[BACK]Return to ChangeLog CVS log [TXT][DIR] Up to [local] / OpenXM / src / kan96xx / gmp-2.0.2-ssh-2

Annotation of OpenXM/src/kan96xx/gmp-2.0.2-ssh-2/ChangeLog, Revision 1.1.1.1

1.1       takayama    1: Thu Jun  6 19:00:53 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                      2:
                      3:        * Version 2.0.2 released.
                      4:
                      5:        * install.sh: New file.
                      6:        * Makefile.in (INSTALL): Use install.sh.
                      7:        (install-normal): New name for target `install'.
                      8:        (install): New dummy target.
                      9:
                     10:        * mpz/pow_ui.c: Swap tests for (e == 0) and (bsize == 0).
                     11:        * mpz/ui_pow_ui.c: Swap tests for (e == 0) and (blimb == 0).
                     12:
                     13:        * config/mt-linux (AR_FLAGS): New file.
                     14:        * configure.in: Use config/mt-linux for all linux systems.
                     15:
                     16: Tue Jun  4 03:42:18 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                     17:
                     18:        * Version 2.0.1 released.
                     19:
                     20:        * mpf/tests/ref.c: Cast result of TMP_ALLOC to the right pointer type.
                     21:
                     22:        * extract-double.c: Test _GMP_IEEE_FLOATS with #if, not plain if.
                     23:
                     24:        * insert-double.c: Don't #include stdlib.h.
                     25:
                     26:        * gmp-impl.h (union ieee_double_extract): Test sparc and __sparc.
                     27:        Do not test __sparc__.
                     28:
                     29:        * mpf/reldiff.c: Change declaration to work around irix5 compiler bug.
                     30:        * mpq/equal.c: Likewise.
                     31:
                     32:        * mpn/generic/gcd.c: Delete spurious comma at end of enumeration.
                     33:
                     34:        * mpn/generic/gcdext.c: Add K&R declaration syntax.
                     35:        * stack-alloc.h: Likewise.
                     36:        * insert-double.c: Likewise.
                     37:        * extract-double.c: Likewise.
                     38:        * mpf/tests/reuse.c: Likewise.
                     39:        * mpz/tests/reuse.c: Likewise.
                     40:        * mpf/tests/t-sub.c: Likewise.
                     41:        * mpf/tests/t-add.c: Likewise.
                     42:        * mpf/tests/t-muldiv.c: Likewise.
                     43:        * mpf/tests/t-conv.c: Likewise.
                     44:        * mpf/tests/ref.c: Likewise.
                     45:
                     46:        * mpn/config/t-oldgas: Renamed from t-freebsd.
                     47:        * mpn/configure.in: Use t-oldgas for freebsd, netbsd, and some linux
                     48:        configurations.
                     49:
                     50:        * mpn/powerpc32/mul_1.s: Really clear cy before entering loop.
                     51:        * mpn/powerpc32/*.s: Fix power/powerpc syntax issues.
                     52:
                     53:        * mpn/config/t-ppc-aix: New file.
                     54:        * mpn/configure.in: Use t-ppc-aix for powerpc like t-pwr-aix for power.
                     55:
                     56: Wed May 29 02:07:31 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                     57:
                     58:        * gmp.h (mp_bits_per_limb): Change qualifier from `const' to
                     59:        __gmp_const.
                     60:
                     61:        * gmp.h (mpf_init_set_str): Add `const' qualifier for 2nd parameter.
                     62:        * mpf/iset_str.c: Likewise.
                     63:
                     64: Mon May 27 00:15:58 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                     65:
                     66:        * gmp-impl.h: Declare __gmp_extract_double.
                     67:
                     68:        * mpz/set_q.c: Delete unused variables.
                     69:
                     70:        * gmp.h (mpq_equal): Declare.
                     71:
                     72:        * mpf/eq.c: mpf_cmp2 -> mpf_eq.
                     73:
                     74: Fri May 24 03:20:44 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                     75:
                     76:        * mpz/iset_d.c: Don't include <math.h>.
                     77:
                     78:        * insert-double.c (__gmp_scale2): New name for scal2.
                     79:        * mpz/get_d.c: Corresponding change.
                     80:        * mpf/get_d.c: Likewise.
                     81:        * mpq/get_d.c: Likewise.
                     82:        * gmp-impl.h: Declare __gmp_scale2.
                     83:
                     84:        * mpn/generic/scan0.c: Clarify comment.
                     85:
                     86:        * mpz/set_q.c: New file.
                     87:        * Makefile.in: Compile it.
                     88:        * make.bat: Likewise.
                     89:        * gmp.h: Declare mpz_set_q.
                     90:
                     91:        * insert-double.c: New file.
                     92:        * Makefile.in: Compile it.
                     93:        * make.bat: Likewise.
                     94:
                     95:        * mpz/get_d.c: New file.
                     96:        * mpz/Makefile.in: Compile it.
                     97:        * make.bat: Likewise.
                     98:        * gmp.h: Declare mpz_get_d.
                     99:
                    100:        * mpf/get_d.c: New file.
                    101:        * mpf/Makefile.in: Compile it.
                    102:        * make.bat: Likewise.
                    103:        * gmp.h: Declare mpf_get_d.
                    104:
                    105:        * make.bat: Compile things in alphabetical order.
                    106:
                    107:        * gmp-impl.h (MP_BASE_AS_DOUBLE): New #define.
                    108:        (LIMBS_PER_DOUBLE): New #define.
                    109:
                    110:        * extract-double.c: New file.
                    111:        * Makefile.in: Compile it.
                    112:        * make.bat: Likewise.
                    113:        * mpz/set_d.c: Rewrite to use __gmp_extract_double.
                    114:        * mpf/set_d.c: Likewise.
                    115:
                    116:        * mpn/configure.in: Use t-pwr-aix also for aix 3.2.4 and up.
                    117:
                    118: Wed May 22 02:48:35 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    119:
                    120:        * gmp-impl.h: Rework code for defining ieee_double_extract.
                    121:        (IEEE_DOUBLE_BIG_ENDIAN): Macro removed.
                    122:        (_GMP_IEEE_FLOATS): New macro.
                    123:        * mpn/vax/gmp-mparam.h: Delete.
                    124:
                    125:        * mpn/config/t-pwr-aix: New file.
                    126:        * mpn/configure.in: Use t-pwr-aix for aix 4 and later.
                    127:
                    128: Mon May 20 16:30:31 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    129:
                    130:        * gmp.h: In code for setting _GMP_H_HAVE_FILE, test more symbols.
                    131:
                    132:        * mpf/tests/t-add.c (oo): Add some `l' printf modifiers.
                    133:        * mpf/tests/t-sub.c (oo): Likewise.
                    134:        * mpf/tests/t-conv.c (oo): Likewise.
                    135:        * mpf/tests/t-sqrt.c (oo): Likewise.
                    136:
                    137:        * mpz/tests/t-mul.c (_mpn_mul_classic): Remove unused variables.
                    138:
                    139:        * mpn/{pyr,i960,clipper}/*.s: Add missing copyright headers.
                    140:
                    141: Fri May 17 02:24:43 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    142:
                    143:        * mpz/set_d.c: Call _mpz_realloc.
                    144:
                    145:        * mpq/set_z.c: New file.
                    146:        * mpq/Makefile.in: Compile it.
                    147:        * make.bat: Likewise.
                    148:        * gmp.h: Declare mpq_set_z.
                    149:
                    150:        * mp?/Makefile.in (libmp?.a): Depend on Makefile, not Makefile.in.
                    151:        * mpf/Makefile.in (test): Delete spurious target.
                    152:        * mpq/Makefile.in (test): Likewise.
                    153:
                    154:        * mpf/out_str.c: Use `e' to separate exponent when base <= 10.
                    155:
                    156:        * mpn/configure.in: Treat ultrasparc just like sparc v8,
                    157:        until 64-bit compilers are ready.
                    158:
                    159:        * mpf/set_d.c: Make it work for 64-bit machines.
                    160:
                    161: Thu May 16 20:53:57 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    162:
                    163:        * gmp-impl.h: Set IEEE_DOUBLE_BIG_ENDIAN to 0 for little-endian
                    164:        machines.
                    165:        * mpn/x86/gmp-mparam.h: Delete file.
                    166:
                    167:        * configure.in: Treat microsparc like sparc8.
                    168:
                    169:        * urandom.h: Test __alpha instead of __alpha__, since the former
                    170:        is the standard symbol.
                    171:        * mpn/generic/random2.c: Likewise.
                    172:        * mpf/random2.c: Likewise.
                    173:
                    174: Tue May 14 13:42:39 1996  Torbjorn Granlund  (tege@tiny.tmg.se)
                    175:
                    176:        * mpz/set_f.c: New file.
                    177:        * mpz/Makefile.in: Compile it.
                    178:        * gmp.h: Declare mpz_set_f.
                    179:
                    180:        * mpf/set_q.c: Simplify expression in rsize == nsize if-then-else arms.
                    181:
                    182: Tue May 14 13:03:07 1996  Torbjorn Granlund  (tege@tiny.tmg.se)
                    183:
                    184:        * make.bat: Add all new files.
                    185:
                    186: Sun May 12 22:24:36 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    187:
                    188:        * mpf/set_z.c: New file.
