Re: Problem building mm-common on MSYS2

[Kjell Ahlstedt <kjellahlstedt gmail com>]

On 2021-04-15 02:24, Oliver Niebuhr wrote:
> On 14/04/2021 22:55, Oliver Niebuhr wrote:
> [cut for readability]
>
> I do not know if this is the Solution but I added 'sys.executable' and 
> the Error Message went away:
>
> result = subprocess.run([sys.executable, 'aclocal', '--print-ac-dir'],
>                           stdout=subprocess.PIPE, 
> stderr=subprocess.DEVNULL,
>                           universal_newlines=True)
>
> I have no Idea if this has any other Impact. I was just following an 
> Example from Digital-Ocean after a Web search about Python and 
> Parameters.
>
> As no additional Files are getting installed (compared to when running 
> into that Error), I assume the Installation itself was fine before 
> already.
>
> BTW: Do I need a mm-common Package for every bitness or can I use the 
> x86_64 Package for x86 builds?
>
> Thanks again for your time!
You need mm-common only if you will build any of the mm packages 
(sigc++, glibmm, gtkmm, etc.) with maintainer-mode=true. That's 
necessary if you build from the git repository. If you build from 
tarballs with maintainer-mode=false, mm-common is not necessary.

If Python's subprocess.run() can't be trusted in MSYS2, or in Windows in 
general, it's not a fatal bug when mm-common is built or installed. 
extra-install-cmd.py is not called, if Meson does not find aclocal. The 
problem for you is that Meson finds aclocal, but subprocess.run() in the 
Python script does not. But if there is a general problem with 
subprocess.run() in MSYS2, you may have trouble elsewhere when you build 
mm modules. subprocess.run() is used in several Python scripts. Perhaps 
you can avoid all the other subprocess.run() if you build from tarballs 
with maintainer-mode=false and build-documentation=false (default when 
building from tarballs).