Re: [PATCH] Add Unicode strings and symbols

2009-08-08 Thread Mike Gran
On Sun, 2009-08-02 at 16:40 -0700, Mike Gran wrote:
> Hi-
> 
> I think I scared Ludo when I committed something, so let me try this
> instead.  Attached please find the next patch toward Unicode strings. It
> is a big patch, but, it is the smallest patch I could make that added
> the next quantum of functionality yet returned master to a working
> state.
> 
> 

Since I see no objection, I'll push this.

Thanks,

Mike






Re: [PATCH] Add Unicode strings and symbols

2009-08-08 Thread Mike Gran
On Sat, 2009-08-08 at 02:44 -0700, Mike Gran wrote:
> On Sun, 2009-08-02 at 16:40 -0700, Mike Gran wrote:
> > Hi-
> > 
> > I think I scared Ludo when I committed something, so let me try this
> > instead.  Attached please find the next patch toward Unicode strings. It
> > is a big patch, but, it is the smallest patch I could make that added
> > the next quantum of functionality yet returned master to a working
> > state.
> > 
> > 
> 
> Since I see no objection, I'll push this.
> 
> Thanks,
> 
> Mike

Freakin' git.  I was surprised to see that a commit named "Merge commit
'origin/master'" appeared in the main tree when I did that.

Sorry.

-Mike





Re: guile performance - Ackermann function: way slower than emacs, slower still if compiled

2009-08-08 Thread Ludovic Courtès
Hello!

Andy Wingo  writes:

> Perhaps the deal is that 1+ is not being inlined. It seems to be calling
> out to the 1+ procedure, which is a primitive, actually making it a
> bit slower. I'll see what I can do.

Currently, invoking subrs conses, which is Bad(TM).  Code that keeps
invoking `1+' and `1-' must suffer from this.

Thanks,
Ludo'.





Re: Elisp performance

2009-08-08 Thread Ludovic Courtès
Hi!

Andy Wingo  writes:

> Dynamic-wind, on the other hand, does pose a problem. The problem is
> that dynamic-wind is a primitive -- to call it, the VM saves its
> registers, conses together its arguments, and passes them to the
> evaluator. The evaluator destructures the consed list, eventually finds
> the C function
[...]

IMO a pressing issue here is to avoid consing when invoking "rest-less"
subrs (à la `scm_i_gsubr_apply ()' [0]).

(FWIW this has been on my to-do list for some time but feel free to give
it a try!)

Thanks,
Ludo'.

[0] http://thread.gmane.org/gmane.lisp.guile.devel/8244