Ticket History #189: warn about untested test cases rather than erroring out

From bjaspan@MIT.EDU Thu Nov 14 12:33:34 1996
Received: from MIT.EDU (PACIFIC-CARRIER-ANNEX.MIT.EDU [18.69.0.28]) by rt-11.MIT.EDU (8.7.5/8.7.3) with SMTP id MAA10138 for <bugs@RT-11.MIT.EDU>; Thu, 14 Nov 1996 12:33:33 -0500
Received: from BEEBLEBROX.MIT.EDU by MIT.EDU with SMTP
id AA09562; Thu, 14 Nov 96 12:33:32 EST
Received: by beeblebrox.MIT.EDU (940816.SGI.8.6.9/4.7) id RAA07116; Thu, 14 Nov 1996 17:33:30 GMT
Message-Id: <199611141733.RAA07116@beeblebrox.MIT.EDU>
Date: Thu, 14 Nov 1996 17:33:30 GMT
From: bjaspan@MIT.EDU
Reply-To: bjaspan@MIT.EDU
To: krb5-bugs@MIT.EDU
Subject: warn about untested test cases rather than erroring out
X-Send-Pr-Version: 3.99

mit

System: IRIX beeblebrox 5.3 11091812 IP22 mips



The krb tests in tests/dejagnu will fail with untested test cases if
the user does not have permission to rlogin in to the local machine as
root; this causes make to exit with an error. This is documented in
the install manual, but it is a good bet that many users will run make
check without reading that section first; they will then send a bug
report to krb5-bugs complaining about the problem.

I think that the tests should not cause make to fail; the
documentation says this error can be safely ignored, so make should
not force people to notice it. We could change the tests to use warn
(which prints WARNING: <msg>) instead of untest, or perhaps something else.




From: "Barry Jaspan" <bjaspan@MIT.EDU>
To: bjaspan@MIT.EDU
Cc: krb5-bugs@MIT.EDU
Subject: Re: krb5-misc/189: warn about untested test cases rather than erroring out
Date: Thu, 14 Nov 1996 14:55:01 -0500

The problem is actually a little more bit complicated than this.
Besides the ability to rlogin as root to the local machine, the user
also must either not have a .k5login or have *both*
<username>@KRBTEST.COM and <username> listed in their .k5login. This
is not documented anywhere that I've seen.

Perhaps the tests should just move .k5login out of the way, do the
test, and move it back (printing a warning in the process). At any
rate, this should be documented. And, the tests should not fail with
an error (causing make to print ***) simply because a test cannot be
performed.

Barry

From: Tom Yu <tlyu@MIT.EDU>
To: bjaspan@MIT.EDU
Cc: krb5-bugs@MIT.EDU, krb5-prs@RT-11.MIT.EDU
Subject: Re: krb5-misc/189: warn about untested test cases rather than erroring out
Date: Mon, 18 Nov 1996 14:14:17 -0500

I believe that dejagnu 1.3 does not have this problem. We can simply
point people at that, or tell them to ignore the make check error if
they're using 1.2.

---Tom

From: Tom Yu <tlyu@MIT.EDU>
To: Unassigned Problem Report <krb5-unassigned@RT-11.MIT.EDU>
Cc: krb5-bugs@MIT.EDU
Subject: Re: krb5-misc/189: warn about untested test cases rather than erroring out
Date: Mon, 18 Nov 1996 14:15:01 -0500

`Tom Yu' made changes to this PR.

State-Changed-From-To: open-closed
State-Changed-By: bjaspan
State-Changed-When: Wed Dec 4 15:35:19 1996
State-Changed-Why:

According to Tom, this problem does not exist with dejagnu 1.3.