This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: g++ 3.1 EH broken for irix6.5
- To: Richard Kenner <kenner at vlsi1 dot ultra dot nyu dot edu>
- Subject: Re: g++ 3.1 EH broken for irix6.5
- From: Richard Henderson <rth at redhat dot com>
- Date: Thu, 28 Jun 2001 18:17:27 -0700
- Cc: gcc-patches at gcc dot gnu dot org
- References: <10106290104.AA02964@vlsi1.ultra.nyu.edu>
On Thu, Jun 28, 2001 at 09:04:42PM -0400, Richard Kenner wrote:
> There is, however, a requirement that there be some way to adjust
> the *maximum* permitted alignment.
>
> I don't follow. What do you mean by that term?
The maximum alignment that the compiler can assume when accessing
the structure.
> I think your change was well-meaning, but clashes with existing
> documented practice,
>
> I don't believe we ever documented anywhere what default alignment we
> gave to packed records.
Certainly we did, or at least implied it strongly enough as to
make no difference. Please re-read the "Type Attributes" node.
In particular:
`aligned (ALIGNMENT)'
The `aligned' attribute can only increase the alignment; but you
can decrease it by specifying `packed' as well. See below.
This asserts that "packed" affects alignment, not just removal of
padding. One has to imply that it affects alignment by setting
it to the minimum, i.e. one unit. Which is what we used to do,
and what folks rely on.
r~