fix compile errors, fix revert-buffer bug on binary/Latin 1 files, Mule-ize some files -------------------- ChangeLog entries follow: -------------------- src/ChangeLog addition: 2010-02-07 Ben Wing <ben@xemacs.org> * emacs.c (make_argc_argv): Cast to Wexttext * to fix compile error. * nt.c (init_user_info): Cast to CIbyte * to fix compile error. * nt.c (open_unc_volume): To fix compile error, store pathname into an Extbyte * variable then write into the dest, casting to LPTSTR. * emacs.c (debug_can_access_memory): * fileio.c (Fmake_temp_name): * fileio.c (a_write): * fns.c: * fns.c (check_losing_bytecode): * fns.c (plists_differ): * fns.c (internal_equal_trapping_problems): * fns.c (base64_value_to_char): * fns.c (base64_conversion_error): * fns.c (STORE_BYTE): * fns.c (vars_of_fns): * lisp.h: * nt.c (init_user_info): * nt.c (mswindows_readdir): * nt.c (mswindows_executable_type): Replace raw `char *' or `unsigned char *' with characterized type -- Rawbyte, Binbyte, Boolbyte, Ibyte or Ascbyte. This should fix at least one real bug -- in a_write(), the "speedy insert" code that checks for an unchanged region declared the file data read in as char[] but then compared the value to an Ichar. Hence, any chars in the range 128-255 would always appear changed -- in particular, this algorithm would fail completely with binary data.

--- a/src/lisp.h	Sat Feb 06 04:27:47 2010 -0600
+++ b/src/lisp.h	Sun Feb 07 05:21:40 2010 -0600
@@ -5292,7 +5292,7 @@
 int internal_equal_0 (Lisp_Object, Lisp_Object, int, int);
 Lisp_Object bytecode_nconc2 (Lisp_Object *);
 int bytecode_arithcompare (Lisp_Object obj1, Lisp_Object obj2);
-void check_losing_bytecode (const char *, Lisp_Object);
+void check_losing_bytecode (const Ascbyte *, Lisp_Object);
 
 Lisp_Object add_suffix_to_symbol (Lisp_Object symbol,
 				  const Ascbyte *ascii_string);