[Fontinst] Re: Bug in fontinstversion{1.927}

[Lars Hellström]

At 17.29 +0100 2005-01-26, Peter Dyballa wrote:
>Am 26.01.2005 um 12:16 schrieb Lars Hellström:
>
>>> Can Dominique Unruh's ucs.sty be useful?
>>
>> That could give you the map Unicode -> LICR, but that's not all there
>> is to
>> font encodings.
>
>What is LICR? Least Insignificant Carriage Return?

LaTeX Internal Character Representation---E, \'{a}, \textemdash, and such.

>>
>>> ttf2pt1 has a nice side-effect in augmenting the glyph names' base:
>>
>> Or maybe not so nice...   Ad hoc renaming is often a bad thing.
>
>It is _obviously_ bad, but doing so it might give an easy rule to not
>add this glyph name to the ever increasing set of 8p.etx and 8p.mtx --
>which brings me to a (not so nice?) question: \installfont can work
>with more than one MTX file, can \transformfont do the same?

The purpose of \transformfont is, in a sense, to create a new base font (it
will appear as such to the DVI driver). Combining two input MTXes does
therefore not make sense. Having two output MTXes could make sense, but the
same ends could always be achieved by doing two \transformfont command, so
it would mostly be feature bloat.

>It would
>be nice having a rather stable 8p.etx file that corresponds to 8p.enc,

With the most common glyphs I presume. Yes indeed.

>and another auxiliary ETX file with all those extra glyph names ...
>
>>> Glyph 551 has the same name as 321: (Omega), changing to _d_551
>>> Glyph 554 has the same name as 301: (Delta), changing to _d_554
>>
>> Here I wonder: Are these glyphs graphically distinct, or is it really
>> the
>> same glyph that for some reason appears twice? (There seems to be both
>> a
>> Greek and a Cyrillic Omega in Unicode.) In the latter case, is that
>> clear
>> in the TTF?
>
>It's obviously this: only one glyph (shape) and one character (slot)
>name where two would be appropriate -- a design flaw of this font!

No, it is not uncommon that the same glyph may be used for several
characters (first example that comes to mind is, if memory serves, that
quoteleft appears twice in the 8y/LY1 encoding). It seems to me the problem
is rather that ttf2pt1 insists on generating one glyph for each character.

>
>Since I fetched on Monday fontinst 2 I'll switch to this before doing
>any further experiments.

That's news to me... Even I'm still only at v1.929 (and that's not final yet).

>Are new releases advertised on this list?

Yes. Most recently on 2004-12-18 (v1.928).

Lars Hellström