New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Use of cmake for build subsystem #102
Comments
|
First of all, for the record, I don't know almost nothing about cmake itself. The sole purpose of this ticket is to find out if it makes sense of removing perl5 from the build process and if cmake can serve as a good replacement. I propose it because it became a de-facto standard for a cross-platform build tool. But before getting any deeper into it I would like to find out more about other opinions about it. Another alternative considered by that discussion was using NQP for the Configure script. But the prerequisites of this approach are such that wouldn't allow as to use it for a long time. One of the most problematic requirements is having pre-compiled and easily available MoarVM for all major platforms. Yet, even in this case MoarVM itself would require perl5 to be compiled. |
|
If you're looking to remove perl5 from the build process, I don't think it makes sense to remove it just from rakudo's build. If it goes forward, I think we'd want this work to cover MoarVM, NQP and Rakudo. |
|
I didn't mention it after all. Yes, it only makes sense if everything is using cmake. |
|
My personal opinion fwiw, but I don't see the point in removing perl5 only to replace it with something that I believe is less commonly available than perl5. Perl5 is available on every platform that Rakudo+NQP+MoarVM supports. I could see removing perl5 and replacing what it was used for with a tool already in use by the toolchain to reduce the total number of different tools needed. But otherwise it seems like a lot of work for too little gain. |
Me either, but I've had to interact with it a few times, iirc on Windows, and I don't remember it being terribly good fun. Not that I've ever found any build system good fun... :-)
If we do that, I'd rather we do it for NQP and Rakudo by:
That way we have it written in something that more members of the community are likely to comprehend and work on, rather than less (I'm quite sure that moving from Perl 5 to cmake would reduce the already-small number of people who are up for working on build stuff). Further, since we have a All of that said: who are we doing this for? If we expect most end users to use a package, MSI, etc. then the source builds are primarily for those of us who are working on it, and for those building packages. So the question is probably what's best for those two groups of people. |
|
Ok, I now have enough reasons for sparing my time for something more useful than learning cmake. Just for the bottom line, after it is agreed that cmake is a bad option:
|
A couple of months ago there was a discussion on IRC about getting rid of perl5 in build subsystem. I would like to find out if cmake could serve as a cross-platform replacement.
The text was updated successfully, but these errors were encountered: