From @lizmat

[13​:33​:31] <lizmat> m​: class A { has $.a = "a"; method b { "aaa".subst(/$!a/,"b") }}; say A.new.a; say A.new.b # why isn't the .subst working ?
[13​:33​:32] <+camelia> rakudo-moar 41574a​: OUTPUT«a␤aaa␤»
[13​:34​:31] <lizmat> m​: class A { has $.a = "a"; method b { my $x = $!a; "aaa".subst(/$x/,"b") }}; say A.new.b # putting it into a lexical first works
[13​:34​:33] <+camelia> rakudo-moar 41574a​: OUTPUT«baa␤»
[13​:34​:50] <lizmat> rakudobug?
[13​:35​:55] <lizmat> FROGGS, jnthn​: ^^^
[13​:36​:47] <jnthn> m​: /$!foo/
[13​:36​:48] <+camelia> rakudo-moar 41574a​: OUTPUT«�[31m===�[0mSORRY!�[31m===�[0m Error while compiling /tmp/GrGGxT5ir_␤Variable $!foo used where no 'self' is available␤at /tmp/GrGGxT5ir_​:1␤------&gt; �[32m/$!foo�[33m⏏�[31m/�[0m␤»
[13​:36​:56] <FROGGS> m​: class A { has $.a = "a"; method b { "aaa".subst(/"$!a"/,"b") }}; say A.new.a; say A.new.b
[13​:36​:57] <+camelia> rakudo-moar 41574a​: OUTPUT«a␤use of uninitialized value of type Any in string context␤baaa␤»
[13​:37​:21] <FROGGS> that seems to be the runtime lexpad problem thingy masak++ had described in his blög
[13​:37​:35] <jnthn> Um
[13​:37​:38] <Ven> class variables are in lexpads ?
[13​:37​:49] <jnthn> Wait...a regex is method-ish, and has its own self...
[13​:37​:56] <FROGGS> eww
[13​:38​:16] <FROGGS> m​: class A { has $.a = "a"; method b { "aaa".subst($!a,"b") }}; say A.new.a; say A.new.b
[13​:38​:17] <lizmat> aha... the plot thickens... rakudobug it ?
[13​:38​:18] <+camelia> rakudo-moar 41574a​: OUTPUT«a␤baa␤»
[13​:38​:24] <FROGGS> yeah
[13​:38​:28] <FROGGS> jnthn​: is right
[13​:38​:56] <jnthn> lizmat​: Um, at this point all we have is "it looks odd" rather than "it should do X"
[13​:38​:59] <FROGGS> m​: class A { has $.a = "a"; method b { "aaa".subst(/{say self}/,"b") }}; say A.new.a; say A.new.b
[13​:39​:02] <+camelia> rakudo-moar 41574a​: OUTPUT«a␤A.new(a => "a")␤baaa␤»
[13​:39​:16] <FROGGS> I have no idea
[13​:39​:25] <FROGGS> m​: class A { has $.a = "a"; method b { "aaa".subst(/"{say self}"/,"b") }}; say A.new.a; say A.new.b
[13​:39​:27] <+camelia> rakudo-moar 41574a​: OUTPUT«a␤A.new(a => "a")␤A.new(a => "a")␤A.new(a => "a")␤A.new(a => "a")␤aaa␤»
[13​:40​:02] <jnthn> m​: class A { has $.a = "a"; method b { "aaa".subst(/​:my $ = say $!a; ./, b') } }; A.new.b
[13​:40​:04] <+camelia> rakudo-moar 41574a​: OUTPUT«�[31m===�[0mSORRY!�[31m===�[0m Error while compiling /tmp/TCkwyckpg_␤Unable to parse expression in argument list; couldn't find final ')' ␤at /tmp/TCkwyckpg_​:1␤------> �[32m b { "aaa".subst(/​:my $ = say $!a; ./, b�[33m⏏�[31m') } }; A.new.b�[0m␤ …»
