add back the forgotten test file for patmat exhuastivity check

[liufengyun]

By adding back the ClassicReporter, I'm not nostalgic, but to avoid rewriting all the check files :)

Review @felixmulder .

[felixmulder]

I'd prefer not to have duplicate reporters (we actually already have a third one in the sbt bridge 😱) - would it be alright with you @liufengyun if I took the time to amend the check files in this PR? (I'll push to staging)

[liufengyun]

@felixmulder I'm afraid that each refactoring of the reporter (even a minor change capitalisation) would require update the check files, which will be very tedious and become a burden of maintenance.

[felixmulder]

@liufengyun - true, let me sleep on it and we'll talk again tomorrow :)

[DarkDimius]

@felixmulder @liufengyun, I'd remind you my suggestion for testing. Have a specific reporter that is only used in tests in order to provide minimal information used to compare test output to the expected value.

./tests/patmat/i947.scala:10: warning: unreachable code
       case ys: List[d18383] => false

would be reported as

UnreachabeCode(line, pos)

I'd also prefer to have this information defined inline in the tests itself, but that's a matter of taste.

[felixmulder]

@DarkDimius - yes, we agreed upon this scheme this morning. Just haven't gotten around to it yet :)

[liufengyun]

According my experience with testing, we often have the need to switch back and forth between check files and source code. Messages like UnreachabeCode(line, pos) is not friendly enough for test and development.

[DarkDimius]

This is why I propose to have this message defined inline inside the test
file itself

On 19 October 2016 1:40:56 p.m. liu fengyun [email protected] wrote:

According my experience with testing, we often have the need to switch back
and forth between check files and source code. Messages like
UnreachabeCode(line, pos) is not friendly enough for test and development.

You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub:
#1609 (comment)

[liufengyun]

@felixmulder what's your opinion on this? Any changes in mind?

[felixmulder]

I have a solution in mind for neg tests, and I will build upon this PR.
Will probably have something to show for this in a couple of hours :)

On Tue, Nov 8, 2016, 10:11 liu fengyun [email protected] wrote:

@felixmulder https://github.com/felixmulder what's your opinion on
this? Any changes in mind?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#1609 (comment), or mute
the thread
https://github.com/notifications/unsubscribe-auth/ABdYweDc-ALCp8wG_U-_cHyszu-Llvywks5q8Dy6gaJpZM4KaC6r
.

[felixmulder]

Latest commit adds a very simplified version of the ClassicReporter that is not likely to change. It basically reports in this fashion:

2: Pattern Match Exhaustivity: List(_, _, _)

WDYT @liufengyun?

[liufengyun]

@felixmulder thanks for working on this. I'm fine for the changes. Good job 👍