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

Re: Automake 2.50 migration strategy, as implemented



On Sat, May 26, 2001 at 07:17:42PM -0400, Ben Collins wrote:
> On Sat, May 26, 2001 at 05:14:15PM -0400, Steve M. Robbins wrote:
> > On Sat, May 26, 2001 at 04:27:41PM -0400, Ben Collins wrote:
> > > On Sat, May 26, 2001 at 04:09:04PM -0400, Ben Pfaff wrote:

> > > > All that autobuilder maintainers, etc., should need to do, is to
> > > > make sure that they install autoconf2.13 as well as the new
> > > > autoconf.  This should not be challenging because the new
> > > > autoconf Recommends: autoconf2.13 for now.  (I intend to demote
> > > > the recommendation to a suggestion, then remove it, over the next
> > > > year or so, as it gradually becomes less necessary.)
> > > 
> > > However, installing autoconf negates having a clean chroot for the
> > > autobuilders. This means that packages with missing build-deps on
> > > autoconf will succeed, even though they are technically broken.
> > > 
> > > I suggest making that recommends a depends for now.
> > 
> > Why wouldn't you just special-case autoconf in the autobuilder
> > code?
> 
> Oh sure, we could special case that, and a million other things. It's
> not as if we work hard enough as it is. I mean going through several
> hundred failed builds every month isn't enough to do.

Ben C,

I don't find the sarcasm very convincing.  

I did not suggest special-casing a million things.

I do think it is legitimate to consider the relative merits of the
various options when making a decision.

In Ben P's original proposal, installing "autoconf" gets you the new
package, and installing both autoconf and autoconf2.13 mostly gets you
the old autoconf.  Some packages that use autoconf make assumptions
that break with version 2.50.  That is inconvenient for the
autobuilder operators.

The autobuilders could have a "special" rule that makes autoconf
depend on autoconf2.13.  That is (arguably) inconvenient for the
autobuilder authors.

Finally, if the autoconf (2.50) package is made to *depend* on
autoconf2.13, one has to go out of one's way to install only the new
autoconf.  That is inconvenient for anyone who wants to try out the
new autoconf.

Which of the three options causes the least total inconvenience?
Well, that depends on how you weight the relative inconvenience.
However, I'd wager that the group affected by option #2 is the
smallest of the three groups, while the group affected by option #3 is
by far the largest.

-S


-- 
by Rocket to the Moon,
by Airplane to the Rocket,
by Taxi to the Airport,
by Frontdoor to the Taxi,
by throwing back the blanket and laying down the legs ...
- They Might Be Giants



Reply to: