Re: Illegal C++

[Han-Wen Nienhuys]

Wiz Aus wrote:
> > > > Because we want to skip the (proxy) function calls for performance 
> > > > reasons.
> > > Um...lilypond uses an interpretive language (Scheme), goes via TEX 
> > > and PS to produce output, and you're worried about a few extra C++ 
> > > function calls?  I would be flabbergasted if this made any measurable 
> > > difference to lilypond's performance.
> >
> >
> > Actually, the optimization was pretty trivial once I had the 
> > precomputed methods going. FWIW, the top of the profile shows a lot of 
> > these "lowlevel" function calls that take a large amount of total 
> > execution time because of how often they are called.
>
>
> Hmm...I reverse the right to remain highly skeptical!  Lilypond is very 
> slow as far as music typesetting software goes - too slow to use in a 
> GUI, for instance (Sibelius claims to reformat entire 100-page scores in 
> milliseconds - no idea how it does it).  

It doesn't, obviously. Try loading a 100-page midi file into it, and see 
how it fares. I suspect it does some clever caching. Also Sibelius' 
algorithms, contrary to what their marketing department says, are not 
that advanced. If you drag notes on-screen, then yes the spacing can be 
changed interactively, but it will happely put different notes on top of 
each other.

> If you really wanted
> performance, you'd scrap the Scheme and TeX layers.

Certainly. Could you send me a patch for that?

> > If we had to do it again, it is doubtful that we would use C++.  I 
> > actually generally hate C++ ,  so I don't care much about conformance. 
> > For the past 9 years, following C++ "conformance" has been a rocky 
> > road.  No compiler is completely standard conformant, so I just track 
> > the latest GCC releases for compatibility.
>
>
> I agree worrying about strict conformance is a bit of a waste of time - 
> but to deliberately make use of a surely undocumented behavioural aspect 
> of a particular compiler is just asking for trouble.

-Wno-pmf-conversion is thorougly documented in the GCC manual.

> I was actually seriously interested in trying to determine what it would 
> take to provide a GUI for lilypond music entry, but I really can't see 
> how it would be done without major changes to the existing code (for a 
> start, there's no easily replaceable interface that's used to do the 
> graphical drawing).

But there is; check out scm/framework-gnome.scm for a backend that 
outputs to a GNOME canvas.


--
 Han-Wen Nienhuys - address@hidden - http://www.xs4all.nl/~hanwen