make :%hash<key> a pair constructor

p6rt commented 8 years ago

Migrated from rt.perl.org#127165 (status was 'open')

Searchable as RT127165$

p6rt commented 8 years ago

From @FROGGS

For a while we can create Pairs from named captures of match variables: m: my %hash = key => 42; say (:$\) rakudo-moar ac4f9f: OUTPUT«key => Nil␤»

Sadly, this does not even parse: m: my %hash = key => 42; say (:$/\) rakudo-moar ac4f9f: OUTPUT«===SORRY!=== Error while compiling /tmp/3GMq24eXRj␤Bogus statement␤at /tmp/3GMq24eXRj:1␤------> my %hash = key => 42; say (:$⏏/\)␤ expecting any of:␤ colon pair␤»

Nor does this create a Pair: m: my %hash = key => 42; say (:%hash\) rakudo-moar ac4f9f: OUTPUT«(Mu)␤»

So I vote for all shown statements to create a Pair of key => Nil or key => Any respectively.

Funnily, Arrays behave differently when used with colonpair syntax: m: my @array = 42, 43; say (:@array[0]) rakudo-moar ac4f9f: OUTPUT«array => [42 43]␤»

Which is also wrong, IMO. This case should either produce 0 => 42, or array => 42. If we go for the latter then the shown Hashes should also be treated as subscripts, and produce hash => 42, if the variable is not anonymous like in $/\ or $\. This would allow %hash\, but then again $\ and $/\<key foo> need to parsefail.

p6rt commented 8 years ago

From @LLFourn

Is the first example meant to be like:

'"foo" ~~ /$\=.+/; say (:$\)'

??

On Tue, Jan 5, 2016 at 6:42 PM Tobias Leich \perl6\-bugs\-followup@perl\.org wrote:

# New Ticket Created by Tobias Leich # Please include the string: [perl #127165] # in the subject line of all future correspondence about this issue. # \<URL: https://rt-archive.perl.org/perl6/Ticket/Display.html?id=127165 >

For a while we can create Pairs from named captures of match variables: m: my %hash = key => 42; say (:$\) rakudo-moar ac4f9f: OUTPUT«key => Nil␤»

Sadly, this does not even parse: m: my %hash = key => 42; say (:$/\) rakudo-moar ac4f9f: OUTPUT«===SORRY!=== Error while compiling /tmp/3GMq24eXRj␤Bogus statement␤at /tmp/3GMq24eXRj:1␤------> my %hash = key => 42; say (:$⏏/\)␤ expecting any of:␤ colon pair␤»

Nor does this create a Pair: m: my %hash = key => 42; say (:%hash\) rakudo-moar ac4f9f: OUTPUT«(Mu)␤»

So I vote for all shown statements to create a Pair of key => Nil or key => Any respectively.

Funnily, Arrays behave differently when used with colonpair syntax: m: my @array = 42, 43; say (:@array[0]) rakudo-moar ac4f9f: OUTPUT«array => [42 43]␤»

Which is also wrong, IMO. This case should either produce 0 => 42, or array => 42. If we go for the latter then the shown Hashes should also be treated as subscripts, and produce hash => 42, if the variable is not anonymous like in $/\ or $\. This would allow %hash\, but then again $\ and $/\<key foo> need to parsefail.

p6rt commented 8 years ago

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

p6rt commented 8 years ago

From @FROGGS

Yes, that's how you get a Match variable.

p6rt commented 8 years ago

From @LLFourn

I was confused because of the preceding hash assignment :)

I also find myself not wanting to do { long-attribute-name => $obj.long-attribute-name }. Maybe $obj.=>long-attribute-name could DWIM and :$obj.my-val becomes ('obj' => 'my-val result').

I'll be quiet now.

On Tue, Jan 5, 2016 at 7:01 PM Tobias Leich \email@froggs\.de wrote:

Yes, that's how you get a Match variable.

Raku / old-issue-tracker