[perl #121940] [@LARRY] say True but False # should be 'True' (advent2010-day19)

[David Warring via RT]

The behavious still seems the same:

% perl6 --v
This is Rakudo version 2016.07.1-135-g77724b2 built on MoarVM version 
2016.07-16-g85b6537
implementing Perl 6.c.
% perl6 -e'my $s = False but True; say $s; say $s.so;'
True
True

This is inconsistent with:

% perl6 -e'my $s = 0 but True; say $s; say $s.so;'
0
True

On Tue Aug 09 08:23:37 2016, coke wrote:
> On Wed May 21 13:36:58 2014, david.warring wrote:
> > [08:09]  r: my $value = True but False; say $value
> > [08:09] <+camelia> rakudo-jvm 196b4f: OUTPUT«(timeout)»
> > [08:09] <+camelia> ..rakudo-{parrot,moar} 196b4f: OUTPUT«False␤»
> > ...
> > [08:10]  rakudo-jvm is all "WARNING. PARADOX. MUST OVERHEAT."
> > [08:11] * dwarring heh
> > ...
> > [08:11]  I'll put in an rakudobug
> > ...
> > [08:18]  dwarring: just be aware, camelia has been known to
> > timeout on rakudo-jvm.
> > [08:18]  dwarring: so it might just be an artifact.
> > [08:21]  n: my $value = True but False; say $value
> > [08:21] <+camelia> niecza v24-109-g48a8de3: OUTPUT«True␤»
> > [08:21]  m: my $value = True but False; say $value
> > [08:21] <+camelia> rakudo-moar 196b4f: OUTPUT«False␤»
> > [08:21]  I guess it is about that
> > [08:22]  this expects 'True' -
> > http://perl6advent.wordpress.com/2010/12/19/day-19-false-truth/
> > [08:24]  I kind of expect True, too.
> > [08:24] == ajr_ [~quas...@96-44-210-169.ip.tor.radiant.net] has quit
> > [Ping timeout: 240 seconds]
> > [08:25]  r: my $value = True but False; say ?$value
> > [08:25] <+camelia> rakudo-{parrot,jvm,moar} 196b4f: OUTPUT«False␤»
> > [08:25]  because the *value* is still True, kind of. it's just
> > the boolean aspect of it that has been overridden.
> > [08:25]  n: my $value = True but False; say ?$value
> > [08:25] <+camelia> niecza v24-109-g48a8de3: OUTPUT«False␤»
> > [08:25]  niecza++
> 
> This is from an old advent post that was pre-Christmas; is this
> actually a bug in current rakudo we should fix, or is the advent post
> out of date?

[perl #131174] [NYI] Atrribute Introspect shape

[David Warring via RT]

I've added a fudged test to S12-introspection/attributes.t
that the container shape is as declared; not a Whatever.

On Wed, 19 Apr 2017 03:14:35 -0700, c...@zoffix.com wrote:
> On Tue, 18 Apr 2017 20:34:15 -0700, david.warring wrote:
> > AFAIK there's currently no way of introspecting the shape of a Class
> > attribute that's a native shaped array.
> > 
> > I'd like this feature for the Native::Packing module. For example to
> > interpret:
> > 
> > class C {
> > has byte @.gif-header[6];
> > has uint16 $.width;
> > has uint16 $height;
> > }
> > 
> > At the moment, I just can't determine the size of shaped native arrays,
> > such as @.gif-header above:
> > 
> > say .type, .container
> > for C.^attributes;
> > 
> > (Positional[byte])(array[byte])
> > (uint16)(uint16)
> > (uint16)(uint16)
> 
> 
> Seems .container does have method .shape, but it returns a Whatever:
> 
> class C {
> has byte @.gif-header[6];
> has uint16 $.width;
> has uint16 $height;
> }
> say .container.shape for C.^attributes[0]; # (*)
> dd C.new.gif-header.shape; # (6,)
> 
> my byte @z[6];
> dd @z.shape; # (6,)

Re: [perl #132073] [BUG] isa method errors when called on a subset type object

[David Warring via RT]

Tests added with roast commit
https://github.com/perl6/roast/commit/d776a06e52c35d6cbb7b7bbade7b7a15b97ecff8

