Re: AC_INIT translates PACKAGE to lower case

[Ralf Corsepius]

Am Mit, 2002-01-30 um 14.15 schrieb Akim Demaille:
> >>>>> "Ralf" == Ralf Corsepius <address@hidden> writes:
> 
> Ralf> If using the new AC_INIT and AM_INIT_AUTOMAKE syntax, PACKAGE
> Ralf> gets translated to lower case letters. - Why this change?
> 
> Because that's the case for most packages.  But that's not true:
> PACKAGE_NAME is not lower cased.
Right, but the problem here is backward compatibility. 

People are using @PACKAGE@ and will be surprized if things will change,
"just because of having upgraded autoconf/automake".

[E.g. I am using @PACKAGE@ in autoconf generated rpm-specs.]

Sorry for having to say this, but in my opinion you are unnecessarily
complicating things here.

> Ralf> Furthermore, PACKAGE_TARNAME also is transformed to lower case
> Ralf> letters - Why this change?
> 
> It _is_ true for TARNAME.  Which is the one that makes sense for
> Automake when it makes dist.

You seem to be missing, that packages which use mixed upper and lower
case letter package and tarball names for whatever reasons for years
will out of a sudden be named differently. 

This will break package managers' infrastructures and a lot more.

[This will cause each and every Linux distributor to rework their
package databases and distributions - Do you really want this?]

> Ralf> IMO, both changes unnecessarily complicate things and avoidably
> Ralf> break well-known features of automake, without further reason.
> 
> Unification.  Let's say that you are used to the present scheme, but
> for a fresh person, there is obviously a problem.
Why? IMHO, autoconf here is going too far. I can't see a single
technical reason why packages should be lowercase only.

Furthermore, I do think that manipulating names should be left to
packagers, maintainers and copyright/tradmark owners and not be subject
to autoconf's activity at.

Do you really think it would be clever if Xt, Xm and TeX suddenly would
have to be renamed, just because autoconf wants it this way?

Actually, I think you are going to open a can of worms and that this
change really means looking for trouble.

> Ralf> In particular this change is in contradiction to this sentence
> Ralf> from info automake.info (1.4):
> 
> Ralf> Automake doesn't do any interpretation of `PACKAGE' or
> Ralf> `VERSION'.
> 
> Ralf> [Strictly speaking this sentence remains valid. It's autoconf
> Ralf> that interferes.]
> 
> Right.
> 
> What do you think about the following patch, that I just applied.

Well, this change is a compromise, but one I actually do not like and do
not want to swallow.

In one sentence: I consider the new behavior to be a very bad idea and
would like to see the old behavior back.

Ralf

PS.: I am cc:-ing Havoc, because his GConf would be directly affected.