This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: [Bug c++/13776] [tree-ssa] Many C++ compile-time regression in3.5-tree-ssa 040120
- From: Diego Novillo <dnovillo at redhat dot com>
- To: Richard Guenther <rguenth at tat dot physik dot uni-tuebingen dot de>
- Cc: "gcc at gcc dot gnu dot org" <gcc at gcc dot gnu dot org>
- Date: Sat, 13 Mar 2004 09:57:22 -0500
- Subject: Re: [Bug c++/13776] [tree-ssa] Many C++ compile-time regression in3.5-tree-ssa 040120
- Organization: Red Hat Canada
- References: <20040120183908.13776.kgardas@objectsecurity.com> <20040313020222.31274.qmail@sources.redhat.com> <4052FFA7.20305@tat.physik.uni-tuebingen.de>
On Sat, 2004-03-13 at 07:33, Richard Guenther wrote:
> I'm trying to look at some dump files now - looking at the optimized
> dump, is it ok to have loads of "duplicate" variable declarations in the
> dump? Like
>
If you use -fdump-tree-...-details you'll get the DECL_UID printed for
every variable. It's quite common to have different variables with the
same DECL_NAME. I've been thinking that we should emit DECL_UID
regardless of -details.
> ...
> struct Domain<3,DomainTraits<Loc<3> > > * const this;
> int d;
> int d;
> int <D1144454>;
> int d;
> int T.4364;
> ...
>
> and a lot more of int d farther away? Also, how are bb's marked? I see
> <bb 0>: but no more, and some gotos reference <bb 18> and <bb 16> (with
> a label, too)?
>
Yes, in GIMPLE, goto statements have a basic block as its destination.
The pretty printed output should also show a label. If you want to see
the block boundaries, use -blocks.
> Can I get summaries somehow here? Or just dump one interesting function
> rather than all of the program?
>
Not at the moment. We dump all or none.
> Also, how do I dump some stuff about the PRE pass? Specifying
> -fdump-tree-pre just dumps the trees after PRE with no information about
> the PRE pass itself.
>
-details is the switch used to enable verbosity.
Diego.