From @pmichaud

I've just removed S29-num/roots.t from spectest_regression,
since all of the tests were being skipped.

If someone cares to write a working 'roots' function for Rakudo,
then we can add it back. :-)

Thanks!

Pm

From @moritz

On Sun Jun 22 10​:04​:48 2008, pmichaud wrote​:

I've just removed S29-num/roots.t from spectest_regression,
since all of the tests were being skipped.

If someone cares to write a working 'roots' function for Rakudo,
then we can add it back. :-)

Not quite working, but I attached two Perl 6 implementation of roots() -
one shiny, lazy that doesn't work with current rakudo, and one stripped
down, ugly that rakudo parses. But it can't execute it yet, because it
relies on Num.polar, which isn't implemented yet.

If somebody feels like converting that to PIR, do that. If not, we'll
have to wait for runtime in Perl 6.

Cheers,
Moritz

From @moritz

roots.pl

From @moritz

roots_reduced.pl

The RT System itself - Status changed from 'new' to 'open'

From @azawawi

Hi moritz,

After looking at the code that you provided, i found that you used
x.polar(y) in @​result. i think you meant x.unpolar(y). Am i missing
something here?

/ahmad

On Wed Jul 09 09​:12​:28 2008, moritz wrote​:

On Sun Jun 22 10​:04​:48 2008, pmichaud wrote​:

I've just removed S29-num/roots.t from spectest_regression,
since all of the tests were being skipped.

If someone cares to write a working 'roots' function for Rakudo,
then we can add it back. :-)

Not quite working, but I attached two Perl 6 implementation of roots()
-
one shiny, lazy that doesn't work with current rakudo, and one
stripped
down, ugly that rakudo parses. But it can't execute it yet, because it
relies on Num.polar, which isn't implemented yet.

If somebody feels like converting that to PIR, do that. If not, we'll
have to wait for runtime in Perl 6.

Cheers,
Moritz

From @azawawi

This is roots_reduced.pl in PIR. It passes only 1 test in t/spec/S29-
num/roots.t

./ahmad &

From @azawawi

Index: src/builtins/any-num.pir
===================================================================
--- src/builtins/any-num.pir    (revision 31763)
+++ src/builtins/any-num.pir    (working copy)
@@ -21,7 +21,7 @@
 .namespace []
 .sub 'onload' :anon :init :load
     $P0 = get_hll_namespace ['Any']
-    '!EXPORT'('abs cis int log chr polar sqrt truncate unpolar', 'from'=>$P0)
+    '!EXPORT'('abs cis int log chr polar sqrt truncate unpolar roots', 'from'=>$P0)

     ##  pre-seed a random number generator
     $P0 = new 'Random'
@@ -89,11 +89,46 @@

 .namespace ['Any']
 .sub 'polar' :method :multi(_)
-    $N0 = self
-    .return 'list'($N0, 0)
+    .local num real, angle
+    real = self
+    angle = atan 0, real
+    .return 'list'(real, angle)
 .end

+=item roots

+=cut
+.sub 'roots' :method :multi(_)
+    .param int n
+    .local num x, mag, angle, abs, twoPI
+    .local pmc retv
+    N1 = pi()
+    N1 *= 2
+    twoPI = N1
+    x = self
+    P1 = 'polar'(x)
+    mag = P1[0]
+    angle = P1[1]
+    unless angle < 0 goto skip
+    angle += twoPI
+  skip:
+    $N0 = 1 / n
+    abs = mag ** $N0
+    P1 = new 'List'
+    retv = P1
+    I1 = 0
+  loop:
+    N1 = twoPI * I1
+    N1 += angle
+    N1 /= n
+    P1 = 'unpolar'(abs, N1)
+    retv.push(P1)
+    I1 += 1
+    if I1 < n goto loop
+  done:
+    .return (retv)
+.end
+
 =item rand()

 =cut

From @jnthn

Thanks to a patch from dukeleto++, roots is now implemented. Resolving
ticket.

@jnthn - Status changed from 'open' to 'resolved'