From @ShimmerFairy

<ShimmerFairy> m​: class Foo { }; multi sub infix​:<>(Foo $a, Foo $b) { 42.6 }; say Foo.new ~~ Foo; say infix​:<>(Foo.new, Foo)
<camelia> rakudo-moar a38b59​: OUTPUT«True␤42.6␤»
<ShimmerFairy> (basically, the Perl 6 grammar treats ~~ specially, so stuff like this and the chaining bug pop up)

To explain, the ~~ infix is one of those operators that src/Perl6/Actions.pm treats specially, in this case constructing an AST for the expression with the 'make_smartmatch' method. The QAST generated by this method bypasses the infix and uses the ACCEPTS method on the arguments, meaning definitions of infix​:<> are never used for the operator. (It's potentially possible that the infix​:<> functions found in the core setting aren't used either.)

Bug #​81336 may also be caused by this special treatment, since the generated QAST (as mentioned) gives no consideration to any defined infix​:<~~> functions and shortcuts to calling ACCEPTS on the left-hand side.

From @moritz

Hi,

On 04.08.2015 00​:52, Faye (via RT) wrote​:

# New Ticket Created by Faye
# Please include the string​: [perl #​125745]
# in the subject line of all future correspondence about this issue.
# <URL​: https://rt-archive.perl.org/perl6/Ticket/Display.html?id=125745 >

<ShimmerFairy> m​: class Foo { }; multi sub infix​:<>(Foo $a, Foo $b) { 42.6 }; say Foo.new ~~ Foo; say infix​:<>(Foo.new, Foo)
<camelia> rakudo-moar a38b59​: OUTPUT«True␤42.6␤»
<ShimmerFairy> (basically, the Perl 6 grammar treats ~~ specially, so stuff like this and the chaining bug pop up)

I'd argue that it's permissible to not consider custom infix​:<~~>,
because we already provide an avenue for customizing smart-match
behavior (method ACCEPTS of the RHS type).

Also note that the standard infix​:<~~> cannot be a normal operator,
because it topicalizes the RHS.

So, not a bug.

(But I cannot reject it in rt.perl.org right now because of login
problems; I hope I'll be able to do so later).

Cheers,
Moritz

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

@moritz - Status changed from 'open' to 'rejected'

From @labster

I'm actually going to reopen this one, because I think it needs the same treatment as #​125427 -- an error message that you're doing something which will never work. See commit 792578840.

On Mon Aug 03 23​:50​:07 2015, moritz wrote​:

Hi,

On 04.08.2015 00​:52, Faye (via RT) wrote​:

# New Ticket Created by Faye
# Please include the string​: [perl #​125745]
# in the subject line of all future correspondence about this issue.
# <URL​: https://rt-archive.perl.org/perl6/Ticket/Display.html?id=125745 >

<ShimmerFairy> m​: class Foo { }; multi sub infix​:<>(Foo $a, Foo $b)
{ 42.6 }; say Foo.new ~~ Foo; say infix​:<>(Foo.new, Foo)
<camelia> rakudo-moar a38b59​: OUTPUT«True␤42.6␤»
<ShimmerFairy> (basically, the Perl 6 grammar treats ~~ specially, so
stuff like this and the chaining bug pop up)

I'd argue that it's permissible to not consider custom infix​:<~~>,
because we already provide an avenue for customizing smart-match
behavior (method ACCEPTS of the RHS type).

Also note that the standard infix​:<~~> cannot be a normal operator,
because it topicalizes the RHS.

So, not a bug.

(But I cannot reject it in rt.perl.org right now because of login
problems; I hope I'll be able to do so later).

Cheers,
Moritz

@labster - Status changed from 'rejected' to 'open'

From @usev6

I created a PR for rakudo (following RT #​125427)​: rakudo/rakudo#487

$ perl6-m -e 'class Foo { }; multi sub infix​:<>(Foo $a, Foo $b) { 42.6 }; say Foo.new ~~ Foo; say infix​:<>(Foo.new, Foo)'
===SORRY!=== Error while compiling -e
Cannot override infix operator '', as it is a special form handled directly by the compiler
at -e​:1
------> class Foo { }; multi sub infix​:<>⏏(Foo $a, Foo $b) { 42.6 }; say Foo.new ~

From @lizmat

Merged.

On 04 Aug 2015, at 23​:16, Christian Bartolomaeus via RT <bugs-comment@​bugs6.perl.org> wrote​:

I created a PR for rakudo (following RT #​125427)​: rakudo/rakudo#487

$ perl6-m -e 'class Foo { }; multi sub infix​:<>(Foo $a, Foo $b) { 42.6 }; say Foo.new ~~ Foo; say infix​:<>(Foo.new, Foo)'
===SORRY!=== Error while compiling -e
Cannot override infix operator '', as it is a special form handled directly by the compiler
at -e​:1
------> class Foo { }; multi sub infix​:<>⏏(Foo $a, Foo $b) { 42.6 }; say Foo.new ~

From @usev6

I added a test with commit d2d4f7a9.

In comments to PR 487 it was suggested to make the error message even better by mentioning .ACCEPTS (rakudo/rakudo#487 (comment)).

Should we leave the ticket open until someone does that or is the ticket closeable?

From @jnthn

On Fri Aug 14 13​:30​:13 2015, bartolin@​gmx.de wrote​:

I added a test with commit d2d4f7a9.

In comments to PR 487 it was suggested to make the error message even
better by mentioning .ACCEPTS
(rakudo/rakudo#487 (comment)).

Should we leave the ticket open until someone does that or is the
ticket closeable?

Did the hint​:

$ perl6-m -e "class Foo { }; multi sub infix​:<>(Foo $a, Foo $b) { 42.6 }"
===SORRY!=== Error while compiling -e
Cannot override infix operator '', as it is a special form handled directly by the compiler
(consider implementing an ACCEPTS method)
at -e​:1
------> class Foo { }; multi sub infix​:<~~><HERE>(Foo $a, Foo $b) { 42.6 }

So categorically closable now. :-)

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