. This is required whenever the test needs to manipulate a UI element even if it's not a UI test in itself.
1 vote -
Remote server: Add some sort of performance index per server
I would like to have some index indicating how well the grid node server performs related to other servers.
For instance, if I run 2 grid servers on Azure and 2 on internal network, I want to know if it's worth adding more servers on Azure or on internal network (or if it helps out scaling Azure servers etc.).One idea might be to have an average processing time on tests, and check when test has been executed on different servers is above or below the average execution time and calculate some sort of performance index from those measurements and…
4 votes -
Test Window Filter
I would like a filter in the test window that "show only tests that should runs automatically due to configured Engine Mode"
11 votes -
Remote server: Conditionally run tests depending on which remote server they are running on.
Scenario:
I have http integration tests which will report results in one way locally (due to IIS custom errors) and one way remotely. Thus for tests verifying the corretc "remote" response I would like them to run on a grid node that is not running on my local machine.The current "Capabilities" is not attractive since that is set on grid node server itself and requires project-specific user configurations on each grid node server (which annoys the developers if they are running them on their dev machines if I will need to interrupt to tweak custom project settings).
Possible solutions…
2 votes -
Make NCrunch grid performance measurable
Provide some kind of programmatic access to performance indicators that are able to report the performance of a ncrunch grid, e.g. provide access to the duration of a full-end-to-end test. This would enable things like auto-scaling a grid.
2 votes -
Add ability to bottleneck ncrunch
Sometimes I have tests which are unstable on the buildserver, since the specs of the buildserver are slower. Even when I crunch the tests locally, they still never fail. (Usually frontend tests, making use of bUnit rendering)
I would find it handy to be able to crunch tests with a forced bottleneck, limiting resources, allowing me to reproduce the buildserver circumstances.
1 vote -
Allow multiple simultaneous instances with different filters
It would be great if I could get multiple risk/progress bars pointed at different categories of tests. I would like to see (for instance) one risk/progress bar for acceptance, one for integration, and one for unit tests. It would be great if these were full parallel instances - like with their own configuration and everything. For instance, having acceptance tests farmed out to servers makes a lot of sense but all my unit tests should run locally because it takes longer to farm out a batch than to run it.
0 votes -
In Test window, Server column for grouping rows should include all servers for children (not just one of them).
In my case Server always says "(Local)" if any child test ran locally. I think this should be a comma-delimited list like Category. Next best thing would be leaving it blank.
If I'm filtering for failed tests only, the current behavior is really a bug, because all of the failing tests may have ran on one node server that isn't (local), yet the grouping row alone still says (local). So it makes scanning the server column to troubleshoot issues (in this case node server dependency config issues) confusing.
0 votes
- Don't see your idea?