Re: %_ - is it available for use?
david nicol wrote:
[EMAIL PROTECTED] perl -le '$_{a}=27; package notmain; print $_{a}'
27
Gosh!
Let's document it! Would it go in perlvar or perlsyn?
It's already documented, in perlvar/Technical Note on the Syntax of Variable Names
(at the end)
Re: %_ - is it available for use?
Odd. I always though that you could do that. Maybe it was just an
assumption I made without understanding. I've used %_ in one liners to
remove duplicates and other such fun.
Michael
On Tue, 2003-08-05 at 21:36, david nicol wrote:
On Sat, 2003-08-02 at 18:33, Michael G Schwern wrote:
Representing the Backwards Compatiblity Police, I've had co-workers use
%_ as the globalist of all global hashes. %_ transends all packages and
scopes and Perl does not localize it, touch it or use it as it does @_ and
$_. In the particular case I'm thinking of, it was used to hold global
arguments for function calls in a template system. Rather insane, really.
Lots of better ways to do it and clearly making use of an undefined
language feature.
I'm not making an argument against %_, just noting that *_ is used
opportunisticly and you will break a few programs.
[EMAIL PROTECTED] perl -le '$_{a}=27; package notmain; print $_{a}'
27
Gosh!
Let's document it! Would it go in perlvar or perlsyn?
The behaviour in question is a side effect of the general magic level of
the other slots in *_, is it not?
Re: %_ - is it available for use?
On Sat, 2003-08-02 at 18:33, Michael G Schwern wrote:
Representing the Backwards Compatiblity Police, I've had co-workers use
%_ as the globalist of all global hashes. %_ transends all packages and
scopes and Perl does not localize it, touch it or use it as it does @_ and
$_. In the particular case I'm thinking of, it was used to hold global
arguments for function calls in a template system. Rather insane, really.
Lots of better ways to do it and clearly making use of an undefined
language feature.
I'm not making an argument against %_, just noting that *_ is used
opportunisticly and you will break a few programs.
[EMAIL PROTECTED] perl -le '$_{a}=27; package notmain; print $_{a}'
27
Gosh!
Let's document it! Would it go in perlvar or perlsyn?
The behaviour in question is a side effect of the general magic level of
the other slots in *_, is it not?
--
David Nicol /
Re: %_ - is it available for use?
Damian Conway [EMAIL PROTECTED] writes: From a Perl 6 perspective, it seems likely that C%_ will be the name commonly used for the slurpy hash of a subroutine. Just as C@_ will often be the name used for the slurpy array. See Exegesis 6 for more details. Indeed, when it comes to object constructors (all of which implicitly have a slurpy hash), C%_ might well be the automatically provided name for that hash. See Exegesis 12 for more details. ;-) Hence, making C%_ mean something different in core Perl 5 might possibly be forwards incompatible. This wouldn't have been a core Perl 5 thing but a (module) Tk thing. But if Perl 6 is giving it a meaning I will avoid it (unless a slurpy hash is what I meant - I haven't kept up the Perl6 jargon.)
Re: %_ - is it available for use?
Damian Conway [EMAIL PROTECTED] writes: Hence, making C%_ mean something different in core Perl 5 might possibly be forwards incompatible. Representing the Backwards Compatiblity Police, I've had co-workers use %_ as the globalist of all global hashes. %_ transends all packages and scopes and Perl does not localize it, touch it or use it as it does @_ and $_. In the particular case I'm thinking of, it was used to hold global arguments for function calls in a template system. Rather insane, really. Lots of better ways to do it and clearly making use of an undefined language feature. I'm not making an argument against %_, just noting that *_ is used opportunisticly and you will break a few programs. -- It's Crack Cocaine time!
Re: %_ - is it available for use?
On Sun, Aug 03, 2003 at 01:37:16AM +0200, Abigail wrote:
I am fond of doing
local %_ = @_;
as one of the first statements of a subroutine. That, or
my %args = @_;
I like the latter because it uses a lexical variable, but I like the
former because %_ fits with @_ and $_.
FWIW, local() is quite a bit slower than my().
$ perl5.8.1 local_vs_my 100
Rate localmy
local 180180/s-- -30%
my257069/s 43%--
$ cat local_vs_my
#!/usr/bin/perl -w
use Benchmark 'cmpthese';
cmpthese( shift || -3, {
local = sub { local %_ = @_; },
my = sub { my %args = @_; },
});
--
Welcome to the Office Of Naval Contemplation
Re: %_ - is it available for use?
On Sat, Aug 02, 2003 at 04:33:19PM -0700, Michael G Schwern wrote: : I'm not making an argument against %_, just noting that *_ is used : opportunisticly and you will break a few programs. Not necessarily. If Perl 6 were to use %_ as parameter name, it would be lexically scoped, and hide any package %_ only in that scope. Larry
Re: %_ - is it available for use?
On Sat, Aug 02, 2003 at 08:16:19PM -0700, Larry Wall wrote: On Sat, Aug 02, 2003 at 04:33:19PM -0700, Michael G Schwern wrote: : I'm not making an argument against %_, just noting that *_ is used : opportunisticly and you will break a few programs. Not necessarily. If Perl 6 were to use %_ as parameter name, it would be lexically scoped, and hide any package %_ only in that scope. Sorry, p5p-cross talk, I was referring only to Perl 5. -- Any sufficiently encapsulated hack is no longer a hack.
%_ - is it available for use?
We have been discussing how to pass data to Tk callbacks. In particular Entry widget validation routines. There are a number of items that they _might_ be interested in but a typical routine would only use a few. Currently it passes them all as positional parameters. One idea that occured to me/us is to have a localized hash for this purpose (which might be tied/magical). One choice of name for this is %_ - which fits neatly with @_ for positional args and $_ as current thing. Also *_ glob already exists to have its GvHV entry (ab-)used. What does p5p think of this use ? Basically @_ says how you been called $_ is what you are working on, and this idea gives %_ the meaning of why. (Caller gives the who ;-) )
Re: %_ - is it available for use?
Nick Ing-Simmons wrote: We have been discussing how to pass data to Tk callbacks. In particular Entry widget validation routines. There are a number of items that they _might_ be interested in but a typical routine would only use a few. Currently it passes them all as positional parameters. One idea that occured to me/us is to have a localized hash for this purpose (which might be tied/magical). One choice of name for this is %_ - which fits neatly with @_ for positional args and $_ as current thing. Also *_ glob already exists to have its GvHV entry (ab-)used. What does p5p think of this use ? Basically @_ says how you been called $_ is what you are working on, and this idea gives %_ the meaning of why. (Caller gives the who ;-) ) From a Perl 6 perspective, it seems likely that C%_ will be the name commonly used for the slurpy hash of a subroutine. Just as C@_ will often be the name used for the slurpy array. See Exegesis 6 for more details. Indeed, when it comes to object constructors (all of which implicitly have a slurpy hash), C%_ might well be the automatically provided name for that hash. See Exegesis 12 for more details. ;-) Hence, making C%_ mean something different in core Perl 5 might possibly be forwards incompatible. Damian
