Re^2: RFC: Compiling C/C++ based Modules under ActiveState using MinGW

[syphilis]

in reply to Re: RFC: Compiling C/C++ based Modules under ActiveState using MinGW
in thread RFC: Compiling C/C++ based Modules under ActiveState using MinGW

I'm surprised you didn't mention Randy Kobes' excellent PPM::Make module

Sorry, I meant to reply to this earlier - got waylaid :-)

I don't follow the relevance of this wrt this particular meditation/tutorial. We all know that randyk does only work of the highest quality (not to mention the admirable character of the man), but I think a treatise on PPM::Make would be more appropriate in a meditation regarding the building of PPM distro's. When it comes to building and installing modules, I think the relevance of PPM::Make is somewhat tenuous (to say the least) - which is the reason that I haven't mentioned it. (And I do want to keep it simple.)

I think I see where you're coming from, btw - doing it that way means that PPM keeps track of everything that's installed.

(Great sig !! - I work in the dairy industry :-)

Cheers,
Rob

[TGI]

I think I see where you're coming from, btw - doing it that way means that PPM keeps track of everything that's installed.

Yep. My rationalization for suggesting PPM::Make is that it a basic part of the module acquisition process ( Download -> Build -> Install ). I guess my thinking lumped build and install together, while you were focusing just on the build portion of the process. I definitely see your point.

Your node is certainly a welcome update. I recall the frustration I went through trying to get some modules compiled and installed while struggling under a ludicrous deadline. A resource like this would have been invaluable. I hope my above response was not taken as an attempt to denigrate your contribution.

Maybe when I get out from under my latest ludicrous deadline, I'll get around to writing a review of PPM:Make or a tutorial on building PPMs. Anyone know where I can purchase a metric ton of tuits?

[syphilis]

you were focusing just on the build portion of the process

Not really ... the difference seems to be that I'm suggesting we run '*make install' (where '*' is either 'd' or 'n'), whereas you're suggesting that, having built the module, one then builds a PPM, then installs it via 'ppm install ...".

I hope my above response was not taken as an attempt to denigrate your contribution

Good Heavens !!, no :-)

I'll get around to writing a review of PPM:Make or a tutorial on building PPMs

Excellent idea, and I encourage you to do so. I've built a few PPM's from time to time - by tar'ing and gzip'ing the blib (using GnuWin32's tar and gzip) and by hand-creating the ppd file. I'm sure that PPM::Make simplifies the process.

One thing I wonder about (having never used PPM::Make) is how it handles the situation where a dll needs to be installed via a post install script. See, for example the line <INSTALL EXEC="PPM_PERL" HREF="http://theoryx5.uwinnipeg.ca/ppms/scripts/install_pdl">install_pdl</INSTALL> at http://theoryx5.uwinnipeg.ca/ppms/PDL.ppd. Does PPM::Make catch the need to insert that 'INSTALL EXEC' entry ? Does it write the post install script ? (Just a couple of issues for you to address when you come to write that tutorial.)

Thanks for your input.

Cheers,
Rob