You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
608 lines
23 KiB
608 lines
23 KiB
--
|
|
-- polyglossia-korean.lua
|
|
--
|
|
|
|
local glyph_id = node.id"glyph"
|
|
local hbox_id = node.id"hlist"
|
|
local vbox_id = node.id"vlist"
|
|
local glue_id = node.id"glue"
|
|
local penalty_id = node.id"penalty"
|
|
local disc_id = node.id"disc"
|
|
|
|
--
|
|
-- attr_korean: variant = plain (0), classic (1), modern (2)
|
|
--
|
|
local attr_korean = luatexbase.attributes["xpg@attr@korean"]
|
|
local attr_josa = luatexbase.attributes["xpg@attr@autojosa"]
|
|
|
|
--
|
|
-- characters after which linebreak is not allowed
|
|
--
|
|
local nobr_after = {
|
|
[0x28] = 1, -- ( LEFT PARENTHESIS
|
|
[0x3C] = 1, -- < LESS-THAN SIGN
|
|
[0x5B] = 1, -- [ LEFT SQUARE BRACKET
|
|
[0x60] = 1, -- ` GRAVE ACCENT
|
|
[0x7B] = 1, -- { LEFT CURLY BRACKET
|
|
[0xAB] = 1, -- « LEFT-POINTING DOUBLE ANGLE QUOTATION MARK
|
|
[0x2018] = 1, -- ‘ LEFT SINGLE QUOTATION MARK
|
|
[0x201C] = 1, -- “ LEFT DOUBLE QUOTATION MARK
|
|
[0x2329] = 1, -- 〈 LEFT-POINTING ANGLE BRACKET
|
|
[0x3008] = 1, -- 〈 LEFT ANGLE BRACKET
|
|
[0x300A] = 1, -- 《 LEFT DOUBLE ANGLE BRACKET
|
|
[0x300C] = 1, -- 「 LEFT CORNER BRACKET
|
|
[0x300E] = 1, -- 『 LEFT WHITE CORNER BRACKET
|
|
[0x3010] = 1, -- 【 LEFT BLACK LENTICULAR BRACKET
|
|
[0x3014] = 1, -- 〔 LEFT TORTOISE SHELL BRACKET
|
|
[0x3016] = 1, -- 〖 LEFT WHITE LENTICULAR BRACKET
|
|
[0x3018] = 1, -- 〘 LEFT WHITE TORTOISE SHELL BRACKET
|
|
[0x301A] = 1, -- 〚 LEFT WHITE SQUARE BRACKET
|
|
[0x301D] = 1, -- 〝 REVERSED DOUBLE PRIME QUOTATION MARK
|
|
[0xFE17] = 1, -- ︗ PRESENTATION FORM FOR VERTICAL LEFT WHITE LENTICULAR BRACKET
|
|
[0xFE35] = 1, -- ︵ PRESENTATION FORM FOR VERTICAL LEFT PARENTHESIS
|
|
[0xFE37] = 1, -- ︷ PRESENTATION FORM FOR VERTICAL LEFT CURLY BRACKET
|
|
[0xFE39] = 1, -- ︹ PRESENTATION FORM FOR VERTICAL LEFT TORTOISE SHELL BRACKET
|
|
[0xFE3B] = 1, -- ︻ PRESENTATION FORM FOR VERTICAL LEFT BLACK LENTICULAR BRACKET
|
|
[0xFE3D] = 1, -- ︽ PRESENTATION FORM FOR VERTICAL LEFT DOUBLE ANGLE BRACKET
|
|
[0xFE3F] = 1, -- ︿ PRESENTATION FORM FOR VERTICAL LEFT ANGLE BRACKET
|
|
[0xFE41] = 1, -- ﹁ PRESENTATION FORM FOR VERTICAL LEFT CORNER BRACKET
|
|
[0xFE43] = 1, -- ﹃ PRESENTATION FORM FOR VERTICAL LEFT WHITE CORNER BRACKET
|
|
[0xFE47] = 1, -- ﹇ PRESENTATION FORM FOR VERTICAL LEFT SQUARE BRACKET
|
|
[0xFE59] = 1, -- ﹙ SMALL LEFT PARENTHESIS
|
|
