From 5f639e1d63bbb39eb0e9cb6d4aca28aff1ac21e3 Mon Sep 17 00:00:00 2001
From: Matthias Kramm <kramm@quiss.org>
Date: Sat, 12 Dec 2009 19:03:19 -0800
Subject: [PATCH] fixed utf8 handling

---
 lib/as3/Makefile      |   10 ++++++++--
 lib/as3/main.c        |    4 +++-
 lib/as3/tokenizer.lex |   21 +++++++++++++++++----
 3 files changed, 28 insertions(+), 7 deletions(-)

diff --git a/lib/as3/Makefile b/lib/as3/Makefile
index be18901..b3ecd23 100644
--- a/lib/as3/Makefile
+++ b/lib/as3/Makefile
@@ -5,17 +5,23 @@ tests: testwrite testrewrite testpaths testreadwrite
 
 D=-g -pg
 
+#BISONDEBUG=yes
+
 MODULES = abc.o opcodes.o code.o pool.o scripts.o expr.o common.o initcode.o
 SOURCES = abc.c abc.h pool.c pool.h files.c files.h code.c code.h registry.c registry.h opcodes.c opcodes.h builtin.c builtin.h compiler.c compiler.h parser.tab.h parser.tab.c tokenizer.yy.c scripts.c import.c import.h expr.c expr.h common.c common.h initcode.c initcode.h
 
 tokenizer.yy.c: tokenizer.lex tokenizer.h
 	flex -Pas3_ -8 -B -otokenizer.yy.c tokenizer.lex
 
+ifeq "$(BISONDEBUG)" "yes"
+BISONDEBUGFLAGS=-t
+BISONDEBUGDEFINE=-DBISONDEBUG
+endif
 parser.tab.h parser.tab.c: parser.y parser.h skeleton.m4 Makefile
-	bison -S ./skeleton.m4 -v --defines -pa3_ parser.y -o parser.tab.c
+	bison $(BISONDEBUGFLAGS) -S ./skeleton.m4 -v --defines -pa3_ parser.y -o parser.tab.c
 
 main.o: main.c parser.tab.h parser.h
-	$(C) main.c -o main.o
+	$(C) $(BISONDEBUGDEFINE) main.c -o main.o
 
 mklib.o: mklib.c parser.tab.h parser.h
 	$(C) mklib.c -o mklib.o
diff --git a/lib/as3/main.c b/lib/as3/main.c
index 2835556..85b36a6 100644
--- a/lib/as3/main.c
+++ b/lib/as3/main.c
@@ -64,7 +64,9 @@ int main(int argn, char*argv[])
     }
     filename=argv[argn-1];
 
-    //a3_debug = 1; //if bison was called with -t
+#ifdef BISONDEBUG
+    a3_debug = 1; //if bison was called with -t
+#endif
    
     as3_add_include_dir(getcwd(buf, 512));
 
diff --git a/lib/as3/tokenizer.lex b/lib/as3/tokenizer.lex
index 4e0495d..05fd407 100644
--- a/lib/as3/tokenizer.lex
+++ b/lib/as3/tokenizer.lex
@@ -481,6 +481,7 @@ void tokenizer_unregister_namespace(const char*id)
 }*/
 static inline char tokenizer_is_namespace(const char*id)
 {
+    if(!active_namespaces) return 0;
     return trie_contains(active_namespaces, (const unsigned char*)id);
 }
 
@@ -514,8 +515,20 @@ static int tokenerror();
 %x XMLTEXT
 %x XML
 
-NAME	 [a-zA-Z_\x80-\xff][a-zA-Z0-9_\\\x80-\xff]*
-_        [^a-zA-Z0-9_\\\x80-\xff]
+X1 parsing identifiers with a non unicode lexer is a knightmare we have to skip all possible
+X2 combinations of byte order markers or utf8 space chars and i dont quite like the fact that
+X3 lex doesnt support proper comments in this section either...
+X4 {NAME_HEAD}{NAME_TAIL} 
+
+NAME_NOC2EF  [a-zA-Z_\x80-\xc1\xc3-\xee\xf0-\xff]
+NAME_EF      [\xef][a-zA-Z0-9_\\\x80-\xba\xbc-\xff]
+NAME_C2      [\xc2][a-zA-Z0-9_\\\x80-\x9f\xa1-\xff]
+NAME_EFBB    [\xef][\xbb][a-zA-Z0-9_\\\x80-\xbe\xc0-\xff]
+NAME_TAIL    [a-zA-Z_0-9\\\x80-\xff]*
+NAME_HEAD    (({NAME_NOC2EF})|({NAME_EF})|({NAME_C2})|({NAME_EFBB}))
+NAME	     {NAME_HEAD}{NAME_TAIL} 
+
+_            [^a-zA-Z0-9_\\\x80-\xff]
 
 HEXINT    0x[a-zA-Z0-9]+
 HEXFLOAT  0x[a-zA-Z0-9]*\.[a-zA-Z0-9]*
@@ -534,7 +547,7 @@ XMLID       [A-Za-z0-9_\x80-\xff]+([:][A-Za-z0-9_\x80-\xff]+)?
 XMLSTRING   ["][^"]*["]
 
 STRING   ["](\\[\x00-\xff]|[^\\"\n])*["]|['](\\[\x00-\xff]|[^\\'\n])*[']
-S 	 ([ \n\r\t\xa0]|\xc2\xa0)
+S 	 ([ \n\r\t\xa0]|[\xc2][\xa0])
 MULTILINE_COMMENT [/][*]+([*][^/]|[^/*]|[^*][/]|[\x00-\x1f])*[*]+[/]
 SINGLELINE_COMMENT \/\/[^\n\r]*[\n\r]
 REGEXP   [/]([^/\n]|\\[/])*[/][a-zA-Z]*
@@ -587,7 +600,7 @@ REGEXP   [/]([^/\n]|\\[/])*[/][a-zA-Z]*
 <REGEXPOK>[\{]               {c(); BEGIN(REGEXPOK);return m(T_DICTSTART);}
 [\{]                         {c(); BEGIN(DEFAULT); return m('{');}
 
-\xef\xbb\xbf                 {/* utf 8 bom */}
+\xef\xbb\xbf                 {/* utf 8 bom (0xfeff) */}
 {S}                          {l();}
 
 {HEXINT}/{_}                 {c(); BEGIN(DEFAULT);return handlehex();}
-- 
1.7.10.4