perl5.9.1 thoughts
Yitzchak Scott-Thoennes
make it happen? Who is going to do it?
Here's what I'd like to see:
mjd's lexical pragmas patch (what more is needed for this?)
Rafael's my $_ patch (ditto)
Module::Build (& dependencies)
CPANPLUS
Some or all of Inline
Ilya had done some work towards making unpack be able to handle any C
structure, but IIRC it came just before a release (and there was some
question as to conflicting with what perl6 will do vis-a-vis pack and
unpack) and so was deferred. I'd like to see this revisited.
I'd like to add:
Hugo's regex rewrite
but don't know what's up with it :)
Nicholas had also said (in the "Snapshot perl@21938" thread):
> will there be a 5.9.0 soon? (Which would be a useful stake in the
> ground, Given that Arthur is keen to break everything by checking in his COW
> code)
Is this still pending?
Rafael Garcia-Suarez
> Here's what I'd like to see:
> mjd's lexical pragmas patch (what more is needed for this?)
Time :)
> Rafael's my $_ patch (ditto)
Rough edges.
> Module::Build (& dependencies)
> CPANPLUS
> Some or all of Inline
Wait wait let's define milestones first.
Regarding Module::Build and CPANPLUS I don't know whether they are stable
and mature enough.
Regarding Inline IIRC Brian Ingerson was saying that a minimal subset
of Inline would be sufficient.
> Nicholas had also said (in the "Snapshot perl@21938" thread):
> > will there be a 5.9.0 soon? (Which would be a useful stake in the
> > ground, Given that Arthur is keen to break everything by checking in his COW
> > code)
>
> Is this still pending?
I think so.
I believe that releasing a 5.9.1 with what we have, what's being
finished, the MakeMaker fixes to come (hint hint) and an updated
perltodo.pod would be a good idea. Nicholas is outpacing blead
currently ;)
Nicholas Clark
> Yitzchak Scott-Thoennes wrote:
> > Nicholas had also said (in the "Snapshot perl@21938" thread):
> > > will there be a 5.9.0 soon? (Which would be a useful stake in the
> > > ground, Given that Arthur is keen to break everything by checking in his COW
> > > code)
> >
> > Is this still pending?
>
> I think so.
This is also my understanding based on a conversation with Arthur on IRC
earlier this week. Arthur should soon be back from a trip away.
> I believe that releasing a 5.9.1 with what we have, what's being
> finished, the MakeMaker fixes to come (hint hint) and an updated
> perltodo.pod would be a good idea. Nicholas is outpacing blead
> currently ;)
My camel's faster than your camel. Neeeur!
:-)
Nicholas Clark
Yitzchak Scott-Thoennes
> Yitzchak Scott-Thoennes wrote:
> > Here's what I'd like to see:
> > mjd's lexical pragmas patch (what more is needed for this?)
>
> Time :)
Does that mean tuits? Anybody's in particular? What needs to be done?
> > Rafael's my $_ patch (ditto)
>
> Rough edges.
>
> > Module::Build (& dependencies)
> > CPANPLUS
> > Some or all of Inline
>
> Wait wait let's define milestones first.
> Regarding Module::Build and CPANPLUS I don't know whether they are stable
> and mature enough.
I think M::B is ready (with the next release anyway). It isn't going
to work for every platform, but I think enough are covered that it's
time to put it in the core and find out where it doesn't work.
About CPANPLUS I don't know.
Tassilo Von Parseval
> What would you all like to see in 5.9.1? What needs to be done to
> make it happen? Who is going to do it?
>
> Here's what I'd like to see:
> mjd's lexical pragmas patch (what more is needed for this?)
Is this the patch that would allow one to write stuff like the
autobox-thing in 'user-space' (that is: in pure Perl)?
Tassilo
--
$_=q#",}])!JAPH!qq(tsuJ[{@"tnirp}3..0}_$;//::niam/s~=)]3[))_$-3(rellac(=_$({
pam{rekcahbus})(rekcah{lrePbus})(lreP{rehtonabus})!JAPH!qq(rehtona{tsuJbus#;
$_=reverse,s+(?<=sub).+q#q!'"qq.\t$&."'!#+sexisexiixesixeseg;y~\n~~dddd;eval
Rafael Garcia-Suarez
> > mjd's lexical pragmas patch (what more is needed for this?)
>
> Is this the patch that would allow one to write stuff like the
> autobox-thing in 'user-space' (that is: in pure Perl)?
More precisely it's an extensive rewrite of the implementation of the
lexical pragmas in the core; it will fix several known problems : the
state of most pragmas isn't known anymore at compile-time, and thus
not propagated into eval(""); more pragmas could be defined; consequently
the pragmas that aren't lexical could be made so (e.g. sort); and I
forget some. MJD is planning to work on this in a couple of months.
Mark Jason Dominus
> Tassilo von Parseval wrote:
> > > mjd's lexical pragmas patch (what more is needed for this?)
> >
> > Is this the patch that would allow one to write stuff like the
> > autobox-thing in 'user-space' (that is: in pure Perl)?
>
> More precisely it's an extensive rewrite
I'm not sure 'extensive' is the right word. The core parts of the
patch turned out to be very tiny; it was much easier to do than I had
initially thought.
The patch I sent was bigger, but that's because it included an
extension module to provide access the core function for pure-Perl
modules, and a demonstration pure-Perl module, and I think some other
stuff.
I am also planning to leave the existing $^H mechanism in place.
I don't know what 'the autobox thing' is. If it's that references to
unblessed arrays can have methods called on them, which would be
dispatched through package ARRAY, then possibly, although it would be
rather a stretch without at least a little more core support than is
presently in there. I should probably avoid speculating.
