[Fontinst] problem with 8y.etx

Lars Hellström Lars.Hellstrom at math.umu.se
Wed Dec 10 18:25:16 CET 2003


At 16.49 +0100 2003-12-10, Paul Pichaureau wrote:
>  I'm trying to use the 8y.etx file (texnansi encoding vector).
>
>  I am very surprised  to find the command
>
>     \ifdirect
>
>in many place. This command seems to be undocumented. I found no
>information in the source

Then what of fimain.dtx lines 760--794? Behold:

% \begin{switch}{direct}
%   The encoding aspect of a font can to a large extent be understood
%   as a mapping from the set of slots to the set of glyphs---such
%   mappings are for example realised by PostScript encoding vectors
%   (mapping slots to glyph names) and virtual fonts (mapping slots to
%   DVI commands for drawing a glyph). In some cases however, encodings
%   are rather used to specify an inverse mapping going the opposite
%   way. Since encodings are in general many-to-one, this means there
%   need not be a proper inverse to a given encoding; instead the
%   inverse image of a glyph is in practice chosen randomly\footnote{In
%   the sense: according to a very deterministic (but not particularly
%   useful) rule which depends completely on details in the
%   implementation, and therefore should not be relied upon anyway.}
%   among the slots that are mapped to it. This is not always desirable.
%
%   The \texttt{direct} switch can be used to check whether an encoding
%   is interpreted in the direct (slots to glyphs) or inverse (glyphs to
%   slots) manner. If those |\setslot| commands which are undesirable
%   under the inverse interpretation are put inside an |\ifdirect|
%   \dots\ |\Fi| construction, then these will be ignored in that
%   circumstance, thus allowing the programmer to exactly specify both
%   the direct and the inverse mappings.
%   \changes{1.924}{2003/02/08}{Command added. (LH)}
%
%   In \package{fontinst}, this command is a switch, which is usually
%   true.
%    \begin{macrocode}
%<pkg|misc>\newif\ifdirect \directtrue
%    \end{macrocode}
%   In \package{fontdoc}, this is a conditional that will be typeset
%   if |\showbranches| is being used.
%    \begin{macrocode}
%<doc>\newcommand\ifdirect{\generic at if{direct}}
%    \end{macrocode}
% \end{switch}

The fisource.dvi index entry for \ifdirect should point you to the above
passage.

>neither in the 'regular' documentation.

Yes, that is out of date (but then it only claims to describe v1.8, does it
not?).

>Can someone explain the purpose and the usage of this command ?
>
>My exact problem is when I try to reencode a font:
>
> \transformfont{putr8y}{\reencodefont{8y}{\fromafm{putr8a}}}
>
>I obtain an uncorrect putr8y.mtx file: many char are in a wrong slot.
>
>It seems that each time a \ifdirect is read in the *.etx file a slot
>is automatically added.

This may well be due to the 8y.etx problem Walter reported earlier this
week. Near some of the \ifdirect there are \setslot commands which should
have been \nextslot commands. They're easily recognizable by having numeric
arguments.

Lars Hellström




More information about the fontinst mailing list