Skip to content
体验新版
项目
组织
正在加载...
登录
切换导航
打开侧边栏
OpenHarmony
Third Party Harfbuzz
提交
a177d027
T
Third Party Harfbuzz
项目概览
OpenHarmony
/
Third Party Harfbuzz
大约 1 年 前同步成功
通知
0
Star
18
Fork
0
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
0
列表
看板
标记
里程碑
合并请求
0
Wiki
0
Wiki
分析
仓库
DevOps
项目成员
Pages
T
Third Party Harfbuzz
项目概览
项目概览
详情
发布
仓库
仓库
文件
提交
分支
标签
贡献者
分支图
比较
Issue
0
Issue
0
列表
看板
标记
里程碑
合并请求
0
合并请求
0
Pages
分析
分析
仓库分析
DevOps
Wiki
0
Wiki
成员
成员
收起侧边栏
关闭侧边栏
动态
分支图
创建新Issue
提交
Issue看板
体验新版 GitCode,发现更多精彩内容 >>
提交
a177d027
编写于
8月 28, 2012
作者:
B
Behdad Esfahbod
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
[GSUB] Move ligation logic over
上级
191fa885
变更
2
隐藏空白更改
内联
并排
Showing
2 changed file
with
89 addition
and
69 deletion
+89
-69
src/hb-ot-layout-gsub-table.hh
src/hb-ot-layout-gsub-table.hh
+9
-69
src/hb-ot-layout-gsubgpos-private.hh
src/hb-ot-layout-gsubgpos-private.hh
+80
-0
未找到文件。
src/hb-ot-layout-gsub-table.hh
浏览文件 @
a177d027
...
...
@@ -507,75 +507,15 @@ struct Ligature
/* Deal, we are forming the ligature. */
c
->
buffer
->
merge_clusters
(
c
->
buffer
->
idx
,
c
->
buffer
->
idx
+
end_offset
);
/*
* - If it *is* a mark ligature, we don't allocate a new ligature id, and leave
* the ligature to keep its old ligature id. This will allow it to attach to
* a base ligature in GPOS. Eg. if the sequence is: LAM,LAM,SHADDA,FATHA,HEH,
* and LAM,LAM,HEH for a ligature, they will leave SHADDA and FATHA wit a
* ligature id and component value of 2. Then if SHADDA,FATHA form a ligature
* later, we don't want them to lose their ligature id/component, otherwise
* GPOS will fail to correctly position the mark ligature on top of the
* LAM,LAM,HEH ligature. See:
* https://bugzilla.gnome.org/show_bug.cgi?id=676343
*
* - If a ligature is formed of components that some of which are also ligatures
* themselves, and those ligature components had marks attached to *their*
* components, we have to attach the marks to the new ligature component
* positions! Now *that*'s tricky! And these marks may be following the
* last component of the whole sequence, so we should loop forward looking
* for them and update them.
*
* Eg. the sequence is LAM,LAM,SHADDA,FATHA,HEH, and the font first forms a
* 'calt' ligature of LAM,HEH, leaving the SHADDA and FATHA with a ligature
* id and component == 1. Now, during 'liga', the LAM and the LAM-HEH ligature
* form a LAM-LAM-HEH ligature. We need to reassign the SHADDA and FATHA to
* the new ligature with a component value of 2.
*
* This in fact happened to a font... See:
* https://bugzilla.gnome.org/show_bug.cgi?id=437633
*/
unsigned
int
klass
=
is_mark_ligature
?
0
:
HB_OT_LAYOUT_GLYPH_CLASS_LIGATURE
;
unsigned
int
lig_id
=
is_mark_ligature
?
0
:
allocate_lig_id
(
c
->
buffer
);
unsigned
int
last_lig_id
=
get_lig_id
(
c
->
buffer
->
cur
());
unsigned
int
last_num_components
=
get_lig_num_comps
(
c
->
buffer
->
cur
());
unsigned
int
components_so_far
=
last_num_components
;
if
(
!
is_mark_ligature
)
set_lig_props_for_ligature
(
c
->
buffer
->
cur
(),
lig_id
,
total_component_count
);
c
->
replace_glyph
(
ligGlyph
,
klass
);
for
(
unsigned
int
i
=
1
;
i
<
count
;
i
++
)
{
while
(
c
->
should_mark_skip_current_glyph
())
{
if
(
!
is_mark_ligature
)
{
unsigned
int
new_lig_comp
=
components_so_far
-
last_num_components
+
MIN
(
MAX
(
get_lig_comp
(
c
->
buffer
->
cur
()),
1u
),
last_num_components
);
set_lig_props_for_mark
(
c
->
buffer
->
cur
(),
lig_id
,
new_lig_comp
);
}
c
->
buffer
->
next_glyph
();
}
last_lig_id
=
get_lig_id
(
c
->
buffer
->
cur
());
last_num_components
=
get_lig_num_comps
(
c
->
buffer
->
cur
());
components_so_far
+=
last_num_components
;
/* Skip the base glyph */
c
->
buffer
->
idx
++
;
}
if
(
!
is_mark_ligature
&&
last_lig_id
)
{
/* Re-adjust components for any marks following. */
for
(
unsigned
int
i
=
c
->
buffer
->
idx
;
i
<
c
->
buffer
->
len
;
i
++
)
{
if
(
last_lig_id
==
get_lig_id
(
c
->
buffer
->
info
[
i
]))
{
unsigned
int
new_lig_comp
=
components_so_far
-
last_num_components
+
MIN
(
MAX
(
get_lig_comp
(
c
->
buffer
->
info
[
i
]),
1u
),
last_num_components
);
set_lig_props_for_mark
(
c
->
buffer
->
info
[
i
],
lig_id
,
new_lig_comp
);
}
else
break
;
}
}
ligate_input
(
c
,
count
,
&
component
[
1
],
ligGlyph
,
match_glyph
,
NULL
,
end_offset
,
is_mark_ligature
,
total_component_count
);
return
TRACE_RETURN
(
true
);
}
...
...
src/hb-ot-layout-gsubgpos-private.hh
浏览文件 @
a177d027
...
...
@@ -496,6 +496,86 @@ static inline bool match_input (hb_apply_context_t *c,
return
true
;
}
static
inline
void
ligate_input
(
hb_apply_context_t
*
c
,
unsigned
int
count
,
/* Including the first glyph (not matched) */
const
USHORT
input
[],
/* Array of input values--start with second glyph */
hb_codepoint_t
lig_glyph
,
match_func_t
match_func
,
const
void
*
match_data
,
unsigned
int
end_offset
,
bool
is_mark_ligature
,
unsigned
int
total_component_count
)
{
/*
* - If it *is* a mark ligature, we don't allocate a new ligature id, and leave
* the ligature to keep its old ligature id. This will allow it to attach to
* a base ligature in GPOS. Eg. if the sequence is: LAM,LAM,SHADDA,FATHA,HEH,
* and LAM,LAM,HEH for a ligature, they will leave SHADDA and FATHA wit a
* ligature id and component value of 2. Then if SHADDA,FATHA form a ligature
* later, we don't want them to lose their ligature id/component, otherwise
* GPOS will fail to correctly position the mark ligature on top of the
* LAM,LAM,HEH ligature. See:
* https://bugzilla.gnome.org/show_bug.cgi?id=676343
*
* - If a ligature is formed of components that some of which are also ligatures
* themselves, and those ligature components had marks attached to *their*
* components, we have to attach the marks to the new ligature component
* positions! Now *that*'s tricky! And these marks may be following the
* last component of the whole sequence, so we should loop forward looking
* for them and update them.
*
* Eg. the sequence is LAM,LAM,SHADDA,FATHA,HEH, and the font first forms a
* 'calt' ligature of LAM,HEH, leaving the SHADDA and FATHA with a ligature
* id and component == 1. Now, during 'liga', the LAM and the LAM-HEH ligature
* form a LAM-LAM-HEH ligature. We need to reassign the SHADDA and FATHA to
* the new ligature with a component value of 2.
*
* This in fact happened to a font... See:
* https://bugzilla.gnome.org/show_bug.cgi?id=437633
*/
unsigned
int
klass
=
is_mark_ligature
?
0
:
HB_OT_LAYOUT_GLYPH_CLASS_LIGATURE
;
unsigned
int
lig_id
=
is_mark_ligature
?
0
:
allocate_lig_id
(
c
->
buffer
);
unsigned
int
last_lig_id
=
get_lig_id
(
c
->
buffer
->
cur
());
unsigned
int
last_num_components
=
get_lig_num_comps
(
c
->
buffer
->
cur
());
unsigned
int
components_so_far
=
last_num_components
;
if
(
!
is_mark_ligature
)
set_lig_props_for_ligature
(
c
->
buffer
->
cur
(),
lig_id
,
total_component_count
);
c
->
replace_glyph
(
lig_glyph
,
klass
);
for
(
unsigned
int
i
=
1
;
i
<
count
;
i
++
)
{
while
(
c
->
should_mark_skip_current_glyph
())
{
if
(
!
is_mark_ligature
)
{
unsigned
int
new_lig_comp
=
components_so_far
-
last_num_components
+
MIN
(
MAX
(
get_lig_comp
(
c
->
buffer
->
cur
()),
1u
),
last_num_components
);
set_lig_props_for_mark
(
c
->
buffer
->
cur
(),
lig_id
,
new_lig_comp
);
}
c
->
buffer
->
next_glyph
();
}
last_lig_id
=
get_lig_id
(
c
->
buffer
->
cur
());
last_num_components
=
get_lig_num_comps
(
c
->
buffer
->
cur
());
components_so_far
+=
last_num_components
;
/* Skip the base glyph */
c
->
buffer
->
idx
++
;
}
if
(
!
is_mark_ligature
&&
last_lig_id
)
{
/* Re-adjust components for any marks following. */
for
(
unsigned
int
i
=
c
->
buffer
->
idx
;
i
<
c
->
buffer
->
len
;
i
++
)
{
if
(
last_lig_id
==
get_lig_id
(
c
->
buffer
->
info
[
i
]))
{
unsigned
int
new_lig_comp
=
components_so_far
-
last_num_components
+
MIN
(
MAX
(
get_lig_comp
(
c
->
buffer
->
info
[
i
]),
1u
),
last_num_components
);
set_lig_props_for_mark
(
c
->
buffer
->
info
[
i
],
lig_id
,
new_lig_comp
);
}
else
break
;
}
}
}
static
inline
bool
match_backtrack
(
hb_apply_context_t
*
c
,
unsigned
int
count
,
...
...
编辑
预览
Markdown
is supported
0%
请重试
或
添加新附件
.
添加附件
取消
You are about to add
0
people
to the discussion. Proceed with caution.
先完成此消息的编辑!
取消
想要评论请
注册
或
登录