> MJD is planning to work on this in a couple of months.
I told the grant committee people March, but now it looks like it
might be April. But I don't think it will take long to finish once I
get started.
The proposal, which describes the current state of the project, and
what I plan to do, is at
http://perl.plover.com/TPF/Pragmas/PROPOSAL
should you care to see it.
Rafael Garcia-Suarez
> > More precisely it's an extensive rewrite
>
> I'm not sure 'extensive' is the right word. The core parts of the
> patch turned out to be very tiny; it was much easier to do than I had
> initially thought.
>
> The patch I sent was bigger, but that's because it included an
> extension module to provide access the core function for pure-Perl
> modules, and a demonstration pure-Perl module, and I think some other
> stuff.
>
> I am also planning to leave the existing $^H mechanism in place.
But %^H could be suppressed probably.
> I don't know what 'the autobox thing' is. If it's that references to
> unblessed arrays can have methods called on them, which would be
> dispatched through package ARRAY, then possibly, although it would be
> rather a stretch without at least a little more core support than is
> presently in there. I should probably avoid speculating.
http://search.cpan.org/~chocolate/autobox-0.10/lib/autobox.pm
Mark Jason Dominus
> Mark Jason Dominus wrote:
> > I am also planning to leave the existing $^H mechanism in place.
>
> But %^H could be suppressed probably.
Either that, or else it would be easy to implement a working version
of it atop the lexical pragma module. One idea I was toying with was
of using %^H as an alternative interface to the low-level features.
Aaron Sherman
> http://search.cpan.org/~chocolate/autobox-0.10/lib/autobox.pm
Would there be any way for a module to "turn on" such a pragma for the
caller? For example, if I had this module:
package ARRAY::Length;
require Exporter;
use autobox;
push @ARRAY::ISA, 'ARRAY::Length';
our @ISA = 'Exporter';
sub length { return scalar @$_ }
Would the caller always have to:
use autobox;
use ARRAY::Length;
? Or to state another way, can I export pragmas in some way? Those with
long memories will recognize this as a feature that I sent in a patch to
the core for a very long time ago. At the time, popular wisdom was that
"objects in the core" were a bad thing, but if we're going that way, I'd
like to write a few utility libraries like the above.
--
Aaron Sherman <a...@ajs.com>
Senior Systems Engineer and Toolsmith
"It's the sound of a satellite saying, 'get me down!'" -Shriekback
Rafael Garcia-Suarez
>
> On Mon, 2004-01-12 at 05:24, Rafael Garcia-Suarez wrote:
>
> > http://search.cpan.org/~chocolate/autobox-0.10/lib/autobox.pm
>
> Would there be any way for a module to "turn on" such a pragma for the
> caller?
Hmm, isn't this how pragmas work ? (strict, warnings, etc.)
Ilya Zakharevich
> Ilya had done some work towards making unpack be able to handle any C
> structure
I may be able to reintegrate this; especially if the consenseus is
nice to me ;-). IIRC, I sent about 5 patches; about 2 of them were
included (see perlpacktut (sp?) - it is mostly about them). I had two
more patches in mind - and with them pack/unpack should become a
smooth operation.
Yours,
Ilya
Aaron Sherman
strict turns on bits in $^H. Pardon my ignorance (had my head buried in
my own code too long to keep up with this) is that how autobox does its
thing? Can my module use autobox and have $^H set for the caller as
well? How does that scoping work?
Tassilo Von Parseval
> On Mon, 2004-01-12 at 08:41, Rafael Garcia-Suarez wrote:
> > Aaron Sherman wrote:
> > >
> > > On Mon, 2004-01-12 at 05:24, Rafael Garcia-Suarez wrote:
> > >
> > > > http://search.cpan.org/~chocolate/autobox-0.10/lib/autobox.pm
> > >
> > > Would there be any way for a module to "turn on" such a pragma for the
> > > caller?
> >
> > Hmm, isn't this how pragmas work ? (strict, warnings, etc.)
>
> strict turns on bits in $^H. Pardon my ignorance (had my head buried in
> my own code too long to keep up with this) is that how autobox does its
> thing? Can my module use autobox and have $^H set for the caller as
> well? How does that scoping work?
If I understand you correctly, then all you need to do is have your
package inherit from autobox. That way, the caller just can do
use your_pragma;
{
no pragma;
...
}
I have an unfinished module in my home-dir where I do that. I override
'import' (it ultimately calls autobox->import). 'no' is inherited
unchanged by my module.
Yitzchak Scott-Thoennes
> On Sun, 2004-01-11 at 15:34, Yitzchak Scott-Thoennes wrote:
> > What would you all like to see in 5.9.1? What needs to be done to
> > make it happen? Who is going to do it?
>
> talked about before implemented. All the pieces exist, according to
> the discussion following the suggestion, and I've applied for a YAS
> grant to sponsor me putting them together.
Could you give any pointers to this discussion?
David Nicol
> make it happen? Who is going to do it?
I'd like to see the reference-as-lvalue aliasing syntax I have
talked about before implemented. All the pieces exist, according to
the discussion following the suggestion, and I've applied for a YAS
grant to sponsor me putting them together.
--
david nicol
"Take your time." -- Allan Quaterman
David Nicol
> > Ilya had done some work towards making unpack be able to handle any C
> > structure
Hooray!
David Nicol
> Could you give any pointers to this discussion?
http://www.nntp.perl.org/group/perl.perl5.porters/70055
Marcus Holland-Moritz
What were those patches actually implementing?
Where can they be found?
> Hooray!
<advertising>
Have you had a look at this one?
http://search.cpan.org/~mhx/Convert-Binary-C/
</advertising>
--
Bender: Hey! What kind of party is this? There's no booze and only one
hooker.