I just discovered this post from jwz. (Jamie, I don't actually read
the autoconf mailing list regularly - please cc me)
>> I'm not proposing that the config.cache feature be removed, only that
>> it not be made the _default_.
>
> Sounds reasonable to me. How about only enabling it when given the
> option --cache-file=<filename>?
jwz> For what it's worth, I think that turning the cache file off by default
jwz> is the right thing. Let people who know what they're doing turn it on
jwz> explicitly.
jwz> On another topic I would guess (but Martin would know better) that a big
jwz> part of why xemacs uses its own autoconf is historical: xemacs has been
jwz> using autoconf for a very long time, and those early versions of
jwz> autoconf really weren't up to the challenge of handling something as
jwz> convoluted as emacs... So at this point there's a lot of historical
jwz> baggage in left over from the bad old days.
Actually xemacs currently uses autoconf 2.13 - I did many major changes
around 1995 to no longer require autoconf 1. However, there is still
lots of junk in there from the early days. Many autoconf macros are
redefined. And removing the config cache unconditionally was my idea,
prompted by my frustration with obscure bug reports and getting burned
by the config caching even when no cache file existed when configure
starts.
jwz> Does RMSmacs use a modern autoconf, or does it have its own too?
I think the configure script in the version of Emacs from the FSF has
evolved less. But it also uses autoconf 2.13.
Martin
