I’ve got my fork of Panda working on Niecza. There are two major issues which remain to be resolved, and they are both related. Panda assumes that the Perl 6 executable is named perl6, and modules should be installed to ~/.perl6. That’s great if you have just one Perl 6 on your system.
But in this crazy modern world, I really want both Rakudo and Niecza on my system. And so assuming that everything can be named “perl6” is a big problem.
I’m hoping the community can put their heads together and figure out a solution. I guess the obvious one (which would only need sorear to buy in) would be installing Niecza modules to ~/.niecza, and creating a “real” niecza executable. (It might well be as simple as a one-line script “mono path-to-Niecza.exe”.)
Good idea or bad idea?
Just Rakudo It 
January 29, 2012 at 3:30 am |
I guess the correct solution for the multiple compiler names is a portable way to get the whole command line for invoking the compiler, so $*COMPILER_CLI would return ‘perl6’ on Rakudo and ‘mono Niecza.exe’ or [‘mono’, ‘Niecza.exe’] on niecza. Not sure which variation makes more sense.
January 29, 2012 at 3:34 am |
I don’t know if that’s quite enough — at least at the moment, the syntax for setting up include paths is different between Rakudo and Niecza, so just knowing how to invoke the compiler isn’t enough. I’ve been figuring these details need to be abstracted away into a module somewhere…
January 30, 2012 at 6:15 am |
Of course, I am a retard and know nix all about this stuff.
But isn’t this sort of problem dealt with by the P5 installer, posssibly in the Strawberry Perl incarnation ?
Maybe alias or one of the Strawberry guns can help out (prob in the AUTHORS or CREDITS in the Strawb dist) ?