Re: Python recompile

On 08/03/2025 23:02, Lawrence D'Oliveiro wrote:

On Sat, 8 Mar 2025 16:46:14 +0000, bart wrote:
 

What Make does is take a mountain of complexity called a 'makefile',
which is special kind of arcane language, and tries to run it. It might
work, it might fail immediately, or it might grind away for several
minutes and then it stops.

 All too common with Windows, I’m afraid.
 I remember when the LibreOffice project was set up as a fork from the
moribund OpenOffice, there was a blog post reporting on initial progress,
from Michael Meeks I think it was, saying that the Windows build still
suffered from mysterious intermittent failures, such that retrying the
build would usually succeed.
 That kind of thing is unheard of on proper *nix systems.

So, what you are saying is that some software that is developed using one particular environment, regarding its tools, resources, and testing, may fail on an alien environment.
That would not be unexpected. But what is wrong is to trash that other environment for being different.
In this case, /why/ was that build failing? Was it even using Windows, or some extra layers to emulate Linux on Windows?
You'd need to pinpoint the exact reason for failure before laying blame. Maybe the build process is 10 times more elaborate than is necessary, and it's going wrong in that 90% that you don't really need anyway.
For example, there's some steps there that might be relevant for some obscure corner of Linux, but are meaningless elsewhere.