Re: automake transition breakages
- To: Russ Allbery <firstname.lastname@example.org>
- Cc: email@example.com
- Subject: Re: automake transition breakages
- From: Ian Jackson <firstname.lastname@example.org>
- Date: Fri, 11 Oct 2013 18:00:42 +0100
- Message-id: <[🔎] email@example.com>
- In-reply-to: <firstname.lastname@example.org>
- References: <1380474955.16141.27835205.4F73148A@webmail.messagingengine.com> <20130930025001.GA3116@gambit> <1380532166.5916.28054821.3053B125@webmail.messagingengine.com> <20130930145817.GD3116@gambit> <5249AD15.email@example.com> <20130930172521.GN3076@mraw.org> <firstname.lastname@example.org> <20130930175117.GO3076@mraw.org> <email@example.com>
Russ Allbery writes ("Re: automake transition breakages"):
> Cyril Brulebois <firstname.lastname@example.org> writes:
> > Meaning one needs to get a fix once; a random NMUer might need to chase
> > upstream patches (never a pleasure when fixing unrelated issues). Given
> > what you wrote, I'm not sure what it buys us to keep -Werror in
> > distribution packages, except for the possibly ticking time bomb.
> Well, the simple answer is that we avoid having to maintain local patches
> against the upstream source to remove -Werror, since generally it's not
> something over which we have a choice without patching the upstream
> source. It's usually expressed in configure.ac.
I don't think this is a sufficiently good reason to take the
Think, for example, about downstreams (including users) who may want
to forward-port a package to a different Debian release and thus a
different automake version.
For such a central tool as automake, we should tolerate lots of
version skew slop.