Re: possible bug?

[Michael Meeks <michael ximian com>]

Hi Dou Wen,

On Sat, 2002-03-09 at 05:48, dou wen wrote:
>       So pain :-( 
>       It's my pleasure :-)

	You masochist you ;-)

> >The best solution is (I think) to have a wrapper for your
> >malloc ( specific to the GIOP allocation code ) that ensures that the
> >memory is 8 byte aligned - but we do need that assertion, could you
> >knock up a simple patch (and comment) for that vs HEAD ORBit2 ?
> 
> Sure, here is my patch, but do not know if it is suitable.:-)

	Ok; I was thinking that we should add a:

	g_assert (memory & 0x3 == 0);

	sort of thing, in ORBit2 - with a nice comment for people who get a
blown assertion - that the memory needs aligning.

	For the Kernel Orbit stuff, yes you need something like the code you
outlined.

	Problem is, moving the pointer like that is going to cause no end of
grief if you try to free that (moved) pointer, so the patch is not so
good as it stands.

	I'll commit a modified fix for you.

	Thanks,

		Michael.

-- 
 mmeeks@gnu.org  <><, Pseudo Engineer, itinerant idiot