[rfc] Wswitch test case (which fails)

[Joseph S. Myers]

On Sat, 2 Feb 2002, Andrew Cagney wrote:

> I'm wondering what GCC's policy is on committing tests that fail, but 
> illustrate regressions, is.

They can go in the testsuite, and should not be marked XFAIL (whereas
tests for non-regression bugs should be marked XFAIL unless there is a
patch available or an immediate intention to fix them).  There should also
be a priority "high" PR concerning the problem, and a comment in the
testcase should reference the PR if pre-existing, or the PR mention the
testcase if a new PR is submitted to accompany the addition of the
testcase.

OK to commit the following patch to bugs.html to state that testcases for 
regressions should not be marked XFAIL?

Index: bugs.html
===================================================================
RCS file: /cvs/gcc/wwwdocs/htdocs/bugs.html,v
retrieving revision 1.49
diff -u -r1.49 bugs.html
--- bugs.html	2002/01/23 15:42:20	1.49
+++ bugs.html	2002/02/03 01:39:03
@@ -212,7 +212,8 @@
 <p>If you find a bug, but you are not fixing it (yet):</p>
 <ol>
 <li>Create a (minimal) test-case.</li>
-<li>Add the test-case to our test-suite, marking it as XFAIL.</li>
+<li>Add the test-case to our test-suite, marking it as XFAIL (unless
+the bug is a regression, in which case it should not be marked XFAIL).</li>
 <li>Add a bug report referencing the test-case to GNATS.</li>
 </ol>
 

-- 
Joseph S. Myers
jsm28@cam.ac.uk