This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: ObjC/ObjC++: Fix try-catch failures on a darwin, second take (this one fixes them all)
- From: Mike Stump <mikestump at comcast dot net>
- To: Nicola Pero <nicola dot pero at meta-innovation dot com>
- Cc: gcc-patches at gnu dot org
- Date: Tue, 30 Nov 2010 12:17:39 -0800
- Subject: Re: ObjC/ObjC++: Fix try-catch failures on a darwin, second take (this one fixes them all)
- References: <E5284BD0-76AF-4A27-A027-B52BBAB19D31@meta-innovation.com>
On Nov 30, 2010, at 10:25 AM, Nicola Pero wrote:
> This is a new patch to fix the sjlj exception handling machinery on Darwin, following
> an idea I had during the recent discussions. And this one fully works :-)
Sweet. Aside from debug information, I don't know of any downside. I think this should be strictly more reliable than the previous code. Could you test debug information and ensure we don't see an extra voltile:
...
int bar;
...
(gdb) ptype bar
We want to ensure we don't get volatile int. If we do, we might need a line or two in the debugging backends to imagine a type that doesn't have volatile set. Also, if volatile is on the decl, there should be an early bailout, so that:
volatile int bar;
(gdb) ptype bar
does print volatile int.
> Ok to commit ?
Ok, and thanks for working on this.