Malyshev's t2 fonts

[Hilmar Schlegel]

Vladimir Volovich wrote:
> It seems that Maksym Polyakov found a bug in definition of glyph
> `hyphenchar' (used in T1 encoding).
> 
> He wrote:
> 
>  MP> T1 and T2 fonts have hyphenchar which is hyphen with metrics
>  MP> width half of its actual width and intended to use for hanging
>  MP> hyphenation.  (BTW, when used, make text look much better!)
> 
> The definition of hyphenchar in latin.mtx (and latinsc.mtx) is:
> 
> \setglyph{hyphenchar}
>    \glyph{hyphen}{1000}
> \endsetglyph
> 
> Which should probably be changed to:
> 
> \setglyph{hyphenchar}
>    \glyph{hyphen}{1000}
>    \resetwidth{\div{\width{hyphen}}{2}}
> \endsetglyph

For some reason I had to use 
\setglyph{hyphenchar}
   \glyph{hyphen}{1000}
   \resetwidth{\div{\width{hyphen}}{2}}
   \resetdepth{\depth{hyphen}}
\endsetglyph

to get the metric correctly - possibly this problem has gone
meanwhile...
 
> What do you think? In EC fonts, the width of hyphenchar (slot #127) is
> exactly half the width of hyphen (slot #45); and the glyph looks the
> same. Hyphenation of texts written with EC fonts when one sets
> \hyphenchar\font=127 is "half-hanging", as compared to the (default)
> case when \hyphenchar\font=45. IMHO, fontinst's definition should
> follow this rule.

In general the optimal lenght the hyphen sticks out is a font-dependent
property. So latin.mtx provides here some useful default value which
should be adjusted for fancy fonts. Using exactly 1/2 of the width is a
rather formalistic choice, though.
 
> BTW, is it okay that latin.mtx (and latinsc.mtx) contains the line
> 
> \setleftkerning{hyphenchar}{hyphen}{1000}
> 
> BEFORE the definition of the `hyphenchar' glyph?

no good idea...

Hilmar Schlegel

###