Re: [RFA] fix disassemble foo::bar::~bar

[Keith Seitz <keiths@redhat.com>]

On 02/22/2013 11:15 AM, Tom Tromey wrote:
> Keith> The above patch fixes Doug's test cases and also is regression-free on
> Keith> native x86_64-unknown-linux (Fedora 15).
>
> I tend to think this is preferable.  At least, as long as we have to
> keep destructor_name_p.

I was only attempting to point out a simpler alternative. As you rightly 
point out, though, it is debatable whether destructor_name_p is needed 
at all. As it is right now (with Doug's or my patch) we get,

(gdb) p foo::~bar
name of destructor must equal name of class

How much more useful is that than

(gdb) p foo::~bar
There is no field named ~bar

IMO, the second error is at least as useful, if not more so, than what 
we have now. Since destructor_name_p is only used in this one place, I 
would urge you/Doug to consider just nuking it and the test in c-exp.y

> This patch is ok with me, if it includes Doug's test case, though I'd
> like to hear from Doug about it as well.

I left the choice up to Doug. It's his ball after all.

Keith