From @moritz

08​:49 < moritz_> rakudo​: $_ = "foo"; s[] = 'bar'; .say
08​:50 <+p6eval> rakudo 989883​: OUTPUT«Could not find sub &s␤ in main
program body at line 22​:/tmp/b7_gMKxUFS␤»
08​:51 < moritz_> rakudo​: $_ = "foo"; s[f] = 'bar'; .say
08​:51 <+p6eval> rakudo 989883​: OUTPUT«baroo␤»
08​:51 < moritz_> so it doesn't commit to the [
08​:54 < TimToady> std​: $_ = "foo"; s[] = 'bar'; .say
08​:54 <+p6eval> std 625303c​: OUTPUT«===SORRY!===␤Null pattern not
allowed at /tmp/zVvv5rwrey line 1​:␤------> $_ = "foo"; s[⏏] = 'bar';
.say␤Parse failed␤FAILED 00​:01 121m␤»
08​:54 < TimToady> std​: $_ = "foo"; s[f] = 'bar'; .say
08​:54 <+p6eval> std 625303c​: OUTPUT«ok 00​:01 123m␤»
08​:54 < TimToady> rakudo should not backtrack there
09​:51 * moritz_ submits rakudobug, justin case

From @FROGGS

the behaviour has changed but the messsage is LTA​:

<FROGGS> rakudo​: $_ = "foo"; s[] = 'bar'; .say
<p6eval> rakudo bf472b​: OUTPUT«===SORRY!===␤Unrecognized regex
metacharacter ] (must be quoted to match literally) at line 2, near " =
'bar'; "␤»

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

From @FROGGS

<FROGGS> rakudo​: $_ = "foo"; s[] = 'bar'; .say
<p6eval> rakudo 4fb07b​: OUTPUT«===SORRY!===␤Null regex not allowed␤at
/tmp/W3qa158pjS​:1␤------> $_ = "foo"; s[⏏] = 'bar'; .say␤ expecting
any of​:␤ postfix␤ infix or meta-infix␤ infix
stopper␤ statement end␤ statement modi…
<FROGGS> rakudo​: $_ = "foo"; s[f] = 'bar'; .say
<p6eval> rakudo 4fb07b​: OUTPUT«baroo␤»

I believe this commit fixed it​:
rakudo/rakudo@d77b765#L1L1623

From @timo

committed a test for this in roast​: 4bd5589

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