From 49ef630936325b2e56a870fcef9aa8473a8f8526 Mon Sep 17 00:00:00 2001 From: Behdad Esfahbod Date: Wed, 4 Nov 2015 17:27:07 -0800 Subject: [PATCH] Adjust the width of various spaces if font does not cover them See discussion here: https://github.com/behdad/harfbuzz/commit/81ef4f407d9c7bd98cf62cef951dc538b13442eb There's no way to disable this fallback, but I don't think it would be needed. Let's hope for the best! Fixes https://github.com/behdad/harfbuzz/issues/153 --- src/hb-buffer-private.hh | 7 +++ src/hb-buffer.cc | 2 + src/hb-ot-shape-fallback-private.hh | 4 ++ src/hb-ot-shape-fallback.cc | 67 +++++++++++++++++++++++++++++ src/hb-ot-shape-normalize.cc | 1 + src/hb-ot-shape.cc | 3 ++ src/hb-unicode-private.hh | 56 +++++++++++++----------- test/shaping/tests/spaces.tests | 24 +++++------ 8 files changed, 126 insertions(+), 38 deletions(-) diff --git a/src/hb-buffer-private.hh b/src/hb-buffer-private.hh index 8e33fc880..c5a116d84 100644 --- a/src/hb-buffer-private.hh +++ b/src/hb-buffer-private.hh @@ -41,6 +41,12 @@ ASSERT_STATIC (sizeof (hb_glyph_info_t) == sizeof (hb_glyph_position_t)); template <> class hb_mark_as_flags_t {}; template <> class hb_mark_as_flags_t {}; +enum hb_buffer_scratch_flags_t { + HB_BUFFER_SCRATCH_FLAG_DEFAULT = 0x00000000u, + HB_BUFFER_SCRATCH_FLAG_HAS_SPACE_FALLBACK = 0x00000001u, +}; +template <> class hb_mark_as_flags_t {}; + /* * hb_buffer_t @@ -55,6 +61,7 @@ struct hb_buffer_t { hb_buffer_flags_t flags; /* BOT / EOT / etc. */ hb_buffer_cluster_level_t cluster_level; hb_codepoint_t replacement; /* U+FFFD or something else. */ + hb_buffer_scratch_flags_t scratch_flags; /* Have space-flallback, etc. */ /* Buffer contents */ hb_buffer_content_type_t content_type; diff --git a/src/hb-buffer.cc b/src/hb-buffer.cc index 50710dd23..f6907693f 100644 --- a/src/hb-buffer.cc +++ b/src/hb-buffer.cc @@ -198,6 +198,7 @@ hb_buffer_t::clear (void) hb_segment_properties_t default_props = HB_SEGMENT_PROPERTIES_DEFAULT; props = default_props; + scratch_flags = HB_BUFFER_SCRATCH_FLAG_DEFAULT; content_type = HB_BUFFER_CONTENT_TYPE_INVALID; in_error = false; @@ -738,6 +739,7 @@ hb_buffer_get_empty (void) HB_BUFFER_FLAG_DEFAULT, HB_BUFFER_CLUSTER_LEVEL_DEFAULT, HB_BUFFER_REPLACEMENT_CODEPOINT_DEFAULT, + HB_BUFFER_SCRATCH_FLAG_DEFAULT, HB_BUFFER_CONTENT_TYPE_INVALID, HB_SEGMENT_PROPERTIES_DEFAULT, diff --git a/src/hb-ot-shape-fallback-private.hh b/src/hb-ot-shape-fallback-private.hh index ec653513f..e134224df 100644 --- a/src/hb-ot-shape-fallback-private.hh +++ b/src/hb-ot-shape-fallback-private.hh @@ -45,5 +45,9 @@ HB_INTERNAL void _hb_ot_shape_fallback_kern (const hb_ot_shape_plan_t *plan, hb_font_t *font, hb_buffer_t *buffer); +HB_INTERNAL void _hb_ot_shape_fallback_spaces (const hb_ot_shape_plan_t *plan, + hb_font_t *font, + hb_buffer_t *buffer); + #endif /* HB_OT_SHAPE_FALLBACK_PRIVATE_HH */ diff --git a/src/hb-ot-shape-fallback.cc b/src/hb-ot-shape-fallback.cc index 3cb3456f1..c9cf737f4 100644 --- a/src/hb-ot-shape-fallback.cc +++ b/src/hb-ot-shape-fallback.cc @@ -484,3 +484,70 @@ _hb_ot_shape_fallback_kern (const hb_ot_shape_plan_t *plan, idx = skippy_iter.idx; } } + + +/* Adjusts width of various spaces. */ +void +_hb_ot_shape_fallback_spaces (const hb_ot_shape_plan_t *plan, + hb_font_t *font, + hb_buffer_t *buffer) +{ + if (!HB_DIRECTION_IS_HORIZONTAL (buffer->props.direction)) + return; + + hb_glyph_info_t *info = buffer->info; + hb_glyph_position_t *pos = buffer->pos; + unsigned int count = buffer->len; + for (unsigned int i = 0; i < count; i++) + if (_hb_glyph_info_is_unicode_space (&info[i]) && !_hb_glyph_info_ligated (&info[i])) + { + hb_unicode_funcs_t::space_t space_type = _hb_glyph_info_get_unicode_space_fallback_type (&info[i]); + hb_codepoint_t glyph; + typedef hb_unicode_funcs_t t; + switch (space_type) + { + case t::NOT_SPACE: /* Shouldn't happen. */ + case t::SPACE: + break; + + case t::SPACE_EM: + case t::SPACE_EM_2: + case t::SPACE_EM_3: + case t::SPACE_EM_4: + case t::SPACE_EM_5: + case t::SPACE_EM_6: + case t::SPACE_EM_16: + pos[i].x_advance = (font->x_scale + ((int) space_type)/2) / (int) space_type; + break; + + case t::SPACE_4_EM_18: + pos[i].x_advance = font->x_scale * 4 / 18; + break; + + case t::SPACE_FIGURE: + for (char u = '0'; u <= '9'; u++) + if (font->get_glyph (u, 0, &glyph)) + { + pos[i].x_advance = font->get_glyph_h_advance (glyph); + break; + } + break; + + case t::SPACE_PUNCTUATION: + if (font->get_glyph ('.', 0, &glyph)) + pos[i].x_advance = font->get_glyph_h_advance (glyph); + else if (font->get_glyph (',', 0, &glyph)) + pos[i].x_advance = font->get_glyph_h_advance (glyph); + break; + + case t::SPACE_NARROW: + /* Half-space? + * Unicode doc http://www.unicode.org/charts/PDF/U2000.pdf says ~1/4 or 1/5 of EM. + * However, in my testing, many fonts have their regular space being about that + * size. To me, a percentage of the space width makes more sense. Half is as + * good as any. */ + pos[i].x_advance /= 2; + break; + } + } +} diff --git a/src/hb-ot-shape-normalize.cc b/src/hb-ot-shape-normalize.cc index c86c634fb..31376b3c4 100644 --- a/src/hb-ot-shape-normalize.cc +++ b/src/hb-ot-shape-normalize.cc @@ -180,6 +180,7 @@ decompose_current_character (const hb_ot_shape_normalize_context_t *c, bool shor { _hb_glyph_info_set_unicode_space_fallback_type (&buffer->cur(), space_type); next_char (buffer, space_glyph); + buffer->scratch_flags |= HB_BUFFER_SCRATCH_FLAG_HAS_SPACE_FALLBACK; } else next_char (buffer, glyph); /* glyph is initialized in earlier branches. */ diff --git a/src/hb-ot-shape.cc b/src/hb-ot-shape.cc index 205be0a50..0b1417283 100644 --- a/src/hb-ot-shape.cc +++ b/src/hb-ot-shape.cc @@ -659,6 +659,8 @@ hb_ot_position_default (hb_ot_shape_context_t *c) &pos[i].y_offset); } + if (c->buffer->scratch_flags & HB_BUFFER_SCRATCH_FLAG_HAS_SPACE_FALLBACK) + _hb_ot_shape_fallback_spaces (c->plan, c->font, c->buffer); } static inline bool @@ -778,6 +780,7 @@ static void hb_ot_shape_internal (hb_ot_shape_context_t *c) { c->buffer->deallocate_var_all (); + c->buffer->scratch_flags = HB_BUFFER_SCRATCH_FLAG_DEFAULT; /* Save the original direction, we use it later. */ c->target_direction = c->buffer->props.direction; diff --git a/src/hb-unicode-private.hh b/src/hb-unicode-private.hh index 43bbed675..129045343 100644 --- a/src/hb-unicode-private.hh +++ b/src/hb-unicode-private.hh @@ -199,21 +199,24 @@ HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS_SIMPLE } } + /* Space estimates based on: + * http://www.unicode.org/charts/PDF/U2000.pdf + * https://www.microsoft.com/typography/developers/fdsspec/spaces.aspx + */ enum space_t { NOT_SPACE = 0, - SPACE_NBSP, - SPACE_EN, - SPACE_EM, - SPACE_EM_3, - SPACE_EM_4, - SPACE_EM_6, + SPACE_EM = 1, + SPACE_EM_2 = 2, + SPACE_EM_3 = 3, + SPACE_EM_4 = 4, + SPACE_EM_5 = 5, + SPACE_EM_6 = 6, + SPACE_EM_16 = 16, + SPACE_4_EM_18, /* 4/18th of an EM! */ + SPACE, SPACE_FIGURE, SPACE_PUNCTUATION, - SPACE_THIN, - SPACE_HAIR, SPACE_NARROW, - SPACE_MEDIUM, - SPACE_IDEOGRAPHIC, }; static inline space_t space_fallback_type (hb_codepoint_t u) @@ -221,22 +224,23 @@ HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS_SIMPLE switch (u) { /* All GC=Zs chars that can use a fallback. */ - default: return NOT_SPACE; /* Shouldn't happen. */ - case 0x00A0u: return SPACE_NBSP; - case 0x2000u: return SPACE_EN; - case 0x2001u: return SPACE_EM; - case 0x2002u: return SPACE_EN; - case 0x2003u: return SPACE_EM; - case 0x2004u: return SPACE_EM_3; - case 0x2005u: return SPACE_EM_4; - case 0x2006u: return SPACE_EM_6; - case 0x2007u: return SPACE_FIGURE; - case 0x2008u: return SPACE_PUNCTUATION; - case 0x2009u: return SPACE_THIN; - case 0x200Au: return SPACE_HAIR; - case 0x202Fu: return SPACE_NARROW; - case 0x205Fu: return SPACE_MEDIUM; - case 0x3000u: return SPACE_IDEOGRAPHIC; + default: return NOT_SPACE; /* Shouldn't happen. */ + case 0x0020u: return SPACE; /* U+0020 SPACE */ + case 0x00A0u: return SPACE; /* U+00A0 NO-BREAK SPACE */ + case 0x2000u: return SPACE_EM_2; /* U+2000 EN QUAD */ + case 0x2001u: return SPACE_EM; /* U+2001 EM QUAD */ + case 0x2002u: return SPACE_EM_2; /* U+2002 EN SPACE */ + case 0x2003u: return SPACE_EM; /* U+2003 EM SPACE */ + case 0x2004u: return SPACE_EM_3; /* U+2004 THREE-PER-EM SPACE */ + case 0x2005u: return SPACE_EM_4; /* U+2005 FOUR-PER-EM SPACE */ + case 0x2006u: return SPACE_EM_6; /* U+2006 SIX-PER-EM SPACE */ + case 0x2007u: return SPACE_FIGURE; /* U+2007 FIGURE SPACE */ + case 0x2008u: return SPACE_PUNCTUATION; /* U+2008 PUNCTUATION SPACE */ + case 0x2009u: return SPACE_EM_5; /* U+2009 THIN SPACE */ + case 0x200Au: return SPACE_EM_16; /* U+200A HAIR SPACE */ + case 0x202Fu: return SPACE_NARROW; /* U+202F NARROW NO-BREAK SPACE */ + case 0x205Fu: return SPACE_4_EM_18; /* U+205F MEDIUM MATHEMATICAL SPACE */ + case 0x3000u: return SPACE_EM; /* U+3000 IDEOGRAPHIC SPACE */ } } diff --git a/test/shaping/tests/spaces.tests b/test/shaping/tests/spaces.tests index d9e5d09c5..cb386de1e 100644 --- a/test/shaping/tests/spaces.tests +++ b/test/shaping/tests/spaces.tests @@ -1,17 +1,17 @@ fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+0020:[gid1=0+560] fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+00A0:[gid1=0+560] fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+1680:[gid0=0+692] -fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2000:[gid1=0+560] -fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2001:[gid1=0+560] -fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2002:[gid1=0+560] -fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2003:[gid1=0+560] -fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2004:[gid1=0+560] -fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2005:[gid1=0+560] -fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2006:[gid1=0+560] +fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2000:[gid1=0+1024] +fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2001:[gid1=0+2048] +fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2002:[gid1=0+1024] +fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2003:[gid1=0+2048] +fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2004:[gid1=0+683] +fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2005:[gid1=0+512] +fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2006:[gid1=0+341] fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2007:[gid1=0+560] fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2008:[gid1=0+560] -fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2009:[gid1=0+560] -fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+200A:[gid1=0+560] -fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+202F:[gid1=0+560] -fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+205F:[gid1=0+560] -fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+3000:[gid1=0+560] +fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+2009:[gid1=0+410] +fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+200A:[gid1=0+128] +fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+202F:[gid1=0+280] +fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+205F:[gid1=0+455] +fonts/sha1sum/1c2c3fc37b2d4c3cb2ef726c6cdaaabd4b7f3eb9.ttf:--font-funcs=ot:U+3000:[gid1=0+2048]