+ 2011-04-10 Paul Eggert <eggert@cs.ucla.edu>
+
++ Fix more problems found by GCC 4.6.0's static checks.
++
+ * xdisp.c (vmessage): Use a better test for character truncation.
+
+ * charset.c (load_charset_map): <, not <=, for optimization,
+ and to avoid potential problems with integer overflow.
+ * chartab.c (sub_char_table_set_range, char_table_set_range): Likewise.
+ * casetab.c (set_identity, shuffle): Likewise.
+ * editfns.c (Fformat): Likewise.
+ * syntax.c (skip_chars): Likewise.
+
+ * xmenu.c (set_frame_menubar): Allocate smaller local vectors.
+ This also lets GCC 4.6.0 generate slightly better loop code.
+
+ * callint.c (Fcall_interactively): <, not <=, for optimization.
+ (Fcall_interactively): Count the number of arguments produced,
+ not the number of arguments given. This is simpler and lets GCC
+ 4.6.0 generate slightly better code.
+
+ * ftfont.c: Distingish more carefully between FcChar8 and char.
+ The previous code passed unsigned char * to a functions like
+ strlen and xstrcasecmp that expect char *, which does not
+ conform to the C standard.
+ (get_adstyle_property, ftfont_pattern_entity): Use FcChar8 for
+ arguments to FcPatternGetString, and explicitly cast FcChar8 * to
+ char * when the C standard requires it.
+
+ * keyboard.c (read_char): Remove unused var.
+
+ * eval.c: Port to Windows vsnprintf (Bug#8435).
+ Include <limits.h>.
+ (SIZE_MAX): Define if the headers do not.
+ (verror): Do not give up if vsnprintf returns a negative count.
+ Instead, grow the buffer. This ports to Windows vsnprintf, which
+ does not conform to C99. Problem reported by Eli Zaretskii.
+ Also, simplify the allocation scheme, by avoiding the need for
+ calling realloc, and removing the ALLOCATED variable.
+
+ * eval.c (verror): Initial buffer size is 4000 (not 200) bytes.
+
+ Remove invocations of doprnt, as Emacs now uses vsnprintf.
+ But keep the doprint source code for now, as we might revamp it
+ and use it again (Bug#8435).
+ * lisp.h (doprnt): Remove.
+ * Makefile.in (base_obj): Remove doprnt.o.
+ * deps.mk (doprnt.o): Remove.
+
+ error: Print 32- and 64-bit integers portably (Bug#8435).
+ Without this change, on typical 64-bit hosts error ("...%d...", N)
+ was used to print both 32- and 64-bit integers N, which relied on
+ undefined behavior.
+ * lisp.h, src/m/amdx86-64.h, src/m/ia64.h, src/m/ibms390x.h (pEd):
+ New macro.
+ * lisp.h (error, verror): Mark as printf-like functions.
+ * eval.c (verror): Use vsnprintf, not doprnt, to do the real work.
+ Report overflow in size calculations when allocating printf buffer.
+ Do not truncate output string at its first null byte.
+ * xdisp.c (vmessage): Use vsnprintf, not doprnt, to do the real work.
+ Truncate the output at a character boundary, since vsnprintf does not
+ do that.
+ * charset.c (check_iso_charset_parameter): Convert internal
+ character to string before calling 'error', since %c now has the
+ printf meaning.
+ * coding.c (Fdecode_sjis_char, Fdecode_big5_char): Avoid int
+ overflow when computing char to be passed to 'error'. Do not
+ pass Lisp_Object to 'error'; pass the integer instead.
+ * nsfns.m (Fns_do_applescript): Use int, not long, since it's
+ formatted with plain %d.
+
+ * eval.c (internal_lisp_condition_case): Don't pass spurious arg.
+
+ * keyboard.c (access_keymap_keyremap): Print func name, not garbage.
+
+ * coding.c (Fdecode_sjis_char): Don't assume CODE fits in int.
+
+ * xterm.c (x_catch_errors): Remove duplicate declaration.
+
+ * term.c (maybe_fatal): Mark its 3rd arg as a printf format, too.
+
+ * xdisp.c, lisp.h (message_nolog): Remove; unused.
+
2011-04-10 Jim Meyering <meyering@redhat.com>
use ssize_t and size_t for read- and write-like emacs_gnutls_* functions