Real.new hangs on .gist

[p6rt]

Migrated from rt.perl.org#126130 (status was 'open')

Searchable as RT126130$

[p6rt]

From zefram@fysh.org

$ ./perl6 -e 'say Real.new'
[hangs]
^C

Like Numeric.new + 0 [perl #​126112], there's a hang on a basically
unimplemented method in a fairly abstract class. This is less than
awesome. It would be nice to have the lack of implementation detected,
resulting in a signalled error.

-zefram

[p6rt]

From @lizmat

“Real.new.Num" infiniloops because​:

  .Num calls .Bridge.Num
  .Bridge calls .Num

This is the ground cause of “say Real.new” infinilooping, because​:

  .gist calls .Str
  .Str calls .Bridge (to call .Str again, but never gets there because​:)
  .Bridge calls .Num
  .Num calls .Bridge

Since I never understood the use of .Bridge in these parts of the setting, I’m hoping that someone else with some more background in this part of the code that is mostly from 2011/2012, and this predates my involvement.

Yary++ for reporting, BenGoldberg++ for following up.

[p6rt]

From @coke

On Wed Feb 10 01​:58​:34 2016, elizabeth wrote​:

“Real.new.Num" infiniloops because​:

.Num calls .Bridge.Num
.Bridge calls .Num

This is the ground cause of “say Real.new” infinilooping, because​:

.gist calls .Str
.Str calls .Bridge (to call .Str again, but never gets there because​:)
.Bridge calls .Num
.Num calls .Bridge

Since I never understood the use of .Bridge in these parts of the
setting, I’m hoping that someone else with some more background in
this part of the code that is mostly from 2011/2012, and this predates
my involvement.

Yary++ for reporting, BenGoldberg++ for following up.

Basically a dupe of #​126130
--
Will "Coke" Coleda

[p6rt]

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

[p6rt]

@coke - Status changed from 'new' to 'open'