mirror of
https://github.com/kovidgoyal/calibre.git
synced 2025-07-08 02:34:06 -04:00
Fix cmap table generation not handling contiguous indices
Also fix tab indentation in code imported from fonttools.
This commit is contained in:
Kovid Goyal
parent
46ed78b892
commit
85d09338d7
@ -17,85 +17,85 @@ from calibre.utils.fonts.utils import read_bmp_prefix
|
|||||||
from calibre.utils.fonts.sfnt import UnknownTable, max_power_of_two
|
from calibre.utils.fonts.sfnt import UnknownTable, max_power_of_two
|
||||||
from calibre.utils.fonts.sfnt.errors import UnsupportedFont
|
from calibre.utils.fonts.sfnt.errors import UnsupportedFont
|
||||||
|
|
||||||
def split_range(start_code, end_code, cmap): # {{{
|
def split_range(start_code, end_code, cmap): # {{{
|
||||||
# Try to split a range of character codes into subranges with consecutive
|
# Try to split a range of character codes into subranges with consecutive
|
||||||
# glyph IDs in such a way that the cmap4 subtable can be stored "most"
|
# glyph IDs in such a way that the cmap4 subtable can be stored "most"
|
||||||
# efficiently.
|
# efficiently.
|
||||||
if start_code == end_code:
|
if start_code == end_code:
|
||||||
return [], [end_code]
|
return [], [end_code]
|
||||||
|
|
||||||
last_id = cmap[start_code]
|
last_id = cmap[start_code]
|
||||||
last_code = start_code
|
last_code = start_code
|
||||||
in_order = None
|
in_order = None
|
||||||
ordered_begin = None
|
ordered_begin = None
|
||||||
sub_ranges = []
|
sub_ranges = []
|
||||||
|
|
||||||
# Gather subranges in which the glyph IDs are consecutive.
|
# Gather subranges in which the glyph IDs are consecutive.
|
||||||
for code in range(start_code + 1, end_code + 1):
|
for code in range(start_code + 1, end_code + 1):
|
||||||
glyph_id = cmap[code]
|
glyph_id = cmap[code]
|
||||||
|
|
||||||
if glyph_id - 1 == last_id:
|
if glyph_id - 1 == last_id:
|
||||||
if in_order is None or not in_order:
|
if in_order is None or not in_order:
|
||||||
in_order = 1
|
in_order = 1
|
||||||
ordered_begin = last_code
|
ordered_begin = last_code
|
||||||
else:
|
else:
|
||||||
if in_order:
|
if in_order:
|
||||||
in_order = 0
|
in_order = 0
|
||||||
sub_ranges.append((ordered_begin, last_code))
|
sub_ranges.append((ordered_begin, last_code))
|
||||||
ordered_begin = None
|
ordered_begin = None
|
||||||
|
|
||||||
last_id = glyph_id
|
last_id = glyph_id
|
||||||
last_code = code
|
last_code = code
|
||||||
|
|
||||||
if in_order:
|
if in_order:
|
||||||
sub_ranges.append((ordered_begin, last_code))
|
sub_ranges.append((ordered_begin, last_code))
