Re: PATCH: Update libiberty/testsuite/demangle-expected

["H. J. Lu" <hjl at lucon dot org>]

On Fri, Oct 17, 2003 at 12:43:19PM -0400, DJ Delorie wrote:
> 
> > For this particular case where we have a failed testcase which is
> > imported from another part of gcc, is keeping it in sync with the
> > original source an obvious change?
> 
> All testsuite changes require approval.
> 
> > Or should we require they should be updated at the same time?
> 
> I think it's a good idea to keep track of changes in the C++
> demangler, and keep the libiberty testsuite up to date.  I just don't
> think that blindly updating it is ever correct.  The whole point of a
> testsuite is to ensure that the code is doing the right thing; every
> change needs to be reviewed to make sure that the new results are more
> correct than the old results; automatic changes to testsuites defeat
> their whole purpose.

I agree we shouldn't change those testcases which are passing. The
failed ones are just reminders for us how they should be fixed.
I don't agree syncing with the original working source is blindly
updating unless the original one is wrong.


H.J.