Re: [PATCH] build-sys: fix configure --without-systemd

[L A Walsh]

On 2020/01/28 04:07, Karel Zak wrote:

> On Fri, Jan 24, 2020 at 12:29:47PM -0500, Theodore Ts'o wrote:
>   
> >  AC_ARG_WITH([systemd],
> >    AS_HELP_STRING([--without-systemd], [do not build with systemd support]),
> > -  [], [with_systemd=check]
> > +  [], [with_systemd=no]
> >     
>
> The current default is to check for the libraries, if installed than
> enable systemd support. This is generic way we use for many libs and
> features. Why do you think that explicit --enable-* will be better?
>   
---

   Perhaps Ted didn't realize that the line he replaced
was only the default action (1st bracket pair empty => no option).
Initially, I thought the same until I read further
down (two lines):

have_systemd=no
AS_IF([test "x$with_systemd" != xno], [ ...
  test for positive case ]

If any of the tests fail, then 'have_systemd=no'
is the option passed along.

I don't _think_ Ted is asking for for the default to be
changed, as the subject of his note states it is
to fix the "--without-systemd" case, but a
fix was added for a bug that wasn't present,
so I don't think he knowingly did what he
intended to do.