-2011-04-20 Paul Eggert <eggert@cs.ucla.edu>
+2011-04-25 Paul Eggert <eggert@cs.ucla.edu>
- * intervals.h (struct interval): Use EMACS_INT for members
- where EMACS_UINT might cause problems. See
- <http://lists.gnu.org/archive/html/emacs-devel/2011-04/msg00514.html>.
- (CHECK_TOTAL_LENGTH): Remove cast to EMACS_INT; no longer needed.
- * intervals.c (interval_deletion_adjustment): Now returns EMACS_INT.
+ * lisp.h: (XVECTOR_SIZE): Remove. All uses replaced with ASIZE.
+ (ASIZE): Now contains previous implementation of XVECTOR_SIZE
+ instead of invoking XVECTOR_SIZE.
+
+ * lisp.h: Say "vectorlike header" rather than "vector header.
+ (struct vectorlike_header): Rename from struct vector_header.
+ (XVECTORLIKE_HEADER_SIZE): Renamed from XVECTOR_HEADER_SIZE.
All uses changed.
- * alloc.c (overrun_check_malloc, overrun_check_realloc): Now static.
- (overrun_check_free): Likewise.
+ lisp.h: Fix a problem with aliasing and vector headers.
+ GCC 4.6.0 optimizes based on type-based alias analysis. For
+ example, if b is of type struct buffer * and v of type struct
+ Lisp_Vector *, then gcc -O2 was incorrectly assuming that &b->size
+ != &v->size, and therefore "v->size = 1; b->size = 2; return
+ v->size;" must therefore return 1. This assumption is incorrect
+ for Emacs, since it type-puns struct Lisp_Vector * with many other
+ types. To fix this problem, this patch adds a new type struct
+ vector_header that documents the constraints on layout of vectors
+ and pseudovectors, and helps optimizing compilers not get fooled
+ by Emacs's type punning. It also adds the macros XSETTYPED_PVECTYPE
+ XSETTYPED_PSEUDOVECTOR, TYPED_PSEUDOVECTORP, for similar reasons.
+ * lisp.h (XVECTOR_SIZE): New convenience macro. All previous uses of
+ XVECTOR (foo)->size replaced to use this macro, to avoid the hassle
+ of writing XVECTOR (foo)->header.size.
+ (XVECTOR_HEADER_SIZE): New macro, for use in XSETPSEUDOVECTOR.
+ (XSETTYPED_PVECTYPE): New macro, specifying the name of the size
+ member.
+ (XSETPVECTYPE): Rewrite in terms of new macro.
+ (XSETPVECTYPESIZE): New macro, specifying both type and size.
+ This is a bit clearer, and further avoids the possibility of
+ undesirable aliasing.
+ (XSETTYPED_PSEUDOVECTOR): New macro, specifying the size.
+ (XSETPSEUDOVECTOR): Rewrite in terms of XSETTYPED_PSEUDOVECTOR
+ and XVECTOR_HEADER_SIZE.
+ (XSETSUBR): Rewrite in terms of XSETTYPED_PSEUDOVECTOR and XSIZE,
+ since Lisp_Subr is a special case (no "next" field).
+ (ASIZE): Rewrite in terms of XVECTOR_SIZE.
+ (struct vector_header): New type.
+ (TYPED_PSEUDOVECTORP): New macro, also specifying the C type of the
+ object, to help avoid aliasing.
+ (PSEUDOVECTORP): Rewrite in terms of TYPED_PSEUDOVECTORP.
+ (SUBRP): Likewise, since Lisp_Subr is a special case.
+
+ * lisp.h (struct Lisp_Vector, struct Lisp_Char_Table):
+ (struct Lisp_Sub_Char_Table, struct Lisp_Bool_Vector):
+ (struct Lisp_Hash_Table): Combine first two members into a single
+ struct vector_header member. All uses of "size" and "next" members
+ changed to be "header.size" and "header.next".
+ * buffer.h (struct buffer): Likewise.
+ * font.h (struct font_spec, struct font_entity, struct font): Likewise.
+ * frame.h (struct frame): Likewise.
+ * process.h (struct Lisp_Process): Likewise.
