Norbert Gruener wrote:
>
> Hi Brian,
>
> these days I had to defend my decision to use Inline instead of XS for
> my Filesys::AFS modules. Finally I came up with the idea described
> below.
>
> It is well known that the usage of Inline is ***THE PERFECT*** choice
> for developing extension modules. But I don't think that the
> "recompile" feature is always desired in a "production environment".
> There are situations when it is not wanted that a "recompile" is
> triggered automatically, e.g. no compiler on the production machine.
>
> Therefore I would suggest a "configuration option" where I can
> prohibit the recompilation of the "extension source code". And when
> suppressing the "recompilation" there should be the possibility to die
> the execution, to get a warning and then run, or to just ignore the
> whole stuff and run with the old library.
>
> What do you think about that idea ?
Good idea. And something that I will definitely support in my major
reworking of Inline coming this April. In fact you'll be able to say
something like:
perl -MInline=makeppd -MMyModule -e1
And it will produce a precompiled binary package that you can ship to
anyone else on the same platform as you.
perl -MInline=makeppd,nosource -MMyModule -e1
Will do the same thing except it will replace your Inline C source code
with the MD5 fingerprint string. (For the more proprietary minded)
Once installed these modules will never trigger a recompile. They will
simply die if they can't find their binary components. (But they always
should anyway)
Cheers, Brian
--
perl -le 'use Inline C=>q{SV*JAxH(char*x){return newSVpvf
("Just Another %s Hacker",x);}};print JAxH+Perl'