[13​:40​:13] <jnthn> m​: class A { has $.a = "a"; method b { "aaa".subst(/​:my $ = say $!a; ./, 'b') } }; A.new.b
[13​:40​:15] <+camelia> rakudo-moar 41574a​: OUTPUT«(Any)␤»
[13​:40​:42] <lizmat> m​: say "aaa".subst(/a$/,"b") # the reason I ran into this
[13​:40​:43] <+camelia> rakudo-moar 41574a​: OUTPUT«aab␤»
[13​:40​:51] <jnthn> m​: class A { has $.a = "a"; method b { "aaa".subst(/​:my $ = say self.WHICH; ./, 'b') } }; A.new.b
[13​:40​:54] <+camelia> rakudo-moar 41574a​: OUTPUT«Cursor|139834212483800␤»
[13​:42​:13] <jnthn> m​: class A { has $.a = "a"; method b { "aaa".subst(/​:my $ = say self.WHICH, $!a; ./, 'b') } }; A.new.b
[13​:42​:15] <+camelia> rakudo-moar 41574a​: OUTPUT«Cursor|139900812459976(Any)␤»
[13​:42​:21] <jnthn> m​: class A { has $.a = "a"; method b { "aaa".subst(/​:my $ = say self.WHICH, $!pos; ./, 'b') } }; A.new.b
[13​:42​:22] <+camelia> rakudo-moar 41574a​: OUTPUT«�[31m===�[0mSORRY!�[31m===�[0m Error while compiling /tmp/DBWEyB7Zo2␤Attribute $!pos not declared in class A␤at /tmp/DBWEyB7Zo2​:1␤------> �[32m $ = say self.WHICH, $!pos; ./, 'b') } }�[33m⏏�[31m; A.new.b�[0m␤ expecting any of​:␤ postfi…»
[13​:42​:51] <jnthn> m​: class A { has $!dummy; has $.a = "a"; method b { "aaa".subst(/​:my $ = say self.WHICH, $!a; ./, 'b') } }; A.new.b
[13​:42​:53] <+camelia> rakudo-moar 41574a​: OUTPUT«Cursor|140307760568368Cursor|140307760671904No such method 'gist' for invocant of type 'ParseShared'␤ in sub say at src/gen/m-CORE.setting​:13805␤ in method match at src/gen/m-CORE.setting​:6280␤ in method subst at src/gen/m-CORE.setting​:6355␤ in…»
[13​:42​:59] <jnthn> Wow.
[13​:43​:22] <jnthn> Yes, that's quite a mis-compile.
[13​:43​:36] <jnthn> I know what it's doing.
[13​:43​:47] <jnthn> I don't know we can make it DWLM though...
[13​:43​:57] <jnthn> Because self in a /.../ thing really is Cursor
[13​:44​:13] <FROGGS> L is for lizmat? :o)
[13​:44​:17] <jnthn> Yes.
[13​:44​:27] <FROGGS> true... and in a code somebody might really want the Cursor
[13​:44​:34] <jnthn> Right
[13​:44​:35] <lizmat> ok
[13​:44​:40] <jnthn> Well, it's clearer with tokens and so forth.
[13​:44​:56] <jnthn> All the same, we shouldn't compile it to stuff that looks up a random attribute on the wrong object. :)
[13​:45​:29] <lizmat> I guess it should at least complain about it being an attribute that doesn't exist on the Regex ?
[13​:45​:38] <jnthn> lizmat​: Yeah.
[13​:45​:44] <lizmat> so rakudobug it ?
[13​:45​:55] <jnthn> Should just not compile
[13​:46​:06] <jnthn> Yes, now we know the expected behavior, it's a resolvable ticket :)
[13​:46​:14] <jnthn> So go for it :)