+ * termhooks.h (struct terminal): Likewise.
+ * window.c (struct save_window_data, struct saved_window): Likewise.
+ * window.h (struct window): Likewise.
+ * alloc.c (allocate_buffer, Fmake_bool_vector, allocate_pseudovector):
+ Use XSETPVECTYPESIZE, not XSETPVECTYPE, to avoid aliasing problems.
+ * buffer.c (init_buffer_once): Likewise.
+ * lread.c (defsubr): Use XSETTYPED_PVECTYPE, since Lisp_Subr is a
+ special case.
+ * process.c (Fformat_network_address): Use local var for size,
+ for brevity.
+
+2011-04-24 Paul Eggert <eggert@cs.ucla.edu>
+
+ * bytecode.c (exec_byte_code): Don't use XVECTOR before CHECK_VECTOR.
+
+2011-04-21 Paul Eggert <eggert@cs.ucla.edu>
Make the Lisp reader and string-to-float more consistent.
* data.c (atof): Remove decl; no longer used or needed.
- (Fstring_to_number): Use new string_to_float function, to be
+ (digit_to_number): Move to lread.c.
+ (Fstring_to_number): Use new string_to_number function, to be
consistent with how the Lisp reader treats infinities and NaNs.
Do not assume that floating-point numbers represent EMACS_INT
without losing information; this is not true on most 64-bit hosts.
Avoid double-rounding errors, by insisting on integers when
parsing non-base-10 numbers, as the documentation specifies.
- Report integer overflow instead of silently converting to
- integers.
- * lisp.h (string_to_float): New decl, replacing ...
+ * lisp.h (string_to_number): New decl, replacing ...
(isfloat_string): Remove.
- * lread.c (read1): Do not accept +. and -. as integers; this
+ * lread.c: Include <inttypes.h>, for uintmax_t and strtoumax.
+ (read1): Do not accept +. and -. as integers; this
appears to have been a coding error. Similarly, do not accept
strings like +-1e0 as floating point numbers. Do not report
- overflow for some integer overflows and not others; instead,
- report them all. Break out the floating-point parsing into a new
- function string_to_float, so that Fstring_to_number parses
+ overflow for integer overflows unless the base is not 10 which
+ means we have no simple and reliable way to continue.
+ Break out the floating-point parsing into a new
+ function string_to_number, so that Fstring_to_number parses
floating point numbers consistently with the Lisp reader.
- (string_to_float): New function, replacing isfloat_string.
+ (digit_to_number): Moved here from data.c. Make it static inline.
+ (E_CHAR, EXP_INT): Remove, replacing with ...
+ (E_EXP): New macro, to solve the "1.0e+" problem mentioned below.
+ (string_to_number): New function, replacing isfloat_string.
This function checks for valid syntax and produces the resulting
- Lisp float number too.
+ Lisp float number too. Rework it so that string-to-number
+ no longer mishandles examples like "1.0e+". Use strtoumax,
+ so that overflow for non-base-10 numbers is reported only when
+ there's no portable and simple way to convert to floating point.
+
+2011-04-20 Paul Eggert <eggert@cs.ucla.edu>
+
+ * textprop.c (set_text_properties_1): Rewrite for clarity,
+ and to avoid GCC warning about integer overflow.
+
+ * intervals.h (struct interval): Use EMACS_INT for members
+ where EMACS_UINT might cause problems. See
+ <http://lists.gnu.org/archive/html/emacs-devel/2011-04/msg00514.html>.
+ (CHECK_TOTAL_LENGTH): Remove cast to EMACS_INT; no longer needed.
+ * intervals.c (interval_deletion_adjustment): Now returns EMACS_INT.
+ All uses changed.
+ (offset_intervals): Tell GCC not to worry about length overflow
+ when negating a negative length.
+
+ * alloc.c (overrun_check_malloc, overrun_check_realloc): Now static.
+ (overrun_check_free): Likewise.
* alloc.c (SDATA_SIZE) [!GC_CHECK_STRING_BYTES]: Avoid runtime check
in the common case where SDATA_DATA_OFFSET is a multiple of Emacs