lefthandedgoat
commented
4 years ago You can make a new module and create your own try-catch wrapped version of before, then make sure in your test the last line of your open statements is for your new module. In F# the most recently defined version of a function is the one that is used in your code.
Something like this: (untested)
module canopyExtensions
open canopy.runner.classic
let before f =
try
before f
with ex -> //do whateverLooking at the runner code the Before and the actual Test are run at the same time : https://github.com/lefthandedgoat/canopy/blob/cfe3fa17526c290c29dfb8f2b8873577bf7d1792/src/canopy/runner.fs#L162
In a try catch: https://github.com/lefthandedgoat/canopy/blob/cfe3fa17526c290c29dfb8f2b8873577bf7d1792/src/canopy/runner.fs#L134-L137
If you can create a sample test on a public site that demonstrates escaping this I can fix it.
Pwntheon
We're changing a few things in our application and because of this, a few of the steps needed to setup a couple of our tests in
beforefunctions fail.This unfortunately raises an exception instead of failing the test - i suspect because the test hasn't "started" yet.
Since this rewrite is taking some time we'd still like for our tests to run, but when we're running a production build of the canopy project on our jenkins server, the exception causes the whole application to crash, thus preventing us from testing.
We could wrap the whole shebang in a try-catch, but ideally i'd like for the tests affected to fail and the rest of the test suites to continue as usual. Is there any way to achieve this?
Also, we're writing quite a lot of tests in many contexts so any solution that prevents a lot of additional code having added to each before and after function would be a big plus as this increases boilerplate.