Removing message formatting rounding #611
Conversation
When defining a coverage threshold (88.45 for example) and the actual coverage is very close (88.449 for example) we will get a failure message. However, the message will be: "FAIL Required test coverage of 88.45% not reached. Total coverage: 88.45%" where it's wrong and confusing since we failed because it's under while it the message it presented as equal. The removing of the rounding will fix it.
|
Hello @AdiKrasin, can this fix get merged and released? |
|
@ionelmc Can you have a look at this? Thanks again |
|
If you are not ok with this change, I'll gladly come up with a PR fixing it as well matching your expectations. This is a long standing issue that is quite painful with long running test suites (because it's a pain to relaunch it to know the coverage) that are distributed (so cannot run with coverage only). I'd argue a more in-line change would be to read the coverage "precision" setting and report it the same way coverage is. Is that something you would be willing to merge and release? |
| @@ -341,7 +341,7 @@ def pytest_terminal_summary(self, terminalreporter): | |||
| markup = {'red': True, 'bold': True} if failed else {'green': True} | |||
| message = ( | |||
| '{fail}Required test coverage of {required}% {reached}. ' | |||
| 'Total coverage: {actual:.2f}%\n' | |||
| 'Total coverage: {actual}%\n' | |||
There was a problem hiding this comment.
I think this would fix the problem you described without changing the output too dramatically.
| 'Total coverage: {actual}%\n' | |
| 'Total coverage: {truncate(actual, 2)}%\n' |
There was a problem hiding this comment.
Would be cleaner imo to rely on what coverage is doing: https://github.com/nedbat/coveragepy/blob/a1d8d2961c9793cf8d3fa09666012190cdec9214/coverage/results.py#L253
There was a problem hiding this comment.
Maybe we could just reuse that from coverage? Hopefully @nedbat won't change it soon :-)
There was a problem hiding this comment.
I am actively refactoring that file to make more of it public: https://github.com/nedbat/coveragepy/tree/nedbat/public-analysis . We could talk about how to make it useful here.
When defining a coverage threshold (88.45 for example) and the actual coverage is very close (88.449 for example) we will get a failure message. However, the message will be: "FAIL Required test coverage of 88.45% not reached. Total coverage: 88.45%" where it's wrong and confusing since we failed because it's under while it the message it presented as equal. The removing of the rounding will fix it.