Re: Error with Sparc builtin setjmp

["Kaveh R. Ghazi" <ghazi at caip dot rutgers dot edu>]

 > From: kenner@vlsi1.ultra.nyu.edu (Richard Kenner)
 > 
 >     There is no situation in which a change won't make things worse.
 >     For instance, this patch might not even compile.  You could also
 >     have made a mistake; perhaps the code was correct previously and
 >     you just missed something.  Anyone can make such mistakes, and the
 >     only cure is careful testing.
 > 
 > Of course I *tested it*!  It's just that the normal bootstrap and
 > regression testing isn't all that useful in a case like this.  I had a
 > test case that failed (it would have been *any* use of the builtin) on
 > a specific port.  I built a cross-compiler for that port, saw the
 > crash, made the change, and verified the crash didn't occur anymore
 > with the fix.

Fine, then please say so in the original message.  Don't wait for
people request it from you.

Not only does the testing information reassure others that the testing
occurred, but in leading by example you encourage others to do so
also.

BTW, did you check the testcase into the testsuite?  I looked in
testsuite/ChangeLog and saw a grand total of 2 entries from you, one
in Aug and one from Feb.  I know a lot of your changes stem from bugs
only reproducible in Ada, but I suspect at least *some* in the last
year could be exposed by C testcases.

		--Kaveh
--
Kaveh R. Ghazi			Engagement Manager / Project Services
ghazi@caip.rutgers.edu		Qwest Internet Solutions