automake-patches
[Top][All Lists]
Advanced

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

Re: FYI: use api version


From: Alexandre Duret-Lutz
Subject: Re: FYI: use api version
Date: Thu, 17 Jan 2002 10:25:57 +0100
User-agent: Gnus/5.090004 (Oort Gnus v0.04) Emacs/21.1 (i386-debian-linux-gnu)

>>> "Tom" == Tom Tromey <address@hidden> writes:

 Tom> Now we use the API version, not the raw version, for versioned
 Tom> installs.  This means we must guarantee compatibility in, say, the 1.5
 Tom> series.  However I think that is easily achieved.

I'm very sceptic about this.  The current definition

  APIVERSION=`echo "$VERSION" | sed -e 's/^\([[0-9]]*\.[[0-9]]*\).*$/\1/'`

means that 1.5 1.5a 1.5b 1.5c 1.5d 1.5e will all be API-compatible.

If 1.5e is the last version before 1.6, makes only documentation
changes, and is API-compatible with 1.5, I don't understand why
1.6 is not deemed compatible with 1.5.

Actually, 1.5b is *NOT* API-compatible with 1.5c.  two reasons:

  1) a Makefile working with 1.5b may fail with 1.5c 
     I can show at least one example of this.

  2) a Makefile which works with 1.5c may fail with 1.5b
     Automake's own Makefile is an example of this.

Probably that's just a matter of including the 'b' or 'c' in the
APIVERSION.  Maybe the APIVERSION could remain the same for 1.5
and its bug fix releases 1.5.1, 1.5.2, etc.  (Actually, that's
not clear to me if this will always works.  To play the devil's
advocate: is fixing a bug someone relies on considered an API
breakage?)


I also have a question regarding the current implementation:
shouldn't `missing' and `init.m4' support versioned `aclocal' too?
-- 
Alexandre Duret-Lutz




reply via email to

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