#197 Updates testcase displayed information
Merged by cipherboy. Opened by cipherboy.
cipherboy/gssproxy fix_testcase_output  into  master

Download 197.patch

Changes SUCCESS/FAILED from reflecting return code value to
expected value with PASS/FAIL, and shows expected value on
testcase failure.

rebased

Some thoughts:

  • Style nit: wrap at 78 characters please; where possible, do not break strings. (One can think of our style as "we worked on krb5 but come on we're all friends here".)
  • Removal of the OO/XX indicators makes this output harder to read in the non-color case (e.g., on builders; Fedora doesn't run this test suite, but Debian does).
  • Printing the expected vs. received is a generally useful thing.
  • PASS/FAIL doesn't show more information by itself than SUCCESS/FAILURE, or any other two-word pairing that I can come up with. It does, however, make this harder to read. This is because the presentation and word choice now match how other test suites would use them - to refer to overall behavior of a test, not individual components.
  • That said, I think our presentation leaves something to be desired, and I would like to see the more invasive change of having these statuses refer to the overall test (i.e., PASS means overall pass, even in the case of intentional failure; FAIL means a problem occurred including an underlying unexpected success.)

Ah okay, was led astray by the commit message. Will merge with minor fixups in a moment.

Commit 39b0ffd4 fixes this pull-request

Pull-Request has been merged by ascheel@redhat.com

Did this commit remove the coloring and the fallback ?
Why!?

I mean, ok to remove OO/XX if it feels it makes it harder to read, but this commit removed all coloring and I am not happy about that.

Simo,

This removes OO/XX only -- note that the branch with colorama remains. Further, rather than relying solely on color/"OO/XX colors" to denote whether the particular test case performed as expected, we use PASS/FAIL to indicate what each test did. This gives a consistent look between having colorama and not.

Then, if the test case didn't perform as expected, after the return value, we denote what should've happened in parenthesis.

TL;DR: coloring remains, unified display with and without coloring, expected result more clearly displayed.

Hopefully that is a little more clear :)

-- Alex

Doh, so sorry, I should have read better, or better yet, just run make tests immediately.

Thanks for clarifying, I liked the PASS/FAIL change just not losing colors

Commit a045a3a0 fixes this pull-request

Pull-Request has been merged by ascheel@redhat.com

Commit 39b0ffd4 fixes this pull-request

Pull-Request has been merged by ascheel@redhat.com

Metadata