                    189:        * mpf/Makefile.in: Compile it.
                    190:        * gmp.h: Declare mpf_set_z.
                    191:
                    192: Sat May 11 19:26:25 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    193:
                    194:        * gmp.h: Declare mpf_set_q.
                    195:
                    196:        * mpf/set_q.c: Compute prec-1 limbs in mpn_divrem call.
                    197:
                    198: Fri May 10 17:37:38 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    199:
                    200:        * mpf/set_q.c: New file.
                    201:        * mpf/Makefile.in: Compile it.
                    202:
                    203:        * config.sub: Recognize sparc8.
                    204:
                    205: Wed May  8 09:19:11 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    206:
                    207:        * mpf/tests/t-dm2exp.c: New file.
                    208:
                    209:        * mpf/tests/t-add.c: Correct header comment.
                    210:        * mpf/tests/t-sub.c: Likewise.
                    211:        * mpf/tests/t-sqrt.c: Likewise.
                    212:
                    213:        * mpf/div.c: Misc variable name cleanups.
                    214:        * mpf/div_ui.c: Base more closely on mpf/div.c.
                    215:        * mpf/ui_div.c: Likewise.
                    216:
                    217:        * mpz/tests/Makefile.in (check): Depend on Makefile.
                    218:        * mpq/tests/Makefile.in (check): Likewise.
                    219:        * mpf/tests/Makefile.in (check): Likewise.
                    220:
                    221:        * mpf/tests/t-muldiv.c: New file.
                    222:        * mpf/tests/Makefile.in: Compile and run `t-muldiv'.
                    223:        (t-ref.o): Delete spurious rule.
                    224:
                    225:        * mpf/sqrt.c: Properly detect negative input operand.
                    226:
                    227:        * mpf/sqrt_ui.c: Delete spurious header comment.
                    228:        * mpf/sqrt.c: Likewise.
                    229:        * mpz/sqrt.c: Likewise.
                    230:
                    231:        * mpz/tests/reuse.c (main): Read `reps' from command line.
                    232:
                    233:        * mpf/tests/reuse.c: New file.
                    234:        * mpf/tests/Makefile.in: Compile and run `reuse'.
                    235:
                    236:        * mpf/mul_ui.c: Disable code for removing low zero limbs.
                    237:
                    238:        * mpf/div.c: Fix condition for when vp and qp overlaps.
                    239:
                    240:        * mpf/add_ui.c: When sum equals u, copy up to prec+1 limbs.
                    241:
                    242:        * mpf/out_str.c: Don't output '\n' after exponent.
                    243:
                    244:        * mpf/add_ui.c: New special case for when U is completely cancelled.
                    245:
                    246: Wed Apr 24 05:33:28 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    247:
                    248:        * Version 2.0 released.
                    249:
                    250:        * All files: Upfate FSF's address.
                    251:
                    252:        * Makefile.in (gmp_toc.html): New name for gmp.html.
                    253:        (TAGS): Depend on force.
                    254:
                    255:        * mpf/tests/t-conv.c: Pass -base to mpf_set_str.
                    256:
                    257: Sat Apr 20 03:54:06 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    258:
                    259:        * Makefile.in (ps): New target, depend on gmp.ps.
                    260:
                    261: Fri Apr 19 14:03:15 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    262:
                    263:        * mpf/out_str.c: Print `@' before exponent, not `e'.
                    264:
                    265:        * make.bat: Update from Makefiles.
                    266:
                    267: Thu Apr 18 01:22:05 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    268:
                    269:        * mpf/set_str.c: If parameter `base' is negative, expect exponent
                    270:        to be decimal, otherwise in the same base as the mantissa.
                    271:
                    272: Wed Apr 17 17:28:36 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    273:
                    274:        * mpf/set_dfl_prec.c: Don't return anything.
                    275:        * gmp.h: Corresponding changes.
                    276:
                    277:        * mpf/set_dfl_prec.c: Use `unsigned long int' for bit counts.
                    278:        * mpf/init2.c: Likewise.
                    279:        * mpf/get_prc.c: Likewise.
                    280:        * mpf/set_prc.c: Likewise.
                    281:        * mpf/set_prc_raw.c: Likewise.
                    282:        * mpz/popcount.c: Likewise.
                    283:        * mpz/hamdist.c: Likewise.
                    284:        * mpz/scan1.c: Likewise.
                    285:        * mpz/scan0.c: Likewise.
                    286:        * mpn/generic/popcount.c: Likewise.
                    287:        * mpn/generic/hamdist.c: Likewise.
                    288:        * mpn/generic/scan1.c: Likewise.
                    289:        * mpn/generic/scan0.c: Likewise.
