Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ncar!gatech!udel!princeton!phoenix!jay
From: jay@phoenix.Princeton.EDU (Jay Plett)
Newsgroups: comp.text
Subject: Re: Questions about (di)troff
Summary: how ditroff searches special fonts
Message-ID: <6455@phoenix.Princeton.EDU>
Date: 16 Feb 89 21:19:07 GMT
References: <5800004@ugun21> <673@pcrat.UUCP>
Organization: Princeton University, NJ
Lines: 31

In article <673@pcrat.UUCP>, rick@pcrat.UUCP (Rick Richardson) writes:
> Actually, ditroff searches the current font and then all of the
> fonts which are marked special (in order) for the special characters.

Close.  Ditroff searches the current font, then all special fonts,
then all regular fonts (starting with font 1), until it finds the
character or exhausts the list of fonts.  Example: say you have
mounted fonts A B C D S T U, where S is the first special font.
If the current font is C, then the search order is C, S, T, U, A,
B, C, D.  I've seen versions of ditroff which are clever enough
to recognize that they needn't search C twice, but they still
didn't quite get it right: they search C, S, T, U, A, B, C (or is
it C, S, T, U, B, C, D? I forget).

By the way, there's another gotcha in there.  It's not "all of the
fonts which are marked special".  All fonts after the first special
font are considered special.  When ditroff reads DESC.out while
starting up, it considers the first font which is marked as special to
be the first of a list of special fonts.  It doesn't care whether
succeeding fonts are marked special or not.  Example: your DESC file
lists A B C D S R I B where "S" is special but R, I and B aren't.
Surprise!  Ditroff considers R, I and B to be special.  What's worse,
the code for .fp doesn't care whether a font is marked as special.  But
it does recognize the name "S" as the name of a special font and
changes the beginning of the special fonts accordingly.  What's even
worse than that is what happens when you use .fp to mount *any* font on
the position which contains the first special font.  I'm running out of
steam here, so I'll leave that one as an exercise for the reader (with
the suggestion that you don't try it except as an experiment).

	...jay@princeton.edu