Hi all,

I've opened an issue on https://github.com/pypa/wheel-builders, but that got 
immediately closed,

so I guess that was the wrong place to ask/contribute, haha.

Here's my situation: I'm not *that* experienced in the "build stuff with gcc" 
department,

i.e. I've done it before but I don't have the in-depth knowledge.

What I'm trying to do is to create a Kivy build that "just works" on as many 
systems as possible.

Motivation: As a new guy, if the first thing you're greeted with upon 
installation of a framework is

some obscure Cython error, then you're probably not going to bother.

(Python + GUI really is a sore spot in the overall ecosystem.)

I've been playing around with different systems and options for a while now 
(including manylinux1), though,

and have started to learn how to properly build RPMs.

My main question at this point is:  should I even bother with manylinux1?

I get that the reason for going back as far as possible is to support as many 
systems as possible.

But RedHat 5 is on extended support only now, that means CentOS 5 also won't 
ever receive any fixes anymore,

if I understand that right.  That is, if I were to use the provided docker 
image, that'd be a potential security issue in itself.

Also, trying to "go back in time" with newer software releases is a major 
effort here (relying on SDL2 and all that involves).

So basically, I think I have 3 options:

- Not even trying to be compatible with manylinux1.

- Backport everything from the latest minor release of CentOS 6 (CentOS only 
supports the latest minor release).

- Try to make it work on CentOS 6, first.  Then try and see if I can find and 
fix the incompatibilities introduced by that.

  But that is over my head, still.

Any advice would be appreciated.

Cheers,

Dominik
_______________________________________________
Wheel-builders mailing list
Wheel-builders@python.org
https://mail.python.org/mailman/listinfo/wheel-builders

Reply via email to