This is the mail archive of the
gdb-patches@sources.redhat.com
mailing list for the GDB project.
Re: [rfa/c++testsuite] The remaining v3 testsuite fixes (or at least most of them)
Committed.
On Fri, Dec 07, 2001 at 04:49:08PM -0500, Jim Blandy wrote:
>
> Please commit this, and the static member test patch above. Thanks!
>
> Daniel Jacobowitz <drow@mvista.com> writes:
> > [This patch includes my previous patch today, for the static member test,
> > but not its changelog; I will commit them separately, but editing it out of
> > this diff would have been tricky.]
> >
> > These are mostly "more of the same" demangler changes:
> > - const & -> const ?&
> > - \\(void\\) -> \\((void|)\\)
> >
> > There are also changes to accept classes without the explicit vtbl member,
> > or the implicit int parameter that v2 seemed to use to construct classes
> > with virtual bases. I also allow the function name in "Breakpoint at "
> > messages to contain the base demangled name, including argument list; makes
> > sense for overloaded functions.
> >
> > I currently except class data printouts with explicit "_vptr.vC = $hex"
> > style entries. I can suppress those if people think that better, but I'd
> > rather get this in first; so I labelled the passes with FIXMEs. Our
> > handling of these is a little inconsistent.
> >
> > These handle all the tests that should pass in my current half-submitted
> > codebase, and only changes the name of one FAIL and one FAIL to PASS in the
> > v2 test results:
> >
> > 111c111
> > < FAIL: gdb.c++/classes.exp: continue to enums2
> > ---
> > > FAIL: gdb.c++/classes.exp: continue to enums2(\(\)|)
> > 1342c1342
> > < FAIL: gdb.c++/namespace.exp: info func xyzq
> > ---
> > > PASS: gdb.c++/namespace.exp: info func xyzq
> >
> > (The fail was because the testsuite expected functions to be sorted by
> > basename and not by full name including namespace).
> >
> > Is this OK to commit?
--
Daniel Jacobowitz Carnegie Mellon University
MontaVista Software Debian GNU/Linux Developer