Re: [rfc / remote protocol] Remote shared library events

[Daniel Jacobowitz <drow@false.org>]

On Fri, May 11, 2007 at 11:02:50AM -0700, Kevin Buettner wrote:
> On Wed, 9 May 2007 16:16:27 -0400
> Daniel Jacobowitz <drow@false.org> wrote:
> 
> > What do you think, Kevin especially?  I'm not entirely thrilled with
> > the way this interacts with other solib_ops vectors, but I'm not
> > unhappy with it either.
> 
> What in particular were you not thrilled with?

I had to add two new methods, add_one_library and remove_one_library,
that don't make sense for any of the other things using solib_ops.
I thought for a bit about making this something completely different,
so that you could have both a solib_ops and target-reported libraries;
I think we'll want multiple sorts of libraries someday, but today
I couldn't make it work sensibly.

Also, the in_dynsym_code method is a bit hokey; calling in_plt_section
happens to be right for SymbianOS, but I have a feeling Windows will
want something more (at least, the relevant test is gdb1555.exp, and
it fails on Windows once I add testsuite support to run it there).
I suspect this should be split into a solib_ops specific bit and an
architecture specific bit, or maybe just an architecture specific bit.

Neither of these are big problems, really.

> I've looked over the solib-specific changes.  They look reasonable to
> me.  I didn't understand the infrun changes.  I too would like to see
> that put into a different patch.

Coming right up (well, probably Monday at this point).  Test case for
it, too.

-- 
Daniel Jacobowitz
CodeSourcery