From @masak

<masak> rakudo​: class A {}; class B { has A $.foo .= new }
<p6eval> rakudo c4857a​: OUTPUT«too few positional arguments​: 1 passed,
2 (or more) expected [...]
<slavik> s1n​: it will take 1 month to do that ...
* masak submits rakudobug
<masak> alpha​: class A {}; class B { has A $.foo .= new }
<p6eval> alpha 30e0ed​: OUTPUT«error​:imcc​:syntax error, unexpected
'\n'␤ in file 'EVAL_25' line 186␤Null PMC access in
find_method('new') [...]

@coke - Status changed from 'new' to 'open'

From sohtil@gmail.com

See also #​77200 for a similar[1] error caused by using .= new with a
lexically declared class.

[1]...regarding the message. Don't know about the internals.

From @coke

On Sun May 02 08​:33​:11 2010, masak wrote​:

<masak> rakudo​: class A {}; class B { has A $.foo .= new }
<p6eval> rakudo c4857a​: OUTPUT«too few positional arguments​: 1 passed,
2 (or more) expected [...]
<slavik> s1n​: it will take 1 month to do that ...
* masak submits rakudobug
<masak> alpha​: class A {}; class B { has A $.foo .= new }
<p6eval> alpha 30e0ed​: OUTPUT«error​:imcc​:syntax error, unexpected
'\n'␤ in file 'EVAL_25' line 186␤Null PMC access in
find_method('new') [...]

This now complains about trying to append to an unitialized string,
which seems reasonable​:

15​:51 < [Coke]> rakudo​: class A {}; class B { has A $.foo .= new }
15​:51 <+p6eval> rakudo 142c41​: OUTPUT«Use of uninitialized value in
string
  context␤Use of uninitialized value in string
  context␤===SORRY!===␤error​:imcc​:syntax error,
unexpected DOT
  ('.')␤ in file '(file unknown)' line 371␤␤»

Closable with tests.

--
Will "Coke" Coleda

From @coke

On Fri, Oct 21, 2011 at 3​:52 PM, Will Coleda via RT
<perl6-bugs-followup@​perl.org> wrote​:

On Sun May 02 08​:33​:11 2010, masak wrote​:

<masak> rakudo​: class A {}; class B { has A $.foo .= new }
<p6eval> rakudo c4857a​: OUTPUT«too few positional arguments​: 1 passed,
2 (or more) expected [...]
<slavik> s1n​: it will take 1 month to do that ...
* masak submits rakudobug
<masak> alpha​: class A {}; class B { has A $.foo .= new }
<p6eval> alpha 30e0ed​: OUTPUT«error​:imcc​:syntax error, unexpected
'\n'␤   in file 'EVAL_25' line 186␤Null PMC access in
find_method('new') [...]

This now complains about trying to append to an unitialized string,
which seems reasonable​:

15​:51 < [Coke]> rakudo​: class A {}; class B { has A $.foo .= new }
15​:51 <+p6eval> rakudo 142c41​: OUTPUT«Use of uninitialized value in
string
               context␤Use of uninitialized value in string
               context␤===SORRY!===␤error​:imcc​:syntax error,
unexpected DOT
               ('.')␤     in file '(file unknown)' line 371␤␤»

Closable with tests.

--
Will "Coke" Coleda

Whoa. much less reasonable is that IMCC error hidden in the output
there. I take back the closable bit.

--
Will "Coke" Coleda

From @diakopter

On Fri Oct 21 12​:55​:20 2011, coke wrote​:

On Fri, Oct 21, 2011 at 3​:52 PM, Will Coleda via RT
<perl6-bugs-followup@​perl.org> wrote​:

On Sun May 02 08​:33​:11 2010, masak wrote​:

<masak> rakudo​: class A {}; class B { has A $.foo .= new }
<p6eval> rakudo c4857a​: OUTPUT«too few positional arguments​: 1
passed,
2 (or more) expected [...]
<slavik> s1n​: it will take 1 month to do that ...
* masak submits rakudobug
<masak> alpha​: class A {}; class B { has A $.foo .= new }
<p6eval> alpha 30e0ed​: OUTPUT«error​:imcc​:syntax error, unexpected
'\n'␤   in file 'EVAL_25' line 186␤Null PMC access in
find_method('new') [...]

