From @masak

<flussence> rn​: multi infix​:<%>($lhs, Str $rhs) is looser(&infix​:<xx>)
{ $lhs.join​: $rhs }; say '........' xx 8 % "\n"
<camelia> rakudo b2072f, niecza v24-66-gaee6525​: OUTPUT«........
........ ........ ........ ........ ........ ........ ........␤»
<flussence> those \s'es should be \n's...
<colomon> that doesn't seem to have worked?
<flussence> rn​: multi infix​:<%>($lhs, Str $rhs) is looser(&infix​:<xx>)
{ $lhs.join​: $rhs }; say ('........' xx 8) % "\n"
<camelia> rakudo b2072f, niecza v24-66-gaee6525​:
OUTPUT«........␤........␤........␤........␤........␤........␤........␤........␤»
<timotimo> yeah, because of $ instead of @​
<colomon> oh, nevermind
<timotimo> for the $lhs ;)
<masak> timotimo​: should still work, no?
<masak> so, why doesn't it work for @​lhs ?
<moritz> ah, I know what's going on
<quietfanatic> r​: multi infix​:<%>(\lhs, Str $rhs) { say "custom op";
lhs.join​: $rhs }; say ( infix​:<%>('........' xx 8, "\n") ).perl
<camelia> rakudo b2072f​: OUTPUT«custom
op␤"........\n........\n........\n........\n........\n........\n........\n........"␤»
<timotimo> oh, facepalm.
<moritz> since there's already an infix %, you can't define a new one
with a different precedence
<flussence> oh!
<moritz> so the precedence spec is simply ignored
<quietfanatic> oh, I think it'd be nice if you could do that
<timotimo> traits can be applied to the proto only?
<quietfanatic> or wait, is it because it's installing it as a multi?
<timotimo> can only be applied to the proto*
<quietfanatic> I see
<benabik> Multis with different precedence is probably a bad idea.
<quietfanatic> r​: sub infix​:<%%%>(\lhs, Str $rhs) is
looser(&infix​:<xx>) { say lhs.perl; lhs.join​: $rhs }; say ('........'
xx 8 %%% "\n" )
<camelia> rakudo b2072f​: OUTPUT«("........", "........", "........",
"........", "........", "........", "........",
"........").list␤........␤........␤........␤........␤........␤........␤........␤........␤»
<quietfanatic> you're right
<masak> could there maybe be a warning if you try to override the precedence?
<masak> it feels wrongish that it just throws it away.
<colomon> masak​: I think an error, even
* masak submits rakudobug
<masak> colomon​: better? :)
<colomon> masak++