Re: [PATCH] Add fullname field for MI -break-info command

[Vladimir Prus <ghost at cs dot msu dot su>]

Bob Rossi wrote:

> On Wed, Jan 25, 2006 at 04:31:19PM +0300, Vladimir Prus wrote:
>> On Wednesday 25 January 2006 00:13, Daniel Jacobowitz wrote:
>> 
>> > > +        {
>> > > +          struct symtab_and_line sal = find_pc_line
>> > > (b->loc->address,
>> > > 0); +          symtab_to_fullname (sal.symtab);
>> > > +
>> > > +          if (sal.symtab->fullname)
>> > > +            {
>> > > +              annotate_field(10);
>> >
>> >              annotate_field (10);
>> 
>> Is that a formatting change, or moving the "annotate_field" call out of
>> "if"?
>> 
>> > But I don't think you need this at all; there's no point emitting
>> > annotations when we're guarded by ui_out_is_mi_like_p.
>> 
>> For a future: can you explain what "annotate_field" does? The annotate.h
>> file has no comments at all, and gdbint has "annotate_field" only inside
>> code examples. And generally, what are "annotations"?
> 
> Annotations were the old interface used between GDB and front ends. It
> was all that was available before MI. For your own sanity, never ever
> try to deal with them.

Does it mean that:

1. No new code should ever contain a call to "annonate_*" functions?
2. It's planned to drop this "annotation" thing completely?

- Volodya