From: Christopher Croughton <crough45 AT amc DOT de>
Message-Id: <98Jan21.150223gmt.27786@internet01.amc.de>
Subject: Re: DJ port of GCC 2.8.0?
To: eliz AT is DOT elta DOT co DOT il (Eli Zaretskii)
Date: Wed, 21 Jan 1998 13:52:15 +0000
Cc: eldredge AT ap DOT net, crough45 AT amc DOT de, djgpp AT delorie DOT com
In-Reply-To: <Pine.SUN.3.91.980121153719.27671D-100000@is> from "Eli Zaretskii" at Jan 21, 98 01:37:52 pm
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Precedence: bulk

Eli Zaretskii wrote:
> 
> With a major new version such as 2.8.0, and with so many different
> platforms supported by GCC, it is no surprise that once an official
> release is made, quite a few of bugs are reported by people who try to
> build and use it.

One of the corollaries to Murphy's Law is:

  "The majority of the bugs will be found just after the release of 
  the software to a major client."

It's true, I've seen it lots of times.  In the case of the GNU software
the build process is so complex that it's not surprising that it has
problems in some configurations.  What I find really surprising is
that it manages to work without modification so often - that autoconf
system is wonderful...

> In fact, it is not uncommon to see much less complicated packages have
> a bugfix release within days of a major version (the last releases of
> `patch' and Make are cases in point).

  "First, apply the patches to 'patch' and 'make'.  To do this, run patch
  to apply the patches and then run make..."

Fortunately, it's not quite as bad as it sounds, unless you do something 
stupid like installing over your old versions, and the old versions did
work adequately.

Incidentally, why do you have to rebuild gcc with itself?  Is it just
to get the better optimisation, or is there some reason that building
it with the old version doesn't work properly?

Chris C