Repository navigation
Conversation
|
This is absolutely AWFUL IMO. Nasty. GROSS. In comparison, I've purged any of this fluent nonsense with assertj and hamcrest from Commons. |
|
This is an awful way to test IMO. So much worse than plain assertions in JUnit, assertNull() is perfect. Assert that bla is Null or whatever the fluent pattern I've seen done with assertj or hamcrest is an anti-pattern for me. Hard to read and maintain. Just AWFUL! |
2c7914e to
5de5a4d
Compare
5de5a4d to
fbe1d2f
Compare
|
@arturobernalg Do you have a preference? |
arturobernalg
left a comment
There was a problem hiding this comment.
As discussed on the dev mailing list back in January, there was broad agreement to adopt AssertJ in core. I gave my +1 then, and I still support the change.
Thanks for taking care of this, @ok2c .
|
Have you ever tried stepping through some assertj mess in a debugger? Good luck. |
|
@garydgregory Out of curiosity, how often do you actually have to step through any of the HttpComponents test cases and when was the last time you had to do that? |
|
Mercifully not often but I certainly never want to deal with assertj unless I'm forced to. And it certainly will not encourage me to write new tests or update existing tests using this awful and overtly VERBOSE API that stinks of literate programming 😢 |
@garydgregory Not that you have done that often in the past. I will not encourage you to write any tests, be sure of that, though one is actually welcome to continue using straight JUnit asserts for new code and new test. However given the fact that between me and @arturobernalg we write pretty much most of the tests and we actually do 100% of bug fixes, so please kindly allow us to make the choice of tools for the job. |
|
I'm not stopping you, I'm voicing my opinion. Notice I didn't say "-1'. |
Folks, I have taken it upon myself to convert all asserts in core from JUnit to AssertJ. This has been discussed on the dev list and I hope no one has objections to using AssertJ.