Brian:
I second Eric's affirmation of the usefulness of Inline::C and
friends! I have been trying to get a number of people here to
start using XS to wrap legacy libraries (instead of depending
on me to do all that work!)
Late last week I pointed one of the other engineers in the direction
of Inline::CPP to wrap the ISearch libraries for a Perl module he was
writing. He had only one or two small problems (including an overlap
with two #defines in the CORE perl headers [POPSTACK() in cop.h and
Copy() in handy.h])! After he got his Inline code working we
converted the generated file into a full-blown XS module in just 5 to
10 minutes!
If Inline becomes part of CORE you might need to add some extra
documentation to specify how to troubleshoot problems like we ran
into. But that is not a major limitation!
Now I am *NOT* the only one here that knows how to wrap libraries
for Perl modules!
Paul E Wilt
Principal Software Engineer
____________________________________________________
XanEdu, Inc. ( a division of Bell+Howell Information&Learning)
http://www.XanEdu.com
mailto:[EMAIL PROTECTED]
mailto:[EMAIL PROTECTED]
300 North Zeeb Rd Phone: (734) 975-6021 (800)
521-0600 x6021
Ann Arbor, MI 48106 Fax: (734) 973-0737
____________________________________________________
-----Original Message-----
From: Kort, Eric [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, March 14,2001 2:46 PM
To: [EMAIL PROTECTED]
Subject: FW: Inline as a 5.8.0 core module.
> -----Original Message-----
> From: Brian Ingerson [mailto:[EMAIL PROTECTED]]
> Sent: Wednesday, March 14, 2001 2:36 PM
> To: [EMAIL PROTECTED]
> Subject: RFC: Inline as a 5.8.0 core module.
>
>
> Jarkko Hietaniemi (the Perl 5.8 "Pumpking") wrote me on
> Monday to ask if
> I would consider putting Inline into Perl release 5.8.0 as a core
> module. This is tenatively scheduled for release mid-summer 2001.
> Although my initial reaction was "Yee Haw!!", I've come to the
> conclusion that I should proceed with care.
I think that Inline has changed the landscape in an important way.
It is now easy to address problems heterogeneously, thereby using
the best tools for each part of a given problem. In fact, I would
like to write an article for a bioinformatics journal demonstrating
how Inline allows the bioinformatician to capitalize on the ease of
string and file processing with Perl while doing the intensive math
in C.
Anyway, this is all by way of saying that I believe Inline should be
a core Perl module.
> 3) The Inline ILSM API will need to be set in stone. The release of
> Inline/Inline::C will be more closely bound to the release of
> Perl.
But there can always be an Inline_dev::C or Inline_dev::Java only
available to Friends of Brian that we can play with right? The number
of people who would be really concerned about this is, by my
calculation, 17, and they all subscribe to this list.
So, from my perspective, this is great news. Good work, Brian, you
have successfully changed the world.
Eric
