[tex-live] TeX Live 2016 binary distribution milestone

[Nelson H. F. Beebe]

Earlier today, I installed TeX Live 2016 binary distribution number

			     _  ___   ___  
			    / |/ _ \ / _ \ 
			    | | | | | | | |
			    | | |_| | |_| |
			    |_|\___/ \___/ 
				   
(100) at the usual site

	http://www.math.utah.edu/pub/texlive-utah/

Unless some really interesting Un*x(-like) operating system variants
show up that I don't already have running here in our software test lab,
I probably won't be adding to that large archive, although I expect to
cover new O/S releases that appear for some of the more common O/S
distributions.

In particular, over the last month, I've added binaries for the
bleeding-edge releases of Debian, Fedora, Mint, and Ubuntu, as well as
for several O/Ses in the BSD family.

The final section of the above Web site has a (currently) 24-item list
of comments about various systems.

It gives a pretty reasonable list of reasons for build failures that
we really should address for next year, to make the build process
easier for us, as well as for the various O/S distribution maintainers
who package TeX Live for their software channels.

Notable among the causes of partial failures on many platforms is the
bogus error message error: "Sorry, you need zlib with compress2" from
dvipdfm-x, and the defective packaging of Asymptote that uses CFLAGS
for C++ compilations instead of CXXFLAGS.

The list of missing executables about 1/3 way down the above Web site
identifies these programs as frequently failing to build:

	 inimflua inimfluajit lualollipop xelollipop
	 tex2xindy texindy xindy xindy.mem xindy.run

The first row has to do with the limited architecture support for LuaJIT
because of its use of assembly-language coding: why isn't there
fall-back portable C or C++ code that works everywhere?

The second row is entirely due to the sad lack of a maintainer for
Clisp, which fails to build on numerous platforms, or builds but fails
to run, and has gone without updates for at least six years.

Many of those systems without Clisp have alternate Common Lisp
implementations, such as gcl and sbcl, and because Common Lisp is
pretty-well defined in Guy Steele's books, it ought to be feasible to
host xindy on different implementations of Common Lisp.  The Reduce and
Maxima symbolic algebra systems can both be built on multiple such
implementations, and xindy is far less complex than they are.

-------------------------------------------------------------------------------
- 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 at math.utah.edu  -
- 155 S 1400 E RM 233                       beebe at acm.org  beebe at computer.org -
- Salt Lake City, UT 84112-0090, USA    URL: http://www.math.utah.edu/~beebe/ -
-------------------------------------------------------------------------------