Inline-0.40-TRIAL3 is available at:
http://www.ttul.org/~ingy/Inline-0.40-TRIAL3.tar.gz
Changes include:
- $o->{API}{location} added back into the Inline object.
- write_inline_file() called automatically by Inline.pm
- ILSMs don't need to be concerned with the .inl file.
- C-Cookbook buglet with pod2man fixed
- Inline verifies $Config::Config{archname} to make sure the object was
compiled for that platform.
- $Config::Config{apiversion} is saved in the .inl, but I don't verify
it yet.
Patrick asked whether to use $o->{ILSM} or $o->{Java} for ILSM values. I
prefer $o->{ILSM} because then someone can more easily subclass your
module. Inline::CPP is now a subclass of Inline::C! That reduces a lot
of redundant code. But I don't see people subclassing Java and Tcl with
anything, so if the authors want to continue using the old standard,
that's OK by me.
Binkley reported problems on Cygwin but I could not reproduce them. He
noticed that my Cygwin was several months older. I'm upgrading Cygwin as
I type, but the process is so damn slow (on VMWare) that I'll have to
put off testing until tommorrow.
I'm pretty sure there's new -T (taint) problems, but I need to test more
and I'm not in the mood right now :)
Jochen Stenzel, are you out there listening? I'd like you to test
Event.pm compatability.
Please keep hammering everybody.
Thanks, Brian
--
perl -le 'use Inline C=>q{SV*JAxH(char*x){return newSVpvf
("Just Another %s Hacker",x);}};print JAxH+Perl'
