[Ecls-list] INLINE now works

Matthew Mondor mm_lists at pulsar-zone.net
Sun Dec 18 20:18:50 UTC 2011


On Sun, 18 Dec 2011 21:04:48 +0100
Juan Jose Garcia-Ripoll <juanjose.garciaripoll at googlemail.com> wrote:

> On Sun, Dec 18, 2011 at 8:33 PM, Juan Jose Garcia-Ripoll <
> juanjose.garciaripoll at googlemail.com> wrote:
> 
> >  Is inlining also possible for setf functions?
> >>
> >
> > Good question... Actually it does not work, but because of a stupid bug
> > that also prevents the compiler from analyzing any kind of compiler macro
> > for SETF forms. Will fix it later on.
> 
> 
> Thanks a lot for pointing out this stupid mistake, Matthew. I fixed the
> code and now SETF function names undergo the same compilation steps as
> ordinary functions. This means compiler macros and inlining works for them;
> so does inlining of slot accessors in sealed classes, which adds another
> 10% speed up in PPRINT, bringing the FPRINT/PRETTY microbenchmark to 20% of
> the original times :-)

Super :)
-- 
Matt




More information about the ecl-devel mailing list