This is the mail archive of the
gdb-patches@sources.redhat.com
mailing list for the GDB project.
Re: [patch/rfc[ Opaque bcache
- From: Daniel Jacobowitz <drow at mvista dot com>
- To: gdb-patches at sources dot redhat dot com
- Date: Wed, 10 Jul 2002 18:22:54 -0400
- Subject: Re: [patch/rfc[ Opaque bcache
- References: <3D2CB213.4050607@ges.redhat.com>
On Wed, Jul 10, 2002 at 06:15:47PM -0400, Andrew Cagney wrote:
> Hello,
>
> This makes the bcache object opaque. Testing so far hasn't shown
> problems. I'll look to commit it in a few days.
>
> comments? Hmm, did I remember to update the makefile ....
Nope to updating the Makefile, and nope to comments:
> +/* The type used to hold a single bcache string. The user data is
> + stored in d.data. Since it can be any type, it needs to have the
> + same alignment as the most strict alignment of any type on the host
> + machine. I don't know of any really correct way to do this in
> + stock ANSI C, so just do it the same way obstack.h does.
> +
> + It would be nicer to have this stuff hidden away in bcache.c, but
> + struct objstack contains a struct bcache directly --- not a pointer
> + to one --- and then the memory-mapped stuff makes this a real pain.
> + We don't strictly need to expose struct bstring, but it's better to
> + have it all in one place. */
That's not accurate any more. Other than that, this looks great; and
it removes a lot of &'s we used to need when accessing a bcache, which
I quite like.
--
Daniel Jacobowitz Carnegie Mellon University
MontaVista Software Debian GNU/Linux Developer