Nicolas Cellier uploaded a new version of Kernel to project The Trunk:
http://source.squeak.org/trunk/Kernel-nice.1237.mcz==================== Summary ====================
Name: Kernel-nice.1237
Author: nice
Time: 23 May 2019, 9:17:06.927326 pm
UUID: 11d57971-88b6-4078-8ac6-91ba023d2ba6
Ancestors: Kernel-nice.1236
Fix sqrtRem in 32bits Squeak.
Since SmallInteger maxVal + 1 has only 4 bytes, self highBit bitShift: -5 may equal 0: we split in a single whole part and engage into an infinite loop.
=============== Diff against Kernel-nice.1236 ===============
Item was changed:
----- Method: LargePositiveInteger>>sqrtRem (in category 'mathematical functions') -----
sqrtRem
"See super. Use a divide and conquer method to perform this operation.
See Paul Zimmermann. Karatsuba Square Root. [Research Report] RR-3805, INRIA. 1999, pp.8. <inria-00072854>
https://hal.inria.fr/inria-00072854/PDF/RR-3805.pdf"
| n qr q s r sr a3a2 a1 a0 |
"Split self in 4 digits a3,a2,a1,a0 in base b,
such that most significant digit a3 >= b/4
It is not a problem to have a3 >= b,
so we can round b down to a whole number of bytes n"
+ n := (self highBit bitShift: -5) max: 1. "bitShift: -2 divide in 4 parts, bitShift: -3 round down in bytes"
- n := self highBit bitShift: -5. "bitShift: -2 divide in 4 parts, bitShift: -3 round down in bytes"
a3a2 := self butLowestNDigits: n * 2.
a1 := self copyDigitsFrom: n + 1 to: n * 2.
a0 := self lowestNDigits: n.
sr := a3a2 sqrtRem.
qr := (sr last bitShift: 8 * n) + a1 divideByInteger: (sr first bitShift: 1).
q := qr first normalize.
s := (sr first bitShift: 8 * n) + q.
r := (qr last normalize bitShift: 8 * n) + a0 - q squared.
r negative
ifTrue:
[r := (s bitShift: 1) + r - 1.
s := s - 1].
sr at: 1 put: s; at: 2 put: r.
^sr
!
Locked
