[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

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

From: Han-Wen Nienhuys
Subject: Re: config.status has been broken by issue 5780 "Accept GUILE 2 without extra configure options"
Date: Sun, 8 Mar 2020 15:17:24 +0100

On Sun, Mar 8, 2020 at 3:04 PM David Kastrup <address@hidden> wrote:

> Ok, I am obviously missing something important here.  You go to a lot of
> pain to use pkgconfig when someone specifies guile-config.  Now let's
> just assume that someone uses guile-config because they either don't
> know better or it's the easiest way to muddle through (which makes being
> transparent about what happens a very good idea either way): is there a
> fundamental drawback in just using guile-config then?
> Now make no mistake: most of the work invested into using pkgconfig in
> this patch actually _is_ beneficial (even should we decide to use
> guile-config in the end) in that it leads to being able to make a very
> pinpointed (and even tested as workable) suggestion to the user what to
> do in future.
> So I am definitely grateful for the work you put in here, and whatever
> we decide, the bulk of it will be put to good use.  But apparently there
> is a good reason for you not wanting to see guile-config used even if
> specified (which technically means that the user can only blame
> themselves if something goes wrong), and that reason likely revolves
> around a technical consideration or necessity, one that likely should
> also apply to whoever feels using GUILE_CONFIG is a good idea.
> Is there any more to it other than "it has been deprecated"?  I feel
> like I am missing some aspect of what is involved here.

we have to keep around and maintain the code that makes supporting
GUILE_CONFIG work. This is a cost which we pay. The upside is that a
certain group of people won't have to adapt some of their build
scripts.  I posit that this is a very small group of people, and that
they're generally expert enough to follow updated instructions. In
other words, I think the cost to them will be small, and it's better
for LilyPond if we make our own lives easier.

(I've been trying to keep out of the discussion, because this
discussion is taking up more time than it seems worth spending.)

Han-Wen Nienhuys - address@hidden -

reply via email to

[Prev in Thread] Current Thread [Next in Thread]