Normalize various spaces to space if font doesn't support
This resurrects the space fallback feature, after I disabled
the compatibility decomposition. Now I can release HarfBuzz
again without breaking Pango!
It also remembers which space character it was, such that later
on we can approximate the width of this particular space
character. That part is not implemented yet.
We normalize all GC=Zs chars except for U+1680 OGHA SPACE MARK,
which is better left alone.
diff --git a/src/hb-ot-layout-private.hh b/src/hb-ot-layout-private.hh
index 9020c89..1759520 100644
--- a/src/hb-ot-layout-private.hh
+++ b/src/hb-ot-layout-private.hh
@@ -237,23 +237,6 @@
UPROPS_MASK_GEN_CAT = 0x1Fu
};
-enum space_t {
- SPACE = 0,
- SPACE_NBSP,
- SPACE_EN,
- SPACE_EM,
- SPACE_EM_3,
- SPACE_EM_4,
- SPACE_EM_6,
- SPACE_FIGURE,
- SPACE_PUNCTUATION,
- SPACE_THIN,
- SPACE_HAIR,
- SPACE_NARROW,
- SPACE_MEDIUM,
- SPACE_IDEOGRAPHIC,
-};
-
static inline void
_hb_glyph_info_set_unicode_props (hb_glyph_info_t *info, hb_unicode_funcs_t *unicode)
{
@@ -331,36 +314,18 @@
HB_UNICODE_GENERAL_CATEGORY_SPACE_SEPARATOR;
}
static inline void
-_hb_glyph_info_set_unicode_space_for_char (hb_glyph_info_t *info, hb_codepoint_t u)
+_hb_glyph_info_set_unicode_space_fallback_type (hb_glyph_info_t *info, hb_unicode_funcs_t::space_t s)
{
if (unlikely (!_hb_glyph_info_is_unicode_space (info)))
return;
-
- space_t s;
- switch (u)
- {
- default: s = SPACE; break; /* Shouldn't happen. */
- case 0x00A0u: s = SPACE_NBSP; break;
- case 0x2002u: s = SPACE_EN; break;
- case 0x2003u: s = SPACE_EM; break;
- case 0x2004u: s = SPACE_EM_3; break;
- case 0x2005u: s = SPACE_EM_4; break;
- case 0x2006u: s = SPACE_EM_6; break;
- case 0x2007u: s = SPACE_FIGURE; break;
- case 0x2008u: s = SPACE_PUNCTUATION;break;
- case 0x2009u: s = SPACE_THIN; break;
- case 0x200Au: s = SPACE_HAIR; break;
- case 0x202Fu: s = SPACE_NARROW; break;
- case 0x205Fu: s = SPACE_MEDIUM; break;
- case 0x3000u: s = SPACE_IDEOGRAPHIC;break;
- }
-
info->unicode_props() = (((unsigned int) s)<<8) | (info->unicode_props() & 0xFF);
}
-static inline space_t
-_hb_glyph_info_get_unicode_space (const hb_glyph_info_t *info)
+static inline hb_unicode_funcs_t::space_t
+_hb_glyph_info_get_unicode_space_fallback_type (const hb_glyph_info_t *info)
{
- return _hb_glyph_info_is_unicode_space (info) ? (space_t) (info->unicode_props()>>8) : SPACE;
+ return _hb_glyph_info_is_unicode_space (info) ?
+ (hb_unicode_funcs_t::space_t) (info->unicode_props()>>8) :
+ hb_unicode_funcs_t::NOT_SPACE;
}
static inline bool _hb_glyph_info_ligated (const hb_glyph_info_t *info);
diff --git a/src/hb-ot-shape-normalize.cc b/src/hb-ot-shape-normalize.cc
index 0a4d404..c86c634 100644
--- a/src/hb-ot-shape-normalize.cc
+++ b/src/hb-ot-shape-normalize.cc
@@ -98,7 +98,7 @@
output_char (hb_buffer_t *buffer, hb_codepoint_t unichar, hb_codepoint_t glyph)
{
buffer->cur().glyph_index() = glyph;
- buffer->output_glyph (unichar);
+ buffer->output_glyph (unichar); /* This is very confusing indeed. */
_hb_glyph_info_set_unicode_props (&buffer->prev(), buffer->unicode);
}
@@ -164,7 +164,8 @@
{
hb_buffer_t * const buffer = c->buffer;
hb_codepoint_t u = buffer->cur().codepoint;
- hb_codepoint_t glyph;
+ hb_codepoint_t glyph, space_glyph;
+ hb_unicode_funcs_t::space_t space_type;
/* Kind of a cute waterfall here... */
if (shortest && c->font->get_glyph (u, 0, &glyph))
@@ -173,6 +174,13 @@
skip_char (buffer);
else if (!shortest && c->font->get_glyph (u, 0, &glyph))
next_char (buffer, glyph);
+ else if (_hb_glyph_info_is_unicode_space (&buffer->cur()) &&
+ (space_type = buffer->unicode->space_fallback_type (u)) != hb_unicode_funcs_t::NOT_SPACE &&
+ c->font->get_glyph (0x0020u, 0, &space_glyph))
+ {
+ _hb_glyph_info_set_unicode_space_fallback_type (&buffer->cur(), space_type);
+ next_char (buffer, space_glyph);
+ }
else
next_char (buffer, glyph); /* glyph is initialized in earlier branches. */
}
diff --git a/src/hb-unicode-private.hh b/src/hb-unicode-private.hh
index e729826..43bbed6 100644
--- a/src/hb-unicode-private.hh
+++ b/src/hb-unicode-private.hh
@@ -199,6 +199,46 @@
}
}
+ enum space_t {
+ NOT_SPACE = 0,
+ SPACE_NBSP,
+ SPACE_EN,
+ SPACE_EM,
+ SPACE_EM_3,
+ SPACE_EM_4,
+ SPACE_EM_6,
+ 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)
+ {
+ 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;
+ }
+ }
struct {
#define HB_UNICODE_FUNC_IMPLEMENT(name) hb_unicode_##name##_func_t name;
