[Chicken-hackers] [PATCH] Speedup in walk-generic and generate less LET statements

2012-02-19 Thread Peter Bex
Hi all, Another week, another patch :) The first patch gets rid of the calls to SRFI-1's EVERY and MAP in favor of a hand-rolled loop. This is slightly more verbose, but by doing it this way we can avoid some recursive consing which MAP performs (if nodes are equal, no need to reverse the list)

Re: [Chicken-hackers] [PATCH] Speedup in walk-generic and generate less LET statements

2012-02-27 Thread Felix
Hi! Patch #1 (walk-generic) signed off and pushed. I still have to review the second one. cheers, felix ___ Chicken-hackers mailing list Chicken-hackers@nongnu.org https://lists.nongnu.org/mailman/listinfo/chicken-hackers

Re: [Chicken-hackers] [PATCH] Speedup in walk-generic and generate less LET statements

2012-03-05 Thread Felix
> The second patch is a bit more delicate and deserves some more > explanation. When a program initially gets translated into a > node tree, it is normalised to have all bodies transformed into > let-statements, as well as the toplevel. Signed off and pushed. cheers, felix

Re: [Chicken-hackers] [PATCH] Speedup in walk-generic and generate less LET statements

2012-03-05 Thread Peter Bex
On Mon, Mar 05, 2012 at 02:00:31PM +0100, Felix wrote: > > The second patch is a bit more delicate and deserves some more > > explanation. When a program initially gets translated into a > > node tree, it is normalised to have all bodies transformed into > > let-statements, as well as the toplevel