steve
/
factor
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

heaps: add comments referring to pypy's heapq.py

freebsd-work
Jon Harper 2019-01-10 21:50:50 +01:00
parent cc70de32a7
commit 1c28d326b1
1 changed files with 9 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
basis/heaps/heaps.factor
View File

@ -6,6 +6,9 @@ kernel.private locals math math.order math.private sequences
sequences.private summary vectors ; sequences.private summary vectors ;
IN: heaps IN: heaps
! names and optimizations copied from pypy's heapq.py,
! refer to it from in depth explanations of the optimizations.
GENERIC: heap-push* ( value key heap -- entry ) GENERIC: heap-push* ( value key heap -- entry )
GENERIC: heap-peek ( heap -- value key ) GENERIC: heap-peek ( heap -- value key )
GENERIC: heap-pop* ( heap -- ) GENERIC: heap-pop* ( heap -- )
@ -86,6 +89,7 @@ M: heap heap-peek ( heap -- value key )
<PRIVATE <PRIVATE
! called bubble-up in the litterature... but we keep pypy's name.
:: sift-down ( heap from to -- ) :: sift-down ( heap from to -- )
heap data>> :> data heap data>> :> data
to data data-nth :> tmp to data data-nth :> tmp
@ -115,6 +119,11 @@ M: heap heap-push*
<PRIVATE <PRIVATE
! called bubble-down in the litterature... but we keep pypy's name.
! A quote from pypy's implementation:
! > We *could* break out of the loop as soon as we find a pos where newitem <=
! > both its children, but turns out that's not a good idea [...]
! Indeed the code is 33% slower if we remove this optmization.
:: sift-up ( heap n -- ) :: sift-up ( heap n -- )
heap data>> :> data heap data>> :> data
data length :> end data length :> end