Mercurial > hg > xemacs-beta
view src/number-mp.c @ 4921:17362f371cc2
add more byte-code assertions and better failure output
-------------------- ChangeLog entries follow: --------------------
src/ChangeLog addition:
2010-02-03 Ben Wing <ben@xemacs.org>
* alloc.c (Fmake_byte_code):
* bytecode.h:
* lisp.h:
* lread.c:
* lread.c (readevalloop):
* lread.c (Fread):
* lread.c (Fread_from_string):
* lread.c (read_list_conser):
* lread.c (read_list):
* lread.c (vars_of_lread):
* symbols.c:
* symbols.c (Fdefine_function):
Turn on the "compiled-function annotation hack". Implement it
properly by hooking into Fdefalias(). Note in the docstring to
`defalias' that we do this. Remove some old broken code and
change code that implemented the old kludgy way of hooking into
the Lisp reader into bracketed by `#ifdef
COMPILED_FUNCTION_ANNOTATION_HACK_OLD_WAY', which is not enabled.
Also enable byte-code metering when DEBUG_XEMACS -- this is a form
of profiling for computing histograms of which sequences of two
bytecodes are used most often.
* bytecode-ops.h:
* bytecode-ops.h (OPCODE):
New file. Extract out all the opcodes and declare them using
OPCODE(), a bit like frame slots and such. This way the file can
be included multiple times if necessary to iterate multiple times
over the byte opcodes.
* bytecode.c:
* bytecode.c (NUM_REMEMBERED_BYTE_OPS):
* bytecode.c (OPCODE):
* bytecode.c (assert_failed_with_remembered_ops):
* bytecode.c (READ_UINT_2):
* bytecode.c (READ_INT_1):
* bytecode.c (READ_INT_2):
* bytecode.c (PEEK_INT_1):
* bytecode.c (PEEK_INT_2):
* bytecode.c (JUMP_RELATIVE):
* bytecode.c (JUMP_NEXT):
* bytecode.c (PUSH):
* bytecode.c (POP_WITH_MULTIPLE_VALUES):
* bytecode.c (DISCARD):
* bytecode.c (UNUSED):
* bytecode.c (optimize_byte_code):
* bytecode.c (optimize_compiled_function):
* bytecode.c (Fbyte_code):
* bytecode.c (vars_of_bytecode):
* bytecode.c (init_opcode_table_multi_op):
* bytecode.c (reinit_vars_of_bytecode):
* emacs.c (main_1):
* eval.c (funcall_compiled_function):
* symsinit.h:
Any time we change either the instruction pointer or the stack
pointer, assert that we're going to move it to a valid location.
This should catch failures right when they occur rather than
sometime later. This requires that we pass in another couple of
parameters into some functions (only with error-checking enabled,
see below).
Also keep track, using a circular queue, of the last 100 byte
opcodes seen, and when we hit an assert failure during byte-code
execution, output the contents of the queue in a nice readable
fashion. This requires that bytecode-ops.h be included a second
time so that a table mapping opcodes to the name of their operation
can be constructed. This table is constructed in new function
reinit_vars_of_bytecode().
Everything in the last two paras happens only when
ERROR_CHECK_BYTE_CODE.
Add some longish comments describing how the arrays that hold the
stack and instructions, and the pointers used to access them, work.
* gc.c:
Import some code from my `latest-fix' workspace to mark the
staticpro's in order from lowest to highest, rather than highest to
lowest, so it's easier to debug when something goes wrong.
* lisp.h (abort_with_message): Renamed from abort_with_msg().
* symbols.c (defsymbol_massage_name_1):
* symbols.c (defsymbol_nodump):
* symbols.c (defsymbol):
* symbols.c (defkeyword):
* symeval.h (DEFVAR_SYMVAL_FWD_OBJECT):
Make the various calls to staticpro() instead call staticpro_1(),
passing in the name of the C var being staticpro'ed, so that it
shows up in staticpro_names. Otherwise staticpro_names just has
1000+ copies of the word `location'.
author | Ben Wing <ben@xemacs.org> |
---|---|
date | Wed, 03 Feb 2010 08:01:55 -0600 |
parents | 2fc0e2f18322 |
children | 16112448d484 |
line wrap: on
line source
/* Numeric types for XEmacs using the MP library. Copyright (C) 2004 Jerry James. This file is part of XEmacs. XEmacs is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2, or (at your option) any later version. XEmacs is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with XEmacs; see the file COPYING. If not, write to the Free Software Foundation, Inc., 51 Franklin St - Fifth Floor, Boston, MA 02111-1301, USA. */ /* Synched up with: Not in FSF. */ #include <config.h> #include <limits.h> #include <math.h> #include "lisp.h" static MINT *bignum_bytesize, *bignum_long_sign_bit, *bignum_one, *bignum_two; MINT *bignum_zero, *intern_bignum; MINT *bignum_min_int, *bignum_max_int, *bignum_max_uint; MINT *bignum_min_long, *bignum_max_long, *bignum_max_ulong; short div_rem; char * bignum_to_string (bignum b, int base) { REGISTER unsigned int i; unsigned int bufsize = 128U, index = 0U; int sign; char *buffer = xnew_array (char, 128), *retval; MINT *quo = MP_ITOM (0); short rem; /* FIXME: signal something if base is < 2 or doesn't fit into a short. */ /* Save the sign for later */ sign = MP_MCMP (b, bignum_zero); if (sign == 0) { XREALLOC_ARRAY (buffer, char, 2); buffer[0] = '0'; buffer[1] = '\0'; return buffer; } /* Copy abs(b) into quo for destructive modification */ else if (sign < 0) MP_MSUB (bignum_zero, b, quo); else MP_MOVE (b, quo); quo = MP_ITOM (0); /* Loop over the digits of b (in BASE) and place each one into buffer */ for (i = 0U; MP_MCMP(quo, bignum_zero) > 0; i++) { MP_SDIV (quo, base, quo, &rem); if (index == bufsize) { bufsize <<= 1; XREALLOC_ARRAY (buffer, char, bufsize); } buffer[index++] = rem < 10 ? rem + '0' : rem - 10 + 'a'; } MP_MFREE (quo); /* Reverse the digits, maybe add a minus sign, and add a null terminator */ bufsize = index + (sign < 0 ? 1 : 0) + 1; retval = xnew_array (char, bufsize); if (sign < 0) { retval[0] = '-'; i = 1; } else i = 0; for (; i < bufsize - 1; i++) retval[i] = buffer[--index]; retval[bufsize - 1] = '\0'; xfree (buffer, char *); return retval; } #define BIGNUM_TO_TYPE(type,accumtype) do { \ MP_MULT (b, quo, quo); \ for (i = 0U; i < sizeof(type); i++) \ { \ MP_SDIV (quo, 256, quo, &rem); \ retval |= ((accumtype) rem) << (8 * i); \ } \ MP_MFREE (quo); \ } while (0) int bignum_to_int (bignum b) { short rem, sign; unsigned int retval = 0; REGISTER unsigned int i; MINT *quo; sign = MP_MCMP (b, bignum_zero) < 0 ? -1 : 1; quo = MP_ITOM (sign); BIGNUM_TO_TYPE (int, unsigned int); return ((int) retval) * sign; } unsigned int bignum_to_uint (bignum b) { short rem; unsigned int retval = 0U; REGISTER unsigned int i; MINT *quo; quo = MP_ITOM (MP_MCMP (b, bignum_zero) < 0 ? -1 : 1); BIGNUM_TO_TYPE (unsigned int, unsigned int); return retval; } long bignum_to_long (bignum b) { short rem, sign; unsigned long retval = 0L; REGISTER unsigned int i; MINT *quo; sign = MP_MCMP (b, bignum_zero) < 0 ? -1 : 1; quo = MP_ITOM (sign); BIGNUM_TO_TYPE (long, unsigned long); return ((long) retval) * sign; } unsigned long bignum_to_ulong (bignum b) { short rem; unsigned long retval = 0UL; REGISTER unsigned int i; MINT *quo; quo = MP_ITOM (MP_MCMP (b, bignum_zero) < 0 ? -1 : 1); BIGNUM_TO_TYPE (unsigned long, unsigned long); return retval; } double bignum_to_double (bignum b) { short rem, sign; double retval = 0.0, factor = 1.0; REGISTER unsigned int i; MINT *quo; sign = MP_MCMP (b, bignum_zero) < 0 ? -1 : 1; quo = MP_ITOM (sign); MP_MULT (b, quo, quo); for (i = 0U; MP_MCMP (quo, bignum_zero) > 0; i++) { MP_SDIV (quo, 256, quo, &rem); retval += rem * factor; factor *= 256.0; } MP_MFREE (quo); return retval * sign; } static short char_to_number (char c) { if (c >= '0' && c <= '9') return c - '0'; if (c >= 'a' && c <= 'z') return c - 'a' + 10; if (c >= 'A' && c <= 'Z') return c - 'A' + 10; return -1; } int bignum_set_string (bignum b, const char *s, int base) { MINT *mbase; short digit; int neg = 0; if (base == 0) { if (s[0] == '0' && (s[1] == 'x' || s[1] == 'X')) { base = 16; s += 2; } else if (*s == '0') { base = 8; s++; } else base = 10; } /* FIXME: signal something if base is < 2 or doesn't fit into a short. */ if (*s == '-') { s++; neg = 1; } mbase = MP_ITOM ((short) base); MP_MOVE (bignum_zero, b); for (digit = char_to_number (*s); digit >= 0 && digit < base; digit = char_to_number (*++s)) { MINT *temp; MP_MULT (b, mbase, b); temp = MP_ITOM (digit); MP_MADD (b, temp, b); MP_MFREE (temp); } if (neg) MP_MSUB (bignum_zero, b, b); return (digit >= 0) ? -1 : 0; } void bignum_set_long (MINT *b, long l) { /* Negative l is hard, not least because -LONG_MIN == LONG_MIN. We pretend that l is unsigned, then subtract off the amount equal to the sign bit. */ bignum_set_ulong (b, (unsigned long) l); if (l < 0L) MP_MSUB (b, bignum_long_sign_bit, b); } void bignum_set_ulong (bignum b, unsigned long l) { REGISTER unsigned int i; MINT *multiplier = MP_ITOM (1); MP_MOVE (bignum_zero, b); for (i = 0UL; l > 0UL; l >>= 8, i++) { MINT *temp = MP_ITOM ((short) (l & 255)); MP_MULT (multiplier, temp, temp); MP_MADD (b, temp, b); MP_MULT (multiplier, bignum_bytesize, multiplier); MP_MFREE (temp); } MP_MFREE (multiplier); } void bignum_set_double (bignum b, double d) { REGISTER unsigned int i; int negative = (d < 0) ? 1 : 0; MINT *multiplier = MP_ITOM (1); MP_MOVE (bignum_zero, b); if (negative) d = -d; for (i = 0UL; d > 0.0; d /= 256, i++) { MINT *temp = MP_ITOM ((short) fmod (d, 256.0)); MP_MULT (multiplier, temp, temp); MP_MADD (b, temp, b); MP_MULT (multiplier, bignum_bytesize, multiplier); MP_MFREE (temp); } MP_MFREE (multiplier); if (negative) MP_MSUB (bignum_zero, b, b); } /* Return nonzero if b1 is exactly divisible by b2 */ int bignum_divisible_p (bignum b1, bignum b2) { int retval; MINT *rem = MP_ITOM (0); MP_MDIV (b1, b2, intern_bignum, rem); retval = (MP_MCMP (rem, bignum_zero) == 0); MP_MFREE (rem); return retval; } void bignum_ceil (bignum quotient, bignum N, bignum D) { MP_MDIV (N, D, quotient, intern_bignum); if (MP_MCMP (intern_bignum, bignum_zero) > 0 && MP_MCMP (quotient, bignum_zero) > 0) MP_MADD (quotient, bignum_one, quotient); } void bignum_floor (bignum quotient, bignum N, bignum D) { MP_MDIV (N, D, quotient, intern_bignum); if (MP_MCMP (intern_bignum, bignum_zero) > 0 && MP_MCMP (quotient, bignum_zero) < 0) MP_MSUB (quotient, bignum_one, quotient); } /* RESULT = N to the POWth power */ void bignum_pow (bignum result, bignum n, unsigned long pow) { MP_MOVE (bignum_one, result); for ( ; pow > 0UL; pow--) MP_MULT (result, n, result); } /* lcm(b1,b2) = b1 * b2 / gcd(b1, b2) */ void bignum_lcm (bignum result, bignum b1, bignum b2) { MP_MULT (b1, b2, result); MP_GCD (b1, b2, intern_bignum); MP_MDIV (result, intern_bignum, result, intern_bignum); } /* FIXME: We can't handle negative args, so right now we just make them positive before doing anything else. How should we really handle negative args? */ #define bignum_bit_op(result, b1, b2, op) \ REGISTER unsigned int i; \ MINT *multiplier = MP_ITOM (1), *n1 = MP_ITOM (0), *n2 = MP_ITOM (0); \ \ if (MP_MCMP (bignum_zero, b1) > 0) \ MP_MSUB (bignum_zero, b1, n1); \ else \ MP_MOVE (b1, n1); \ if (MP_MCMP (bignum_zero, b2) > 0) \ MP_MSUB (bignum_zero, b2, n2); \ else \ MP_MOVE (b2, n2); \ \ MP_MOVE (bignum_zero, result); \ \ for (i = 0UL; MP_MCMP (bignum_zero, n1) < 0 && \ MP_MCMP (bignum_zero, n2) < 0; i++) \ { \ short byte1, byte2; \ MINT *temp; \ \ MP_SDIV (n1, 256, n1, &byte1); \ MP_SDIV (n2, 256, n2, &byte2); \ temp = MP_ITOM (byte1 op byte2); \ MP_MULT (multiplier, temp, temp); \ MP_MADD (result, temp, result); \ MP_MULT (multiplier, bignum_bytesize, multiplier); \ MP_MFREE (temp); \ } \ MP_MFREE (n2); \ MP_MFREE (n1); \ MP_MFREE (multiplier) void bignum_and (bignum result, bignum b1, bignum b2) { bignum_bit_op (result, b1, b2, &); } void bignum_ior (bignum result, bignum b1, bignum b2) { bignum_bit_op (result, b1, b2, |); } void bignum_xor (bignum result, bignum b1, bignum b2) { bignum_bit_op (result, b1, b2, ^); } /* NOT is not well-defined for bignums ... where do you stop flipping bits? We just flip until we see the last one. This is probably a bad idea. */ void bignum_not (bignum result, bignum b) { REGISTER unsigned int i; MINT *multiplier = MP_ITOM (1), *n = MP_ITOM (0); if (MP_MCMP (bignum_zero, b) > 0) MP_MSUB (bignum_zero, b, n); else MP_MOVE (b, n); MP_MOVE (bignum_zero, result); for (i = 0UL; MP_MCMP (bignum_zero, n) < 0; i++) { short byte; MINT *temp; MP_SDIV (n, 256, n, &byte); temp = MP_ITOM (~byte); MP_MULT (multiplier, temp, temp); MP_MADD (result, temp, result); MP_MULT (multiplier, bignum_bytesize, multiplier); MP_MFREE (temp); } MP_MFREE (n); MP_MFREE (multiplier); } void bignum_setbit (bignum b, unsigned long bit) { bignum_pow (intern_bignum, bignum_two, bit); bignum_ior (b, b, intern_bignum); } /* This is so evil, even I feel queasy. */ void bignum_clrbit (bignum b, unsigned long bit) { MINT *num = MP_ITOM (0); /* See if the bit is already set, and subtract it off if not */ MP_MOVE (b, intern_bignum); bignum_pow (num, bignum_two, bit); bignum_ior (intern_bignum, intern_bignum, num); if (MP_MCMP (b, intern_bignum) == 0) MP_MSUB (b, num, b); MP_MFREE (num); } int bignum_testbit (bignum b, unsigned long bit) { bignum_pow (intern_bignum, bignum_two, bit); bignum_and (intern_bignum, b, intern_bignum); return MP_MCMP (intern_bignum, bignum_zero); } void bignum_lshift (bignum result, bignum b, unsigned long bits) { bignum_pow (intern_bignum, bignum_two, bits); MP_MULT (b, intern_bignum, result); } void bignum_rshift (bignum result, bignum b, unsigned long bits) { bignum_pow (intern_bignum, bignum_two, bits); MP_MDIV (b, intern_bignum, result, intern_bignum); } void bignum_random_seed(unsigned long seed) { /* FIXME: Implement me */ } void bignum_random(bignum result, bignum limit) { /* FIXME: Implement me */ MP_MOVE (bignum_zero, result); } void init_number_mp () { REGISTER unsigned int i; bignum_zero = MP_ITOM (0); bignum_one = MP_ITOM (1); bignum_two = MP_ITOM (2); /* intern_bignum holds throwaway values from macro expansions in number-mp.h. Its value is immaterial. */ intern_bignum = MP_ITOM (0); /* bignum_bytesize holds the number of bits in a byte. */ bignum_bytesize = MP_ITOM (256); /* bignum_long_sign_bit holds an adjustment for negative longs. */ bignum_long_sign_bit = MP_ITOM (256); for (i = 1UL; i < sizeof (long); i++) MP_MULT (bignum_bytesize, bignum_long_sign_bit, bignum_long_sign_bit); /* The MP interface only supports turning short ints into MINTs, so we have to set these the hard way. */ bignum_min_int = MP_ITOM (0); bignum_set_long (bignum_min_int, INT_MIN); bignum_max_int = MP_ITOM (0); bignum_set_long (bignum_max_int, INT_MAX); bignum_max_uint = MP_ITOM (0); bignum_set_ulong (bignum_max_uint, UINT_MAX); bignum_min_long = MP_ITOM (0); bignum_set_long (bignum_min_long, LONG_MIN); bignum_max_long = MP_ITOM (0); bignum_set_long (bignum_max_long, LONG_MAX); bignum_max_ulong = MP_ITOM (0); bignum_set_ulong (bignum_max_ulong, ULONG_MAX); }