On 23 May 2015 at 12:08, Andrew Barnert via Python-ideas <python-ideas@python.org> wrote:

Sorry, meant to include this in my previous reply, but I accidentally cut and didn't paste...

Sent from my iPhone

On May 22, 2015, at 17:08, André Freitas <p.andrefreitas@gmail.com> wrote:

Hi,

Thanks for sharing Ryan Gonzalez :)

It just could be another alternative and not a replacement of autotools.

If the problem is that the autotools build system is a nightmare to maintain, how is having two completely different complex build systems that have to be kept perfectly in sync not going to be an even bigger nightmare?

Three - we already have to keep autotools and the MSVS solution in sync (except where they're deliberately different, such as always bundling OpenSSL on Windows).

I don't think there's actually any active *opposition* to replacing autotools, there just aren't currently any sufficiently compelling alternatives out there to motivate someone to do all the work involved in proposing a change, working through all the build requirements across all the different redistributor channels (including the nascent iOS and Android support being pursued on mobile-sig), and figuring out how to get from point A to point B without breaking the world at any point in the process.

That said, I'll admit that to someone interested in the alternatives, listing some of the problems that autotools is currently solving for us may *sound* like opposition, rather than accurately scoping out the problem requirements and the transition to be managed :)

Cheers,

Nick.

--

Nick Coghlan   |   ncoghlan@gmail.com   |   Brisbane, Australia