From quester.pm@gmail.com

Hello,

I noticed finanalyst having problems with Rakudo on #perl6 on 2010-04-20​:

05​:36 finanalyst alpha​: my @​a = 1,2,3,4; my %h; %h{@​a[0]} = @​a[1 ..^ +@​a];
%h.perl.say
05​:36 p6eval alpha 30e0ed​: OUTPUT«{"1" => [2, 3, 4]}␤»
05​:36 finanalyst rakudo​: my @​a = 1,2,3,4; my %h; %h{@​a[0]} = @​a[1 ..^ +@​a];
%h.perl.say
05​:36 p6eval rakudo b05155​: OUTPUT«{"1" => Seq.new()}␤»

The fix was just to copy the perl method from Array.pm to Seq.pm (and change
the output from square brackets to parentheses.)

I also remembered how many times I have been annoyed with Rakudo going into
an infinite loop when trying to cope with Iterator.perl. For example,

my $a=(1,1,* + * ... *); $a.perl.say;

There was some discussion around 10​:54 UTC 2010-04-20 on #perl6 about the
Iterator.perl case, most of which doesn't lend itself to implementation
without drastic work on the compiler (the ability to serialize code and have
iterators be able to memoize and unget their values.) To make life somewhat
more tolerable before utopia arrives, I took one of TimToady's many
suggestions and cut the output short at forty entries.

I include some comments in Iterator.pm to indicate that the code is a
placeholder for future, better code; you might want to edit that for style.

Please let me know if you have any questions or suggestions.

Best regards,
Ira Kevin Byerly
quester_ on #perl6
quester on perlmonks
quester.pm@​gmail.com

From quester.pm@gmail.com

From 54ab1068bff26928c2920598a8d53ac1cab8ef21 Mon Sep 17 00:00:00 2001
From: quester <quester.pm@gmail.com>
Date: Tue, 20 Apr 2010 00:20:12 -1000
Subject: [PATCH] Fix Seq.perl to not say Seq.new().  Interim fix for Iterator.perl infinite loops.

---
 src/core/Iterator.pm |   13 +++++++++++--
 src/core/Seq.pm      |    4 ++++
 2 files changed, 15 insertions(+), 2 deletions(-)

diff --git a/src/core/Iterator.pm b/src/core/Iterator.pm
index b4c4d77..6545dd6 100644
--- a/src/core/Iterator.pm
+++ b/src/core/Iterator.pm
@@ -1,6 +1,15 @@
 augment class Iterator {
-    multi method perl() {
-        $.eager.perl;
+    multi method perl(:$limit = 40) {
+        # Doing this right is probably beyond mortal dwimmery,
+        # and even in simple cases it needs the ability to
+        # serialize code to have any hope of producing
+        # output like (1, 1, * + * ... *)
+        # rather than (1, 1, 2, 3, 5, {???} ... ???).
+        # Try to do something useful for debugging, for now.
+        my @a = self.batch($limit);
+        my $and_more = '';
+        $and_more = ', {???} ... ???' if @a == $limit;
+        '(' ~ @a.map({ $^a.perl }).join(', ') ~ $and_more ~ ')';
     }
 
     multi method elems() {
diff --git a/src/core/Seq.pm b/src/core/Seq.pm
index a7b95c2..b3459ff 100644
--- a/src/core/Seq.pm
+++ b/src/core/Seq.pm
@@ -88,6 +88,10 @@ augment class Seq {
     multi method fmt($format = '%s', $seperator = ' ') {
         self.map({ .fmt($format)}).join($seperator);
     }
+
+    method perl() {
+        '(' ~ self.map({ $^a.perl }).join(', ') ~ ')';
+    }
 }
 
 multi sub sort (@x, :&by = &infix:<cmp>) { @x.sort(&by) }
-- 
1.6.6.1

From @quester

This is a fix for RT #​74518 "[BUG] rakudo regression .perl". Sorry, I
didn't notice this bug report until after I submitted the patch. - quester

From @moritz

Applied by colomon++ as d952fc4b1d052c0b2f6cce2ae2f26ab813776a61

@moritz - Status changed from 'new' to 'resolved'