gmane.comp.lang.perl.perl6.compiler

[perl #113078] [BUG] $*PROGRAM_NAME erroneously reports the first argument to 'perl6', even when this argument isn't a program name in Rakudo

Carl Mäsak — 2012-05-24T18:54:32

# New Ticket Created by  "Carl Mäsak" 
# Please include the string:  [perl #113078]
# in the subject line of all future correspondence about this issue. 
# <URL: https://rt.perl.org:443/rt3/Ticket/Display.html?id=113078 >


<moritz> $ perl6 -Ilib -e 'say $*PROGRAM_NAME'
<moritz> -Ilib
<moritz> /o\
* masak submits $*PROGRAM_NAME bug

[perl #113042] [BUG] [WEIRD] LTA internals-leaky error message when doing 'for our $:: ();' in Rakudo

Carl Mäsak — 2012-05-22T20:19:12

# New Ticket Created by  "Carl Mäsak" 
# Please include the string:  [perl #113042]
# in the subject line of all future correspondence about this issue. 
# <URL: https://rt.perl.org:443/rt3/Ticket/Display.html?id=113042 >


<masak> r: for our $:: ();
<p6eval> rakudo dca0fa: OUTPUT«===SORRY!===␤ResizablePMCArray: index
out of bounds!␤»
* masak submits LTA error diakopter++ rakudobug

[perl #113040] [BUG] [WEIRD] LTA error message when doing 'for my $/ ();' in Rakudo

Carl Mäsak — 2012-05-22T20:17:31

# New Ticket Created by  "Carl Mäsak" 
# Please include the string:  [perl #113040]
# in the subject line of all future correspondence about this issue. 
# <URL: https://rt.perl.org:443/rt3/Ticket/Display.html?id=113040 >


<masak> r: for my $/ ();
<p6eval> rakudo dca0fa: OUTPUT«===SORRY!===␤Contextual %*PARAM_INFO not found␤»
* masak submits LTA error diakopter++ rakudobug

[perl #113026] array iterator does not track a growing array

Patrick R. Michaud via RT — 2012-05-21T18:15:07

I suspect this bug may be related to #112716 
(https://rt.perl.org/rt3/Ticket/Display.html?id=112716); the .map() 
operation (which 'for' uses) tends to be a bit too eager in evaluating 
its invocant list.

I've taken both bugs and will work on a fix shortly.

Pm

[perl #113026] a much better bug report

Siddhant Saraf — 2012-05-21T16:52:45

(so please ignore my initial report, because it was lazily written,
here is the proper bugreport, restated with clarity)

The following loop should iterate infinitely, but it does not:

    my < at >array = 1, 2;

    for < at >array -> $n {
        $n.say;
        < at >array.push: $n + 2;
    }

    say < at >rray;

$ perl6 --version
This is perl6 version 2012.05-15-gdca0fa6 built on parrot 4.4.0
revision RELEASE_4_4_0

bug discovered: http://irclog.perlgeek.de/perl6/2012-05-21#i_5619326

[perl #113026] array iterator does not track a growing array

Siddhant Saraf — 2012-05-21T16:06:27

# New Ticket Created by  Siddhant Saraf 
# Please include the string:  [perl #113026]
# in the subject line of all future correspondence about this issue. 
# <URL: https://rt.perl.org:443/rt3/Ticket/Display.html?id=113026 >


$ perl6 -v
This is perl6 version 2012.05-15-gdca0fa6 built on parrot 4.4.0
revision RELEASE_4_4_0

<sisar> so here is the same code in perl6 and python:
https://gist.github.com/2762953. While the perl6 the loop terminates,
in pyhton it is an infinite loop. Though both behaviour have their
reasons, I can't decide which is better.
<sisar> why python's behaviour is logical, perl6's behaviour is trying
to prevent one from potentially infinite loops
...
<sisar> so, perl6 makes a temporary copy of < at >array before it even
starts looping. Why ?
...
<flussence> it counts the array then iterates over it
...
<flussence> copying every array just to loop over it would be crazy
inefficient, I think only PHP does that (and only because it uses copy
on write)
<sisar> flussence: what are merits of that, counting before iterating ?
<sisar> flussence: aye, making a copy would be super inefficient.
<flussence> if you have an SQL result pretending to be an array, you
don't want it going off and calling a .count that does "select
count(*)" on every iteration
<sisar> flussence: i'm sorry i'm not familiar with SQL
<geekosaur> don't worry about it.  point is that an array might not be
an array, it might be a big chunk of code pretending to be an array
<geekosaur> and you don't want to hammer that big chunk of code pointlessly
<flussence> it'd be like calling strlen() in C on a really big string
...
<TimToady> the whole point of lazy lists is that you never ask for the
size of something you're iterating over if you can help it
<TimToady> and I think an array iterator should be able to track a
growing array without snapshotting the old length
<flussence> that makes sense, but I can't imagine it being fun to implement :)
...
<sisar> TimToady: so it is a genuine bug? fwiw, both niecza and rakudo
iterate only twice.
<TimToady> I think so

[perl #112996] sequence deduction tries to convert strings to numbers, should not

Moritz Lenz — 2012-05-19T05:31:18

# New Ticket Created by  Moritz Lenz 
# Please include the string:  [perl #112996]
# in the subject line of all future correspondence about this issue. 
# <URL: https://rt.perl.org:443/rt3/Ticket/Display.html?id=112996 >


< moritz> r: say ('1a' ... *)[^5]
<+p6eval> rakudo 45679a: OUTPUT«1a 1b 1c 1d 1e␤»
< moritz> r: say ('1a', '1b' ... *)[^5]
<+p6eval> rakudo 45679a: OUTPUT«Cannot convert string to number:
trailing characters after number in '1⏏b' (indicated by ⏏)␤
in method Numeric at src/gen/CORE.setting:9058␤
in sub infix:<-> at src/gen/CORE.setting:2381␤
in sub infix:<-> at src/gen/CORE.setting:2381␤

I'd argue that the sequence operator should never try to do arithmetic
with strings; it can fail, and it does.

[perl #112986] error 'Use of uninitialized $_ of type Any in numeric context" throws extra errors

Siddhant Saraf — 2012-05-18T20:05:09

# New Ticket Created by  Siddhant Saraf 
# Please include the string:  [perl #112986]
# in the subject line of all future correspondence about this issue. 
# <URL: https://rt.perl.org:443/rt3/Ticket/Display.html?id=112986 >


<sisar> r: $_**2;
<p6eval> rakudo 45679a: OUTPUT«use of uninitialized variable $_ of
type Any in numeric context  in block <anon> at /tmp/uUxfmOEKJJ:1␤␤»
<sisar> that gives a different error locally
<sisar> it says "use of uninitialized variable $_ of type Any in
numeric contextMethod 'message' not found for invocant of class 'Any'
"
<moritz> the second half looks like an error during error reporting :/
<moritz> ah, you run it in the REPL
<sisar> yup
...
<moritz> but it's a bug that needs fixing
<moritz> r: say ($_ ** 2).WHAT
<p6eval> rakudo 45679a: OUTPUT«use of uninitialized variable $_ of
type Any in numeric context  in block <anon> at
/tmp/eWivcSD3YY:1␤␤Int()␤»
<sisar> so should I report a rakudobug, or are you fixing it right now ?
<moritz> please report

[perl #112988] excessive errors when foo.pl does not exist

Siddhant Saraf — 2012-05-18T20:24:36

# New Ticket Created by  Siddhant Saraf 
# Please include the string:  [perl #112988]
# in the subject line of all future correspondence about this issue. 
# <URL: https://rt.perl.org:443/rt3/Ticket/Display.html?id=112988 >


$ perl6 foo
Error while reading from file: Unable to open filehandle from path 'foo.pl'
current instr.: 'nqp;HLL;Compiler;_block1994' pc 35586
(src/stage2/gen/NQPHLL.pir:13627) (src/stage2/gen/NQPHLL.pm:1453)
called from Sub 'nqp;HLL;Compiler;evalfiles' pc 35251
(src/stage2/gen/NQPHLL.pir:13472) (src/stage2/gen/NQPHLL.pm:1439)
called from Sub 'nqp;HLL;Compiler;_block1868' pc 34069
(src/stage2/gen/NQPHLL.pir:12887) (src/stage2/gen/NQPHLL.pm:1387)
called from Sub 'nqp;HLL;Compiler;command_eval' pc 33637
(src/stage2/gen/NQPHLL.pir:12701) (src/stage2/gen/NQPHLL.pm:1402)
called from Sub 'nqp;Perl6;Compiler;command_eval' pc 1205
(src/gen/perl6-compiler.pir:485) (src/Perl6/Compiler.nqp:18)
called from Sub 'nqp;HLL;Compiler;command_line' pc 33259
(src/stage2/gen/NQPHLL.pir:12525) (src/stage2/gen/NQPHLL.pm:1353)
called from Sub 'MAIN' pc 1029 (src/gen/perl6.pir:392) (src/main.nqp
src/gen/main-version.nqp:35)
called from Sub '_block1000' pc 105 (src/gen/perl6.pir:73)
(src/main.nqp src/gen/main-version.nqp:1)

now, that's too much information. Ideally it should be something like
what perl(5) says :

$ perl foo.pl
Can't open perl script "foo.pl": No such file or directory

It clearly says "no such file".

[Announce] Rakudo 2012.05 "MadMongers"

Tadeusz Sośnierz — 2012-05-17T17:59:28

    Announce: Rakudo Perl 6 compiler development release #52 ("MadMongers")

On behalf of the Rakudo development team, I'm glad to announce the
May 2012 release of Rakudo Perl #52 "MadMongers". Rakudo is an
implementation of Perl 6 on the Parrot Virtual Machine (see 
<http://www.parrot.org>). The tarball for this release
is available from <http://github.com/rakudo/rakudo/downloads>.

Please note: This announcement is not for the Rakudo Star
distribution[*] -- it's announcing a new release of the compiler only.
For the latest Rakudo Star release, see
<http://github.com/rakudo/star/downloads>.

The Rakudo Perl compiler follows a monthly release cycle, with each
release named after a Perl Mongers group. The May 2012 release is
code named after MadMongers.

This release includes a whole lot of changes since the last one, including:

* -I and -M command-line options
* support for non-Int enums
* 'use' now accepts positional arguments and is able to import by tag name
* 'import' now works
* basic support for Version literals
* %*ENV now propagates into subprocesses
* basic implementation of pack and unpack ported from 'ng' branch
* fff flip-flop operator is now implemented, ff has been improved
* various new regex features and improvements

Rakudo now also includes the lib.pm module.

This is only a small peek at the changes in this release. For a more
detailed list, see "docs/ChangeLog".

The development team thanks all of our contributors and sponsors for
making Rakudo Perl possible, as well as those people who worked on 
Parrot, the Perl 6 test suite and the specification.

The following people contributed to this release:

Moritz Lenz, Jonathan Worthington, Patrick R. Michaud,
Jonathan Scott Duff, Tadeusz Sośnierz, Carl Masak, Will "Coke" Coleda,
Marcus Ramberg, kboga, TimToady, Kyle Hasselbacher, Geir Amdal, JimmyZ,
benabik and gfldex.

If you would like to contribute, see <http://rakudo.org/how-to-help>,
ask on the perl6-compiler< at >perl.org mailing list, or ask on IRC #perl6
on freenode.

The next release of Rakudo (#53) is scheduled for June 21, 2012. A
list of the other planned release dates and code names for 2012 is
available in the "docs/release_guide.pod" file. In general, Rakudo
development releases are scheduled to occur soon after each Parrot
monthly release. Parrot releases the third Tuesday of each month.

On behalf of the development team, I encourage you to try the new release.
Have fun, and let us know about your experience.

Kind regards,
Tadeusz Sośnierz

[*] What's the difference between the Rakudo compiler and the Rakudo
    Star distribution?  

    The Rakudo compiler is a compiler for the Perl 6 language.
    Nothing else.

    The Rakudo Star distribution is the Rakudo compiler plus a selection
    of useful Perl 6 modules, the most recent incarnation of the "Using
    Perl 6" book, and other software that can be used with the Rakudo
    compiler to enhance its utility.  Rakudo Star is meant for early
    adopters who wish to explore what's possible with Rakudo Perl 6 and
    provide feedback on what works, what doesn't, and what else they
    would like to see included in the distribution.

[perl #111286] [BUG] Name capture happens in the presence of a quantifier but not in its absence in Rakudo

Patrick R. Michaud via RT — 2012-05-17T15:50:46

Now fixed in 0ed00f0.... probably needs spectests to close ticket.

Thanks!

Pm

[perl #112970] [BUG] LTA, misleading error message when a typed array in a signature fails to bind in Rakudo

Carl Mäsak — 2012-05-17T15:06:23

# New Ticket Created by  "Carl Mäsak" 
# Please include the string:  [perl #112970]
# in the subject line of all future correspondence about this issue. 
# <URL: https://rt.perl.org:443/rt3/Ticket/Display.html?id=112970 >


<spider-mario> r: sub f(Int < at >x) {}; f([])
<p6eval> rakudo d43876: OUTPUT«Nominal type check failed for parameter
'< at >x'; expected Positional but got Array instead [...]
<jnthn> grr, really need to fix that to say Positional[Int]
<masak> spider-mario: the error message is misleading.
* masak submits rakudobug

[perl #72822] [BUG] Strange error when doing 0..WHAT in Rakudo

Carl Mäsak via RT — 2012-05-17T13:53:27

This "works" now.

<masak> rakudo: say 0..WHAT
<p6eval> rakudo d43876: OUTPUT«===SORRY!===␤CHECK FAILED:␤Calling 'WHAT' 
will never work with no arguments (line 1)␤ Expected: :(\$x)␤»
* masak closes https://rt.perl.org/rt3/Ticket/Display.html?id=72822

[perl #70229] [BUG] Failure stringifies to '' in Rakudo, shouldn't

Carl Mäsak via RT — 2012-05-17T13:49:51

This looks fixed to me, and tested. Resolving ticket.

[perl #68112] [BUG] .sort doesn't accept &rand (or zero-arg subs in general) in Rakudo

Carl Mäsak via RT — 2012-05-17T13:45:54

already
sort
transform.

Changing my mind on this one. use .pick(*) if you want .sort(&rand) 
semantics. Wrap your nullary function in a unary or binary function if 
you want something more advanced.

Rejecting ticket.

[perl #61764] [TODO] Implement ff and fff

Carl Mäsak via RT — 2012-05-17T13:37:34

And now they're both implemented, and I'm told they have tests.

[perl #112956] [BUG] Misleading error message when trying to augment enums in Rakudo

Carl Mäsak — 2012-05-15T21:23:57

# New Ticket Created by  "Carl Mäsak" 
# Please include the string:  [perl #112956]
# in the subject line of all future correspondence about this issue. 
# <URL: https://rt.perl.org:443/rt3/Ticket/Display.html?id=112956 >


<masak> r: use MONKEY_TYPING; enum Weekday <Mon Tue Wed Thu Fri Sat
Sun>; augment class Weekday { multi method succ($  where Sun:) { Mon }
}; say "alive"
<p6eval> rakudo f69e58: OUTPUT«===SORRY!===␤Cannot auto-generate a
proto method in the setting␤»
<masak> why am I getting this error message?
<jnthn> masak: Because it failed to detect the meta-object mismatch at
the point you started augmenting.
<jnthn> masak: And then EnumHOW ain't set up to support multi-methods
being added.
<masak> where does the setting enter into the picture?
<jnthn> masak: The code path that issues that is the one that you hit
if you don't explicitly write a proto when declaring a multi in the
setting.
<masak> so it tries to figure out what went wrong, and arrives at a
common cause which isn't applicable in this case?
<jnthn> Right.
* masak submits LTA rakudobug
<jnthn> Well, what's really LTA is it let you get that far.
<masak> aye.
<jnthn> Which I do want to fix. So yes, please file.

[perl #112922] [TODO] Add compiler errors for the case of non-typechecking defaults in signatures in Rakudo

Carl Mäsak — 2012-05-14T12:59:07

# New Ticket Created by  "Carl Mäsak" 
# Please include the string:  [perl #112922]
# in the subject line of all future correspondence about this issue. 
# <URL: https://rt.perl.org:443/rt3/Ticket/Display.html?id=112922 >


< masak> r: sub foo(Bool $b = sub { False }) {}; say "alive"
<p6eval> rakudo abb8f3: OUTPUT«alive␤»
<jnthn> r: sub foo(Bool $b = sub { False }) {}; foo(); say "alive"
<p6eval> rakudo abb8f3: OUTPUT«Nominal type check failed for parameter
'$b'; expected Bool but got Sub instead [...]
<jnthn> phew
<masak> :)
<masak> jnthn: would it be possible to get a compiler warning (or even
an error) for the case of non-smartmatching defaults?
<jnthn> masak: If you mean "non-typechecking"?
<jnthn> Then probably, yes.
<jnthn> If we know the type of the RHS, which we do in this case.
<jnthn> Guess an error is fine.
<jnthn> I mean, if it's gonna fail anyway... :)
<masak> not necessarily.
<masak> not if you always call it with the argument provided.
<masak> but it's a ticking bomb, for sure.
<masak> jnthn: will you be helped by me submitting a TODO RT ticket 'bout it?
<masak> should I make a spec change saying that copmpilers should
explicitly check for this?
<jnthn> masak: We don't tend to spec such things as "must". But a TODO
RT is fine.
<jnthn> Or just implement it...it's only gonna be 3 or so lines of code. :)
* masak submits TODO rakudobug, and might come around to implementing it leytur
<jnthn> (in parameter action method...you already have the
nominal_type to hand, and can see if the default has a known compile
time value, which we already consider anyway)

[perl #112910] enum mixin leads to non-transitive type checking

Moritz Lenz — 2012-05-13T09:29:56

# New Ticket Created by  Moritz Lenz 
# Please include the string:  [perl #112910]
# in the subject line of all future correspondence about this issue. 
# <URL: https://rt.perl.org:443/rt3/Ticket/Display.html?id=112910 >


< moritz> r: enum A <b c>; say b ~~ A
<+p6eval> rakudo 2ac569: OUTPUT«True␤»
< moritz> r: enum A <b c>; say ("foo" but b) ~~ A
<+p6eval> rakudo 2ac569: OUTPUT«False␤»
< moritz> r: enum A <b c>; say ("foo" but b) ~~ b
<+p6eval> rakudo 2ac569: OUTPUT«True␤»
< moritz> wait. b ~~ A, but ("foo" but b) ~~ b  and ("foo" but b) !~~ A?
< moritz> that seems... odd
< moritz> I kinda thought our type system was transitive :-)
 * moritz submits rakudobug

[perl #112908] cannot use result of enum invocation with infix:, even though it is identical to the enum value

Moritz Lenz — 2012-05-13T09:21:42

# New Ticket Created by  Moritz Lenz 
# Please include the string:  [perl #112908]
# in the subject line of all future correspondence about this issue. 
# <URL: https://rt.perl.org:443/rt3/Ticket/Display.html?id=112908 >


< moritz> r: enum day <a b c d e>;say day(3) === d; "foo" but d; say
"alive"; "foo" but day(3)
<+p6eval> rakudo 2ac569: OUTPUT«True␤alive␤Type check failed
          in assignment to '$!day'; expected 'day' but got 'Int'␤
          in block  <anon> at src/gen/CORE.setting:650␤
          in method BUILD_LEAST_DERIVED at src/gen/CORE.setting:637␤
          in sub infix:<but> at src/gen/CORE.setting:10579␤
< moritz> day(3) === d  is True, but '"foo" but d' and '"foo" but
day(3)'  dies?
< moritz> r: enum A <b c>; nqp::say nqp::iscont(b); nqp::say
                nqp::iscont(A(0))
<+p6eval> rakudo 2ac569: OUTPUT«0␤0␤»

So that's not the problem either.

[perl #112906] Enum type object invocation leads to segmentation fault

Moritz Lenz — 2012-05-13T08:41:20

# New Ticket Created by  Moritz Lenz 
# Please include the string:  [perl #112906]
# in the subject line of all future correspondence about this issue. 
# <URL: https://rt.perl.org:443/rt3/Ticket/Display.html?id=112906 >


< moritz> r: enum day (:Sun(1), 'Mon', 'Tue', 'Wed', 'Thu', 'Fri',
'Sat'); say day(Tue).Str
<+p6eval> rakudo 2ac569: OUTPUT«Method 'Str' not found for invocant of
class 'Undef'␤  in block <anon> at /tmp/dR0d80NyvO:1␤␤»
< moritz> r: enum day (:Sun(1), 'Mon', 'Tue', 'Wed', 'Thu', 'Fri',
Sat'); say day(Tue) eq day(3)
<+p6eval> rakudo 2ac569: OUTPUT«(signal SEGV)»
* moritz submits rakudobug

Further golfed by diakopter:
10:38 < diakopter> r: enum day (); ~day()
10:38 <+p6eval> rakudo 2ac569: OUTPUT«(signal SEGV)»


that's what S12-enums/pseudo-functional.t does.