Re: CHECK_STRUCTS/dmpstruct.h mechanism is broken.

[Eli Zaretskii]

> Cc: address@hidden
> From: Paul Eggert <address@hidden>
> Date: Thu, 11 Apr 2019 15:23:15 -0700
> 
> On 4/11/19 12:15 PM, Eli Zaretskii wrote:
> > You want to test for system-dependent features with Make?  How does
> > one do that?
> 
> You execute the test as a 'make' rule that records the test's result as
> a makefile snippet (or as a .h file, or whatever).

Sure, but that means we'll need to rewrite all the Autoconf tests as
Make scripts, complete with the test programs and other stuff.  Why is
such an effort worth our while?

> > Wouldn't we end up with heaps of Make wizardry only a few understand?
> Sure, just as we now have heaps of .m4 and shell etc. wizardry that only
> a few understand.

There's enough high-level, well documented Autoconf infrastructure
built on top of m4 that even I can write tests quite easily.  If we
move to Make, we will have to reinvent that wheel entirely from
scratch.

> However, an advantage of using Make instead of M4 is that we already
> need to use Make anyway

I'm sure you know that we currently use about 10% of what GNU Make can
do.  Moreover, using sophisticated Make facilities would mean we have
to require relatively recent versions of Make, which will be a
complication for some platforms.

> Another advantage is that Make can easily run in parallel whereas
> 'configure' does not.

The disadvantages are so many that this advantage drowns in them.

> 'cd lisp; make -j16 compile-always' took 38 real-time seconds

That's not the full build, you miss regenerating many files that
happens in a fresh checkout.