|
Boost : |
Subject: Re: [boost] [gil] GIL_FORCEINLINE replaced with BOOST_FORCEINLINE?
From: Emil Dotchevski (emildotchevski_at_[hidden])
Date: 2013-03-22 15:36:56
On Fri, Mar 22, 2013 at 4:38 AM, Mathias Gaunard
<mathias.gaunard_at_[hidden]> wrote:
> On 12/03/13 18:41, Mateusz Loskot wrote:
>> In boost/gil/gil_config.hpp, there is GIL_FORCEINLINE defined,
>> but there is no easy way to undef/redefine it.
> I initially added the ability to redefine BOOST_FORCEINLINE to allow people
> to disable force-inlining for certain debugging scenarios.
Shouldn't GIL_FORCEINLINE be defined as BOOST_FORCEINLINE by default?
That way BOOST_FORCEINLINE would take effect by default but could be
overridden specifically for GIL.
Another thing is, I like defining different macros that match the
function they're used with, e.g. BOOST_INLINE_CRITICAL,
BOOST_INLINE_TRIVIAL, etc.which have a default mapping to inline or
BOOST_FORCEINLINE. That allows for a little bit more flexibility when
tinkering with inlining as a result of profiling.
-- Emil Dotchevski Reverge Studios, Inc. http://www.revergestudios.com/reblog/index.php?n=ReCode
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk