lua-check-hyphen (2apr16)

git-svn-id: svn://tug.org/texlive/trunk@40211 c570f23f-e606-0410-a88d-b1316a301751

2 files changed, 87 insertions, 13 deletions

diff --git a/Master/texmf-dist/tex/lualatex/lua-check-hyphen/lua-check-hyphen.lua b/Master/texmf-dist/tex/lualatex/lua-check-hyphen/lua-check-hyphen.lua
index e0eee9b88cf..148699382d3 100644
--- a/Master/texmf-dist/tex/lualatex/lua-check-hyphen/lua-check-hyphen.lua
+++ b/Master/texmf-dist/tex/lualatex/lua-check-hyphen/lua-check-hyphen.lua
@@ -1,13 +1,13 @@
--- Copyright 2012 Patrick Gundlach (patrick@gundla.ch)
+-- Copyright 2012-2016 Patrick Gundlach (patrick@gundla.ch)
 -- Public repository: https://github.com/pgundlach/lua-check-hyphen
 -- Licensed under the MIT license. See the file 'mit-license.txt' for exact terms.
 
--- Version: 0.3
+-- Version: 0.4
 
 
 -- for debugging purpuse:
 -- function w( ... )
--- 	texio.write_nl(string.format(...))
+-- 	texio.write_nl("--->" .. string.format(...))
 -- end
 
 local explode = function(s,p)
@@ -38,8 +38,9 @@ local a_whatsit_node = node.id("whatsit")
 local subtype_rightskip = 9
 
 
+local sln = unicode.utf8
+
 luacheckhyphen.deligature = function ( glyph_node )
-	local sln = unicode.utf8
 	local head = glyph_node.components
 	local str = ""
 	while head do
@@ -56,6 +57,12 @@ luacheckhyphen.deligature = function ( glyph_node )
 end
 
 
+-- This functions analyzes the list beginning at head. If it encounters a box,
+-- it recurses into the box. If it finds a disc node, it goes back until it finds the word
+-- start. Then it analyzes the word und finds all hyphenation points. For example the German
+-- word "Salpetersäure" has these disc nodes: , "Sal-petersäure", "Salpe-tersäure",
+-- "Salpeter-säure", and "Salpetersäu-re". Each of these "word with hyphen" gets stored in
+-- the hash hyphenwords
 luacheckhyphen.collect_discs = function(head)
 	local word_start
 	local word
@@ -95,6 +102,7 @@ luacheckhyphen.collect_discs = function(head)
 		end
 		if thisbreakpoint then
 			word_with_hyphen = sln.sub(word,1,thisbreakpoint) .. "-" .. sln.sub(word,thisbreakpoint+1,-1)
+			-- word with hyphen has all possible hyphenation points
 			luacheckhyphen.hyphenwords[hyphencounter] = word_with_hyphen
 			hyphencounter = #luacheckhyphen.hyphenwords + 1
 		end
@@ -104,6 +112,12 @@ luacheckhyphen.collect_discs = function(head)
 	return true
 end
 
+-- Remove '-' from word
+local function removedash( word )
+	local ret = sln.gsub(word,"-","")
+	return ret
+end
+
 luacheckhyphen.check_discs = function (head,parent)
 	local c
 	local word
@@ -119,7 +133,14 @@ luacheckhyphen.check_discs = function (head,parent)
 		if luacheckhyphen.word_whitelist[word] then
 			-- word found, but OK (whitelisted)
 		else
-			luacheckhyphen.all_hyphenatedwords[word] = true
+			if luachekchyphen.compact == nil or luachekchyphen.compact == "true" then
+				local word_without_hyphen = sln.lower(removedash(word))
+				local tmp = luacheckhyphen.all_hyphenatedwords[word_without_hyphen] or {}
+				tmp[word] = true
+				luacheckhyphen.all_hyphenatedwords[word_without_hyphen] = tmp
+			else
+				luacheckhyphen.all_hyphenatedwords[word] = true
+			end
 			if luacheckhyphen.drawmarks then
 				tmp = node.copy(luacheckhyphen.rectangle)
 				node.insert_after(parent,head,tmp)
@@ -131,6 +152,21 @@ luacheckhyphen.check_discs = function (head,parent)
 	return true
 end
 
+-- http://www.lua.org/pil/19.3.html
+local function pairsByKeys (t)
+	local a = {}
+  	for n in pairs(t) do table.insert(a, n) end
+  	table.sort(a)
+  	local i = 0      -- iterator variable
+  	local iter = function ()   -- iterator function
+  		i = i + 1
+  		if a[i] == nil then return nil
+  		else return a[i], t[a[i]]
+  		end
+  	end
+  return iter
+end
+
 luacheckhyphen.listhyphenatedwords = function()
 	if luacheckhyphen.final == "true" then
 		return
@@ -139,9 +175,31 @@ luacheckhyphen.listhyphenatedwords = function()
 	if not luacheckhyphen.nofile then
 		local unknown_hyphenation_filename = tex.jobname .. ".uhy"
 		local unknown_hyphenation_file = io.open(unknown_hyphenation_filename,"w")
-
-		for k,v in pairs(luacheckhyphen.all_hyphenatedwords) do
-			unknown_hyphenation_file:write(k .. "\n")
+		for k,v in pairsByKeys(luacheckhyphen.all_hyphenatedwords) do
+			if luachekchyphen.compact == "true" or luachekchyphen.compact == nil then
+				local hyphenationlist = {}
+				local hyphenpos = {}
+				for l,_ in pairs(v) do
+					local tmp = string.find(l,"-")
+					if tmp then
+						hyphenpos[#hyphenpos + 1] = tmp
+					end
+					hyphenationlist[#hyphenationlist + 1] = l
+				end
+				table.sort(hyphenpos)
+				local word_with_all_hyphenationpoints = {}
+				local cur = 1
+				for i=1,string.len(k) do
+					if hyphenpos[cur] == i then
+						word_with_all_hyphenationpoints[#word_with_all_hyphenationpoints + 1] =  "-"
+						cur = cur + 1
+					end
+					word_with_all_hyphenationpoints[#word_with_all_hyphenationpoints + 1]  = string.sub(k,i, i)
+				end
+				unknown_hyphenation_file:write(table.concat(word_with_all_hyphenationpoints,"") .. "\n")
+			else
+				unknown_hyphenation_file:write(k .. "\n")
+			end
 		end
 		unknown_hyphenation_file:close()
 	end
@@ -166,7 +224,22 @@ luacheckhyphen.enable = function()
 			else
 				filecontents = whitelistfile:read("*a")
 				for _,entry in ipairs(explode(filecontents,"[^%s]+")) do
-					luacheckhyphen.word_whitelist[entry] = true
+					parts = string.explode(entry,"-")
+					if #parts > 2 then
+						local c = 1
+						for c=1,#parts - 1 do
+							local word = {}
+							for i=1,#parts do
+								word[#word + 1] = parts[i]
+								if i == c then
+									word[#word + 1] = "-"
+								end
+							end
+							luacheckhyphen.word_whitelist[table.concat(word,"")] = true
+						end
+					else
+						luacheckhyphen.word_whitelist[entry] = true
+					end
 				end
 			end
 		end
diff --git a/Master/texmf-dist/tex/lualatex/lua-check-hyphen/lua-check-hyphen.sty b/Master/texmf-dist/tex/lualatex/lua-check-hyphen/lua-check-hyphen.sty
index b919bd3b654..d38ebeab8cb 100644
--- a/Master/texmf-dist/tex/lualatex/lua-check-hyphen/lua-check-hyphen.sty
+++ b/Master/texmf-dist/tex/lualatex/lua-check-hyphen/lua-check-hyphen.sty
@@ -1,9 +1,9 @@
-% Copyright 2012 Patrick Gundlach (patrick@gundla.ch)  
-% Public repository: https://github.com/pgundlach/lua-check-hyphen  
+% Copyright 2012 Patrick Gundlach (patrick@gundla.ch)
+% Public repository: https://github.com/pgundlach/lua-check-hyphen
 % Licensed under the MIT license. See the file 'mit-license.txt' for exact terms.
 
-\def\luachekchyphenpkgdate{2014/07/30}
-\def\luachekchyphenversion{0.3}
+\def\luachekchyphenpkgdate{2016/04/02}
+\def\luachekchyphenversion{0.4}
 
 
 \RequirePackage{ifluatex}
@@ -23,6 +23,7 @@
 \def\luachekchyphendothings{\directlua{luachekchyphen.enable() }}
 
 \define@key{luacheckhyphen}{nofile}[true]{\directlua{luachekchyphen.nofile  = "\luatexluaescapestring{#1}"}}
+\define@key{luacheckhyphen}{compact}[true]{\directlua{luachekchyphen.compact  = "\luatexluaescapestring{#1}"}}
 \define@key{luacheckhyphen}{final}[true]{\directlua{luachekchyphen.final  = "\luatexluaescapestring{#1}"}}
 \define@key{luacheckhyphen}{mark}[true]{\directlua{luachekchyphen.mark     = "\luatexluaescapestring{#1}"}}
 \define@key{luacheckhyphen}{whitelist}{\directlua{luachekchyphen.whitelist = "\luatexluaescapestring{#1}"}}