One remaining todo test for the subset of a subset case, ie:

subset S of Int; subset S2 of S; say S2.isa(S)

On Sun, Sep 17, 2017 at 6:34 AM, David Warring 
wrote:

> After that commit: subset S of Int; S.isa(True) returns true as expected.
>
> I've noticed a quibble with subset of a subset:
>
> perl6 -e'subset S of Int; subset S2 of S; say S2.isa(S)'
> False
>
> Should be True.
>
> On Sun, Sep 17, 2017 at 3:17 AM, Aleks-Daniel Jakimenko-Aleksejev via RT <
> perl6-bugs-follo...@perl.org> wrote:
>
>> What about this commit?
>> https://github.com/rakudo/rakudo/commit/0704cd97226e63001943
>> 42c88cef1c5fe711
>>
>> On 2017-09-12 13:55:51, david.warring wrote:
>> > current behavior of isa method on a subset:
>> >
>> > % perl6 -v
>> > This is Rakudo version 2017.08-110-g5f3350656 built on MoarVM version
>> > 2017.08.1-156-g4
>> > 9b90b99
>> > implementing Perl 6.c.
>> > % perl6 -e'subset S of Int; say S.isa(Int)'
>> > Cannot resolve caller isa(Perl6::Metamodel::SubsetHOW: S, Int); none of
>> > these signatur
>> > es match:
>> > (Mu \SELF: Mu $type, *%_)
>> > (Mu \SELF: Str:D $name, *%_)
>> > in block  at -e line 1
>> >
>> > The documentation https://docs.perl6.org/routine/isa implies this
>> should
>> > return True.
>>
>>
>

Re: [perl #132073] [BUG] isa method errors when called on a subset type object

[David Warring via RT]

After that commit: subset S of Int; S.isa(True) returns true as expected.

I've noticed a quibble with subset of a subset:

perl6 -e'subset S of Int; subset S2 of S; say S2.isa(S)'
False

Should be True.

On Sun, Sep 17, 2017 at 3:17 AM, Aleks-Daniel Jakimenko-Aleksejev via RT <
perl6-bugs-follo...@perl.org> wrote:

> What about this commit?
> https://github.com/rakudo/rakudo/commit/0704cd97226e6300194342c88c
> ef1c5fe711
>
> On 2017-09-12 13:55:51, david.warring wrote:
> > current behavior of isa method on a subset:
> >
> > % perl6 -v
> > This is Rakudo version 2017.08-110-g5f3350656 built on MoarVM version
> > 2017.08.1-156-g4
> > 9b90b99
> > implementing Perl 6.c.
> > % perl6 -e'subset S of Int; say S.isa(Int)'
> > Cannot resolve caller isa(Perl6::Metamodel::SubsetHOW: S, Int); none of
> > these signatur
> > es match:
> > (Mu \SELF: Mu $type, *%_)
> > (Mu \SELF: Str:D $name, *%_)
> > in block  at -e line 1
> >
> > The documentation https://docs.perl6.org/routine/isa implies this should
> > return True.
>
>

[perl #132073] [BUG] isa method errors when called on a subset type object

[David Warring via RT]

Final issues resolved with 
https://github.com/rakudo/rakudo/commit/cee1be22cff6153506e31df2916f8a0be27b5fc8

On Sat, 16 Sep 2017 12:12:35 -0700, david.warring wrote:
> Tests added with roast commit
> https://github.com/perl6/roast/commit/d776a06e52c35d6cbb7b7bbade7b7a15b97ecff8
> 
> One remaining todo test for the subset of a subset case, ie:
> 
> subset S of Int; subset S2 of S; say S2.isa(S)
> 
> On Sun, Sep 17, 2017 at 6:34 AM, David Warring
> 
> wrote:
> 
> > After that commit: subset S of Int; S.isa(True) returns true as
> > expected.
> >
> > I've noticed a quibble with subset of a subset:
> >
> > perl6 -e'subset S of Int; subset S2 of S; say S2.isa(S)'
> > False
> >
> > Should be True.
> >
> > On Sun, Sep 17, 2017 at 3:17 AM, Aleks-Daniel Jakimenko-Aleksejev via
> > RT <
> > perl6-bugs-follo...@perl.org> wrote:
> >
> >> What about this commit?
> >> https://github.com/rakudo/rakudo/commit/0704cd97226e63001943
> >> 42c88cef1c5fe711
> >>
> >> On 2017-09-12 13:55:51, david.warring wrote:
> >> > current behavior of isa method on a subset:
> >> >
> >> > % perl6 -v
> >> > This is Rakudo version 2017.08-110-g5f3350656 built on MoarVM
> >> > version
> >> > 2017.08.1-156-g4
> >> > 9b90b99
> >> > implementing Perl 6.c.
> >> > % perl6 -e'subset S of Int; say S.isa(Int)'
> >> > Cannot resolve caller isa(Perl6::Metamodel::SubsetHOW: S, Int);
> >> > none of
> >> > these signatur
> >> > es match:
> >> > (Mu \SELF: Mu $type, *%_)
> >> > (Mu \SELF: Str:D $name, *%_)
> >> > in block  at -e line 1
> >> >
> >> > The documentation https://docs.perl6.org/routine/isa implies this
> >> should
> >> > return True.
> >>
> >>
> >

Re: [perl #132222] [BUG] 'HAS' Embedded C-Structs not working as documented

[David Warring via RT]

Thanks for that. As a worka-around Rakudo seems to do a better job, if I
give it a helping hand viz a TWEAK method:

use NativeCall;
class Point is repr('CStruct') {
has uint8 $.x;
has uint8 $.y;
}

class MyStruct2 is repr('CStruct') {
HAS Point $.point;  # <-- embedded
has int8 $.flags;
method TWEAK {
$!point := Point.new;
}
}

say .name, ': ', .inlined for MyStruct2.^attributes;

my $s = nativecast(MyStruct2, buf8.new: 42,69,11);
say $s.flags;
say $s.point.defined;
say $s.point.x;
say $s.point.y;

Produces:

$!point: 1
$!flags: 0
11
True
42
69



On Fri, Oct 6, 2017 at 5:12 AM, Brian S. Julin via RT <
perl6-bugs-follo...@perl.org> wrote:

> On Wed, 04 Oct 2017 23:19:19 -0700, david.warring wrote:
> > From
> > https://docs.perl6.org/language/nativecall#Embedding_
> CStructs_and_CUnions
> >
> > class Point is repr('CStruct') {
> > has uint16 $.x;
> > has uint16 $.y;
> > }
> >
> > class MyStruct2 is repr('CStruct') {
> > HAS Point $.point;  # <-- embedded
> > has int32 $.flags;
> > }
> >
> > my $s = MyStruct2.new;
> > say $s.flags;
> > say $s.point.defined;
> > say $s.point.x;
> >
> > Produces:
> >
> > 0
> > False
> > Cannot look up attributes in a Point type object
> >  in block  at /tmp/tst.pl line 15
> >
> > Ie, the embedded struct is not automatically created and defined, when a
> > new containing struct is created.
> >
> > Rakudo version 2017.09-199-gc91c40115 built on MoarVM version
> > 2017.09.1-62-g89ca8eb0
> > implementing Perl 6.c.
>
>
> This is more an NYI than a bug... the code for that is more or less
> a first-pass draft.  There are workarounds (see my xcb module), but they
> require surrendering your sanity to dark forces.
>
>

[perl #128000] [NativeCall] Feature Request: Pointer Arithmetic

[David Warring via RT]

On Tue, 26 Apr 2016 15:24:57 -0700, sortiz wrote:
> Hi David,
> 
> Can you give a try to NativeHelpers::Pointer, part of NativeHelpers::Blob?
> 
> When well tested I plan push it to core.
> 
> Regards.
> 
> Salvador Ortiz. 

Hi Salvador,
That looks. OK. Can it go straight into core?

Re: [perl #131965] Shaped arrays can't have zero size

[David Warring via RT]

Attached is my use case which is parsing of PDF cross reference indices.
There are normally three numeric entries per line. e.g.

xref
0 8
00 65535 f
09 0 n
74 0 n
000120 0 n

Which populates nicely into an array of 'n' lines of shape 3.

There's the rare, but possible case of an xref section with zero entries

xref
0 0

(I've encountered real-world PDF's like this).

The attached grammar/actions attempted to handle it as an array of shape
[0;3],
but wont work on current Rakudo.

Just my preferred solution for the above case.
- David


On Sun, Aug 27, 2017 at 12:38 PM, Brandon Allbery via RT <
perl6-bugs-follo...@perl.org> wrote:

> > Well, what do you mean? Of course you can't put anything into it, and any
> > attempt to index it will throw. It may seem useless, however, if you can
> > have
> > an empty array, why can't you have a shaped empty array?
> >
>
> Only if all dimensions are unindexable. Otherwise you have something with
> inaccessible slots (and, potentially thereby, not truly leaked but
> unusable, memory).
>
> And an empty unshaped array, in addition to being usable as such
> (degenerate case of all shaped dimensions being size 0), can be made
> non-empty. Shaped arrays cannot.
>
>

grammar XRef {
rule  TOP  { xref\n+ }
rule  xref-section 

[perl #128000] [NativeCall] Feature Request: Pointer Arithmetic

[David Warring via RT]

Thanks Salvador,

Now in Rakudo core 
https://github.com/rakudo/rakudo/commit/3ca6554fdd8ff91da5423e85d4a2b7d309949531
On Tue, 03 Oct 2017 10:41:41 -0700, david.warring wrote:
> On Tue, 26 Apr 2016 15:24:57 -0700, sortiz wrote:
> > Hi David,
> > 
> > Can you give a try to NativeHelpers::Pointer, part of NativeHelpers::Blob?
> > 
> > When well tested I plan push it to core.
> > 
> > Regards.
> > 
> > Salvador Ortiz. 
> 
> Hi Salvador,
> That looks. OK. Can it go straight into core?

[perl #131919] [RFC] Returning Failure from failed P6-level .parse

[David Warring via RT]

On Mon, 21 Aug 2017 07:38:56 -0700, alex.jakime...@gmail.com wrote:
> It was removed completely for 2017.08 release.
> 
> Rakudo commit:
> https://github.com/rakudo/rakudo/commit/465d91abdfda038cb7feda35f7966be4ec39acf3
> Discussion: https://irclog.perlgeek.de/perl6-dev/2017-08-21#i_15048995
> On 2017-08-19 15:31:01, alex.jakime...@gmail.com wrote:
> > I had to revert one of the nqp commits related to .parse:
> > https://github.com/perl6/nqp/commit/d4d77b66c46c57de800b147df61fe486b4486acd
> >
> > Here's a ticket for the module that was affected by the change:
> > https://github.com/p6-css/CSS-Module-p6/issues/10
> >
> > Most certainly we want this change back ASAP, but it had to be
> > reverted for the
> > release.
> >
> > ♥
> >
> > On 2017-08-19 05:36:36, alex.jakime...@gmail.com wrote:
> > > The change was moved to v6.d in this commit:
> > >
> >
> https://github.com/rakudo/rakudo/commit/d2278b471cb0bd691dc7a1235fbcb9318ff5d61f
> > >
> > > On 2017-08-17 08:50:48, alex.jakime...@gmail.com wrote:
> > > > See this commit:
> > > >
> > >
> >
> https://github.com/rakudo/rakudo/commit/9501edae4f73a970e3270e3b0336a7b3045d3329
> > > >
> > > > These roast commits:
> > > > *
> > > >
> > >
> >
> https://github.com/perl6/roast/commit/1fb68c4b7a7c975f26fc81ad79f000958d1b4afd
> > > > *
> > > >
> > >
> >
> https://github.com/perl6/roast/commit/b53616f8e67f9b19366008b3abf55400a3d6cd2b
> > > >
> > > > And this justification:
> > > > * https://irclog.perlgeek.de/perl6-dev/2017-08-16#i_15021994
> > > >
> > > > This blog post noticing the breakage due to the change:
> > > > * https://gfldex.wordpress.com/2017/08/17/parsing-nothing/
> > > >
> > > > And these thoughts about postponing it to v6.d:
> > > > * https://irclog.perlgeek.de/perl6-dev/2017-08-17#i_15032160
> > > >
> > > >
> > > > I am confident that we will not be able to process this change
> > > > and
> > > > all
> > > > potential breakage associated with it in ≈3 days before the
> > > > release,
> > > > so the revert is coming.
> > > >
> > > > Personally, I don't mind rereverting it afterwards for inclusion
> > > > in
> > > > 2017.09, assuming we make sure to fix all modules that were
> > > > relying
> > > > on
> > > > Nils being returned. However, I do see the point for postponing
> > > > it
> > > > till v6.d.
> > > >
> > > >
> > > > So should we feel adventurous and push this change as early as we
> > > > can
> > > > (in 2017.09, aftear at least one month)? Or is it better to be
> > > > safe
> > > > and wait for v6.d? Please discuss.

