Weird internal error when parsing some very simple Pod with '-' in the first column of a table in Rakudo

[p6rt]

Migrated from rt.perl.org#128221 (status was 'resolved')

Searchable as RT128221$

[p6rt]

From @masak

<tbrowder> continuing my problems with p6 pod​: I have scrubbed a bad
pod file down to I think a minimum, and used the suggested method to
test it
<masak> tbrowder​: sounds like great news. do you have a gist for us?
<tbrowder> I'll put it in a gist shortly, but besides the bad pod, the
worst part is perl6's response to it leaves no cluse as to what the
problem is
<tbrowder> for example, executing 'perl6 bad.pod' yields exactly
"===SORRY!=== Cannot iterate object with P6opaque representation"
<masak> tbrowder​: yes, that's an unacceptable error message. I agree.
<tbrowder> the gist with the short good and bad Perl 6 pod examples is
here​: https://gist.github.com/tbrowder/23c5ef38cccab30a32a7f7a2ca61659d
<masak> nice. short.
<masak> problem reproduced.
* masak submits rakudobug

For completeness, gist reproduced below​:

## bad-pod-example.pod

=begin pod
# notice leading hyphen '-' in column 1 text
=table
-col-1 col-2
=end pod
say $=pod.perl

## good-pod-example.pod

=begin pod
# notice leading hyphen '-' in column 1 text has been escaped
=table
\-col-1 col-2
=end pod
say $=pod.perl

[p6rt]

From @tbrowder

I have been able to get a good debug trace which should enable an experienced dev to fix the problem.

See the details in my github gist​:

  https://gist.github.com/tbrowder/5e7649384c3338d516ecde8a2451b4a3

[p6rt]

From @tbrowder

On Fri May 27 07​:43​:52 2016, tom.browder@​gmail.com wrote​:

I have been able to get a good debug trace which should enable an
experienced dev to fix the problem.
...

See my new comment above (IMHO, RT is not a good issue tracker!!).

[p6rt]

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

[p6rt]

From @tbrowder

I am working on this issue and would like ownership of it.

[p6rt]

From @tbrowder

I am working on this issue and would like ownership of it.

[p6rt]

From @tbrowder

On Fri Jul 01 14​:11​:44 2016, tbrowder wrote​:

I am working on this issue and would like ownership of it.

I am making progress, albeit slowly. It is clear to me now that the two major types of tables (whitespace-delineated columns versus visible-character-delineated columns) must be treated separately, and that is the direction I have taken.

[p6rt]

From @zoffixznet

♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥
🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁

SKIP-fudged tests added in Raku/roast@f1ab1c7

🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁🏁
♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥♥

[p6rt]

From @tbrowder

Fixed by PR #​1240.

[p6rt]

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

[p6rt]

From @tbrowder

tests added in roast PR #​353