Re: [rt.cpan.org #125472] error with coderefs passes as scalar args

At 03:07 AM 7/3/2018, you wrote:
><URL: https://rt.cpan.org/Ticket/Display.html?id=125472 >
>
>we very much appreciate the effort;
>
>unclear for me - why $sub1 and $sub2 will have same slot for this statement,
>$interp->call('if,'1',$sub1,$sub2);
>
>there is:
>...
># stringify sub, becomes "CODE(0x######)" in ::perl namespace
>
>so $sub1 and $sub2 will recieve different slots in %anon_refs hash


in the v1.05 code
$current_r will be set once at the top of the loop by
   my $current_r = join ' ', grep {defined} grep {!ref} @args;
so it will be 'if 1'
and is passed into create_tcl_sub() in the 4th parm where it becomes $rname
and $anon_refs gets set with
   $anon_refs{$rname} = bless [\$sub, $interp], 'Tcl::Code';
so both calls get sequentially created into the same $anon_refs slot 
$anon_refs{'if 1'}
And before the second gets assigned Tcl::Code::DESTROY is triggered 
on the first.


>all in all, if you'll find a correct solution - nice. if isn't - 
>also not bad. for mee, we already have too much code for the case, 
>which I would prefer to be simplier.

This kind of reference tracking is never simple, i have tried to keep 
it as simple as it needs to be.

>also, what a need for
>sub _anon_refs_cheat { return \%anon_refs;}
>
>?

it serves to expose %anon_refs to external programs so they can test what has
been done to it by the process in the module. I often do this to be 
able to test that things are working right without having to add 
other exposing code to the primary module.

>can we switch to mailing list?

i think i got subscribed, so this is cc'd to the list.

for those first seeing this on the list there is historical background at
https://rt.cpan.org/Public/Bug/Display.html?id=125472