Re: Patches to fix GCC?s C++ exception handling on NetBSD/VAX

[Jake Hamby <jehamby420%me.com@localhost>]

I see we have been working along similar lines. :-) I have rebased my previous patches on top of GCC 5.3 and will send them as an attachment on the more recent post to the list. I also have some ideas about bugs that may be lurking in the optimizer backend due to the way that condition codes work, and how that could be fixed.

Regards,
Jake


> On Mar 23, 2016, at 06:21, Valery Ushakov <uwe%stderr.spb.ru@localhost> wrote:
> 
> Hi, Jake.
> 
> On Mon, Mar 21, 2016 at 14:03:51 -0700, Jake Hamby wrote:
> 
>>> http://jakehamby.com/?p=7
> 
> Cool, thanks for working on that.
> 
> Have you seen my old http://gnats.netbsd.org/44155 pr?  It's been 5
> years, so I completely forgot the details.  You seems to have
> identified more or less the same problems, so I would appreciate if
> you can cast a quick look at my old notes before I close the pr.
> 
> My change to vax.c:
> 
> -      dwarf2out_def_cfa (label, FRAME_POINTER_REGNUM, -(offset - 4));
> +      dwarf2out_def_cfa (label, FRAME_POINTER_REGNUM, -(offset - 8));
> 
> is probably not needed in your patch b/c you fix STARTING_FRAME_OFFSET
> properly.
> 
> The first part of that hunk (use mask instead of repeating the test)
> is not needed since upstreem made the same change already it seems.
> 
> The first hunk - as far as I can tell your change does more or less
> the same except that it uses existing loop instead of just assingning
> the constant.  I think the latter is more explicit and is easier to
> read.
> 
> Thanks again.
> 
> -uwe