atom feed24 messages in org.perl.perl5-portersRe: fixing ~~ harder. a lot harder
FromSent OnAttachments
Ricardo SignesJul 6, 2011 2:24 pm 
Leon TimmermansJul 6, 2011 3:19 pm 
ZeframJul 6, 2011 3:25 pm 
Reverend ChipJul 6, 2011 3:31 pm 
ZeframJul 6, 2011 3:49 pm 
Reverend ChipJul 6, 2011 3:53 pm 
Ricardo SignesJul 6, 2011 4:14 pm 
ZeframJul 6, 2011 4:30 pm 
Ricardo SignesJul 6, 2011 4:36 pm 
Reverend ChipJul 6, 2011 7:05 pm 
Ed AvisJul 7, 2011 9:15 am 
Jesse LuehrsJul 7, 2011 9:19 am 
Ed AvisJul 7, 2011 9:27 am 
Reverend ChipJul 7, 2011 9:33 am 
Ricardo SignesJul 7, 2011 9:43 am 
ZeframJul 7, 2011 9:43 am 
Reverend ChipJul 7, 2011 9:50 am 
Salvador FandinoJul 7, 2011 10:42 am 
Ricardo SignesJul 7, 2011 11:56 am 
Reverend ChipJul 7, 2011 12:29 pm 
Ricardo SignesJul 7, 2011 1:02 pm 
Reverend ChipJul 7, 2011 1:15 pm 
Randal L. SchwartzJul 10, 2011 3:42 pm 
Yuval KogmanJul 16, 2011 8:13 am 
Subject:Re: fixing ~~ harder. a lot harder
From:Ricardo Signes (perl@rjbs.manxome.org)
Date:Jul 6, 2011 4:36:22 pm
List:org.perl.perl5-porters

* Zefram <zef@fysh.org> [2011-07-06T19:31:01]

Ricardo Signes wrote:

I think that "was enreferenced" marker is not so bad, but I'm not really married to it, if everyone else thinks it is unneeded.

I see a big problem with it. Meta-predicate objects, such as Smart::Match provide, will want to apply arbitrary predicate to arbitrary input. Currently they can do that by "$candidate ~~ $predicate". If enreferencement is significant, they'll need a big branching set of options to *de*reference $candidate just so it can be implicitly re-enreferenced to get the flag right. Of course, you need the *right kind of* dereferencement.

Your point is very well taken, thanks for raising it. Given that I couldn't yet see why we *would* want that flag, I am pretty happy to not worry about it.