[0xFE5B] = 1, -- ﹛ SMALL LEFT CURLY BRACKET
|
|
[0xFE5D] = 1, -- ﹝ SMALL LEFT TORTOISE SHELL BRACKET
|
|
[0xFF08] = 1, -- （ FULLWIDTH LEFT PARENTHESIS
|
|
[0xFF3B] = 1, -- ［ FULLWIDTH LEFT SQUARE BRACKET
|
|
[0xFF5B] = 1, -- ｛ FULLWIDTH LEFT CURLY BRACKET
|
|
[0xFF5F] = 1, -- ｟ FULLWIDTH LEFT WHITE PARENTHESIS
|
|
[0xFF62] = 1, -- ｢ HALFWIDTH LEFT CORNER BRACKET
|
|
}
|
|
|
|
--
|
|
-- characters before which linebreak is not allowed
|
|
-- (currently, not much differences among the followings)
|
|
-- 1: normal chars
|
|
-- 2: hangul jamo vowels and trailing consonants
|
|
-- 3: kana small letters
|
|
-- 0: dashes (supress visible spacing)
|
|
--
|
|
local nobr_before = setmetatable({
|
|
[0x21] = 1, -- ! EXCLAMATION MARK
|
|
[0x22] = 1, -- " QUOTATION MARK
|
|
[0x27] = 1, -- ' APOSTROPHE
|
|
[0x29] = 1, -- ) RIGHT PARENTHESIS
|
|
[0x2C] = 1, -- , COMMA
|
|
[0x2D] = 0, -- - HYPHEN-MINUS
|
|
[0x2E] = 1, -- . FULL STOP
|
|
[0x2F] = 0, -- / SOLIDUS
|
|
[0x3A] = 0, -- : COLON
|
|
[0x3B] = 1, -- ; SEMICOLON
|
|
[0x3E] = 1, -- > GREATER-THAN SIGN
|
|
[0x3F] = 1, -- ? QUESTION MARK
|
|
[0x5C] = 0, -- \ REVERSE SOLIDUS
|
|
[0x5D] = 1, -- ] RIGHT SQUARE BRACKET
|
|
[0x7D] = 1, -- } RIGHT CURLY BRACKET
|
|
[0x7E] = 0, -- ~ TILDE
|
|
[0xB7] = 1, -- · MIDDLE DOT
|
|
[0xBB] = 1, -- » RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK
|
|
[0x2013] = 0, -- – EN DASH
|
|
[0x2014] = 0, -- — EM DASH
|
|
[0x2015] = 1, -- ― HORIZONTAL BAR
|
|
[0x2019] = 1, -- ’ RIGHT SINGLE QUOTATION MARK
|
|
[0x201D] = 1, -- ” RIGHT DOUBLE QUOTATION MARK
|
|
[0x2025] = 1, -- ‥ TWO DOT LEADER
|
|
[0x2026] = 1, -- … HORIZONTAL ELLIPSIS
|
|
[0x232A] = 1, -- 〉 RIGHT-POINTING ANGLE BRACKET
|
|
[0x3001] = 1, -- 、 IDEOGRAPHIC COMMA
|
|
[0x3002] = 1, -- 。 IDEOGRAPHIC FULL STOP
|
|
[0x3005] = 1, -- 々 IDEOGRAPHIC ITERATION MARK
|
|
[0x3009] = 1, -- 〉 RIGHT ANGLE BRACKET
|
|
[0x300B] = 1, -- 》 RIGHT DOUBLE ANGLE BRACKET
|
|
[0x300D] = 1, -- 」 RIGHT CORNER BRACKET
|
|
[0x300F] = 1, -- 』 RIGHT WHITE CORNER BRACKET
|
|
[0x3011] = 1, -- 】 RIGHT BLACK LENTICULAR BRACKET
|
|
[0x3015] = 1, -- 〕 RIGHT TORTOISE SHELL BRACKET
|
|
[0x3017] = 1, -- 〗 RIGHT WHITE LENTICULAR BRACKET
|
|
[0x3019] = 1, -- 〙 RIGHT WHITE TORTOISE SHELL BRACKET
|
|
[0x301B] = 1, -- 〛 RIGHT WHITE SQUARE BRACKET
|
|
[0x301C] = 1, -- 〜 WAVE DASH
|
|
[0x301E] = 1, -- 〞 DOUBLE PRIME QUOTATION MARK
|
|
[0x301F] = 1, -- 〟 LOW DOUBLE PRIME QUOTATION MARK
|
|
[0x3035] = 1, -- 〵 VERTICAL KANA REPEAT MARK LOWER HALF
|
|
[0x303B] = 1, -- 〻 VERTICAL IDEOGRAPHIC ITERATION MARK
|
|
[0x303C] = 1, -- 〼 MASU MARK
|
|
[0x3041] = 3, -- ぁ HIRAGANA LETTER SMALL A
|
|
[0x3043] = 3, -- ぃ HIRAGANA LETTER SMALL I
|
|
[0x3045] = 3, -- ぅ HIRAGANA LETTER SMALL U
|
|
[0x3047] = 3, -- ぇ HIRAGANA LETTER SMALL E
|
|
[0x3049] = 3, -- ぉ HIRAGANA LETTER SMALL O
|
|
[0x3063] = 3, -- っ HIRAGANA LETTER SMALL TU
|
|
[0x3083] = 3, -- ゃ HIRAGANA LETTER SMALL YA
|
|
[0x3085] = 3, -- ゅ HIRAGANA LETTER SMALL YU
|
|
[0x3087] = 3, -- ょ HIRAGANA LETTER SMALL YO
|
|
[0x308E] = 3, -- ゎ HIRAGANA LETTER SMALL WA
|
|
[0x3095] = 3, -- ゕ HIRAGANA LETTER SMALL KA
|
|
[0x3096] = 3, -- ゖ HIRAGANA LETTER SMALL KE
|
|
[0x3099] = 1, -- COMBINING KATAKANA-HIRAGANA VOICED SOUND MARK
|
|
[0x309A] = 1, -- COMBINING KATAKANA-HIRAGANA SEMI-VOICED SOUND MARK
|
|
[0x309B] = 1, -- ゛ KATAKANA-HIRAGANA VOICED SOUND MARK
|
|
[0x309C] = 1, -- ゜ KATAKANA-HIRAGANA SEMI-VOICED SOUND MARK
|
|
[0x309D] = 1, -- ゝ HIRAGANA ITERATION MARK
|
|
[0x309E] = 1, -- ゞ HIRAGANA VOICED ITERATION MARK
|
|
[0x30A0] = 1, -- ゠ KATAKANA-HIRAGANA DOUBLE HYPHEN
|
|
[0x30A1] = 3, -- ァ KATAKANA LETTER SMALL A
|
|
[0x30A3] = 3, -- ィ KATAKANA LETTER SMALL I
|
|
[0x30A5] = 3, -- ゥ KATAKANA LETTER SMALL U
|
|
[0x30A7] = 3, -- ェ KATAKANA LETTER SMALL E
|
|
[0x30A9] = 3, -- ォ KATAKANA LETTER SMALL O
|
|
[0x30C3] = 3, -- ッ KATAKANA LETTER SMALL TU
|
|
[0x30E3] = 3, -- ャ KATAKANA LETTER SMALL YA
|
|
[0x30E5] = 3, -- ュ KATAKANA LETTER SMALL YU
|
|
[0x30E7] = 3, -- ョ KATAKANA LETTER SMALL YO
|
|
[0x30EE] = 3, -- ヮ KATAKANA LETTER SMALL WA
|
|
[0x30F5] = 3, -- ヵ KATAKANA LETTER SMALL KA
|
|
[0x30F6] = 3, -- ヶ KATAKANA LETTER SMALL KE
|
|
[0x30FB] = 1, -- ・ KATAKANA MIDDLE DOT
|
|
[0x30FC] = 1, -- ー KATAKANA-HIRAGANA PROLONGED SOUND MARK
|
|
[0x30FD] = 1, -- ヽ KATAKANA ITERATION MARK
|
|
[0x30FE] = 1, -- ヾ KATAKANA VOICED ITERATION MARK
|
|
[0xFE30] = 1, -- ︰ PRESENTATION FORM FOR VERTICAL TWO DOT LEADER
|
|
[0xFE31] = 1, -- ︱ PRESENTATION FORM FOR VERTICAL EM DASH
|
|
[0xFE32] = 1, -- ︲ PRESENTATION FORM FOR VERTICAL EN DASH
|
|
[0xFE36] = 1, -- ︶ PRESENTATION FORM FOR VERTICAL RIGHT PARENTHESIS
|
|
[0xFE38] = 1, -- ︸ PRESENTATION FORM FOR VERTICAL RIGHT CURLY BRACKET
|
|
[0xFE3A] = 1, -- ︺ PRESENTATION FORM FOR VERTICAL RIGHT TORTOISE SHELL BRACKET
|
|
[0xFE3C] = 1, -- ︼ PRESENTATION FORM FOR VERTICAL RIGHT BLACK LENTICULAR BRACKET
|
|
[0xFE3E] = 1, -- ︾ PRESENTATION FORM FOR VERTICAL RIGHT DOUBLE ANGLE BRACKET
|
|
[0xFE40] = 1, -- ﹀ PRESENTATION FORM FOR VERTICAL RIGHT ANGLE BRACKET
|
|
[0xFE42] = 1, -- ﹂ PRESENTATION FORM FOR VERTICAL RIGHT CORNER BRACKET
|
|
[0xFE44] = 1, -- ﹄ PRESENTATION FORM FOR VERTICAL RIGHT WHITE CORNER BRACKET
|
|
[0xFE48] = 1, -- ﹈ PRESENTATION FORM FOR VERTICAL RIGHT SQUARE BRACKET
|
|
[0xFE5A] = 1, -- ﹚ SMALL RIGHT PARENTHESIS
|
|
[0xFE5C] = 1, -- ﹜ SMALL RIGHT CURLY BRACKET
|
|
[0xFE5E] = 1, -- ﹞ SMALL RIGHT TORTOISE SHELL BRACKET
|
|
[0xFF01] = 1, -- ！ FULLWIDTH EXCLAMATION MARK
|
|
[0xFF09] = 1, -- ） FULLWIDTH RIGHT PARENTHESIS
|
|
[0xFF0C] = 1, -- ， FULLWIDTH COMMA
|
|
[0xFF0E] = 1, -- ． FULLWIDTH FULL STOP
|
|
[0xFF1A] = 1, -- ： FULLWIDTH COLON
|
|
[0xFF1B] = 1, -- ； FULLWIDTH SEMICOLON
|
|
[0xFF1F] = 1, -- ？ FULLWIDTH QUESTION MARK
|
|
[0xFF3D] = 1, -- ］ FULLWIDTH RIGHT SQUARE BRACKET
|
|
[0xFF5D] = 1, -- ｝ FULLWIDTH RIGHT CURLY BRACKET
|
|
[0xFF60] = 1, -- ｠ FULLWIDTH RIGHT WHITE PARENTHESIS
|
|
[0xFF61] = 1, -- ｡ HALFWIDTH IDEOGRAPHIC FULL STOP
|
|
[0xFF63] = 1, -- ｣ HALFWIDTH RIGHT CORNER BRACKET
|
|
[0xFF64] = 1, -- ､ HALFWIDTH IDEOGRAPHIC COMMA
|
|
[0xFF65] = 1, -- ･ HALFWIDTH KATAKANA MIDDLE DOT
|
|
[0xFF9E] = 1, -- ﾞ HALFWIDTH KATAKANA VOICED SOUND MARK
|
|
[0xFF9F] = 1, -- ﾟ HALFWIDTH KATAKANA SEMI-VOICED SOUND MARK
|
|
}, { __index = function(_,c)
|
|
if c >= 0x1160 and c <= 0x11FF then return 2 end
|
|
if c >= 0xD7B0 and c <= 0xD7FF then return 2 end
|
|
if c >= 0x302A and c <= 0x302F then return 1 end
|
|
if c >= 0x31F0 and c <= 0x31FF then return 3 end
|
|
if c >= 0xFF67 and c <= 0xFF70 then return 3 end
|
|
if c >= 0xFE00 and c <= 0xFE0F then return 1 end
|
|
if c >= 0xFE10 and c <= 0xFE19 and not (c == 0xFE17) then return 1 end
|
|
if c >= 0xFE50 and c <= 0xFE58 then return 1 end
|
|
if c >= 0xE0100 and c <= 0xE01EF then return 1 end
|
|
end
|
|
})
|
|
|
|
--
|
|
-- whether 'c' is a cjk character
|
|
--
|
|
local function is_cjk (c)
|
|
return c >= 0xAC00 and c <= 0xD7FF
|
|
or c >= 0x1100 and c <= 0x11FF
|
|
or c >= 0xA960 and c <= 0xA97F
|
|
or c >= 0x2E80 and c <= 0x9FFF
|
|
or c >= 0xF900 and c <= 0xFAFF
|
|
or c >= 0xFE10 and c <= 0xFE1F
|
|
or c >= 0xFE30 and c <= 0xFE6F
|
|
or c >= 0xFF00 and c <= 0xFFEF
|
|
or c >= 0x1F100 and c <= 0x1F2FF
|
|
or c >= 0x20000 and c <= 0x2FA1F
|
|
or nobr_after[c] and c > 0x2014
|
|
or nobr_before[c] and c > 0x2014
|
|
end
|
|
|
|
--
|
|
-- classify cjk characters
|
|
-- 1: openings
|
|
-- 2: closings
|
|
-- 3: centered chars
|
|
-- 4: full stops
|
|
-- 5: ellipses
|
|
-- 6: exclamation and question marks
|
|
-- 0: all others
|
|
--
|
|
local charclass = setmetatable({
|
|
[0x2018] = 1, [0x201C] = 1, [0x2329] = 1, [0x3008] = 1,
|
|
[0x300A] = 1, [0x300C] = 1, [0x300E] = 1, [0x3010] = 1,
|
|
[0x3014] = 1, [0x3016] = 1, [0x3018] = 1, [0x301A] = 1,
|
|
[0x301D] = 1, [0xFE17] = 1, [0xFE35] = 1, [0xFE37] = 1,
|
|
[0xFE39] = 1, [0xFE3B] = 1, [0xFE3D] = 1, [0xFE3F] = 1,
|
|
[0xFE41] = 1, [0xFE43] = 1, [0xFE47] = 1, [0xFF08] = 1,
|
|
[0xFF3B] = 1, [0xFF5B] = 1, [0xFF5F] = 1, [0xFF62] = 1,
|
|
[0x2019] = 2, [0x201D] = 2, [0x232A] = 2, [0x3001] = 2,
|
|
[0x3009] = 2, [0x300B] = 2, [0x300D] = 2, [0x300F] = 2,
|
|
[0x3011] = 2, [0x3015] = 2, [0x3017] = 2, [0x3019] = 2,
|
|
[0x301B] = 2, [0x301E] = 2, [0x301F] = 2, [0xFE10] = 2,
|
|
[0xFE11] = 2, [0xFE18] = 2, [0xFE36] = 2, [0xFE38] = 2,
|
|
[0xFE3A] = 2, [0xFE3C] = 2, [0xFE3E] = 2, [0xFE40] = 2,
|
|
[0xFE42] = 2, [0xFE44] = 2, [0xFE48] = 2, [0xFF09] = 2,
|
|
[0xFF0C] = 2, [0xFF3D] = 2, [0xFF5D] = 2, [0xFF60] = 2,
|
|
[0xFF63] = 2, [0xFF64] = 2, [0x00B7] = 3, [0x30FB] = 3,
|
|
[0xFF1A] = 3, [0xFF1B] = 3, [0xFF65] = 3, [0x3002] = 4,
|
|
[0xFE12] = 4, [0xFF0E] = 4, [0xFF61] = 4, [0x2015] = 5,
|
|
[0x2025] = 5, [0x2026] = 5, [0xFE19] = 5, [0xFE30] = 5,
|
|
[0xFE31] = 5, [0xFE15] = 6, [0xFE16] = 6, [0xFF01] = 6,
|
|
[0xFF1F] = 6,
|
|
}, { __index = function() return 0 end })
|
|
|
|
--
|
|
-- table for spacing between char classes
|
|
-- 1 stands for 0.5*fontsize when variant=classic
|
|
--
|
|
local intercharclass = { [0] =
|
|
{ [0] = nil, {1,1}, nil, {.5,.5} },
|
|
{ [0] = nil, nil, nil, {.5,.5} },
|
|
{ [0] = {1,1}, {1,1}, nil, {.5,.5}, nil, {1,1}, {1,1} },
|
|
{ [0] = {.5,.5},{.5,.5},{.5,.5},{1,.5}, {.5,.5},{.5,.5},{.5,.5} },
|
|
{ [0] = {1,0}, {1,0}, nil, {1.5,.5},nil, {1,0}, {1,0} },
|
|
{ [0] = nil, {1,1}, nil, {.5,.5} },
|
|
{ [0] = {1,1}, {1,1}, nil, {.5,.5} },
|
|
}
|
|
|
|
--
|
|
-- get a new penalty node
|
|
--
|
|
local function get_new_penalty (p)
|
|
local penalty = node.new("penalty")
|
|
penalty.penalty = p
|
|
return penalty
|
|
end
|
|
|
|
--
|
|
-- get a new glue node
|
|
--
|
|
local function get_new_glue (...)
|
|
local glue = node.new("glue")
|
|
node.setglue(glue, ...)
|
|
return glue
|
|
end
|
|
|
|
--
|
|
-- return 0.5*fontsize of given fontid
|
|
-- space: true if variant=modern; then 0.5*interword_space
|
|
--
|
|
local function get_font_size (fid, space)
|
|
local size = font.getparameters(fid)
|
|
if space then
|
|
size = size and size.space or 196608
|
|
else
|
|
size = size and size.quad or 655360
|
|
end
|
|
return size/2
|
|
end
|
|
|
|
--
|
|
-- charclass 1 thru 4 will be packed in \hbox to 0.5em{\hss? curr \hss?}
|
|
-- when variant=classic/modern
|
|
--
|
|
local function glyph_to_box (head, curr, class)
|
|
local g, h = curr
|
|
local size = get_font_size(g.font)
|
|
head, curr = node.remove(head, curr)
|
|
g.next, g.prev = nil, nil
|
|
local hss = get_new_glue(0, 65536, 65536, 2, 2)
|
|
if class == 1 then
|
|
h, hss.next, g.prev = hss, g, hss
|
|
elseif class == 2 or class == 4 then
|
|
h, g.next, hss.prev = g, hss, g
|
|
else
|
|
local hss2 = node.copy(hss)
|
|
h, hss.next, g.prev, g.next, hss2.prev = hss, g, hss, hss2, g
|
|
end
|
|
h = nodes.simple_font_handler(h)
|
|
local box = node.hpack(h, size, "exactly")
|
|
if curr then
|
|
head, curr = node.insert_before(head, curr, box)
|
|
else
|
|
head, curr = node.insert_after(head, node.tail(head), box)
|
|
end
|
|
return head, curr
|
|
end
|
|
|
|
--
|
|
-- insert spacing defined as charclass[a][b] between a and b
|
|
-- f: fontid
|
|
-- var: variant = plain, classic, modern
|
|
-- cc: charclass of current char
|
|
-- nc: charclass of next char
|
|
-- nobr: linebreak is not allowed
|
|
--
|
|
local function insert_cjk_penalty_glue (head, curr, f, var, cc, nc, nobr)
|
|
if nobr or cc == 1 or nc > 1 then
|
|
local penalty = get_new_penalty(10000)
|
|
head, curr = node.insert_after(head, curr, penalty)
|
|
end
|
|
local factor = get_font_size(f, var == 2)
|
|
local t = intercharclass[cc][nc]
|
|
local glue = get_new_glue(t[1]*factor, nil, t[2]*factor)
|
|
head, curr = node.insert_after(head, curr, glue)
|
|
return head, curr
|
|
end
|
|
|
|
--
|
|
-- insert inter-character spacing in other normal cases
|
|
-- f: fontid
|
|
-- var: variant = plain, classic, modern
|
|
-- x: true between cjk and non-cjk (a little more spacing)
|
|
--
|
|
local function insert_penalty_glue (head, curr, f, var, x)
|
|
if var ~= 1 then
|
|
local penalty = get_new_penalty(50)
|
|
head, curr = node.insert_after(head, curr, penalty)
|
|
end
|
|
local size, glue = get_font_size(f, x and var == 2)
|
|
if x then
|
|
glue = get_new_glue(size/2, size/4, size/8)
|
|
else
|
|
glue = get_new_glue(0, size/10, size/50)
|
|
end
|
|
head, curr = node.insert_after(head, curr, glue)
|
|
return head, curr
|
|
end
|
|
|
|
--
|
|
-- main process for linebreak and inter-character spacing
|
|
-- lb: true if pre_linebreak_filter
|
|
--
|
|
local function korean_break (head, lb)
|
|
local curr = head
|
|
while curr do
|
|
if curr.id == glyph_id then
|
|
local var = node.has_attribute(curr, attr_korean)
|
|
if var then
|
|
local c, f = curr.char or 0, curr.font or 0
|
|
local cc, cjkc = charclass[c], is_cjk(c)
|
|
|
|
-- compress cjk punctuations when charclass is 1 thru 4
|
|
if var > 0 and cc > 0 and cc < 5 then
|
|
head, curr = glyph_to_box(head, curr, cc)
|
|
end
|
|
|
|
local next = curr.next
|
|
if next and next.id == glyph_id then
|
|
local n = next.char or 0
|
|
local nc = charclass[n]
|
|
local nobr = nobr_before[n] or nobr_after[c]
|
|
|
|
-- insert spacing as of intercharclass
|
|
if var > 0 and intercharclass[cc][nc] then
|
|
head, curr = insert_cjk_penalty_glue(head, curr, f, var, cc, nc, nobr)
|
|
|
|
-- or insert spacing when linebreak is allowed
|
|
elseif not nobr then
|
|
local cjkn = is_cjk(n)
|
|
|
|
-- if curr or next is cjk char
|
|
if cjkc or cjkn then
|
|
|
|
-- if between cjk and non-cjk
|
|
if var > 0 and not (cjkc and cjkn) and nobr_before[c] ~= 0 then
|
|
head, curr = insert_penalty_glue(head, curr, f, var, true)
|
|
|
|
-- or under pre_linebreak_filter
|
|
elseif lb then
|
|
head, curr = insert_penalty_glue(head, curr, f, var)
|
|
end
|
|
end
|
|
end
|
|
end
|
|
end
|
|
end
|
|
curr = curr.next
|
|
end
|
|
return head
|
|
end
|
|
|
|
--
|
|
-- process for reordering hangul tone marks (U+302E, U+302F)
|
|
-- some hangul fonts (eg. Noto CJK) are so designed that hangul tone marks
|
|
-- should be moved to the first position of a syllable.