|
||||||
assert last_code == end_code
|
assert last_code == end_code
|
||||||
|
|
||||||
# Now filter out those new subranges that would only make the data bigger.
|
# Now filter out those new subranges that would only make the data bigger.
|
||||||
# A new segment cost 8 bytes, not using a new segment costs 2 bytes per
|
# A new segment cost 8 bytes, not using a new segment costs 2 bytes per
|
||||||
# character.
|
# character.
|
||||||
new_ranges = []
|
new_ranges = []
|
||||||
for b, e in sub_ranges:
|
for b, e in sub_ranges:
|
||||||
if b == start_code and e == end_code:
|
if b == start_code and e == end_code:
|
||||||
break # the whole range, we're fine
|
break # the whole range, we're fine
|
||||||
if b == start_code or e == end_code:
|
if b == start_code or e == end_code:
|
||||||
threshold = 4 # split costs one more segment
|
threshold = 4 # split costs one more segment
|
||||||
else:
|
else:
|
||||||
threshold = 8 # split costs two more segments
|
threshold = 8 # split costs two more segments
|
||||||
if (e - b + 1) > threshold:
|
if (e - b + 1) > threshold:
|
||||||
new_ranges.append((b, e))
|
new_ranges.append((b, e))
|
||||||
sub_ranges = new_ranges
|
sub_ranges = new_ranges
|
||||||
|
|
||||||
if not sub_ranges:
|
if not sub_ranges:
|
||||||
return [], [end_code]
|
return [], [end_code]
|
||||||
|
|
||||||
if sub_ranges[0][0] != start_code:
|
if sub_ranges[0][0] != start_code:
|
||||||
sub_ranges.insert(0, (start_code, sub_ranges[0][0] - 1))
|
sub_ranges.insert(0, (start_code, sub_ranges[0][0] - 1))
|
||||||
if sub_ranges[-1][1] != end_code:
|
if sub_ranges[-1][1] != end_code:
|
||||||
sub_ranges.append((sub_ranges[-1][1] + 1, end_code))
|
sub_ranges.append((sub_ranges[-1][1] + 1, end_code))
|
||||||
|
|
||||||
# Fill the "holes" in the segments list -- those are the segments in which
|
# Fill the "holes" in the segments list -- those are the segments in which
|
||||||
# the glyph IDs are _not_ consecutive.
|
# the glyph IDs are _not_ consecutive.
|
||||||
i = 1
|
i = 1
|
||||||
while i < len(sub_ranges):
|
while i < len(sub_ranges):
|
||||||
if sub_ranges[i-1][1] + 1 != sub_ranges[i][0]:
|
if sub_ranges[i-1][1] + 1 != sub_ranges[i][0]:
|
||||||
sub_ranges.insert(i, (sub_ranges[i-1][1] + 1, sub_ranges[i][0] - 1))
|
sub_ranges.insert(i, (sub_ranges[i-1][1] + 1, sub_ranges[i][0] - 1))
|
||||||
i = i + 1
|
i = i + 1
|
||||||
i = i + 1
|
i = i + 1
|
||||||
|
|
||||||
# Transform the ranges into start_code/end_code lists.
|
# Transform the ranges into start_code/end_code lists.
|
||||||
start = []
|
start = []
|
||||||
end = []
|
end = []
|
||||||
for b, e in sub_ranges:
|
for b, e in sub_ranges:
|
||||||
start.append(b)
|
start.append(b)
|
||||||
end.append(e)
|
end.append(e)
|
||||||
start.pop(0)
|
start.pop(0)
|
||||||
|
|
||||||
assert len(start) + 1 == len(end)
|
assert len(start) + 1 == len(end)
|
||||||
return start, end
|
return start, end
|
||||||
# }}}
|
# }}}
|
||||||
|
|
||||||
def set_id_delta(id_delta): # {{{
|
def set_id_delta(id_delta): # {{{
|
||||||
# The lowest gid in glyphIndexArray, after subtracting id_delta, must be 1.
|
# The lowest gid in glyphIndexArray, after subtracting id_delta, must be 1.
|
||||||
# id_delta is a short, and must be between -32K and 32K
|
# id_delta is a short, and must be between -32K and 32K
|
||||||
# startCode can be between 0 and 64K-1, and the first glyph index can be between 1 and 64K-1
|
# startCode can be between 0 and 64K-1, and the first glyph index can be between 1 and 64K-1
|
||||||
@ -237,57 +237,53 @@ class CmapTable(UnknownTable):
|
|||||||
start_code = [last_code]
|
start_code = [last_code]
|
||||||
|
|
||||||
for code in codes[1:]:
|
for code in codes[1:]:
|
||||||
if code == last_code + 1:
|
if code == last_code + 1:
|
||||||
last_code = code
|
last_code = code
|
||||||
continue
|
continue
|
||||||
start, end = split_range(start_code[-1], last_code, cmap)
