Message-Id: <199401281011.AA05568@mail.cs.tu-berlin.de>
Reply-To: Mailing list for the LaTeX3 project <LATEX-L@DHDURZ1.BITNET>
Date:         Fri, 28 Jan 1994 10:06:59 LCL
From: Mike Piff <M.Piff@sheffield.ac.uk>
Sender: Mailing list for the LaTeX3 project <LATEX-L@DHDURZ1.BITNET>
To: Multiple recipients of list LATEX-L <LATEX-L@DHDURZ1.BITNET>
Subject:      patterns
Status: R

Thought I would forward this from NTS-L, as it seems relevant to the
discussions here about languages, and whether the language could be decided
in the document rather than in the format, at great expense in time.

Mike Piff

------- Forwarded Message Follows -------

Date:          Fri, 28 Jan 1994 00:07:30 -0500
Reply-to:      NTS-L Distribution list <NTS-L@DHDURZ1.EARN>
From:          Laurent Siebenmann <laurent@MATH.TORONTO.EDU>
Subject:       patterns
X-To:          NTS-L@DHDURZ1.bitnet
To:            Multiple recipients of list NTS-L <NTS-L@DHDURZ1.EARN>


patterns

     Peter Breitenlohner stated (Thu Jan 27)

 > Initex can load patterns (cumulative with TeX3)
 > only if
 >    1. no .fmt is loaded
 >    2. no paragraph has been typeset

     I want to thank Peter for thus confirming an
assertion (part 1.) that I recently made in TUGboat (Oct93).
Unfortunately it is FALSE (!!)  I had better
grab this opportunity to undo my mischief while you are all
listening.

     Peter's qualitative explanation is roughly correct:

 > When the first paragraph is typeset or a \dump occurs
 > the patterns are brought to their final compacted
 > form.  And when initex reads a .fmt file it gets the
 > compacted patterns from that file. Once this has
 > happened even initex can not load additional
 > patterns, instead one gets a "Too late" error message.

What gives? Look closely and you see a loophole big enough for
a pregnant donkey.  When a format is compiled with no
patterns, then no patterns are compacted and no obstruction is
created to the assimilation of further patterns!

      This loophole has recently been of great service to me
when playing with patterns, and I am sure some of you will want
to use it until e-TeX does better. I compile my favorite format
with no patterns at all and then dump it, say to my.fmt.  Then

initex &my.fmt  mypatterns.tex \dump <return>

produces a format "mypatterns.fmt" in a big hurry and I use it
under  virtex  to test patterns introduced in mypatterns.tex.

     Whether this loophole is of much use to non-gourous
inder TeX 3 is moot.

      Correct me if the following is still inaccurate :

 (*) Initex (version 3.xx) can load patterns
  if and only if no patterns have been compacted on that
  run, and no compacted patterns exist in the format
  (if any) that initex loads.

     Bernd Raichle commented (TeX-Euro list 1993) that
TeX only compacts pending patterns loaded by
\patterns when it REALLY needs the hyphenation trie.
Thus (sorry Peter), NEITHER 1. NOR 2. is a necessary
condition for loading patterns.

      Laurent S

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% Dr M J Piff, School of Mathematics and Statistics, University of %%
%% Sheffield, UK.                    e-mail: M.Piff@sheffield.ac.uk %%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%