fonttosfnt: Changes to 'master'

[GitLab Mirror]

 fonttosfnt.c |    6 +
 fonttosfnt.h |   37 ++++++---
 read.c       |   50 ++++++------
 struct.c     |   93 +++++++---------------
 util.c       |   20 ++++
 write.c      |  239 ++++++++++++++++++++++++++++++++++++++++++++---------------
 6 files changed, 289 insertions(+), 156 deletions(-)

New commits:
commit ca9ad454496bebbc37125959d049cf2173c07c2a
Author: Christopher Zimmermann <madroach at gmerlin.de>
Date:   Sun Aug 16 21:56:15 2020 +0200

    use PIXEL_SIZE to calculate lineGap

commit 0447b81053d4ccaa1e1b0087bf6076abae431cae
Author: Christopher Zimmermann <madroach at gmerlin.de>
Date:   Fri Jul 31 00:07:26 2020 +0200

    fix sign of sTypoDescender

commit 747c58aa17b9adc016aed0caab049417a20f0b22
Author: Christopher Zimmermann <madroach at gmerlin.de>
Date:   Thu Jul 30 22:46:50 2020 +0200

    calculate capHeight and xHeight as recommended

commit 64e7f386baf69fa6857047fc8ad9b4cd9a8185f1
Author: Christopher Zimmermann <madroach at gmerlin.de>
Date:   Thu Jul 30 21:34:59 2020 +0200

    move metrics calculation to write.c

commit 19412d9ac3fb7f00346273cfe914601c844f5872
Author: Christopher Zimmermann <madroach at gmerlin.de>
Date:   Sat Jul 18 08:39:01 2020 +0200

    Add warnings when wrapping multiple fonts and on PCF fonts

commit 1fa97fdc6dc900728be8523be89096b4cb117d4d
Author: Christopher Zimmermann <madroach at gmerlin.de>
Date:   Sun Jul 12 15:36:18 2020 +0200

    provide version 5 OS/2 table
    
    this was an attempt to fix linespacing, which did not help, but
    would still keep it around to avoid double efforts.

commit facf71184163d50adf0b8fccffcaffa18b7bd277
Author: Christopher Zimmermann <madroach at gmerlin.de>
Date:   Sun Jul 12 15:12:59 2020 +0200

    use design metrics in eblc instead of calculated metrics
    
    This will probably not work for a font providing multiple strikes.
    But .bdf fonts don't provide multiple strikes (or do they?!?)
    I don't know about .pcf fonts, but I would recommend against running fonttosfnt
    on .pcf fonts because bdf2pcf seems to loose some attributes during conversion.

commit 9fb05de7d6c57d045b4a88516f3c28cf3cf98722
Author: Christopher Zimmermann <madroach at gmerlin.de>
Date:   Tue Jul 7 21:02:56 2020 +0200

    use more properties according to XLFD

commit 58cbf737557f34744b950668e3354ec2f2dae766
Author: Christopher Zimmermann <madroach at gmerlin.de>
Date:   Mon Jul 6 22:50:23 2020 +0200

    use standard C rounding functions

commit f8fa7919397659fb0ac5dade75152343d57f20ce
Author: Christopher Zimmermann <madroach at gmerlin.de>
Date:   Tue Jun 16 18:43:12 2020 +0200

    correctly set OS2 fsSelection for italic and bold fonts

commit 1df352c9e862686692ac053b838d76a36d721805
Author: Christopher Zimmermann <madroach at gmerlin.de>
Date:   Mon Jun 15 14:08:49 2020 +0200

    truncate foundry if there is no known abbreviation.
    
    Honestly I have no idea why the foundry needs to fit in four bytes.
    But anyway truncation might indeed be better then "UNKN" ?
    
    Anyone more knowledgeable than me out there?

commit 8532812f1eddfadc0bcfad220b283bf5990bba10
Author: Christopher Zimmermann <madroach at gmerlin.de>
Date:   Sun May 31 08:02:24 2020 +0200

    use zero lineGap, sTypoDescender should be negative
    
    According to
    https://simoncozens.github.io/fonts-and-layout/opentype.html#vertical-metrics-hhea-and-os2
    sTypoDescender should be negative and lineGap should be zero.

commit de8068f9c4251fb6cb9407c07bf245e1937270bc
Author: Christopher Zimmermann <madroach at gmerlin.de>
Date:   Sun May 31 07:58:10 2020 +0200

    don't add arbitrary number to the glypth metrics
    
    This caused pango to calculate wrong, too large font extents and in
    consequence gvim used too large character cells.