Re: distinguish automake 1.11 from 1.12+ at autoconf time

[Ralf Corsepius]

On 08/08/2012 05:13 AM, Bob Friesenhahn wrote:
> On Mon, 6 Aug 2012, Eric Blake wrote:
> > What _specific_ feature are you using from 1.12 that wasn't present in
> > 1.11?  Or put another way, either your configure.ac works equally well
> > with both versions (so you don't care which version), or there's
> > something you do with 1.12 that doesn't work with 1.11 and therefore you
> > are trying to make it conditional so that 1.11 can still process the
> > rest of the file.  Determine that feature, and we are that much closer
> > to helping you conditionalize your configure.ac to work with both
> > versions.
>
> This logic is reversed.  I am mostly worried about features that my
> package currently depends on which have been announced to be deprecated
> for no apparent reason, and without a sufficiently functional
> work-around, and with only the sound of crickets chirping when these
> issues are brought up on the list.

That's one point ... worse are such cases, when automake changes 
behavior of something which had worked for many years, between of minor 
automake releases. Such cases simply must not happen.

> Most likely others are also worried that their project will stop working
> properly if a newer Automake is used.
This worry is justified.

Esp. with packages, whose upstream maintainers do not ship prebuilt 
auto*tools generated files and want their users to rerun the autotools 
(IMO, this qualifies as improper use of the autotools), and which 
packages which for one or the other reason need the auto*tools to be 
rerun during building a package, is problem is for real.

In Fedora, after mass-rebuilts (Rebuilding all 10000+ packages the 
Fedora distributions consists of), we typically are facing larger 
numbers of packages, failing to build just because of some auto*tool 
version-jump having introduced behavioral or API changes and have broken 
something.

> Automake provides a means to
> prevent a lesser version of Automake from being used, but not a greater
> version because it was always assumed that documented Automake
> functionality was like a contract which would never be broken without
> good cause.
Exactly.

Ralf