-2011-06-02 Paul Eggert <eggert@cs.ucla.edu>
++2011-06-03 Paul Eggert <eggert@cs.ucla.edu>
+
+ Minor fixes for signed vs unsigned integers.
+ * character.h (MAYBE_UNIFY_CHAR):
+ * charset.c (maybe_unify_char):
+ * keyboard.c (read_char, reorder_modifiers):
+ XINT -> XFASTINT, since the integer must be nonnegative.
+ * ftfont.c (ftfont_spec_pattern):
+ * keymap.c (access_keymap, silly_event_symbol_error):
+ XUINT -> XFASTINT, since the integer must be nonnegative.
+ (Fsingle_key_description, preferred_sequence_p): XUINT -> XINT,
+ since it makes no difference and we prefer signed.
+ * keyboard.c (record_char): Use XUINT when all the neighbors do.
+ (access_keymap): NATNUMP -> INTEGERP, since the integer must be
+ nonnegative.
+
+2011-06-02 Paul Eggert <eggert@cs.ucla.edu>
+
+ Malloc failure behavior now depends on size of allocation.
+ * alloc.c (buffer_memory_full, memory_full): New arg NBYTES.
+ * lisp.h: Change signatures accordingly.
+ * alloc.c, buffer.c, editfns.c, menu.c, minibuf.c, xterm.c:
+ All callers changed. (Bug#8762)
+
+ * gnutls.c: Use Emacs's memory allocators.
+ Without this change, the gnutls library would invoke malloc etc.
+ directly, which causes problems on non-SYNC_INPUT hosts, and which
+ runs afoul of improving memory_full behavior. (Bug#8761)
+ (fn_gnutls_global_set_mem_functions): New macro or function pointer.
+ (emacs_gnutls_global_init): Use it to specify xmalloc, xrealloc,
+ xfree instead of the default malloc, realloc, free.
+ (Fgnutls_boot): No need to check for memory allocation failure,
+ since xmalloc does that for us.
+
+ Remove arbitrary limit of 2**31 entries in hash tables. (Bug#8771)
+ * category.c (hash_get_category_set):
+ * ccl.c (ccl_driver):
+ * charset.c (Fdefine_charset_internal):
+ * charset.h (struct charset.hash_index):
+ * composite.c (get_composition_id, gstring_lookup_cache)
+ (composition_gstring_put_cache):
+ * composite.h (struct composition.hash_index):
+ * dispextern.h (struct image.hash):
+ * fns.c (next_almost_prime, larger_vector, cmpfn_eql)
+ (cmpfn_equal, cmpfn_user_defined, hashfn_eq, hashfn_eql)
+ (hashfn_equal, hashfn_user_defined, make_hash_table)
+ (maybe_resize_hash_table, hash_lookup, hash_put)
+ (hash_remove_from_table, hash_clear, sweep_weak_table, SXHASH_COMBINE)
+ (sxhash_string, sxhash_list, sxhash_vector, sxhash_bool_vector)
+ (Fsxhash, Fgethash, Fputhash, Fmaphash):
+ * image.c (make_image, search_image_cache, lookup_image)
+ (xpm_put_color_table_h):
+ * lisp.h (struct Lisp_Hash_Table):
+ * minibuf.c (Ftry_completion, Fall_completions, Ftest_completion):
+ * print.c (print): Use 'EMACS_UINT' and 'EMACS_INT'
+ for hashes and hash indexes, instead of 'unsigned' and 'int'.
+ * alloc.c (allocate_vectorlike):
+ Check for overflow in vector size calculations.
+ * ccl.c (ccl_driver):
+ Check for overflow when converting EMACS_INT to int.
+ * fns.c, image.c: Remove unnecessary static decls that would otherwise
+ need to be updated by these changes.
+ * fns.c (make_hash_table, maybe_resize_hash_table):
+ Check for integer overflow with large hash tables.
+ (make_hash_table, maybe_resize_hash_table, Fmake_hash_table):
+ Prefer the faster XFLOAT_DATA to XFLOATINT where either will do.
+ (SXHASH_REDUCE): New macro.
+ (sxhash_string, sxhash_list, sxhash_vector, sxhash_bool_vector):
+ Use it instead of discarding useful hash info with large hash values.
+ (sxhash_float): New function.
+ (sxhash): Use it. No more need for "& INTMASK" due to above changes.
+ * lisp.h (FIXNUM_BITS): New macro, useful for SXHASH_REDUCE etc.
+ (MOST_NEGATIVE_FIXNUM, MOST_POSITIVE_FIXNUM, INTMASK):
+ Rewrite to use FIXNUM_BITS, as this simplifies things.
+ (next_almost_prime, larger_vector, sxhash, hash_lookup, hash_put):
+ Adjust signatures to match updated version of code.
+ (consing_since_gc): Now EMACS_INT, since a single hash table can
+ use more than INT_MAX bytes.
+
2011-06-01 Dan Nicolaescu <dann@ics.uci.edu>
Make it possible to build with GCC-4.6+ -O2 -flto.