Re: [patch] make --disable-gdbmi work

[Aleksandar Ristovski <aristovski@qnx.com>]

Doug Evans wrote:
> On Jan 31, 2008 1:01 PM, Aleksandar Ristovski <aristovski@qnx.com> wrote:
>> Daniel Jacobowitz wrote:
>>> On Thu, Jan 31, 2008 at 03:50:27PM -0500, Aleksandar Ristovski wrote:
>>>> The attached patch fixes the problem.
>>> My opinion is that we don't want this functionality any more.  MI used
>>> to be optional, but it is a central part of GDB nowadays, and
>>> conditional compilation is hard to maintain.
>>>
>> It certainly adds some effort to maintenance, but it's not as bad as it may
>> sound, especially since the change is not extensive at all (see the patch).
>>
>> For embedded systems, it is important to have as small a binary as possible, and
>> if we can reduce it with reasonable effort, I would say, why not?
> 
> [fwiw]
> 
> The patch appeals to the minimalist side in me.  And given its trivial
> nature, it could go in without formally committing to the intended
> goal.  If later things get unwieldy we can revisit whether to support
> minimal builds then.
> 
> btw, do we have automated builds+tests for gdb?
> 

Yes, Vladimir correctly pointed out - my argument about reducing the code size 
is not really a good one...

I still think the patch is useful, if for no other reason, to keep dependencies 
single directional (mi -> gdb).