Jarkko Hietaniemi wrote:
>
> I guess my basic question is: are you game? Can you get Inline
> into such a shape by 5.8 that you will be proud? Can I mention
> publicly Inline to be under consideration for inclusion?
Jarkko,
I've been going back and forth on this all week. I've talked to a lot of
people both on the mailing list and here at ActiveState. Here are my
thoughts for and against including Inline in the 5.8.0 distribution:
For:
1) It gives Inline the exposure and access to the 'other' platforms and
their porters. I am *really* interested in seeing this happen because I
think Inline can add a lot of value to Perl on these platforms, and that
in turn boosts the value of Perl overall.
2) If Inline is to eventually be the successor to XS, it needs to be a
core module.
Against:
1) Inline is only 8 months old. It is young and evolving rapidly.
"You'll rewrite it 2 or 3 more times", sayeth Sarathy. Jan also thinks
it's too early. I tend to agree.
2) Inline is also slightly different than the average module. It's a
meta-module of sorts. A framework for building extension modules. I
don't want to limit it's definition artificially. It needs to
settle/mature in its own time.
3) There are currently no extension modules that require Inline. I'd
like to be filling a real need rather than an expected one.
4) I would like to see a few more ILSMs implemented, before I lock down
the Inline API.
My Decision:
I definitely want to see Inline/Inline::C as a core module. Eventually.
I
would like to put Inline into the 5.7 development tree _now_, and start
down the road of getting Inline ready for core distribution. But I would
like to _not_ have a deadline of 5.8.0. It will be evident when Inline
is ready, and that will most likely be sometime after 5.8.0.
Again, I really appreciate the opportunity. I just want to do the right
thing for the long term. I hope this plan makes sense to you as well.
Cheers, Brian
--
perl -le 'use Inline C=>q{SV*JAxH(char*x){return newSVpvf
("Just Another %s Hacker",x);}};print JAxH+Perl'
