From @masak

<diakopter> rnp​: my %a{%a}
<camelia> rakudo 13f2f9​: OUTPUT«===SORRY!=== Error while compiling
/tmp/K2iV1Xdt96␤Variable '%a' is not declared [...]
<camelia> ..pugs​: ( no output )
<camelia> ..niecza v24-88-g1f87209​: OUTPUT«===SORRY!===␤␤Variable %a
is not predeclared [...]
<diakopter> std​: my %a{%a}
<camelia> std c2215f0​: OUTPUT«ok 00​:00 43m␤»
<moritz> wait, what? It's supposed to be visible before the
declaration is finished?
<jnthn> stdbug :P
<masak> yeah, stdbug.
<TimToady> how so?
<masak> what? really?
<diakopter> :}
<jnthn> dude!
<TimToady> it's just not caring about the semantics of {%a}
<FROGGS> ahhh
* masak reluctantly submits rakudobug
<jnthn> masak​: um, why? :)
<TimToady> it's the same one we've already had
<TimToady> %a isn't introduced soon enough
<masak> TimToady​: different case, though.
<masak> and this one might well be harder.
<diakopter> good luck finding the other(s)...
<TimToady> not if you fix the first one right...
<masak> yeah, right :)
<diakopter> seems legit
* diakopter is feeling MeanFuzzer personality today
<masak> diakopter++ # MeanFuzzer

If you cannot follow the confused negotiation above... I don't blame
you. Anyway, the conclusion is that STD.pm6 is correct and Rakudo is
wrong. (And Niecza is wrong too, but this is a rakudobug ticket.)

@diakopter - Status changed from 'new' to 'resolved'