On Fri, Jul 1, 2016 at 8:21 AM, Johan Vromans <jvrom...@squirrel.nl> wrote:
> Hmm. This smells like PAR is factually a discontinued product. > Would you call your Getopt::Long a "product"? The PAR ecosystem is certainly not discontinued, it's being used and I intend to support it, just check the mailing list archives. That its internals are crufty shouldn't distract you from using it, otherwise why use Perl at all since the same holds for it. The official PAR website par.perl.org is dysfunctional as well (it links to > the perl6 web site). > Where did you get this "official PAR website" info from, I thought I purged all references to it. BTW, regarding you summary of PerlApp, Cava and PAR: PAR Packager on the other hand produces a Perl-agnostic binary. All Perl > stuff (libperl, modules, XS libs) is packaged. Everything else needs to be > present on the end user system. This leaves the burden of installing > wxWidgets and such to the end user, something she may not be able to do. > That is not true. PAR::Packer is able to package non-XS DLLs, too. But it lacks the tools to automagically discover and locate them and puts this burden on the packager. But judging from the mailing list, there are people that do just that. And if you know any CPAN module that can do this, just drop me line. The makers of PerlApp and Cava have obviously solved this (and charge you for it), but it's highly non-trivial even for their limited set of supported platforms and in fact has nothing to do with Perl anyway. Handwaving arguments like "there's ldd" don't cut it. And when it comes to packaging complete "apps" containing modern GUIs, using Perl based solutions is probably the wrong direction anyway. Cheers, Roderich
