From @japhb

Rakudo doesn't parse Foo​::_foo() properly, while Foo​::foo() works fine.

Before patch​:

$ ./perl6 -e 'module Foo { sub _foo { say "foo" } }; Foo​::_foo()'
Statement not terminated properly at line 1, near "_foo()"

current instr.​: 'parrot;PGE;Util;die' pc 129 (runtime/parrot/library/PGE/Util.pir​:85)
called from Sub 'parrot;Perl6;Grammar;eat_terminator' pc 86107 (src/gen_grammar.pir​:2292)
called from Sub 'parrot;Perl6;Grammar;statementlist' pc 84764 (src/gen_grammar.pir​:1770)
called from Sub 'parrot;Perl6;Grammar;statement_block' pc 82197 (src/gen_grammar.pir​:750)
called from Sub 'parrot;Perl6;Grammar;TOP' pc 76936 (src/gen_setting.pir​:16404)
called from Sub 'parrot;PCT;HLLCompiler;parse' pc 665 (src/PCT/HLLCompiler.pir​:402)
called from Sub 'parrot;PCT;HLLCompiler;compile' pc 428 (src/PCT/HLLCompiler.pir​:303)
called from Sub 'parrot;PCT;HLLCompiler;eval' pc 920 (src/PCT/HLLCompiler.pir​:521)
called from Sub 'parrot;PCT;HLLCompiler;command_line' pc 1510 (src/PCT/HLLCompiler.pir​:800)
called from Sub 'parrot;Perl6;Compiler;main' pc 23787 (perl6.pir​:164)

After patch​:

$ ./perl6 -e 'module Foo { sub _foo { say "foo" } }; Foo​::_foo()'
foo

The patch is trivial​:

diff --git a/src/parser/grammar.pg b/src/parser/grammar.pg
index 50b95e7..9c0b5b2 100644
--- a/src/parser/grammar.pg
+++ b/src/parser/grammar.pg
@@ -815,7 +815,7 @@ token name {
 token morename {
     '::'
     [
-        <?before '(' | <alpha> >
+        <?before '(' | <ident> >
         [
         | <identifier>
         | '(' <EXPR> ')'

-'f

From @moritz

On Sat Apr 25 16​:33​:44 2009, japhb wrote​:

token morename {
'​::'
[
- <?before '(' | <alpha> >
+ <?before '(' | <ident> >
[
| <identifier>
| '(' <EXPR> ')'

In STD.pm the token alpha is define is <ident> \w* - so maybe even
<?before '(' | \w >
would be enough?
At least it makes your test pass, and the test suite is also clean.

Cheers,
Moritz

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

From @pmichaud

On Sat, Apr 25, 2009 at 04​:33​:44PM -0700, Geoffrey Broadwell wrote​:

Rakudo doesn't parse Foo​::_foo() properly, while Foo​::foo() works fine.
[...]
$ ./perl6 -e 'module Foo { sub _foo { say "foo" } }; Foo​::_foo()'
Statement not terminated properly at line 1, near "_foo()"

While the patch appears to solve the issue, it also deviates from
the definition given in STD.pm, which uses <alpha> instead of <ident>.
(Also, testing for <ident> is a fair bit more expensive than simply
testing for <alpha>.)

STD.pm appears to parse the above okay, though, so we need to
determine why STD.pm is succeeding here and see if Rakudo should
match that.

Patch rejected (for now), but we'll leave the ticket open until
we can find a solution to the leading underscore problem.

Thanks!

Pm

From @pmichaud

On Mon, Apr 27, 2009 at 08​:31​:38AM -0500, Patrick R. Michaud wrote​:

On Sat, Apr 25, 2009 at 04​:33​:44PM -0700, Geoffrey Broadwell wrote​:

Rakudo doesn't parse Foo​::_foo() properly, while Foo​::foo() works fine.
[...]
$ ./perl6 -e 'module Foo { sub _foo { say "foo" } }; Foo​::_foo()'
Statement not terminated properly at line 1, near "_foo()"

While the patch appears to solve the issue, it also deviates from
the definition given in STD.pm, which uses <alpha> instead of <ident>.
(Also, testing for <ident> is a fair bit more expensive than simply
testing for <alpha>.)

STD.pm appears to parse the above okay, though, so we need to
determine why STD.pm is succeeding here and see if Rakudo should
match that.

Answered on #perl6​:
  16​:14 <TimToady> pmichaud​: _ is considered an alpha

So, we need to patch PGE so that <alpha> matches underscore, and
we need some tests to match.

Pm

From @coke

On Mon, Apr 27, 2009 at 12​:26 PM, Patrick R. Michaud <pmichaud@​pobox.com> wrote​:

Answered on #perl6​:
   16​:14 <TimToady> pmichaud​: _ is considered an alpha

So, we need to patch PGE so that <alpha> matches underscore, and
we need some tests to match.

Pm

Created TT #​585 for parrot to track this issue.

--
Will "Coke" Coleda

From @pmichaud

Now fixed in f93154a.

We need to verify there's a spectest for this particular
problem. We should also add spectests to make sure that <alpha>
matches underscores.

Assigning to moritz++ for test updates.

Thanks!

Pm

From @moritz

Test added to t/spec/S02-names_and_variables/names.t

Cheers,
Moritz

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