From @masak

<TuxCM> http://pastebin.com/PpDKvrd9 <= I do not understand
<Ven> pmurias​: right -- I agree with that! I was just wondering why
angular of all the template systems
<moritz> TuxCM​: seems to be a scoping bug of sorts
<masak> m​: my $sep = "|"; class C { method foo { .say for
"foo|bar".split(/$sep/).map(*) } }; C.new.foo
<masak> m​: class C { has $.sep = "|"; method foo { .say for
"foo|bar".split(/$!sep/).map(*) } }; C.new.foo
<camelia> rakudo-moar da3aae​: OUTPUT«foo|bar␤»
<masak> TuxCM​: interesting.
* masak submits rakudobug

From @pmichaud

On Fri, Oct 03, 2014 at 06​:48​:50AM -0700, Carl Mäsak wrote​:

<masak> m​: class C { has $.sep = "|"; method foo { .say for
"foo|bar".split(/$!sep/).map(~*) } }; C.new.foo
<camelia> rakudo-moar da3aae​: OUTPUT«foo|bar␤»
<masak> TuxCM​: interesting.
* masak submits rakudobug

14​:09 <pmichaud> m​: class C { has $.rx = 'bc'; method foo { say 'abcd' ~~ /$!rx/; } }; C.new.foo
14​:09 <camelia> rakudo-moar da3aae​: OUTPUT«Nil␤»
14​:10 <pmichaud> I'm guessing that regexes don't know how to interpolate $!vars
14​:10 <pmichaud> m​: class C { has $.rx = 'bc'; method foo { say $!rx; say 'abcd' ~~ /$!rx/; } }; C.new.foo
14​:10 <camelia> rakudo-moar da3aae​: OUTPUT«bc␤Nil␤»
14​:11 <moritz> or maybe the attribute lookup gets the cursor object as invocant?

Pm

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

From @jnthn

On Wed Jul 09 04​:48​:43 2014, elizabeth wrote​:

[13​:33​:31] <lizmat> m​: class A { has $.a = "a"; method b {
"aaa".subst(/$!a/,"b") }}; say A.new.a; say A.new.b # why isn't the
.subst working ?
[13​:33​:32] <+camelia> rakudo-moar 41574a​: OUTPUT«a␤aaa␤»
[13​:34​:31] <lizmat> m​: class A { has $.a = "a"; method b { my $x
= $!a; "aaa".subst(/$x/,"b") }}; say A.new.b # putting it into a
lexical first works
[13​:34​:33] <+camelia> rakudo-moar 41574a​: OUTPUT«baa␤»
[13​:34​:50] <lizmat> rakudobug?
[13​:35​:55] <lizmat> FROGGS, jnthn​: ^^^
<snip>
[13​:42​:59] <jnthn> Wow.
[13​:43​:22] <jnthn> Yes, that's quite a mis-compile.
[13​:43​:36] <jnthn> I know what it's doing.
[13​:43​:47] <jnthn> I don't know we can make it DWLM though...
[13​:43​:57] <jnthn> Because self in a /.../ thing really is
Cursor
[13​:44​:13] <FROGGS> L is for lizmat? :o)
[13​:44​:17] <jnthn> Yes.
[13​:44​:27] <FROGGS> true... and in a code somebody might really
want the Cursor
[13​:44​:34] <jnthn> Right
[13​:44​:35] <lizmat> ok
[13​:44​:40] <jnthn> Well, it's clearer with tokens and so forth.
[13​:44​:56] <jnthn> All the same, we shouldn't compile it to
stuff that looks up a random attribute on the wrong object. :)
[13​:45​:29] <lizmat> I guess it should at least complain about it
being an attribute that doesn't exist on the Regex ?
[13​:45​:38] <jnthn> lizmat​: Yeah.
[13​:45​:44] <lizmat> so rakudobug it ?
[13​:45​:55] <jnthn> Should just not compile
[13​:46​:06] <jnthn> Yes, now we know the expected behavior, it's
a resolvable ticket :)
[13​:46​:14] <jnthn> So go for it :)

And it now gives a decent error and a suggestion of what to do​:

$ perl6-m -e "class C { has $!a = 'a'; method m() { say 'a' ~~ /$!a/ } }; C.new.m"
===SORRY!=== Error while compiling -e
Attribute $!a not available inside of a regex, since regexes are methods on Cursor.
Consider storing the attribute in a lexical, and using that in the regex.
at -e​:1
------> $!a = 'a'; method m() { say 'a' ~~ /$!a<HERE>/ } }; C.new.m

Made it a typed exception, and tested it in S05-interpolation/regex-in-variable.t. Resolving.

/jnthn

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

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