This now complains about trying to append to an unitialized string,
which seems reasonable​:

15​:51 < [Coke]> rakudo​: class A {}; class B { has A $.foo .= new }
15​:51 <+p6eval> rakudo 142c41​: OUTPUT«Use of uninitialized value in
string
               context␤Use of uninitialized value in string
               context␤===SORRY!===␤error​:imcc​:syntax error,
unexpected DOT
               ('.')␤     in file '(file unknown)' line 371␤␤»

Closable with tests.

--
Will "Coke" Coleda

Whoa. much less reasonable is that IMCC error hidden in the output
there. I take back the closable bit.

it is now much more reasonable. marking testneeded.

13​:53 <diakopter> rakudo​: class A {}; class B { has A $.foo .= new }
13​:53 <p6eval> rakudo e2d876​: OUTPUT«===SORRY!===␤Cannot use .= to
initialize an attribute
  at line 2, near " }"␤»

From @bbkr

Test in

Raku/roast@84fa0c2df0

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

From @bbkr

Reopening because this is limitation of Rakudo, not Spec​:

moritz bbkr​: is that behavior specced? 22​:19
moritz or is that a limitation in rakudo? 22​:19
moritz nr​: class A { has Array $x .= new }; A.new 22​:20
p6eval niecza v19-13-g442e075​: OUTPUT«Unhandled exception​: Unable to resolve
method ctxzyg in type Method␤ at /home/p6eval/niecza/src/NieczaPassSimplifier.pm6 line
162 (run_optree @​ 8) ␤ at /home/p6eval/niecza/src/NieczaPassSimplifier.pm6 line 19
(NieczaPassSimplifier.invoke_incr @​ 5) ␤ at … 22​:20
p6eval ..rakudo 1fe39c​: OUTPUT«===SORRY!===␤Cannot use .= to initialize an
attribute at line 2, near " }; A.new"␤» 22​:20
sorear bbkr​: proper form these days is to smartmatch on types like X​::Foo
22​:20
sorear bbkr​: this doesn't work on niecza, I know :| but it's clearly the right way going
forward 22​:20
moritz sorear​: do you think the example above (has Type $.x .= new) should work?
22​:21
sorear moritz​: Yes 22​:21
moritz then the proper thing is to test that it works 22​:21
moritz and then both rakudo and niecza fail the test 22​:21
moritz except that rakudo has the nicer error message 22​:22
bbkr I'll reverse test logic and fudge it for Rakudo and Niecza 22​:22
sorear good point, I wasn't even looking at the test itself :D 22​:22
sorear moritz​: saying that Rakudo has the nicer error message implies that Niecza
has an error message at all, which I would dispute :D 22​:22
moritz sorear​: it prints something in the case of an error, so I'd call it an "error
message" :-) 22​:23
bbkr also RT should be reopened here, am I right? 22​:23
moritz yes

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

From @bbkr

fudged test in

Raku/roast@506d0a59c5

From @Mouq

Fixed in Rakudo commit​:

commit df5b31eae0cdecd36d7aa92ea07fdec4843de051
Author​: Mouq <alexmoquin@​gmail.com>
Date​: Sat Feb 14 20​:57​:56 2015 -0500

  Implement #​74850 "has Array $.x .= new(1, 2, 3);"

And the test bbkr++ mentioned was unfudged.

Closing as resolved :)

On Mon Jul 16 13​:33​:54 2012, bbkr wrote​:

fudged test in

Raku/roast@506d0a59c5

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