Re: [PATCH 1/3] define and check itset

[Pedro Alves <palves@redhat.com>]

On 05/31/2012 03:23 PM, Yao Qi wrote:

> On 05/31/2012 09:34 PM, Pedro Alves wrote:
>> What good does it do to put this in ahead of its prerequisites?
>> Can we please stop trying to put the cart before the horse?
>> I'm trying to help with the async stuff, but the constant push
>> in trying to put other bits in first frustrates me.  :-/
>>
> 
> I don't want to frustrates anyone here.  This bit doesn't have any
> prerequisites, as I said, it is quite isolated from other parts.  I
> don't see anything wrong this bits go in first.  


It's useless on its own.

> The review process to
> patch series 'run all-stop on top of non-stop' is not smooth, which
> forces me to start to push this part first.


Sorry, but nothing should be forcing you to do that.  I wrote most of
that series, so its natural that others expect me to review it first; and most
probably nobody else feels very qualified to review it.  But I never considered
that code final, and I want to get back to play with the series a bit more,
not just review what I already had, but unfortunately, I keep getting distracted
with reviews, etc., so I'm progressing quite slowly.  Apologies.  But it'll
get done.  E.g., my recent new_thread_event patch is preparatory work.  I'm
working on some other cleanup that ends up addressing PR11692.  I need to
get back to the MI async issue.

>> > The itsets bits need to be last.  I'm not even sure the syntax is
>> > what we want to end up with at all.  Last I touched them, I had
>> > wanted to spend a while trying to unify the concept of "current
>> > thread" with the itset.  There's a disconnect that gets in the
>> > way a bit.

> I don't know your plan, and you are free to change whatever you'd like
> to.  AFAICS, the syntax looks fine to me, so I posted them.

My plan is to make async good enough to flip it on; hack on
all-stop-on-top-of-non-stop some more, and put that in.  And only
then we'll get back to itsets.

-- 
Pedro Alves