Re: [PATCH 16/22] Class-ify ui_out_level

[Pedro Alves <palves at redhat dot com>]

On 11/30/2016 01:47 PM, Antoine Tremblay wrote:
> 
> Pedro Alves writes:
> 
>> On 11/30/2016 12:40 PM, Antoine Tremblay wrote:

>> In this case it looks like it's documented?
> 
> Yes this case yes.  I was thinking of
> https://sourceware.org/ml/gdb-patches/2016-11/msg00973.html
> 

GCC does not require C++11, so it's natural their document
won't be saying anything about >= C++11 things.  Maybe they
won't mind it if we propose C++11 bits, if we want to.
(But let's leave that particular style discussion that that thread.)

> Also we allow the use of dynamic_cast while GCC doesn't...

We're already requiring exceptions, which requires (some kind of) rtti
under to hood, so allowing dynamic_cast would seem like a natural
consequence.  Not that I'd generally think that a design
relying on dynamic_cast is a sane design, TBC.

> 
> I'm sure other things will come up ?

Sure, in which case we can document them.  Here:

 https://sourceware.org/gdb/wiki/Internals%20GDB-C-Coding-Standards#C.2B-.2B--specific_coding_conventions

I had added the note about exceptions a while ago.  We can
certainly improve this while we go.

Thanks,
Pedro Alves