|
start, end = split_range(start_code[-1], last_code, cmap)
|
||||||
start_code.extend(start)
|
start_code.extend(start)
|
||||||
end_code.extend(end)
|
end_code.extend(end)
|
||||||
start_code.append(code)
|
start_code.append(code)
|
||||||
last_code = code
|
last_code = code
|
||||||
end_code.append(last_code)
|
end_code.append(last_code)
|
||||||
start_code.append(0xffff)
|
start_code.append(0xffff)
|
||||||
end_code.append(0xffff)
|
end_code.append(0xffff)
|
||||||
|
|
||||||
id_delta = []
|
id_delta = []
|
||||||
id_range_offset = []
|
id_range_offset = []
|
||||||
glyph_index_array = []
|
glyph_index_array = []
|
||||||
for i in xrange(len(end_code)-1): # skip the closing codes (0xffff)
|
for i in xrange(len(end_code)-1): # skip the closing codes (0xffff)
|
||||||
indices = []
|
indices = list(cmap[char_code] for char_code in xrange(start_code[i], end_code[i] + 1))
|
||||||
for char_code in xrange(start_code[i], end_code[i] + 1):
|
if indices == list(xrange(indices[0], indices[0] + len(indices))):
|
||||||
indices.append(cmap[char_code])
|
# indices is a contiguous list
|
||||||
if (indices == xrange(indices[0], indices[0] + len(indices))):
|
id_delta_temp = set_id_delta(indices[0] - start_code[i])
|
||||||
id_delta_temp = set_id_delta(indices[0] - start_code[i])
|
id_delta.append(id_delta_temp)
|
||||||
id_delta.append(id_delta_temp)
|
id_range_offset.append(0)
|
||||||
id_range_offset.append(0)
|
else:
|
||||||
else:
|
id_delta.append(0)
|
||||||
id_delta.append(0)
|
id_range_offset.append(2 * (len(end_code) + len(glyph_index_array) - i))
|
||||||
id_range_offset.append(2 * (len(end_code) +
|
glyph_index_array.extend(indices)
|
||||||
len(glyph_index_array) - i))
|
id_delta.append(1) # 0xffff + 1 == 0. So this end code maps to .notdef
|
||||||
glyph_index_array.extend(indices)
|
id_range_offset.append(0)
|
||||||
id_delta.append(1) # 0xffff + 1 == 0. So this end code maps to .notdef
|
|
||||||
id_range_offset.append(0)
|
|
||||||
|
|
||||||
seg_count = len(end_code)
|
seg_count = len(end_code)
|
||||||
max_exponent = max_power_of_two(seg_count)
|
max_exponent = max_power_of_two(seg_count)
|
||||||
search_range = 2 * (2 ** max_exponent)
|
search_range = 2 * (2 ** max_exponent)
|
||||||
entry_selector = max_exponent
|
entry_selector = max_exponent
|
||||||
range_shift = 2 * seg_count - search_range
|
range_shift = 2 * seg_count - search_range
|
||||||
|
|
||||||
char_code_array = end_code + [0] + start_code
|
char_code_array = end_code + [0] + start_code
|
||||||
char_code_array = pack(b'>%dH'%len(char_code_array), *char_code_array)
|
char_code_array = pack(b'>%dH'%len(char_code_array), *char_code_array)
|
||||||
id_delta_array = pack(b'>%dh'%len(id_delta), *id_delta)
|
id_delta_array = pack(b'>%dh'%len(id_delta), *id_delta)
|
||||||
rest_array = id_range_offset + glyph_index_array
|
rest_array = id_range_offset + glyph_index_array
|
||||||
rest_array = pack(b'>%dH'%len(rest_array), *rest_array)
|
rest_array = pack(b'>%dH'%len(rest_array), *rest_array)
|
||||||
data = char_code_array + id_delta_array + rest_array
|
data = char_code_array + id_delta_array + rest_array
|
||||||
|
|
||||||
length = calcsize(fmt) + len(data)
|
length = calcsize(fmt) + len(data)
|
||||||
header = pack(fmt, 4, length, 0,
|
header = pack(fmt, 4, length, 0, 2*seg_count, search_range, entry_selector, range_shift)
|
||||||
2*seg_count, search_range, entry_selector, range_shift)
|
self.bmp_table = header + data
|
||||||
self.bmp_table = header + data
|
|
||||||
|
|
||||||
fmt = b'>4HL'
|
fmt = b'>4HL'
|
||||||
offset = calcsize(fmt)
|
offset = calcsize(fmt)
|
||||||
self.raw = pack(fmt, self.version, self.num_tables, 3, 1, offset) + \
|
self.raw = pack(fmt, self.version, self.num_tables, 3, 1, offset) + self.bmp_table
|
||||||
self.bmp_table
|
|
||||||
|
|
||||||
|
|||||||
Loading…
x
Reference in New Issue
Block a user