bug#20218: emacs-24.5-rc1 build successes, and one failure

["Nelson H. F. Beebe" <beebe@math.utah.edu>]

I've now completed 111 build attempts for emacs-24.5-rc1 on 38 flavors
of Unix systems, and am pleased to report success on 37 of them:

	DragonFlyBSD 3.6, 3.8, 3.9, 4.0		x86-64
	GNU/Linux ArchLinux			x86-64 and ARM
	GNU/Linux Debian 6.0, 7.7, 8.0		x86-64
	GNU/Linux Debian 7			SPARC
	GNU/Linux Fedora 20			x86
	GNU/Linux Fedora 21 and 22		x86-64
	GNU/Linux OpenSUSE 11.4, 12.3, 13.2	x86-64
	GNU/Linux Red Hat 5			IA-64 and x86
	GNU/Linux Red Hat 5, 6, 7		x86-64
	GNU/Linux Scientific Linux 6.5		x86-64
	GNU/Linux Slackware 14			x86-64
	GNU/Linux Ubuntu 14			x86
	kFreeBSD 7.7				x86-64
	Mac OS X 10.5.8				PowerPC-32
	Mac OS X 10.7.5				x86-64
	Mac OS X 10.10.2			x86-64
	MirBSD 10				x86
	NetBSD 5.0, 6.1				x86-64
	OpenBSD 4.9, 5.1, 5.4, 5.5, 5.6		x86-64
	OpenIndiana 11				x86-64
	SGI IRIX				MIPS R10000
	Solaris 10				SPARC, x86, x86-64
	Solaris 11				x86, x86-64

On about half, the build succeeded out of the box with my standard
automated build procedures.  On the remainder, I had to manually add
various --with-XXX=no options to deal with missing, or out-of-date,
graphics libraries. 

I therefore recommend that the configure scripts should be adjusted so
that failure to use any graphics library should just be prominently
recorded in the configure output, and then --with-XXX=no silently
assumed for that library.  Doing so would significantly increase the
number of builds that succeed on the first try.

In three or four cases, I had to temporarily rename the /usr/local
tree during the build to hide it from gcc, which obnoxiously insists
on searching /usr/local/include before /usr/include, even when you
don't want it to.  For that reason, and the fact that some of the BSD
distributions incorrectly use /usr/local for their package management
installations, on all new systems that we install, we replace
/usr/local by /usr/uumath as the tree for packages locally built and
installed from source code.

The one failure for my emacs-24.5-rc1 builds is this system:

	GNU Hurd 8 (jessie)			x86-64

I have previously successfully built emacs-24.3.90, emacs-24.3.92, and
emacs-24.3.94 on that system, and after several failures for
emacs-24.5-rc1, I retried with the identical PATH and other
environment variables and configure flags as for 24.3.94.  Some of my
several attempts succeeded in building src/temacs, but all died like
this:

	Loading version...
	Loading widget...
	Loading custom...
	Required feature `widget' was not provided
	Makefile:815: recipe for target 'bootstrap-emacs' failed

I'm puzzled at what to do about this.  We keep all of our systems at
current patch levels, so there have certainly been many system updates
on Hurd between 24.3.94 (8-Oct-2014) and 24.5-rc1.  I plan to go back
and try to build 24.3.94 again with the same build environment as
before.

In any event, I consider emacs-24.5-rc1 as yet another in long line of
outstanding successes in the quality and portability of GNU emacs
releases.  Congratulations to the development team --- you've done a
great job in keeping this fantastic editor potentially available
virtually everywhere!

-------------------------------------------------------------------------------
- Nelson H. F. Beebe                    Tel: +1 801 581 5254                  -
- University of Utah                    FAX: +1 801 581 4148                  -
- Department of Mathematics, 110 LCB    Internet e-mail: beebe@math.utah.edu  -
- 155 S 1400 E RM 233                       beebe@acm.org  beebe@computer.org -
- Salt Lake City, UT 84112-0090, USA    URL: http://www.math.utah.edu/~beebe/ -
-------------------------------------------------------------------------------