Re: config.status has been broken by issue 5780 "Accept GUILE 2 without extra configure options"

[Jonas Hahnfeld]

Am Sonntag, den 08.03.2020, 11:54 +0100 schrieb David Kastrup:
> Han-Wen Nienhuys <
> address@hidden
> > writes:
> 
> > On Sun, Mar 8, 2020 at 11:33 AM David Kastrup <
> > address@hidden
> > > wrote:
> > > Jonas Hahnfeld <
> > > address@hidden
> > > > writes:
> > > 
> > > > Am Samstag, den 07.03.2020, 23:20 +0100 schrieb David Kastrup:
> > > > > So assuming GUILE_CONFIG is set, that should be tried in preference
> > > > > to a .pc file, giving a warning (an error would be a nuisance when
> > > > > trying to have a common configuration for both 2.20 and 2.21, a
> > > > > consideration that will not really be relevant any more for 2.22 and
> > > > > 2.23).
> > > > > 
> > > > 
> > > > The following prints an error and directs the integrators into the
> > > > right direction:
> > > > diff --git a/configure.ac b/configure.ac
> > > > index 29e4e5680f..80a34f7b09 100644
> > > > --- a/configure.ac
> > > > +++ b/configure.ac
> > > > @@ -189,6 +189,11 @@ STEPMAKE_MSGFMT(REQUIRED)
> > > >  STEPMAKE_TEXMF(REQUIRED)
> > > >  STEPMAKE_TEXMF_DIRS(REQUIRED)
> > > 
> > > What about "an error would be a nuisance when trying to have a common
> > > configuration for both 2.20 and 2.21" was unclear?

There will never be a shared configuration for both 2.20 and 2.21:
Current master requires Python 3 which 2.20 not even attempts to be
compatible with.

> > I don't understand this whole discussion. Who are the system
> > integrators that we are doing all this work for? You yourself have
> > said we'll probably never have a 2.20.1, so under what conditions is
> > it useful to have shared configuration?
> 
> This would concern things like running Patchy, and also things like
> checking out pretests of stable releases for system packages.  If the
> spec files of the stable release fails mysteriously, most users will
> give up.

With the patch it doesn't fail "mysteriously" - there's a clear error
saying what the tester is supposed to do. And from my understanding
"unstable" releases really means that.

> I cannot believe the resistance against creating a few dozen lines for
> making the life for users and testers of LilyPond easier and insisting
> on a configuration that will fail for everything except a single
> painstakingly "correct" use that is not documented.

As I wrote yesterday, the whole thing wasn't documented before. I
politely ask to take a step back and try to understand the point of
view shared by Werner, Han-Wen and me.

Jonas

signature.asc
Description: This is a digitally signed message part