I've had a chance to golf this (somewhat) from CSS::Module. The following fails 
to parse if I revert 
https://github.com/perl6/nqp/commit/d4d77b66c46c57de800b147df61fe486b4486acd

grammar Tst {
rule local {:i('local')'(' [ \'\' ] ')'}
rule declaration-list {  * }
rule declarations {
'{'  '}'
}
rule test-prop { (src) ':'   +% ',' }
}
say Tst.parse('{  src: local(\'Gentium\') }', :rule )

[perl #131666] NativeCall MoarVM panic

[David Warring via RT]

Can confirm this is still current behaviour with Rakudo 2019. I have came 
across exactly the same issue and found this ticket. I've noticed, If I change 
the example to catch exceptions. It then runs:

sub start-element($, $elem, $attr)
{
say "open $elem".indent($depth * 4);
say $attr.elems;
++$depth;
CATCH { default { warn "whoops: $_" } }
}

Produces:

open calendar
whoops: Don't know how many elements a C array returned from a library
  in block  at /tmp/tst.p6 line 27
open advent
whoops: Don't know how many elements a C array returned from a library
  in block  at /tmp/tst.p6 line 27
...etc

Underlying issue seems to be that any uncaught exception in a NativeCall Perl 
callback currently results in a unfriendly NativeCall MoarVM panic. The 
exception itself is lost.

On Tue, 27 Jun 2017 05:59:08 -0700, vladimir.ma...@oracle.com wrote:
> I took the NativeCall sample from here:
> 
> https://perl6advent.wordpress.com/2015/12/21/day-21-nativecall-backs-
> and-beyond-c/
> 
> 
> use NativeCall;
> 
> sub XML_SetElementHandler(OpaquePointer $parser,
>  (OpaquePointer, Str, CArray[Str]),
>(OpaquePointer, Str))
> is native('expat') { ... }
> 
> sub XML_ParserCreate(Str --> OpaquePointer)   is
> native('expat') { ... }
> sub XML_ParserFree(OpaquePointer) is
> native('expat') { ... }
> sub XML_Parse(OpaquePointer, Buf, int32, int32 --> int32) is
> native('expat') { ... }
> 
> my $xml = q:to/XML/;
> 
> 
> 
> 
> 
> XML
> 
> my $depth = 0;
> 
> sub start-element($, $elem, $attr)
> {
> say "open $elem".indent($depth * 4);
> ++$depth;
> }
> 
> sub end-element($, $elem)
> {
> --$depth;
> say "close $elem".indent($depth * 4);
> }
> 
> my $parser = XML_ParserCreate('UTF-8');
> XML_SetElementHandler($parser, , );
> 
> my $buf = $xml.encode('UTF-8');
> XML_Parse($parser, $buf, $buf.elems, 1);
> 
> XML_ParserFree($parser);
> 
> 
> 
> All works fine, unless I try to use the $attr in start-element
> 
> sub start-element($, $elem, $attr)
> {
> say "open $elem".indent($depth * 4);
> say $attr.elems;
> ++$depth;
> }
> 
> 
> The program then terminates with:
> 
> MoarVM panic: Internal error: Unwound entire stack and missed handler
> 
> 
> I have seen that on Solaris, but Linux has the same issue. In both
> cases
> custom build perl6 using 'rakudobrew build moar' so I suppose latest
> git revision.
> 
> Thank you