Christopher Adams skrev:
Hi Lars,
Thanks for your help as always.
2010/5/14 Lars Hellström <[email protected]>
Christopher Adams skrev:
What I don't see support for in fontinst is MM instances generated along
the
width axis for use in font expansion.
If you're using \installfont level commands, you simply need to write more
of them (because you want to generate more fonts).
What should the \installfont commands look like? Something like this?
\installfont{pmnr8t12}{pmnr8r12,newlatin}{t1}{T1}{pmn}{m}{n}{}
\installfont{pmnr8t12-20}{pmnr8r12-20,newlatin}{t1}{T1}{pmn}{m-20}{n-20}{}
\installfont{pmnr8t12+20}{pmnr8r12+20,newlatin}{t1}{T1}{pmn}{m+20}{n+20}{}
I'd say n rather than n-20 and n+20, since there is no shape variation
involved; the -20 or +20 in the series suffice for distinguishing the
font (thus preventing confusion caused by declaring the same font
twice). Hence:
\installfont{pmnr8t12}{pmnr8r12,newlatin}{t1}{T1}{pmn}{m}{n}{}
\installfont{pmnr8t12-20}{pmnr8r12-20,newlatin}{t1}{T1}{pmn}{m-20}{n}{}
\installfont{pmnr8t12+20}{pmnr8r12+20,newlatin}{t1}{T1}{pmn}{m+20}{n}{}
An alternative would be
\installfont{pmnr8t12}{pmnr8r12,newlatin}{t1}{T1}{pmn}{m}{n}{}
\offcommand\installfontas
\installfont{pmnr8t12-20}{pmnr8r12-20,newlatin}{t1}{}{}{}{}{}
\installfont{pmnr8t12+20}{pmnr8r12+20,newlatin}{t1}{}{}{}{}{}
\oncommand\installfontas
where the last two \installfont commands generate the fonts but make no
FD entries for them. (This however relies on the fact that \installfont
calls \installfontas, which has not always been the case.)
The only thing I can think of that might be potentially problematic is that
the above generates NFSS declarations for all the font widths: there's an m
series, a b series, an m+5 series, a b+5 series, an m+10 series, and so
on...
I also took note of the fact that the Thanh states:
Similar to Type 1 fonts, only the entry for the base font is needed in map
files:
pmnr8y12 MinionMM_400_535_12_ <texnansi.enc <pmnr8a12.pfb
Following this instruction, I removed the other two widths from the map file
before installing it.
I saw that too. Since he still _is_ generating PFBs for every width, I
suspect "base" here means "not virtual" rather than "default MM
instance". In that case, the ordinary fontinst-generated mapfile should
be just right.
Of course, I can get Minion installed, but microtype doesn't seem to
discover the expanded instances. If I use the auto=false option, which is
supposed to force the package to use the really-existing expansion steps, as
in:
\usepackage[auto=false,shrink=20,stretch=20,step=20]{microtype}
pdfLaTex tries and fails to generate some different font on the fly using
MetaFont. I've pasted the error below.
Hmm... It seems to complain already about the TFM, and it's a TFM for
an EC font. Maybe you need to tell microtype to not apply for those
fonts (or make sure they're not used anywhere in the document)?
Lars Hellström
