Re: [Xen-devel] [PATCH 3/7] vm-event: introduce vm_event_vcpu_enter

[Corneliu ZUZU <czuzu@xxxxxxxxxxxxxxx>]

On 6/17/2016 2:27 PM, Jan Beulich wrote:
> > > > On 17.06.16 at 13:13, <czuzu@xxxxxxxxxxxxxxx> wrote:
> > On 6/17/2016 10:17 AM, Jan Beulich wrote:
> > > (And to be clear, I much appreciate any form of reduction of the
> > > sometimes extremely long lists of #include-s, just not [apparently
> > > or really] randomly mixed with other, substantial changes. That's
> > > namely because it's not clear whether source files should explicitly
> > > include everything they need, or instead be allowed to rely on
> > > headers they include to include further headers they also
> > > _explicitly_ rely on.
> > Personally I prefer the former since I think it also cuts down
> > compilation time.
> > Having header H include every header Ni needed by source S makes H
> > unnecessarily bulky at compilation time for other sources <> S that
> > don't need headers Ni but which depend on H nonetheless.
> I nowhere said every _header_ should include everything any of
> its _consumers_ would require. My point was solely about source
> files. For example, if S depends on both H1 and H2, and H2
> already includes H1, whether S then needs to just include H2, or
> should also explicitly include H1 (such that S doesn't need changing
> when the inclusion of H1 by H2 goes away).
>
> Jan
Ah, ok got it.
I restate my view of treating these "issues" with automation tools 
rather than leaving the programmer to do primitive work that he 
shouldn't be required to do with a nowadays programming language.
Clang for example offers a powerful parsing library (can parse GCC too) 
with python bindings, it would be nice to take more advantage of that.
Corneliu.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel