| Current File : //usr/share/texlive/texmf-dist/tex/luatex/luaotfload/fontloader-font-osd.lua |
if not modules then modules = { } end modules ['font-osd'] = { -- script devanagari
version = 1.001,
comment = "companion to font-ini.mkiv",
author = "Kai Eigner, TAT Zetwerk / Hans Hagen, PRAGMA ADE",
copyright = "TAT Zetwerk / PRAGMA ADE / ConTeXt Development Team",
license = "see context related readme files"
}
-- we need to check nbsphash (context only)
-- A few remarks:
--
-- This code is a partial rewrite of the code that deals with devanagari. The data
-- and logic is by Kai Eigner and based based on Microsoft's OpenType specifications
-- for specific scripts, but with a few improvements. More information can be found
-- at:
--
-- deva: http://www.microsoft.com/typography/OpenType%20Dev/devanagari/introO.mspx
-- dev2: http://www.microsoft.com/typography/OpenType%20Dev/devanagari/intro.mspx
--
-- Rajeesh Nambiar provided patches for the malayalam variant. Thanks to feedback
-- from the mailing list some aspects could be improved.
--
-- As I touched nearly all code, reshuffled it, optimized a lot, etc. etc. (imagine
-- how much can get messed up in over a week work) it could be that I introduced
-- bugs. There is more to gain (esp in the functions applied to a range) but I'll do
-- that when everything works as expected. Kai's original code is kept in
-- font-odk.lua as a reference so blame me (HH) for bugs. (We no longer ship that
-- file as the code below has diverted too much and in the meantime has more than
-- doubled in size.)
--
-- Interesting is that Kai managed to write this on top of the existing otf handler.
-- Only a few extensions were needed, like a few more analyzing states and dealing
-- with changed head nodes in the core scanner as that only happens here. There's a
-- lot going on here and it's only because I touched nearly all code that I got a
-- bit of a picture of what happens. For in-depth knowledge one needs to consult
-- Kai.
--
-- The rewrite mostly deals with efficiency, both in terms of speed and code. We
-- also made sure that it suits generic use as well as use in ConTeXt. I removed
-- some buglets but can as well have messed up the logic by doing this. For this we
-- keep the original around as that serves as reference. Due to the lots of
-- reshuffling glyphs quite some leaks occur(red) but once I'm satisfied with the
-- rewrite I'll weed them. I also integrated initialization etc into the regular
-- mechanisms.
--
-- In the meantime, we're down from 25.5-3.5=22 seconds to 17.7-3.5=14.2 seconds for
-- a 100 page sample (mid 2012) with both variants so it's worth the effort. Some
-- more speedup is to be expected. Due to the method chosen it will never be real
-- fast. If I ever become a power user I'll have a go at some further speed up. I
-- will rename some functions (and features) once we don't need to check the
-- original code. We now use a special subset sequence for use inside the analyzer
-- (after all we could can store this in the dataset and save redundant analysis).
--
-- By now we have yet another incremental improved version. In the end I might
-- rewrite the code.
--
-- Hans Hagen, PRAGMA-ADE, Hasselt NL
-- Todo:
--
-- Matras: according to Microsoft typography specifications "up to one of each type:
-- pre-, above-, below- or post- base", but that does not seem to be right. It could
-- become an option.
--
-- Resources:
--
-- The tables that we had here are now generated from char-def.lua or in the case of
-- generic usage loaded from luatex-basics-chr.lua. Still a couple of entries need
-- to be added to char-def.lua but finally I moved the indic specific tables there.
-- For generic usage one can create the relevant resources by running:
--
-- context luatex-basics-prepare.tex
--
-- and an overview with:
--
-- context --global s-fonts-basics.mkiv
--
-- For now we have defined: bengali, devanagari, gujarati, gurmukhi, kannada,
-- malayalam, oriya, tamil and tolugu but not all are checked. Also, some of the
-- code below might need to be adapted to the extra scripts.
local insert, imerge, copy, tohash = table.insert, table.imerge, table.copy, table.tohash
local next, type = next, type
local report = logs.reporter("otf","devanagari")
fonts = fonts or { }
fonts.analyzers = fonts.analyzers or { }
fonts.analyzers.methods = fonts.analyzers.methods or { node = { otf = { } } }
local otf = fonts.handlers.otf
local handlers = otf.handlers
local methods = fonts.analyzers.methods
local otffeatures = fonts.constructors.features.otf
local registerotffeature = otffeatures.register
local nuts = nodes.nuts
local getnext = nuts.getnext
local getprev = nuts.getprev
local getboth = nuts.getboth
local getid = nuts.getid
local getchar = nuts.getchar
local getfont = nuts.getfont
local getsubtype = nuts.getsubtype
local setlink = nuts.setlink
local setnext = nuts.setnext
local setprev = nuts.setprev
local setchar = nuts.setchar
local getprop = nuts.getprop
local setprop = nuts.setprop
local getstate = nuts.getstate
local setstate = nuts.setstate
local ischar = nuts.ischar
local insert_node_after = nuts.insert_after
local copy_node = nuts.copy
local remove_node = nuts.remove
local flush_list = nuts.flush_list
local flush_node = nuts.flush_node
local copyinjection = nodes.injections.copy -- KE: is this necessary? HH: probably not as positioning comes later and we rawget/set
local unsetvalue = attributes.unsetvalue
local fontdata = fonts.hashes.identifiers
local a_syllabe = attributes.private('syllabe')
local dotted_circle = 0x25CC
local c_nbsp = 0x00A0
local c_zwnj = 0x200C
local c_zwj = 0x200D
local states = fonts.analyzers.states -- not features
local s_rphf = states.rphf
local s_half = states.half
local s_pref = states.pref
local s_blwf = states.blwf
local s_pstf = states.pstf
local s_init = states.init
local replace_all_nbsp = nil
replace_all_nbsp = function(head) -- delayed definition
replace_all_nbsp = typesetters and typesetters.characters and typesetters.characters.replacenbspaces or function(head)
return head
end
return replace_all_nbsp(head)
end
local processcharacters = nil
if context then
local fontprocesses = fonts.hashes.processes
function processcharacters(head,font)
local processors = fontprocesses[font]
for i=1,#processors do
head = processors[i](head,font,0)
end
return head
end
else
function processcharacters(head,font)
local processors = fontdata[font].shared.processes
for i=1,#processors do
head = processors[i](head,font,0)
end
return head
end
end
-- We can assume that script are not mixed in the source but if that is the case
-- we might need to have consonants etc per script and initialize a local table
-- pointing to the right one. But not now.
local indicgroups = characters and characters.indicgroups
if not indicgroups and characters then
local indic = {
c = { }, -- consonant
i = { }, -- independent vowel
d = { }, -- dependent vowel
m = { }, -- vowel modifier
s = { }, -- stress tone mark
o = { }, -- other
}
local indicmarks = {
l = { }, -- left | pre_mark
t = { }, -- top | above_mark
b = { }, -- bottom | below_mark
r = { }, -- right | post_mark
s = { }, -- split | twopart_mark
}
local indicclasses = {
nukta = { },
halant = { },
ra = { },
anudatta = { },
}
local indicorders = {
bp = { }, -- before_postscript
ap = { }, -- after_postscript
bs = { }, -- before_subscript
as = { }, -- after_subscript
bh = { }, -- before_half
ah = { }, -- after_half
bm = { }, -- before_main
am = { }, -- after_main
}
for k, v in next, characters.data do
local i = v.indic
if i then
indic[i][k] = true
i = v.indicmark
if i then
if i == "s" then
local s = v.specials
indicmarks[i][k] = { s[2], s[3] }
else
indicmarks[i][k] = true
end
end
i = v.indicclass
if i then
indicclasses[i][k] = true
end
i = v.indicorder
if i then
indicorders[i][k] = true
end
end
end
indicgroups = {
consonant = indic.c,
independent_vowel = indic.i,
dependent_vowel = indic.d,
vowel_modifier = indic.m,
stress_tone_mark = indic.s,
-- other = indic.o,
pre_mark = indicmarks.l,
above_mark = indicmarks.t,
below_mark = indicmarks.b,
post_mark = indicmarks.r,
twopart_mark = indicmarks.s,
nukta = indicclasses.nukta,
halant = indicclasses.halant,
ra = indicclasses.ra,
anudatta = indicclasses.anudatta,
before_postscript = indicorders.bp,
after_postscript = indicorders.ap,
before_half = indicorders.bh,
after_half = indicorders.ah,
before_subscript = indicorders.bs,
after_subscript = indicorders.as,
before_main = indicorders.bm,
after_main = indicorders.am,
}
indic = nil
indicmarks = nil
indicclasses = nil
indicorders = nil
characters.indicgroups = indicgroups
end
local consonant = indicgroups.consonant
local independent_vowel = indicgroups.independent_vowel
local dependent_vowel = indicgroups.dependent_vowel
local vowel_modifier = indicgroups.vowel_modifier
local stress_tone_mark = indicgroups.stress_tone_mark
local pre_mark = indicgroups.pre_mark
local above_mark = indicgroups.above_mark
local below_mark = indicgroups.below_mark
local post_mark = indicgroups.post_mark
local twopart_mark = indicgroups.twopart_mark
local nukta = indicgroups.nukta
local halant = indicgroups.halant
local ra = indicgroups.ra
local anudatta = indicgroups.anudatta
local before_postscript = indicgroups.before_postscript
local after_postscript = indicgroups.after_postscript
local before_half = indicgroups.before_half
local after_half = indicgroups.after_half
local before_subscript = indicgroups.before_subscript
local after_subscript = indicgroups.after_subscript
local before_main = indicgroups.before_main
local after_main = indicgroups.after_main
local mark_four = table.merged (
pre_mark,
above_mark,
below_mark,
post_mark
)
local mark_above_below_post = table.merged (
above_mark,
below_mark,
post_mark
)
-- We use some pseudo features as we need to manipulate the nodelist based
-- on information in the font as well as already applied features. We can
-- probably replace some of the code below by injecting 'real' features
-- using the extension mechanism.
local zw_char = { -- both_joiners_true
[c_zwnj] = true,
[c_zwj ] = true,
}
local dflt_true = {
dflt = true,
}
local two_defaults = { }
local one_defaults = { }
local false_flags = { false, false, false, false }
local sequence_reorder_matras = {
features = { dv01 = two_defaults },
flags = false_flags,
name = "dv01_reorder_matras",
order = { "dv01" },
type = "devanagari_reorder_matras",
nofsteps = 1,
steps = {
{
coverage = pre_mark,
}
}
}
local sequence_reorder_reph = {
features = { dv02 = two_defaults },
flags = false_flags,
name = "dv02_reorder_reph",
order = { "dv02" },
type = "devanagari_reorder_reph",
nofsteps = 1,
steps = {
{
coverage = { },
}
}
}
local sequence_reorder_pre_base_reordering_consonants = {
features = { dv03 = one_defaults },
flags = false_flags,
name = "dv03_reorder_pre_base_reordering_consonants",
order = { "dv03" },
type = "devanagari_reorder_pre_base_reordering_consonants",
nofsteps = 1,
steps = {
{
coverage = { },
}
}
}
local sequence_remove_joiners = {
features = { dv04 = one_defaults },
flags = false_flags,
name = "dv04_remove_joiners",
order = { "dv04" },
type = "devanagari_remove_joiners",
nofsteps = 1,
steps = {
{
coverage = zw_char, -- both_joiners_true
},
}
}
-- Looping over feature twice as efficient as looping over basic forms (some
-- 350 checks instead of 750 for one font). This is something to keep an eye on
-- as it might depends on the font. Not that it's a bottleneck.
local basic_shaping_forms = {
akhn = true,
blwf = true,
cjct = true,
half = true,
nukt = true,
pref = true,
pstf = true,
rkrf = true,
rphf = true,
vatu = true,
locl = true,
}
local valid = {
abvs = true,
akhn = true,
blwf = true,
calt = true,
cjct = true,
half = true,
haln = true,
nukt = true,
pref = true,
pres = true,
pstf = true,
psts = true,
rkrf = true,
rphf = true,
vatu = true,
pres = true,
abvs = true,
blws = true,
psts = true,
haln = true,
calt = true,
locl = true,
}
local scripts = { }
local scripts_one = { "deva", "mlym", "beng", "gujr", "guru", "knda", "orya", "taml", "telu" }
local scripts_two = { "dev2", "mlm2", "bng2", "gjr2", "gur2", "knd2", "ory2", "tml2", "tel2" }
local nofscripts = #scripts_one
for i=1,nofscripts do
local one = scripts_one[i]
local two = scripts_two[i]
scripts[one] = true
scripts[two] = true
two_defaults[two] = dflt_true
one_defaults[one] = dflt_true
one_defaults[two] = dflt_true
end
local function valid_one(s) for i=1,nofscripts do if s[scripts_one[i]] then return true end end end
local function valid_two(s) for i=1,nofscripts do if s[scripts_two[i]] then return true end end end
local function initializedevanagi(tfmdata)
local script, language = otf.scriptandlanguage(tfmdata,attr) -- todo: take fast variant
if scripts[script] then
local resources = tfmdata.resources
local devanagari = resources.devanagari
if not devanagari then
--
report("adding devanagari features to font")
--
local gsubfeatures = resources.features.gsub
local sequences = resources.sequences
local sharedfeatures = tfmdata.shared.features
--
gsubfeatures["dv01"] = two_defaults -- reorder matras
gsubfeatures["dv02"] = two_defaults -- reorder reph
gsubfeatures["dv03"] = one_defaults -- reorder pre base reordering consonants
gsubfeatures["dv04"] = one_defaults -- remove joiners
--
local reorder_pre_base_reordering_consonants = copy(sequence_reorder_pre_base_reordering_consonants)
local reorder_reph = copy(sequence_reorder_reph)
local reorder_matras = copy(sequence_reorder_matras)
local remove_joiners = copy(sequence_remove_joiners)
local lastmatch = 0
for s=1,#sequences do -- classify chars and make sure basic_shaping_forms come first
local features = sequences[s].features
if features then
for k, v in next, features do
if k == "locl" then
local steps = sequences[s].steps
local nofsteps = sequences[s].nofsteps
for i=1,nofsteps do
local step = steps[i]
local coverage = step.coverage
if coverage then
for k, v in next, pre_mark do
local locl = coverage[k]
if locl then
if #locl > 0 then --contextchain; KE: is this right?
for j=1,#locl do
local ck = locl[j]
local f = ck[4]
local chainlookups = ck[6]
if chainlookups then
local chainlookup = chainlookups[f]
for j=1,#chainlookup do
local chainstep = chainlookup[j]
local steps = chainstep.steps
local nofsteps = chainstep.nofsteps
for i=1,nofsteps do
local step = steps[i]
local coverage = step.coverage
if coverage then
locl = coverage[k]
end
end
end
end
end
end
if locl then
reorder_matras.steps[1].coverage[locl] = true
end
end
end
end
end
end
if basic_shaping_forms[k] then
lastmatch = lastmatch + 1
if s ~= lastmatch then
table.insert(sequences, lastmatch, table.remove(sequences, s))
end
end
end
end
end
local insertindex = lastmatch + 1
--
if tfmdata.properties.language then
dflt_true[tfmdata.properties.language] = true
end
--
insert(sequences,insertindex,reorder_pre_base_reordering_consonants)
insert(sequences,insertindex,reorder_reph)
insert(sequences,insertindex,reorder_matras)
insert(sequences,insertindex,remove_joiners)
--
local blwfcache = { }
local vatucache = { }
local pstfcache = { }
local seqsubset = { }
local rephstep = {
coverage = { } -- will be adapted each work
}
local devanagari = {
reph = false,
vattu = false,
blwfcache = blwfcache,
vatucache = vatucache,
pstfcache = pstfcache,
seqsubset = seqsubset,
reorderreph = rephstep,
}
--
reorder_reph.steps = { rephstep }
--
local pre_base_reordering_consonants = { }
reorder_pre_base_reordering_consonants.steps[1].coverage = pre_base_reordering_consonants
--
resources.devanagari = devanagari
--
for s=1,#sequences do
local sequence = sequences[s]
local steps = sequence.steps
local nofsteps = sequence.nofsteps
local features = sequence.features
local has_rphf = features.rphf
local has_blwf = features.blwf
local has_vatu = features.vatu
local has_pstf = features.pstf
if has_rphf and has_rphf[script] then
devanagari.reph = true
elseif (has_blwf and has_blwf[script] ) or (has_vatu and has_vatu[script] ) then
devanagari.vattu = true
for i=1,nofsteps do
local step = steps[i]
local coverage = step.coverage
if coverage then
for k, v in next, coverage do
for h, w in next, halant do
if v[h] then
if not blwfcache[k] then
blwfcache[k] = v
end
end
if has_vatu and has_vatu[script] and not vatucache[k] then
vatucache[k] = v
end
end
end
end
end
elseif has_pstf and has_pstf[script] then
for i=1,nofsteps do
local step = steps[i]
local coverage = step.coverage
if coverage then
for k, v in next, coverage do
if not pstfcache[k] then
pstfcache[k] = v
end
end
for k, v in next, ra do
local r = coverage[k]
if r then
local found = false
if #r > 0 then -- contextchain; KE: is this right?
for j=1,#r do
local ck = r[j]
local f = ck[4]
local chainlookups = ck[6]
if chainlookups and chainlookups[f] then --KE: why is check for chainlookups[f] necessacy???
local chainlookup = chainlookups[f]
for j=1,#chainlookup do
local chainstep = chainlookup[j]
local steps = chainstep.steps
local nofsteps = chainstep.nofsteps
for i=1,nofsteps do
local step = steps[i]
local coverage = step.coverage
if coverage then
local h = coverage[k]
if h then
for k, v in next, h do
found = v and v.ligature
if found then
pre_base_reordering_consonants[found] = true
break
end
end
if found then
break
end
end
end
end
end
end
end
else
for k, v in next, r do
found = v and v.ligature
if found then
pre_base_reordering_consonants[found] = true
break
end
end
end
if found then
break
end
end
end
end
end
end
for kind, spec in next, features do
if valid[kind] and valid_two(spec)then
for i=1,nofsteps do
local step = steps[i]
local coverage = step.coverage
if coverage then
local reph, rephbase = false, false
if kind == "rphf" then
-- rphf acts on consonant + halant
for k, v in next, ra do
local r = coverage[k]
if r then
rephbase = k
local h = false
if #r > 0 then --contextchain; KE: is this right?
for j=1,#r do
local ck = r[j]
local f = ck[4]
local chainlookups = ck[6]
if chainlookups then
local chainlookup = chainlookups[f]
for j=1,#chainlookup do
local chainstep = chainlookup[j]
local steps = chainstep.steps
local nofsteps = chainstep.nofsteps
for i=1,nofsteps do
local step = steps[i]
local coverage = step.coverage
if coverage then
local r = coverage[k]
if r then
for k, v in next, halant do
local h = r[k]
if h then
reph = h.ligature or false
break
end
end
if h then
break
end
end
end
end
end
end
end
else
for k, v in next, halant do
local h = r[k]
if h then
reph = h.ligature or false
break
end
end
end
if reph then
break
end
end
end
end
seqsubset[#seqsubset+1] = { kind, coverage, reph, rephbase }
end
end
end
if kind == "pref" then
local steps = sequence.steps
local nofsteps = sequence.nofsteps
for i=1,nofsteps do
local step = steps[i]
local coverage = step.coverage
if coverage then
for k, v in next, halant do
local h = coverage[k]
if h then
local found = false
if #h > 0 then -- contextchain; KE: is this right?
for j=1,#h do
local ck = h[j]
local f = ck[4]
local chainlookups = ck[6]
if chainlookups then
local chainlookup = chainlookups[f]
for j=1,#chainlookup do
local chainstep = chainlookup[j]
local steps = chainstep.steps
local nofsteps = chainstep.nofsteps
for i=1,nofsteps do
local step = steps[i]
local coverage = step.coverage
if coverage then
local h = coverage[k]
if h then
for k, v in next, h do
found = v and v.ligature
if found then
pre_base_reordering_consonants[found] = true
break
end
end
if found then
break
end
end
end
end
end
end
end
else
for k, v in next, h do
found = v and v.ligature
if found then
pre_base_reordering_consonants[found] = true
break
end
end
end
if found then
break
end
end
end
end
end
end
end
end
--
if two_defaults[script] then
sharedfeatures["dv01"] = true -- dv01_reorder_matras
sharedfeatures["dv02"] = true -- dv02_reorder_reph
sharedfeatures["dv03"] = true -- dv03_reorder_pre_base_reordering_consonants
sharedfeatures["dv04"] = true -- dv04_remove_joiners
elseif one_defaults[script] then
sharedfeatures["dv03"] = true -- dv03_reorder_pre_base_reordering_consonants
sharedfeatures["dv04"] = true -- dv04_remove_joiners
end
if script == "mlym" or script == "taml" then
devanagari.left_matra_before_base = true
end
end
end
end
registerotffeature {
name = "devanagari",
description = "inject additional features",
default = true,
initializers = {
node = initializedevanagi,
},
}
local show_syntax_errors = false
local function inject_syntax_error(head,current,char)
local signal = copy_node(current)
copyinjection(signal,current)
if pre_mark[char] then
setchar(signal,dotted_circle)
else
setchar(current,dotted_circle)
end
return insert_node_after(head,current,signal)
end
-- hm, this is applied to one character:
local function initialize_one(font,attr) -- we need a proper hook into the dataset initializer
local tfmdata = fontdata[font]
local datasets = otf.dataset(tfmdata,font,attr) -- don't we know this one?
local devanagaridata = datasets.devanagari
if not devanagaridata then
devanagaridata = {
reph = false,
vattu = false,
blwfcache = { },
vatucache = { },
pstfcache = { },
}
datasets.devanagari = devanagaridata
local resources = tfmdata.resources
local devanagari = resources.devanagari
for s=1,#datasets do
local dataset = datasets[s]
if dataset and dataset[1] then -- value
local kind = dataset[4]
if kind == "rphf" then
-- deva
devanagaridata.reph = true
elseif kind == "blwf" or kind == "vatu" then
-- deva
devanagaridata.vattu = true
-- dev2
devanagaridata.blwfcache = devanagari.blwfcache
devanagaridata.vatucache = devanagari.vatucache
devanagaridata.pstfcache = devanagari.pstfcache
end
end
end
end
return devanagaridata.reph, devanagaridata.vattu, devanagaridata.blwfcache, devanagaridata.vatucache, devanagaridata.pstfcache
end
local function contextchain(contexts, n)
local char = getchar(n)
for k=1,#contexts do
local ck = contexts[k]
local seq = ck[3]
local f = ck[4]
local l = ck[5]
if (l - f) == 1 and seq[f+1][char] then
local ok = true
local c = n
for i=l+1,#seq do
c = getnext(c)
if not c or not seq[i][ischar(c)] then
ok = false
break
end
end
if ok then
c = getprev(n)
for i=1,f-1 do
c = getprev(c)
if not c or not seq[f-i][ischar(c)] then
ok = false
end
end
end
if ok then
return true
end
end
end
return false
end
local function order_matras(c)
local cn = getnext(c)
local char = getchar(cn)
while dependent_vowel[char] do
local next = getnext(cn)
local cc = c
local cchar = getchar(cc)
while cc ~= cn do
if (above_mark[char] and (below_mark[cchar] or post_mark[cchar])) or (below_mark[char] and (post_mark[cchar])) then
local prev, next = getboth(cn)
if next then
setprev(next,prev)
end
-- todo: setlink
setnext(prev,next)
setnext(getprev(cc),cn)
setprev(cn,getprev(cc))
setnext(cn,cc)
setprev(cc,cn)
break
end
cc = getnext(cc)
cchar = getchar(cc)
end
cn = next
char = getchar(cn)
end
end
local function reorder_one(head,start,stop,font,attr,nbspaces)
local reph, vattu, blwfcache, vatucache, pstfcache = initialize_one(font,attr) -- todo: a hash[font]
local devanagari = fontdata[font].resources.devanagari
local current = start
local n = getnext(start)
local base = nil
local firstcons = nil
local lastcons = nil
local basefound = false
if reph and ra[getchar(start)] and halant[getchar(n)] then
-- if syllable starts with Ra + H and script has 'Reph' then exclude Reph
-- from candidates for base consonants
if n == stop then
return head, stop, nbspaces
end
if getchar(getnext(n)) == c_zwj then
current = start
else
current = getnext(n)
setstate(start,s_rphf)
end
end
if getchar(current) == c_nbsp then
-- Stand Alone cluster
if current == stop then
stop = getprev(stop)
head = remove_node(head,current)
flush_node(current)
return head, stop, nbspaces
else
nbspaces = nbspaces + 1
base = current
firstcons = current
lastcons = current
current = getnext(current)
if current ~= stop then
local char = getchar(current)
if nukta[char] then
current = getnext(current)
char = getchar(current)
end
if char == c_zwj and current ~= stop then
local next = getnext(current)
if next ~= stop and halant[getchar(next)] then
current = next
next = getnext(current)
local tmp = next and getnext(next) or nil -- needs checking
local changestop = next == stop
local tempcurrent = copy_node(next)
copyinjection(tempcurrent,next)
local nextcurrent = copy_node(current)
copyinjection(nextcurrent,current) -- KE: necessary? HH: probably not as positioning comes later and we rawget/set
setlink(tempcurrent,nextcurrent)
setstate(tempcurrent,s_blwf)
tempcurrent = processcharacters(tempcurrent,font)
setstate(tempcurrent,unsetvalue)
if getchar(next) == getchar(tempcurrent) then
flush_list(tempcurrent)
if show_syntax_errors then
head, current = inject_syntax_error(head,current,char)
end
else
setchar(current,getchar(tempcurrent)) -- we assumes that the result of blwf consists of one node
local freenode = getnext(current)
setlink(current,tmp)
flush_node(freenode)
flush_list(tempcurrent)
if changestop then
stop = current
end
end
end
end
end
end
end
while not basefound do
-- find base consonant
local char = getchar(current)
if consonant[char] then
setstate(current,s_half)
if not firstcons then
firstcons = current
end
lastcons = current
if not base then
base = current
elseif blwfcache[char] then
-- consonant has below-base form
setstate(current,s_blwf)
elseif pstfcache[char] then
-- consonant has post-base form
setstate(current,s_pstf)
else
base = current
end
end
basefound = current == stop
current = getnext(current)
end
if base ~= lastcons then
-- if base consonant is not last one then move halant from base consonant to last one
local np = base
local n = getnext(base)
local ch = getchar(n)
if nukta[ch] then
np = n
n = getnext(n)
ch = getchar(n)
end
if halant[ch] then
if lastcons ~= stop then
local ln = getnext(lastcons)
if nukta[getchar(ln)] then
lastcons = ln
end
end
-- local np = getprev(n)
local nn = getnext(n)
local ln = getnext(lastcons) -- what if lastcons is nn ?
setlink(np,nn)
setnext(lastcons,n)
if ln then
setprev(ln,n)
end
setnext(n,ln)
setprev(n,lastcons)
if lastcons == stop then
stop = n
end
end
end
n = getnext(start)
if n ~= stop and ra[getchar(start)] and halant[getchar(n)] and not zw_char[getchar(getnext(n))] then
-- if syllable starts with Ra + H then move this combination so that it follows either:
-- the post-base 'matra' (if any) or the base consonant
local matra = base
if base ~= stop then
local next = getnext(base)
if dependent_vowel[getchar(next)] then
matra = next
end
end
-- [sp][start][n][nn] [matra|base][?]
-- [matra|base][start] [n][?] [sp][nn]
local sp = getprev(start)
local nn = getnext(n)
local mn = getnext(matra)
setlink(sp,nn)
setlink(matra,start)
setlink(n,mn)
if head == start then
head = nn
end
start = nn
if matra == stop then
stop = n
end
end
local current = start
while current ~= stop do
local next = getnext(current)
if next ~= stop and halant[getchar(next)] and getchar(getnext(next)) == c_zwnj then
setstate(current,unsetvalue)
end
current = next
end
if base ~= stop and getstate(base) then -- state can also be init
local next = getnext(base)
if halant[getchar(next)] and not (next ~= stop and getchar(getnext(next)) == c_zwj) then
setstate(base,unsetvalue)
end
end
-- split two- or three-part matras into their parts. Then, move the left 'matra' part to the beginning of the syllable.
-- classify consonants and 'matra' parts as pre-base, above-base (Reph), below-base or post-base, and group elements of the syllable (consonants and 'matras') according to this classification
local current, allreordered, moved = start, false, { [base] = true }
local a, b, p, bn = base, base, base, getnext(base)
if base ~= stop and nukta[getchar(bn)] then
a, b, p = bn, bn, bn
end
while not allreordered do
-- current is always consonant
local c = current
local n = getnext(current)
local l = nil -- used ?
if c ~= stop then
local ch = getchar(n)
if nukta[ch] then
c = n
n = getnext(n)
ch = getchar(n)
end
if c ~= stop then
if halant[ch] then
c = n
n = getnext(n)
ch = getchar(n)
end
local tpm = twopart_mark[ch]
while tpm do
local extra = copy_node(n)
copyinjection(extra,n)
ch = tpm[1]
setchar(n,ch)
setchar(extra,tpm[2])
head = insert_node_after(head,current,extra)
tpm = twopart_mark[ch]
end
while c ~= stop and dependent_vowel[ch] do
c = n
n = getnext(n)
ch = getchar(n)
end
if c ~= stop then
if vowel_modifier[ch] then
c = n
n = getnext(n)
ch = getchar(n)
end
if c ~= stop and stress_tone_mark[ch] then
c = n
n = getnext(n)
end
end
end
end
local bp = getprev(firstcons)
local cn = getnext(current)
local last = getnext(c)
while cn ~= last do
-- move pre-base matras...
if pre_mark[getchar(cn)] then
if devanagari.left_matra_before_base then
local prev, next = getboth(cn)
setlink(prev,next)
if cn == stop then
stop = getprev(cn)
end
if base == start then
if head == start then
head = cn
end
start = cn
end
setlink(getprev(base),cn)
setlink(cn,base)
-- setlink(getprev(base),cn,base) -- maybe
cn = next
else
if bp then
setnext(bp,cn)
end
local prev, next = getboth(cn)
if next then
setprev(next,prev)
end
setnext(prev,next)
if cn == stop then
stop = prev
end
setprev(cn,bp)
setlink(cn,firstcons)
if firstcons == start then
if head == start then
head = cn
end
start = cn
end
cn = next
end
elseif current ~= base and dependent_vowel[getchar(cn)] then
local prev, next = getboth(cn)
if next then
setprev(next,prev)
end
setnext(prev,next)
if cn == stop then
stop = prev
end
setlink(b,cn,getnext(b))
order_matras(cn)
cn = next
elseif current == base and dependent_vowel[getchar(cn)] then
local cnn = getnext(cn)
order_matras(cn)
cn = cnn
while cn ~= last and dependent_vowel[getchar(cn)] do
cn = getnext(cn)
end
else
cn = getnext(cn)
end
end
allreordered = c == stop
current = getnext(c)
end
if reph or vattu then
local current, cns = start, nil
while current ~= stop do
local c = current
local n = getnext(current)
if ra[getchar(current)] and halant[getchar(n)] then
c = n
n = getnext(n)
local b, bn = base, base
while bn ~= stop do
local next = getnext(bn)
if dependent_vowel[getchar(next)] then
b = next
end
bn = next
end
if getstate(current,s_rphf) then
-- position Reph (Ra + H) after post-base 'matra' (if any) since these
-- become marks on the 'matra', not on the base glyph
if b ~= current then
if current == start then
if head == start then
head = n
end
start = n
end
if b == stop then
stop = c
end
local prev = getprev(current)
setlink(prev,n)
local next = getnext(b)
setlink(c,next)
setlink(b,current)
end
elseif cns and getnext(cns) ~= current then -- todo: optimize next
-- position below-base Ra (vattu) following the consonants on which it is placed (either the base consonant or one of the pre-base consonants)
local cp = getprev(current)
local cnsn = getnext(cns)
setlink(cp,n)
setlink(cns,current) -- cns ?
setlink(c,cnsn)
if c == stop then
stop = cp
break
end
current = getprev(n)
end
else
local char = getchar(current)
if consonant[char] then
cns = current
local next = getnext(cns)
if halant[getchar(next)] then
cns = next
end
if not vatucache[char] then
next = getnext(cns)
while dependent_vowel[getchar(next)] do
cns = next
next = getnext(cns)
end
end
elseif char == c_nbsp then
nbspaces = nbspaces + 1
cns = current
local next = getnext(cns)
if halant[getchar(next)] then
cns = next
end
if not vatucache[char] then
next = getnext(cns)
while dependent_vowel[getchar(next)] do
cns = next
next = getnext(cns)
end
end
end
end
current = getnext(current)
end
end
if getchar(base) == c_nbsp then
nbspaces = nbspaces - 1
if base == stop then
stop = getprev(stop)
end
head = remove_node(head,base)
flush_node(base)
end
return head, stop, nbspaces
end
-- If a pre-base matra character had been reordered before applying basic features,
-- the glyph can be moved closer to the main consonant based on whether half-forms had been formed.
-- Actual position for the matra is defined as “after last standalone halant glyph,
-- after initial matra position and before the main consonant”.
-- If ZWJ or ZWNJ follow this halant, position is moved after it.
-- so we break out ... this is only done for the first 'word' (if we feed words we can as
-- well test for non glyph.
function handlers.devanagari_reorder_matras(head,start) -- no leak
local current = start -- we could cache attributes here
local startfont = getfont(start)
local startattr = getprop(start,a_syllabe)
while current do
local char = ischar(current,startfont)
local next = getnext(current)
if char and getprop(current,a_syllabe) == startattr then
if halant[char] then -- state can also be init
if next then
local char = ischar(next,startfont)
if char and zw_char[char] and getprop(next,a_syllabe) == startattr then
current = next
next = getnext(current)
end
end
-- can be optimzied
local startnext = getnext(start)
head = remove_node(head,start)
setlink(start,next)
setlink(current,start)
-- setlink(current,start,next) -- maybe
start = startnext
break
-- elseif consonant[char] and (not getstate(current) or getstate(current,s_init) then
-- startnext = getnext(start)
-- head = remove_node(head,start)
-- if current == head then
-- setlink(start,current)
-- head = start
-- else
-- setlink(getprev(current),start)
-- setlink(start,current)
-- end
-- start = startnext
-- break
end
else
break
end
current = next
end
return head, start, true
end
-- Reph’s original position is always at the beginning of the syllable, (i.e. it is
-- not reordered at the character reordering stage). However, it will be reordered
-- according to the basic-forms shaping results. Possible positions for reph,
-- depending on the script, are; after main, before post-base consonant forms, and
-- after post-base consonant forms.
-- In Devanagari reph has reordering position 'before postscript' and dev2 only
-- follows step 2, 4, and 6.
local rephbase = { }
function handlers.devanagari_reorder_reph(head,start)
local current = getnext(start)
local startnext = nil
local startprev = nil
local startfont = getfont(start)
local startattr = getprop(start,a_syllabe)
--
::step_1::
--
-- If reph should be positioned after post-base consonant forms, proceed to step 5.
--
local char = ischar(start,startfont)
local rephbase = rephbase[startfont][char]
if char and after_subscript[rephbase] then
goto step_5
end
--
::step_2::
--
-- If the reph repositioning class is not after post-base: target position is after
-- the first explicit halant glyph between the first post-reph consonant and last
-- main consonant. If ZWJ or ZWNJ are following this halant, position is moved after
-- it. If such position is found, this is the target position. Otherwise, proceed to
-- the next step. Note: in old-implementation fonts, where classifications were
-- fixed in shaping engine, there was no case where reph position will be found on
-- this step.
--
if char and not after_postscript[rephbase] then
while current do
local char = ischar(current,startfont)
if char and getprop(current,a_syllabe) == startattr then
if halant[char] then
local next = getnext(current)
if next then
local nextchar = ischar(next,startfont)
if nextchar and zw_char[nextchar] and getprop(next,a_syllabe) == startattr then
current = next
next = getnext(current)
end
end
startnext = getnext(start)
head = remove_node(head,start)
setlink(start,next)
setlink(current,start)
-- setlink(current,start,next) -- maybe
start = startnext
startattr = getprop(start,a_syllabe)
break
end
current = getnext(current)
else
break
end
end
end
--
::step_3::
--
-- If reph should be repositioned after the main consonant: find the first consonant
-- not ligated with main, or find the first consonant that is not a potential
-- pre-base reordering Ra.
--
if not startnext then
if char and after_main[rephbase] then
current = getnext(start)
while current do
local char = ischar(current,startfont)
if char and getprop(current,a_syllabe) == startattr then
if consonant[char] and not getstate(current,s_pref) then
startnext = getnext(start)
head = remove_node(head,start)
setlink(current,start)
setlink(start,getnext(current))
-- setlink(current,start,getnext(current)) -- maybe
start = startnext
startattr = getprop(start,a_syllabe)
break
end
current = getnext(current)
else
break
end
end
end
end
--
::step_4::
--
-- If reph should be positioned before post-base consonant, find first post-base
-- classified consonant not ligated with main. If no consonant is found, the target
-- position should be before the first matra, syllable modifier sign or vedic sign.
--
if not startnext then
if char and before_postscript[rephbase] then
current = getnext(start)
local c = nil
while current do
local char = ischar(current,startfont)
if char and getprop(current,a_syllabe) == startattr then
if getstate(current,s_pstf) then -- post-base
startnext = getnext(start)
head = remove_node(head,start)
setlink(getprev(current),start)
setlink(start,current)
-- setlink(getprev(current),start,current) -- maybe
start = startnext
startattr = getprop(start,a_syllabe)
break
elseif not c and ( vowel_modifier[char] or stress_tone_mark[char] ) then
c = current
end
current = getnext(current)
else
if c then
startnext = getnext(start)
head = remove_node(head,start)
setlink(getprev(c),start)
setlink(start,c)
-- setlink(getprev(c),start,c) -- maybe
start = startnext
startattr = getprop(start,a_syllabe)
end
break
end
end
end
end
--
::step_5::
--
-- If no consonant is found in steps 3 or 4, move reph to a position immediately
-- before the first post-base matra, syllable modifier sign or vedic sign that has a
-- reordering class after the intended reph position. For example, if the reordering
-- position for reph is post-main, it will skip above-base matras that also have a
-- post-main position.
--
if not startnext then
current = getnext(start)
local c = nil
while current do
local char = ischar(current,startfont)
if char and getprop(current,a_syllabe) == startattr then
local state = getstate(current)
if before_subscript[rephbase] and (state == s_blwf or state == s_pstf) then
c = current
elseif after_subscript[rephbase] and (state == s_pstf) then
c = current
end
current = getnext(current)
else
break
end
end
-- here we can loose the old start node: maybe best split cases
if c then
startnext = getnext(start)
head = remove_node(head,start)
setlink(getprev(c),start)
setlink(start,c)
-- setlink(getprev(c),start,c) -- maybe
-- end
start = startnext
startattr = getprop(start,a_syllabe)
end
end
--
::step_6::
--
-- Otherwise, reorder reph to the end of the syllable.
--
if not startnext then
current = start
local next = getnext(current)
while next do
local nextchar = ischar(next,startfont)
if nextchar and getprop(next,a_syllabe) == startattr then
current = next
next = getnext(current)
else
break
end
end
if start ~= current then
startnext = getnext(start)
head = remove_node(head,start)
setlink(start,getnext(current))
setlink(current,start)
-- setlink(current,start,getnext(current)) -- maybe
start = startnext
end
end
--
return head, start, true
end
-- If a pre-base reordering consonant is found, reorder it according to the following rules:
--
-- 1 Only reorder a glyph produced by substitution during application of the feature. (Note
-- that a font may shape a Ra consonant with the feature generally but block it in certain
-- contexts.)
-- 2 Try to find a target position the same way as for pre-base matra. If it is found, reorder
-- pre-base consonant glyph.
-- 3 If position is not found, reorder immediately before main consonant.
-- Here we implement a few handlers:
--
-- function(head,start,dataset,sequence,lookupmatch,rlmode,skiphash,step)
-- return head, start, done
-- end
local reordered_pre_base_reordering_consonants = { } -- shared ? not reset ?
function handlers.devanagari_reorder_pre_base_reordering_consonants(head,start)
if reordered_pre_base_reordering_consonants[start] then
return head, start, true
end
local current = start -- we could cache attributes here
local startfont = getfont(start)
local startattr = getprop(start,a_syllabe)
while current do
local char = ischar(current,startfont)
local next = getnext(current)
if char and getprop(current,a_syllabe) == startattr then
if halant[char] then -- state can also be init
if next then
local char = ischar(next,startfont)
if char and zw_char[char] and getprop(next,a_syllabe) == startattr then
current = next
next = getnext(current)
end
end
-- can be optimzied
local startnext = getnext(start)
head = remove_node(head,start)
setlink(start,next)
setlink(current,start)
-- setlink(current,start,next) -- maybe
reordered_pre_base_reordering_consonants[start] = true
start = startnext
return head, start, true
-- elseif consonant[char] and (not getstate(current) or getstate(current,s_init)) then
-- startnext = getnext(start)
-- head = remove_node(head,start)
-- if current == head then
-- setlink(start,current)
-- head = start
-- else
-- setlink(getprev(current),start)
-- setlink(start,current)
-- end
-- start = startnext
-- break
end
else
break
end
current = next
end
local startattr = getprop(start,a_syllabe)
local current = getprev(start)
while current and getprop(current,a_syllabe) == startattr do
local char = ischar(current)
if (not dependent_vowel[char] and (not getstate(current) or getstate(current,s_init))) then
startnext = getnext(start)
head = remove_node(head,start)
if current == head then
setlink(start,current)
head = start
else
setlink(getprev(current),start)
setlink(start,current)
end
reordered_pre_base_reordering_consonants[start] = true
start = startnext
break
end
current = getprev(current)
end
return head, start, true
end
function handlers.devanagari_remove_joiners(head,start,kind,lookupname,replacement)
local stop = getnext(start)
local font = getfont(start)
local last = start
while stop do
local char = ischar(stop,font)
if char and (char == c_zwnj or char == c_zwj) then
last = stop
stop = getnext(stop)
else
break
end
end
local prev = getprev(start)
if stop then
setnext(last)
setlink(prev,stop)
elseif prev then
setnext(prev)
end
if head == start then
head = stop
end
flush_list(start)
return head, stop, true
end
local function initialize_two(font,attr)
local devanagari = fontdata[font].resources.devanagari
if devanagari then
return devanagari.seqsubset or { }, devanagari.reorderreph or { }
else
return { }, { }
end
end
-- this one will be merged into the caller: it saves a call, but we will then make function
-- of the actions
local function reorder_two(head,start,stop,font,attr,nbspaces) -- maybe do a pass over (determine stop in sweep)
local seqsubset, reorderreph = initialize_two(font,attr)
local halfpos = nil
local basepos = nil
local subpos = nil
local postpos = nil
reorderreph.coverage = { }
rephbase[font] = { }
for i=1,#seqsubset do
-- this can be done more efficient, the last test and less getnext
local subset = seqsubset[i]
local kind = subset[1]
local lookupcache = subset[2]
if kind == "rphf" then
reorderreph.coverage[subset[3]] = true -- neat
rephbase[font][subset[3]] = subset[4]
local current = start
local last = getnext(stop)
while current ~= last do
if current ~= stop then
local c = getchar(current)
local found = lookupcache[c]
if found then
local next = getnext(current)
if found[getchar(next)] or contextchain(found, next) then --above-base: rphf Consonant + Halant
local afternext = next ~= stop and getnext(next)
if afternext and zw_char[getchar(afternext)] then -- ZWJ and ZWNJ prevent creation of reph
current = afternext -- getnext(next)
elseif current == start then
setstate(current,s_rphf)
current = next
else
current = next
end
end
end
end
current = getnext(current)
end
elseif kind == "pref" then
local current = start
local last = getnext(stop)
while current ~= last do
if current ~= stop then
local c = getchar(current)
local found = lookupcache[c]
if found then -- pre-base: pref Halant + Consonant
local next = getnext(current)
if found[getchar(next)] or contextchain(found, next) then
if (not getstate(current) and not getstate(next)) then --KE: state can also be init...
setstate(current,s_pref)
setstate(next,s_pref)
current = next
end
end
end
end
current = getnext(current)
end
elseif kind == "half" then -- half forms: half / Consonant + Halant
local current = start
local last = getnext(stop)
while current ~= last do
if current ~= stop then
local c = getchar(current)
local found = lookupcache[c]
if found then
local next = getnext(current)
if found[getchar(next)] or contextchain(found, next) then
if next ~= stop and getchar(getnext(next)) == c_zwnj then -- zwnj prevent creation of half
current = next
elseif (not getstate(current)) then --KE: state can also be init...
setstate(current,s_half)
if not halfpos then
halfpos = current
end
end
current = getnext(current)
end
end
end
current = getnext(current)
end
elseif kind == "blwf" or kind == "vatu" then -- below-base: blwf / Halant + Consonant
local current = start
local last = getnext(stop)
while current ~= last do
if current ~= stop then
local c = getchar(current)
local found = lookupcache[c]
if found then
local next = getnext(current)
if found[getchar(next)] or contextchain(found, next) then
if (not getstate(current) and not getstate(next)) then --KE: state can also be init...
setstate(current,s_blwf)
setstate(next,s_blwf)
current = next
subpos = current
end
end
end
end
current = getnext(current)
end
elseif kind == "pstf" then -- post-base: pstf / Halant + Consonant
local current = start
local last = getnext(stop)
while current ~= last do
if current ~= stop then
local c = getchar(current)
local found = lookupcache[c]
if found then
local next = getnext(current)
if found[getchar(next)] or contextchain(found, next) then
if (not getstate(current) and not getstate(next)) then -- KE: state can also be init...
setstate(current,s_pstf)
setstate(next,s_pstf)
current = next
postpos = current
end
end
end
end
current = getnext(current)
end
end
end
local current, base, firstcons = start, nil, nil
if getstate(start,s_rphf) then
-- if syllable starts with Ra + H and script has 'Reph' then exclude Reph from candidates for base consonants
current = getnext(getnext(start))
end
if current ~= getnext(stop) and getchar(current) == c_nbsp then
-- Stand Alone cluster
if current == stop then
stop = getprev(stop)
head = remove_node(head,current)
flush_node(current)
return head, stop, nbspaces
else
nbspaces = nbspaces + 1
base = current
current = getnext(current)
if current ~= stop then
local char = getchar(current)
if nukta[char] then
current = getnext(current)
char = getchar(current)
end
if char == c_zwj then
local next = getnext(current)
if current ~= stop and next ~= stop and halant[getchar(next)] then
current = next
next = getnext(current)
local tmp = getnext(next)
local changestop = next == stop
setnext(next)
setstate(current,s_pref)
current = processcharacters(current,font)
setstate(current,s_blwf)
current = processcharacters(current,font)
setstate(current,s_pstf)
current = processcharacters(current,font)
setstate(current,unsetvalue)
if halant[getchar(current)] then
setnext(getnext(current),tmp)
if show_syntax_errors then
head, current = inject_syntax_error(head,current,char)
end
else
setnext(current,tmp) -- assumes that result of pref, blwf, or pstf consists of one node
if changestop then
stop = current
end
end
end
end
end
end
else -- not Stand Alone cluster
local last = getnext(stop)
while current ~= last do -- find base consonant
local next = getnext(current)
if consonant[getchar(current)] then
if not (current ~= stop and next ~= stop and halant[getchar(next)] and getchar(getnext(next)) == c_zwj) then
if not firstcons then
firstcons = current
end
-- check whether consonant has below-base or post-base form or is pre-base reordering Ra
local a = getstate(current)
if not (a == s_blwf or a == s_pstf or (a ~= s_rphf and a ~= s_blwf and ra[getchar(current)])) then
base = current
end
end
end
current = next
end
if not base then
base = firstcons
end
end
if not base then
if getstate(start,s_rphf) then
setstate(start,unsetvalue)
end
return head, stop, nbspaces
else
if getstate(base) then -- state can also be init
setstate(base,unsetvalue)
end
basepos = base
end
if not halfpos then
halfpos = base
end
if not subpos then
subpos = base
end
if not postpos then
postpos = subpos or base
end
-- Matra characters are classified and reordered by which consonant in a conjunct they have affinity for
local moved = { }
local current = start
local last = getnext(stop)
while current ~= last do
local char = getchar(current)
local target = nil
local cn = getnext(current)
-- not so efficient (needed for malayalam)
local tpm = twopart_mark[char]
while tpm do
local extra = copy_node(current)
copyinjection(extra,current)
char = tpm[1]
setchar(current,char)
setchar(extra,tpm[2])
head = insert_node_after(head,current,extra)
tpm = twopart_mark[char]
end
--
if not moved[current] and dependent_vowel[char] then
if pre_mark[char] then -- or: if before_main or before_half
moved[current] = true
-- can be helper to remove one node
local prev, next = getboth(current)
setlink(prev,next)
if current == stop then
stop = getprev(current)
end
local pos
if before_main[char] then
pos = basepos
-- basepos = current -- is this correct?
else
-- must be before_half
pos = halfpos
-- halfpos = current -- is this correct?
end
local ppos = getprev(pos) -- necessary?
while ppos and getprop(ppos,a_syllabe) == getprop(pos,a_syllabe) do
if getstate(ppos,s_pref) then
pos = ppos
end
ppos = getprev(ppos)
end
local ppos = getprev(pos) -- necessary?
while ppos and getprop(ppos,a_syllabe) == getprop(pos,a_syllabe) and halant[ischar(ppos)] do
ppos = getprev(ppos)
if ppos and getprop(ppos,a_syllabe) == getprop(pos,a_syllabe) and consonant[ischar(ppos)] then
pos = ppos
ppos = getprev(ppos)
else
break
end
end
if pos == start then
if head == start then
head = current
end
start = current
end
setlink(getprev(pos),current)
setlink(current,pos)
-- setlink(getprev(pos),current,pos) -- maybe
elseif above_mark[char] then
-- after main consonant
target = basepos
if subpos == basepos then
subpos = current
end
if postpos == basepos then
postpos = current
end
basepos = current
elseif below_mark[char] then
-- after subjoined consonants
target = subpos
if postpos == subpos then
postpos = current
end
subpos = current
elseif post_mark[char] then
-- after post-form consonant
local n = getnext(postpos) -- nukta and vedic sign come first - is that right? and also halant+ra
while n do
local v = ischar(n,font)
if nukta[v] or stress_tone_mark[v] or vowel_modifier[v] then
postpos = n
else
break
end
n = getnext(n)
end
target = postpos
postpos = current
end
if mark_above_below_post[char] then
local prev = getprev(current)
if prev ~= target then
local next = getnext(current)
setlink(prev,next)
if current == stop then
stop = prev
end
setlink(current,getnext(target))
setlink(target,current)
-- setlink(target,current,getnext(target)) -- maybe
end
end
end
current = cn
end
-- reorder halant+Ra
local current = getnext(start)
local last = getnext(stop)
while current ~= last do
local char = getchar(current)
local cn = getnext(current)
if halant[char] and ra[ischar(cn)] and (not getstate(cn,s_rphf)) and (not getstate(cn,s_blwf)) then
if after_main[ischar(cn)] then
local prev = getprev(current)
local next = getnext(cn)
local bpn = getnext(basepos)
while bpn and dependent_vowel[ischar(bpn)] do
basepos = bpn
bpn = getnext(bpn)
end
if basepos ~= prev then
setlink(prev,next)
setlink(cn, getnext(basepos))
setlink(basepos, current)
if cn == stop then
stop = prev
end
cn = next
end
end
-- after_postscript
-- after_subscript
-- before_postscript
-- before_subscript
end
current = cn
end
-- Reorder marks to canonical order: Adjacent nukta and halant or nukta and vedic sign are always repositioned if necessary, so that the nukta is first.
local current = start
local c = nil
while current ~= stop do
local char = getchar(current)
if halant[char] or stress_tone_mark[char] then
if not c then
c = current
end
else
c = nil
end
local next = getnext(current)
if c and nukta[getchar(next)] then
if head == c then
head = next
end
if stop == next then
stop = current
end
setlink(getprev(c),next)
local nextnext = getnext(next)
setnext(current,nextnext)
local nextnextnext = getnext(nextnext)
if nextnextnext then
setprev(nextnextnext,current)
end
setlink(nextnext,c)
end
if stop == current then break end
current = getnext(current)
end
if getchar(base) == c_nbsp then
if base == stop then
stop = getprev(stop)
end
nbspaces = nbspaces - 1
head = remove_node(head, base)
flush_node(base)
end
return head, stop, nbspaces
end
-- cleaned up and optimized ... needs checking (local, check order, fixes, extra hash, etc)
local separator = { }
imerge(separator,consonant)
imerge(separator,independent_vowel)
imerge(separator,dependent_vowel)
imerge(separator,vowel_modifier)
imerge(separator,stress_tone_mark)
for k, v in next, nukta do separator[k] = true end
for k, v in next, halant do separator[k] = true end
local function analyze_next_chars_one(c,font,variant) -- skip one dependent vowel
-- why two variants ... the comment suggests that it's the same ruleset
local n = getnext(c)
if not n then
return c
end
if variant == 1 then
local v = ischar(n,font)
if v and nukta[v] then
n = getnext(n)
if n then
v = ischar(n,font)
end
end
if n and v then
local nn = getnext(n)
if nn then
local vv = ischar(nn,font)
if vv then
local nnn = getnext(nn)
if nnn then
local vvv = ischar(nnn,font)
if vvv then
if vv == c_zwj and consonant[vvv] then
c = nnn
elseif (vv == c_zwnj or vv == c_zwj) and halant[vvv] then
local nnnn = getnext(nnn)
if nnnn then
local vvvv = ischar(nnnn,font)
if vvvv and consonant[vvvv] then
c = nnnn
end
end
end
end
end
end
end
end
elseif variant == 2 then
local v = ischar(n,font)
if v and nukta[v] then
c = n
end
n = getnext(c)
if n then
v = ischar(n,font)
if v then
local nn = getnext(n)
if nn then
local vv = ischar(nn,font)
if vv and zw_char[v] then
n = nn
v = vv
nn = getnext(nn)
vv = nn and ischar(nn,font)
end
if vv and halant[v] and consonant[vv] then
c = nn
end
end
end
end
end
-- c = ms_matra(c)
local n = getnext(c)
if not n then
return c
end
local v = ischar(n,font)
if not v then
return c
end
local already_pre_mark -- = false
local already_above_mark -- = false
local already_below_mark -- = false
local already_post_mark -- = false
while dependent_vowel[v] do
local vowels = twopart_mark[v] or { v }
for k, v in next, vowels do
if pre_mark[v] and not already_pre_mark then
already_pre_mark = true
elseif above_mark[v] and not already_above_mark then
already_above_mark = true
elseif below_mark[v] and not already_below_mark then
already_below_mark = true
elseif post_mark[v] and not already_post_mark then
already_post_mark = true
else
return c
end
end
c = getnext(c)
n = getnext(c)
if not n then
return c
end
v = ischar(n,font)
if not v then
return c
end
end
if nukta[v] then
c = getnext(c)
n = getnext(c)
if not n then
return c
end
v = ischar(n,font)
if not v then
return c
end
end
if halant[v] then
c = getnext(c)
n = getnext(c)
if not n then
return c
end
v = ischar(n,font)
if not v then
return c
end
end
if vowel_modifier[v] then
c = getnext(c)
n = getnext(c)
if not n then
return c
end
v = ischar(n,font)
if not v then
return c
end
end
if stress_tone_mark[v] then
c = getnext(c)
n = getnext(c)
if not n then
return c
end
v = ischar(n,font)
if not v then
return c
end
end
if stress_tone_mark[v] then
return n
else
return c
end
end
local function analyze_next_chars_two(c,font)
local n = getnext(c)
if not n then
return c
end
local v = ischar(n,font)
if v and nukta[v] then
c = n
end
n = c
while true do
local nn = getnext(n)
if nn then
local vv = ischar(nn,font)
if vv then
if halant[vv] then
n = nn
local nnn = getnext(nn)
if nnn then
local vvv = ischar(nnn,font)
if vvv and zw_char[vvv] then
n = nnn
end
end
elseif vv == c_zwnj or vv == c_zwj then
-- n = nn -- not here (?)
local nnn = getnext(nn)
if nnn then
local vvv = ischar(nnn,font)
if vvv and halant[vvv] then
n = nnn
end
end
else
break
end
local nn = getnext(n)
if nn then
local vv = ischar(nn,font)
if vv and consonant[vv] then
n = nn
local nnn = getnext(nn)
if nnn then
local vvv = ischar(nnn,font)
if vvv and nukta[vvv] then
n = nnn
end
end
c = n
else
break
end
else
break
end
else
break
end
else
break
end
end
--
if not c then
-- This shouldn't happen I guess.
return
end
local n = getnext(c)
if not n then
return c
end
local v = ischar(n,font)
if not v then
return c
end
if anudatta[v] then
c = n
n = getnext(c)
if not n then
return c
end
v = ischar(n,font)
if not v then
return c
end
end
if halant[v] then
c = n
n = getnext(c)
if not n then
return c
end
v = ischar(n,font)
if not v then
return c
end
if v == c_zwnj or v == c_zwj then
c = n
n = getnext(c)
if not n then
return c
end
v = ischar(n,font)
if not v then
return c
end
end
else
-- c = ms_matra(c)
-- same as one
local already_pre_mark -- = false
local already_above_mark -- = false
local already_below_mark -- = false
local already_post_mark -- = false
while dependent_vowel[v] do
local vowels = twopart_mark[v] or { v }
for k, v in next, vowels do
if pre_mark[v] and not already_pre_mark then
already_pre_mark = true
elseif above_mark[v] and not already_above_mark then
already_above_mark = true
elseif below_mark[v] and not already_below_mark then
already_below_mark = true
elseif post_mark[v] and not already_post_mark then
already_post_mark = true
else
return c
end
end
c = n
n = getnext(c)
if not n then
return c
end
v = ischar(n,font)
if not v then
return c
end
end
if nukta[v] then
c = n
n = getnext(c)
if not n then
return c
end
v = ischar(n,font)
if not v then
return c
end
end
if halant[v] then
c = n
n = getnext(c)
if not n then
return c
end
v = ischar(n,font)
if not v then
return c
end
end
end
-- same as one
if vowel_modifier[v] then
c = n
n = getnext(c)
if not n then
return c
end
v = ischar(n,font)
if not v then
return c
end
end
if stress_tone_mark[v] then
c = n
n = getnext(c)
if not n then
return c
end
v = ischar(n,font)
if not v then
return c
end
end
if stress_tone_mark[v] then
return n
else
return c
end
end
-- It looks like these two analyzers were written independently but they share
-- a lot. Common code has been synced.
local function method_one(head,font,attr)
local current = head
local start = true
local done = false
local nbspaces = 0
local syllabe = 0
while current do
local char = ischar(current,font)
if char then
done = true
local syllablestart = current
local syllableend = nil
local c = current
local n = getnext(c)
local first = char
if n and ra[first] then
local second = ischar(n,font)
if second and halant[second] then
local n = getnext(n)
if n then
local third = ischar(n,font)
if third then
c = n
first = third
end
end
end
end
local standalone = first == c_nbsp
if standalone then
local prev = getprev(current)
if prev then
local prevchar = ischar(prev,font)
if not prevchar then
-- different font or language so quite certainly a different word
elseif not separator[prevchar] then
-- something that separates words
else
standalone = false
end
else
-- begin of paragraph or box
end
end
if standalone then
-- stand alone cluster (at the start of the word only): #[Ra+H]+NBSP+[N]+[<[<ZWJ|ZWNJ>]+H+C>]+[{M}+[N]+[H]]+[SM]+[(VD)]
local syllableend = analyze_next_chars_one(c,font,2)
current = getnext(syllableend)
if syllablestart ~= syllableend then
head, current, nbspaces = reorder_one(head,syllablestart,syllableend,font,attr,nbspaces)
current = getnext(current)
end
else
-- we can delay the getsubtype(n) and getfont(n) and test for say halant first
-- as an table access is faster than two function calls (subtype and font are
-- pseudo fields) but the code becomes messy (unless we make it a function)
if consonant[char] then
-- syllable containing consonant
local prevc = true
while prevc do
prevc = false
local n = getnext(current)
if not n then
break
end
local v = ischar(n,font)
if not v then
break
end
if nukta[v] then
n = getnext(n)
if not n then
break
end
v = ischar(n,font)
if not v then
break
end
end
if halant[v] then
n = getnext(n)
if not n then
break
end
v = ischar(n,font)
if not v then
break
end
if v == c_zwnj or v == c_zwj then
n = getnext(n)
if not n then
break
end
v = ischar(n,font)
if not v then
break
end
end
if consonant[v] then
prevc = true
current = n
end
end
end
local n = getnext(current)
if n then
local v = ischar(n,font)
if v and nukta[v] then
-- nukta (not specified in Microsft Devanagari OpenType specification)
current = n
n = getnext(current)
end
end
syllableend = current
current = n
if current then
local v = ischar(current,font)
if not v then
-- skip
elseif halant[v] then
-- syllable containing consonant without vowels: {C + [Nukta] + H} + C + H
local n = getnext(current)
if n then
local v = ischar(n,font)
if v and zw_char[v] then
-- code collapsed, probably needs checking with intention
syllableend = n
current = getnext(n)
else
syllableend = current
current = n
end
else
syllableend = current
current = n
end
else
-- syllable containing consonant with vowels: {C + [Nukta] + H} + C + [M] + [VM] + [SM]
if dependent_vowel[v] then
syllableend = current
current = getnext(current)
v = ischar(current,font)
end
if v and vowel_modifier[v] then
syllableend = current
current = getnext(current)
v = ischar(current,font)
end
if v and stress_tone_mark[v] then
syllableend = current
current = getnext(current)
end
end
end
if syllablestart ~= syllableend then
if syllableend then
syllabe = syllabe + 1
local c = syllablestart
local n = getnext(syllableend)
while c ~= n do
setprop(c,a_syllabe,syllabe)
c = getnext(c)
end
end
head, current, nbspaces = reorder_one(head,syllablestart,syllableend,font,attr,nbspaces)
current = getnext(current)
end
elseif independent_vowel[char] then
-- syllable without consonants: VO + [VM] + [SM]
syllableend = current
current = getnext(current)
if current then
local v = ischar(current,font)
if v then
if vowel_modifier[v] then
syllableend = current
current = getnext(current)
v = ischar(current,font)
end
if v and stress_tone_mark[v] then
syllableend = current
current = getnext(current)
end
end
end
else
if show_syntax_errors then
local mark = mark_four[char]
if mark then
head, current = inject_syntax_error(head,current,char)
end
end
current = getnext(current)
end
end
else
current = getnext(current)
end
start = false
end
if nbspaces > 0 then
head = replace_all_nbsp(head)
end
current = head
local n = 0
while current do
local char = ischar(current,font)
if char then
if n == 0 and not getstate(current) then
setstate(current,s_init)
end
n = n + 1
else
n = 0
end
current = getnext(current)
end
return head, done
end
-- there is a good change that when we run into one with subtype < 256 that the rest is also done
-- so maybe we can omit this check (it's pretty hard to get glyphs in the stream out of the blue)
local function method_two(head,font,attr)
local current = head
local start = true
local done = false
local syllabe = 0
local nbspaces = 0
while current do
local syllablestart = nil
local syllableend = nil
local char = ischar(current,font)
if char then
done = true
syllablestart = current
local c = current
local n = getnext(current)
if n and ra[char] then
local nextchar = ischar(n,font)
if nextchar and halant[nextchar] then
local n = getnext(n)
if n then
local nextnextchar = ischar(n,font)
if nextnextchar then
c = n
char = nextnextchar
end
end
end
end
if independent_vowel[char] then
-- vowel-based syllable: [Ra+H]+V+[N]+[<[<ZWJ|ZWNJ>]+H+C|ZWJ+C>]+[{M}+[N]+[H]]+[SM]+[(VD)]
current = analyze_next_chars_one(c,font,1)
syllableend = current
else
local standalone = char == c_nbsp
if standalone then
nbspaces = nbspaces + 1
local p = getprev(current)
if not p then
-- begin of paragraph or box
elseif ischar(p,font) then
-- different font or language so quite certainly a different word
elseif not separator[getchar(p)] then
-- something that separates words
else
standalone = false
end
end
if standalone then
-- Stand Alone cluster (at the start of the word only): #[Ra+H]+NBSP+[N]+[<[<ZWJ|ZWNJ>]+H+C>]+[{M}+[N]+[H]]+[SM]+[(VD)]
current = analyze_next_chars_one(c,font,2)
syllableend = current
elseif consonant[getchar(current)] then
-- WHY current INSTEAD OF c ?
-- Consonant syllable: {C+[N]+<H+[<ZWNJ|ZWJ>]|<ZWNJ|ZWJ>+H>} + C+[N]+[A] + [< H+[<ZWNJ|ZWJ>] | {M}+[N]+[H]>]+[SM]+[(VD)]
current = analyze_next_chars_two(current,font) -- not c !
syllableend = current
end
end
end
if syllableend then
syllabe = syllabe + 1
local c = syllablestart
local n = getnext(syllableend)
while c ~= n do
setprop(c,a_syllabe,syllabe)
c = getnext(c)
end
end
if syllableend and syllablestart ~= syllableend then
head, current, nbspaces = reorder_two(head,syllablestart,syllableend,font,attr,nbspaces)
end
if not syllableend and show_syntax_errors then
local char = ischar(current,font)
if char and not getstate(current) then -- state can also be init
local mark = mark_four[char]
if mark then
head, current = inject_syntax_error(head,current,char)
end
end
end
start = false
current = getnext(current)
end
if nbspaces > 0 then
head = replace_all_nbsp(head)
end
current = head
local n = 0
while current do
local char = ischar(current,font)
if char then
if n == 0 and not getstate(current) then -- state can also be init
setstate(current,s_init)
end
n = n + 1
else
n = 0
end
current = getnext(current)
end
return head, done
end
for i=1,nofscripts do
methods[scripts_one[i]] = method_one
methods[scripts_two[i]] = method_two
end