|
|
-- Currently, this functionality is not provided by luaotfload.
|
|
--
|
|
local function reorder_tm (head)
|
|
local curr, tone = node.slide(head)
|
|
while curr do
|
|
if curr.id == glyph_id and node.has_attribute(curr, attr_korean) then
|
|
local f = font.getfont(curr.font) or font.fonts[curr.font]
|
|
if f and f.hb then -- harfbuzz do the right thing
|
|
tone = nil
|
|
else
|
|
local c, wd = curr.char or 0, curr.width or 0
|
|
if (c == 0x302E or c == 0x302F) and wd > 0 then
|
|
tone = curr
|
|
elseif tone and not nobr_before[c] then
|
|
head = node.remove(head, tone)
|
|
tone.next, tone.prev = nil, nil
|
|
head, curr = node.insert_before(head, curr, tone)
|
|
tone = nil
|
|
end
|
|
end
|
|
end
|
|
curr = curr.prev
|
|
end
|
|
return head
|
|
end
|
|
|
|
--
|
|
-- automatic josa selection
|
|
--
|
|
local josa_table = {
|
|
-- consonant ㄹ, vowel, other consonants
|
|
[0xAC00] = {0xC774, 0xAC00, 0xC774}, -- 가 => 이, 가, 이
|
|
[0xC740] = {0xC740, 0xB294, 0xC740}, -- 은 => 은, 는, 은
|
|
[0xC744] = {0xC744, 0xB97C, 0xC744}, -- 을 => 을, 를, 을
|
|
[0xC640] = {0xACFC, 0xC640, 0xACFC}, -- 와 => 과, 와, 과
|
|
[0xC73C] = {nil, nil, 0xC73C}, -- 으(로) => , , 으
|
|
[0xC774] = {0xC774, nil, 0xC774}, -- 이(라) => 이, , 이
|
|
}
|
|
|
|
--
|
|
-- helper function for number-like characters
|
|
--
|
|
local function josa_char_num (t, c)
|
|
c = c % 10 + 0x30
|
|
return t[c] or 2
|
|
end
|
|
|
|
--
|
|
-- decide josa selection
|
|
--
|
|
local josa_code = setmetatable({
|
|
[0x30] = 3, [0x31] = 1, [0x33] = 3, [0x36] = 3, [0x37] = 1,
|
|
[0x38] = 1, [0x4C] = 1, [0x4D] = 3, [0x4E] = 3, [0x6C] = 1,
|
|
[0x6D] = 3, [0x6E] = 3, [0xFB02] = 1, [0xFB04] = 1,
|
|
},{ __index = function(t,c)
|
|
if c >= 0xAC00 and c <= 0xD7A3 then
|
|
c = (c - 0xAC00) % 28 + 0x11A7
|
|
end
|
|
if c >= 0x11A8 and c <= 0x11FF then
|
|
if c == 0x11AF then return 1 end
|
|
return 3
|
|
end
|
|
if c >= 0xD7CB and c <= 0xD7FB then return 3 end
|
|
if c >= 0x2170 and c <= 0x217F then c = c - 0x10 end
|
|
if c >= 0x2160 and c <= 0x216F then
|
|
if c >= 0x216C then return 3 end
|
|
return josa_char_num(t, c - 0x215F)
|
|
end
|
|
if c >= 0x2460 and c <= 0x2473 then return josa_char_num(t, c - 0x245F) end
|
|
if c >= 0x2474 and c <= 0x2487 then return josa_char_num(t, c - 0x2473) end
|
|
if c >= 0x2488 and c <= 0x249B then return josa_char_num(t, c - 0x2487) end
|
|
if c >= 0x249C and c <= 0x24B5 then return t[c - 0x249C + 0x61] or 2 end
|
|
if c >= 0x24B6 and c <= 0x24CF then return t[c - 0x24B6 + 0x61] or 2 end
|
|
if c >= 0x24D0 and c <= 0x24E9 then return t[c - 0x24D0 + 0x61] or 2 end
|
|
if c >= 0x3131 and c <= 0x318E then
|
|
if c == 0x3139 then return 1 end
|
|
if c >= 0x314F and c <= 0x3163 or c >= 0x3187 then return 2 end
|
|
return 3
|
|
end
|
|
if c >= 0x3260 and c <= 0x327E then c = c - 0x60 end
|
|
if c >= 0x3200 and c <= 0x321E then
|
|
if c == 0x3203 then return 1 end
