utf character handling; optimized iso9 parser code

1 files changed, 17 insertions, 1 deletions

diff --git a/tex/context/third/transliterator/transliterator.lua b/tex/context/third/transliterator/transliterator.lua
index 110d862..36eb804 100644
--- a/tex/context/third/transliterator/transliterator.lua
+++ b/tex/context/third/transliterator/transliterator.lua
@@ -93,6 +93,22 @@ do
     end
 end
 
+-- Modified version of Hans’s utf pattern (l-lpeg.lua).
+
+do
+    local P, R, V = lpeg.P, lpeg.R, lpeg.V
+
+    translit.utfchar = P{
+        V"utf8one" + V"utf8two" + V"utf8three" + V"utf8four",
+
+        utf8next  = R("\128\191"),
+        utf8one   = R("\000\127"),
+        utf8two   = R("\194\223") * V"utf8next",
+        utf8three = R("\224\239") * V"utf8next" * V"utf8next",
+        utf8four  = R("\240\244") * V"utf8next" * V"utf8next" * V"utf8next",
+    }
+end
+
 -- We might want to have all the table data nicely formatted by \CONTEXT\ 
 -- itself, here's how we'll do it.  \type{translit.show_tab(t)} handles a
 -- single table \type{t}, builds a Natural TABLE out of its content and
@@ -220,7 +236,7 @@ function translit.transliterate (method, text)
     elseif  method == "ru_transcript_de"      or
             method == "ru_transcript_de_exp"  or -- experimental lpeg
             method == "ru_transcript_en"      or
-            method == "ru_transcript_en_sub"  or -- old multiple substitution
+            method == "ru_transcript_en_exp"  or
             method == "ru_cz"                 or
             method == "ocs_cz"
             then