## Font not loadable - fontools problem

tex

### Next

• 1. Citation font
Hi there, Does anyone know a simple way to change the font of a citation? I am using the natbib plainnat style because I require harvard-like citations, but I am willing to change if there is a better option. Thanks Best Regards Edd
• 2. BCOR length
Hi NG, does anybody know the name of the BCOR length of the KOMA-Script and if I can call this length? Thanks, Stefan Pinnow
• 3. confused about interaction between verbatim environment and new command

### Font not loadable - fontools problem

I'm trying to install Myriad Pro Opentype fonts with the help of
fontools scripts:

Everything seems to be ok until I launch Yap previewer and see that no
fonts are load, the error message says:

-----------------

Making PK font:
720 600 magstep(1.0) ljfour
Trying to make PK font "LY1--MyriadPro-Regular--base" (at 720 DPI)...
"ttf2pk" -q -t "LY1--MyriadPro-Regular--base"
makepk: "LY1--MyriadPro-Regular--base.pk" could not be created.

-----------------

What is missing? I've put all pfb, fd, vf, enc, tfm, map and sty files
in the correct place (I think), added the MyriadPro.map to updmap.cfg
and issued mktexlsr before updmap.

Thanks,

Bernhard Enders.



### Re: Font not loadable - fontools problem

Problem found: the MyriadPro.map file was missing from texmf tree. Now
it works. Sorry for the inconvenience.

Bernhard Enders.



Hello,

I started to use CJK in LaTeX recently, and I spend a lot of time
trying to fix the following problem:

! Font C70/song/m/n/9/65=cyberb65 at 9.0pt not loadable:

relax
l.74 ^^e6^^97^^a0
\par
?

I searched Internet to find a fix. I'm not the only one who got such
error, but the only recommendation was to correctly install CJK fonts.
Unfortunately, it wasn't my case. My CJK fonts were installed and
worked.

Finally, I've traced the source of the problem and fixed it. I share
the solution in hope that someone somewhen found it useful.

I use background texts which are produced in \EveryShipout. The text
should be displayed in Helvetica font, so the shipout code contains:

... \fontfamily{phv}\selectfont ...

Now let's trace TeX execution. It processes CJK text, and the current
encoding is some "CXX". A page is full, so shipout procedure is
activated. "Selectfont" tries to load "phv" for "CXX" encoding (font
shape "CXX/phv/m/n"). It's undefined, so fallback is "CXX/song/m/n",
which somehow is translated to "cyberbXX" (my CJK font), and for some
reason it is not found (altough exists).

The fix is simple. The shipout code should also define the font
encoding:

... \fontfamily{phv}\fontencoding{T1}\selectfont ...

All this seems obvious now, but it wasn't easy to realize. TeX error
message is very misleading: error context is the source code, not the
shipout code.



Thanks for being so responsive, Derek! I am mailing you the file in
question.

The Debian freetype package I use claims to be 2.0.9-1 (I have also
installed freetype1-tools version 1.4pre20011029-1, but I doubt this
matters). xpdf is 1.00-3.1.

Meanwhile, I have run a couple of tests using different options for
the PostScript output from Word. The (I suppose quite popular) way I
use to convert Word to PostScript is to "print" it to a PostScript
printer but redirect this to a file. I have discovered there are a few
switches in the printer settings (I am translating these since my
stuff is in German; the wording might not be exactly what is said in
English) that do influence the result in a more or less pronounced
way:

- you can choose between "download TrueType fonts to the printer as
softfonts" and "replace TrueType by printer fonts"

- "TrueType download options" offers you "automatically", "bitmap",
"outline" (or "shape"?), and "TrueType"

- a third option the name of which I have forgotten offers you the
choice between "optimise speed", "optimise portability",
"encapsulated postscript", and "archive (the) format".

I have run tests combining almost all these options (I have NOT tried,
however: 1) "replace TT by printer fonts", because this was not what I
was concerned about (I assume that would produce more portable
documents if you are using Helvetica and Times and the like), 2)
"download TT automatically", because unpredictable, and 3) "optimise
speed", which does not make sense when printing to a file), amounting
to a total of 9 combinations for the PostScript output from Word.

That postscript output was then, still under Windows, converted to PDF
1.3 by FreePDF, which really uses AFPL Ghostscript 8.14. Because I had
become curious, I also ran this step under Linux, using the ps2pdf13
script which invokes GNU Ghostscript 6.53.

I have tried viewing both the PS files and the PDF files. The former
with gsview under Windows, and gv under Linux, both of which use the
respective Ghostscript versions. I tried viewing the resultant PDF
files with Acrobat Reader 5.1 and Adobe Reader 6, as well as gsview
under Windows. Under Linux, I tried xpdf and gv.

Of course the results fill a table with 9 rows and 12 columns, which
I'll mail to anyone who's interested, but I'll extract here what
pertains to the question I have posted here:

- the problem with the boxes instead of glyphs arises only if the TT
fonts are downloaded as "TrueType" (the default is "automatic",
supposedly meaning the same).

Interestingly, both xpdf and gv have this problem if the PDF was
produced under Windows (though the boxes look somewhat different in
both). Neither gsview nor Acrobat/Adobe reader have a problem then.

Very interestingly, if the PS->PDF conversion step was performed
under Linux instead, xpdf displays PART of the file (roughly the
second part), while gv displays the whole. The Windows viewers have
again no problem, but there is a difference: with Acrobat Reader 5,
the result looks very pixelish, if not to say unpleasant to read.

- when the TT fonts are downloaded to the "printer" as "outline" (or
"shape"?), the results are fine (and very similar) on all viewers I
tried out (and regardless of whether PS-->PDF was done under Win or
Linux). This was apparently the only PDF deserving the attribute
"portable" one could produce with the tools I had.

- the "illegal bfrange" message seems to lead on the wrong path: it is
issued if and only if the PDF was produced by the win combo
(FreePDF/AFPL gs 8.14), but not if it was produced by ps2pdf13 under
Linux (GNU gs 6.53). This is so regardless of other factors, and
does not seem to have an impact on the display of the file.

Although I now know (by trial and error) HOW to produce one kind of
"P"DF, I have no idea why this is the case. If in both cases, TrueType
fonts are embedded (but are they?), there should be no problem with
either. But what is the difference between PostScript with embedded
TrueType "as TrueType", and with TrueType as "outline" (or "shape")?
And in what way do PDF documents normally contain/embed/use
TrueType? Do they normally need the external .ttf files?

The reason for my concern is that I expect from a "portable" document
format to be portable. Hence I assume there must be a standard way of
implementing things, not several possibilities. At least if using
TrueType in PDF _is_ to be considered standard.

--

Florian v. Savigny

If you are going to reply in private, please be patient, as I only
check for mail something like once a week. - Si vous allez rondre
personellement, patientez s.v.p., car je ne lis les courriels
qu'environ une fois par semaine.


Hi!

When installing a new LaTeX font from Postscript type1 I put the following
line in my fontinst driver file:

\substitutesilent{bx}{b}

But I?m still getting the LaTeX font warning:

Font shape T1/pbbx/bx/n' in size <12> not available
Font shape T1/pbbx/b/n' tried instead on input line 11.

So, what is \substitutesilent for? If we have to put
\renewcommand*{\bfdefault}{b} in every font package we create, why use
\substitutesilent?

Another question: Is a verbose fontinst file really necessary to install
semibold weight available in a font set. Can \latinfamily macro do this
task, how? I've got the following warning when using \latinfamily:

LaTeX Font Warning: Font shape T1/pbbx/sb/n' undefined

TIA,

Bernhard.
`