[tex-live] "Bug": problems with dmacron in pdfTeX

Mojca Miklavec mojca.miklavec.lists at gmail.com
Wed Jan 18 14:10:31 CET 2012


Hello,

I have some weird problems with ttf2afm and pdftex.

My main problem is that if I want to support both Type1 and TTF fonts
at the same time, I can have a single encoding file and two map files,
for example:

gentium-truetype.map:
qx-gentiumbasic-bold GentiumBasic-Bold <GenBasB.ttf gentium-qx.enc

gentium-type1.map:
qx-gentiumbasic-bold GentiumBasic-Bold "gentium-qxEncoding
ReEncodeFont" <gentium-qx.enc <GenBasB.pfb


This all works nice and perfect, with two exceptions. pdfTeX refuses
to recognize two glyphs from .enc file:
    /dcroat
    /Idotaccent
It would recognize both /dmacron and /uni0111 for "dcroat" and both
/Idot and /uni0130 for "Idotaccent" (dcroat and Idotaccent are both
standard Adobe names), but not the other two.

This is highly annoying for multiple reasons:
- I need two different encoding files, one for Type1 and one for TTF,
but only for T1 and T5, all others work perfectly fine with a single
file.
- Even more annoying: with any given tool that generates tfm files, I
need to have /dcroat and /Idotaccent in enc file, else the glyph won't
be found the the slot in tfm would remain empty. However, when I want
to use the font, I need to have /dmacron and /Idot in enc file. This
means that I need to have two different enc files for the sake of
using TTF fonts alone. The same enc file that works for generating tfm
doesn't work when using that tfm.

I checked all *.enc files in TeX Live. The only place where /dmacron
is being used is "my own" package (gentium). The sole reason is that
nothing else worked. The only places where /uni0111 is being used are:
- dejavu fonts (by Pavel Farář, for exactly the same reason as I
described he has to use different enc files just for T1)
- vntex (t5d.enc, t5uni.enc)
- cm-unicode
- TeX Gyre-related (not included in any map file):
% Public domain - version of 22dec08.
% Created by Han The Thanh from the corresponding TeX Gyre encoding
% file, by translating the human-readable names into their Unicode
% equivalents, primarily for use with TrueType fonts.

To make it even more fun, here is what I get with pdfTeX's ttf2afm:

GentiumPlus-R:

 C -1 ; WX 529 ; N dcroat.BarBowl ; B -4 -14 522 756 ;
 C -1 ; WX 520 ; N dmacron ; B 39 -15 519 756 ;
 C -1 ; WX 592 ; N Eth ; B 13 0 548 629 ;
 C -1 ; WX 592 ; N Dcroat ; B 13 0 548 629 ;
 C -1 ; WX 557 ; N dcroat.BarBowl.SC ; B 10 0 511 520 ;
 C -1 ; WX 557 ; N dcroat.SC ; B 10 0 511 520 ;

Gentium-R:

C -1 ; WX 592 ; N Dcroat ; B 13 0 548 629 ;
C -1 ; WX 520 ; N dmacron ; B 39 -15 519 756 ;

Complete inconsistency (apart from the fact that dcroat and dcroat
only is very consistently replaced by dmacron, but no other name like
Dcroat or dcroat.SC is).

Is there any chance to convince pdfTeX to accept /dcroat as a glyph
name? (Of course I could put /uni0111 into .enc file, but then it is
useless for Type1.)

Mojca



More information about the tex-live mailing list