Re: Interesting combining character issue

[James Cloos <cloos@jhcloos.com>]

>>>>> "Kenichi" == Kenichi Handa <handa@m17n.org> writes:

Kenichi> The last vector is the information about the glyph of U+304
Kenichi> (COMBINING MACRON).  Yours says that the glyph width is -9
Kenichi> (5th element), which is a little bit strange.

Agreed.

Kenichi> Please show me these info:
Kenichi> % freetype-config --version

Kenichi> % libotf-config --version
Kenichi> % m17n-config --version

Kenichi> Mine are 9.18.3, 0.9.8, 1.5.3 respectively.

9.18.3, 0.9.8, 1.5.1

I'm upgrading m17n now.

Kenichi> And, please run the attached program as this:
Kenichi> % fttest .../DejaVuSansMono.ttf 647

Kenichi> I got this:

Kenichi> width=8, lbearing=2, rbearing=6, acsent=10, descent=-9

I got the same.

Ignoring the glyf instructions, uni0304 looks like:

,----< excerpt from DejaVuSansMono.sfd >
| StartChar: uni0304
| Encoding: 772 772 772
| Width: 1233
| Flags: W
| AnchorPoint: "above" 616 1120 mark 0
| LayerCount: 2
| Fore
| SplineSet
| 317 1526 m 1,0,-1
|  915 1526 l 1,1,-1
|  915 1378 l 1,2,-1
|  317 1378 l 1,3,-1
|  317 1526 l 1,0,-1
| EndSplineSet
| Position2: "'mark' Zero-Width Marks lookup" dx=0 dy=0 dh=-1233 dv=0
| EndChar
`----

So it is a spacing glyph with a positioning command to make it work like
a zero-width glyph.  uni0304 in DejaVu Serif, OTOH, is a conventional
zero-width glyph.  I believe it was done this way to ensure that DejaVu
Sans Mono gets listed as a monowidth font everywhere.

Perhaps that is where the -9 comes from?

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6