|
|
if c >= 0x320E then return 2 end
|
|
return 3
|
|
end
|
|
if c >= 0xFF10 and c <= 0xFF19 then return josa_char_num(t, c - 0xFF10) end
|
|
if c >= 0xFF21 and c <= 0xFF3A then return t[c - 0xFF21 + 0x61] or 2 end
|
|
if c >= 0xFF41 and c <= 0xFF5A then return t[c - 0xFF41 + 0x61] or 2 end
|
|
return 2
|
|
end
|
|
})
|
|
|
|
--
|
|
-- obtain char that comes just before the josa
|
|
--
|
|
local function get_prev_char (p)
|
|
while p do
|
|
if p.id == glyph_id then
|
|
local pc = p.char or 0
|
|
if not nobr_after[pc] then
|
|
if not nobr_before[pc] or nobr_before[pc] >= 2 then
|
|
return pc
|
|
end
|
|
end
|
|
elseif p.id == hbox_id or p.id == vbox_id then
|
|
local pc = get_prev_char(node.slide(p.head))
|
|
if pc then return pc end
|
|
end
|
|
p = p.prev
|
|
end
|
|
end
|
|
|
|
--
|
|
-- main process of josa selection
|
|
--
|
|
local function auto_josa (head)
|
|
local curr, tofree = head, {}
|
|
while curr do
|
|
if curr.id == glyph_id then
|
|
local josa = node.has_attribute(curr, attr_josa)
|
|
if josa then
|
|
local cc = curr.char or 0
|
|
if josa == 0 then
|
|
josa = josa_code[get_prev_char(curr.prev) or 0x30]
|
|
end
|
|
if cc == 0xC774 then
|
|
local n = curr.next
|
|
if n and n.char and n.char >= 0xAC00 and n.char <= 0xD7A3 then
|
|
else
|
|
cc = 0xAC00
|
|
end
|
|
end
|
|
local new = josa_table[cc]
|
|
if new then
|
|
cc = new[josa]
|
|
if cc then
|
|
curr.char = cc
|
|
else
|
|
head = node.remove(head, curr)
|
|
table.insert(tofree, curr)
|
|
end
|
|
end
|
|
node.unset_attribute(curr, attr_josa)
|
|
end
|
|
end
|
|
curr = curr.next
|
|
end
|
|
for _,v in ipairs(tofree) do node.free(v) end
|
|
return head
|
|
end
|
|
|
|
--
|
|
-- now register to luatex callbacks
|
|
-- As char value of glyphs can be changed by opentype GSUB process,
|
|
-- we have to occupy the first position among callback functions.
|
|
--
|
|
local prepend_to_callback
|
|
if luatexbase.base_add_to_callback then
|
|
prepend_to_callback = function(name, func, desc)
|
|
luatexbase.add_to_callback(name, func, desc, 1)
|
|
end
|
|
else
|
|
prepend_to_callback = function(name, func, desc)
|
|
local t = { {func, desc} }
|
|
for _,v in ipairs(luatexbase.callback_descriptions(name)) do
|
|
table.insert(t, {luatexbase.remove_from_callback(name, v)})
|
|
end
|
|
for _,v in ipairs(t) do
|
|
luatexbase.add_to_callback(name, v[1], v[2])
|
|
end
|
|
end
|
|
end
|
|
|
|
prepend_to_callback ("pre_linebreak_filter",
|
|
function(head)
|
|
head = auto_josa(head)
|
|
head = korean_break(head, true)
|
|
head = reorder_tm(head)
|
|
return head
|
|
end,
|
|
"polyglossia.lang_korean")
|
|
|
|
prepend_to_callback ("hpack_filter",
|
|
function(head)
|
|
head = auto_josa(head)
|
|
head = korean_break(head)
|
|
head = reorder_tm(head)
|
|
return head
|
|
end,
|
|
"polyglossia.lang_korean")
|
|
|
|
-- vim:ft=lua:tw=0:sw=4:ts=4:expandtab
|
|
|