configure separate from make or not (was: Using LS/LE)

[Alejandro Colomar <alx@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 2398 bytes --]

Hi Ingo,

> Date: 2026-08-29 11:39:20+0200
> From: Ingo Schwarze <schwarze@usta.de>
>
[...]
> 
> I think the main reason why build system developers don't usually
> run configure automatically from the Makefile is that from a user
> perspective, it makes sense to keep the two steps "inspect the
> configuration and make some decisions, but do not start building
> anything just yet" and "build the stuff, using the variables and
> decisions made at the configure stage" separate.  Users may even
> want to pause between the two steps, consider the decisions made,
> for example by inspecting configure output files like config.h
> or Makefile.local or, in case serious debugging is needed, config.log
> and maybe even change some of the decisions manually and re-run
> configure, even though autoconf(1) makes both steps, inspection
> and manipulation of decisions, gratuitiously hard, both by spewing
> vast amounts of noise and by the ./configure script being next to
> unreadable.  But saner, better-written configure scripts can make
> both inspection and manipulation quite simple and pleasant.
>
> Running the configure script automatically from the Makefile
> would remove this opportunity for inspection and tweaking.

In the Linux man-pages, the configure step is done entirely within
make(1) --there's no separate ./configure step--.  But as you say,
a separate configure-only step is important for serious debugging.

The solution I implemented is a 'nothing' target in the Makefile, which
as the name implies, does nothing.  However, it still runs the
configuration step, since that runs as part of parsing the makefiles.
Here's the rule:

	nothing:;

When I need to debug the build system itself, I run that target; often
with special flags for debugging.  One common command I use when I need
to do this kind of debugging is:

	$ make -p nothing | less

And if I need to see stderr output from the commands used by the
configure scripts, here's a common command:

	$ make HIDE_ERR= nothing

I'm not saying that this should be done, but this _can_ be done.  With
BSD make(1), it might be more difficult to debug, since it doesn't have
the -p flag.  But the output from configure tests can still be read
without building anything, with a dummy 'nothing' rule.


Have a lovely day!
Alex

-- 
<https://www.alejandro-colomar.es>

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]