                    290:        * gmp.h: Likewise.
                    291:
                    292:        * mpf/eq.c: New file, based on mpf/diff.c.
                    293:        * mpf/diff.c: Delete.
                    294:        * mpf/Makefile.in: Corresponding changes.
                    295:        * gmp.h: Likewise.
                    296:
                    297:        * mpf/reldiff.c: New file.
                    298:        * mpf/Makefile.in: Compile it.
                    299:        * gmp.h: Declare mpf_reldiff.
                    300:
                    301:        * mpz/iset_d.c: New file.
                    302:        * mpz/Makefile.in: Compile it.
                    303:        * gmp.h: Declare mpz_init_set_d.
                    304:
                    305: Tue Apr 16 16:28:31 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    306:
                    307:        * Makefile.in (gmp.html): Pass -acc to texi2html.
                    308:
                    309: Mon Apr 15 16:20:24 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    310:
                    311:        * mpf/set_str.c: Switch off code for defaulting the base from the
                    312:        leading characters.
                    313:
                    314:        * gmp.h (mp?_sign): Delete.
                    315:        (mp?_sgn): New macros.
                    316:
                    317: Fri Apr 12 17:23:33 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    318:
                    319:        * Makefile.in (gmp.dvi): Delete tmp.* at end of rule.
                    320:
                    321: Wed Apr 10 22:52:02 1996  Torbjorn Granlund  (tege@tiny.tmg.se)
                    322:
                    323:        * mpf/random2.c: Change of `exp' param, mp_size_t => mp_exp_t.
                    324:        * gmp.h: Corresponding change.
                    325:
                    326:        * gmp.h (mp_bits_per_limb): Make it const.
                    327:
                    328: Sat Mar 30 01:20:23 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    329:
                    330:        * configure.in: Re-enable recognition of with_gcc.
                    331:
                    332:        * mpf/Makefile.in (.c.o): Pass XCFLAGS.
                    333:        * mpn/Makefile.in (.c.o): Likewise.
                    334:        * mpz/Makefile.in (.c.o): Likewise.
                    335:        * mpq/Makefile.in (.c.o): Likewise.
                    336:        * mpbsd/Makefile.in (.c.o): Likewise.
                    337:        * mpf/tests/Makefile.in (.c.o): Likewise.
                    338:        * mpz/tests/Makefile.in (.c.o): Likewise.
                    339:        * mpq/tests/Makefile.in (.c.o): Likewise.
                    340:
                    341:        * Makefile.in (XCFLAGS): Default to empty.
                    342:        (FLAGS_TO_PASS): Pass on XCFLAGS.
                    343:        (.c.o): Pass XCFLAGS.
                    344:
                    345:        * config/mt-m88110 (XCFLAGS): Define instead of CC.
                    346:        * config/mt-sprc8-gcc (XCFLAGS): Likewise.
                    347:        * config/mt-supspc-gcc (XCFLAGS): Likewise.
                    348:
                    349:        * configure: Don't default CC to "gcc -O2" is -with-gcc=no was
                    350:        specified.
                    351:
                    352: Mon Mar 25 01:07:54 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    353:
                    354:        * urandom.h: Test for __SVR4 in addition to __svr4__.
                    355:
                    356:        * mp_bpl.c (mp_bits_per_limb): Declare as `const'.
                    357:
                    358:        * Makefile.in (CFLAGS): `-O2' => `-O'.
                    359:        * mpn/Makefile.in (CFLAGS): Likewise.
                    360:
                    361:        * gmp-impl.h: Get rid of obsolete field access macros.
                    362:
                    363:        * mpn/mp_bases.c (__mp_bases): 1e39 => 1e38 to work around Solaris
                    364:        cc compiler bug.
                    365:
                    366:        * gmp.h (__MPN): Make it work also for non-ANSI compilers.
                    367:
                    368: Thu Mar 21 01:07:54 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    369:
                    370:        * mpf/sub.c: New special case for ediff <= 1 before generic code.
                    371:        Simplify generic code for ediff == 0.
                    372:        Rename uexp => exp.
                    373:
                    374: Mon Mar 11 18:24:57 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    375:
                    376:        * mpf/tests/*.c: Use ref_mpf_sub for error calculation.
                    377:        * mpf/tests/Makefile.in: Link ref.o to all executables.
                    378:
                    379:        * mpf/tests/t-sub.c: Make u = v + 1 with 50% probability.
                    380:
                    381: Sun Mar 10 21:03:17 1996  Torbjorn Granlund  (tege@tiny.tmg.se)
                    382:
                    383:        * mpf/get_str.c: In digit development loop for fractions, change
                    384:        loop condition from `<' to `<='.
                    385:
                    386: Thu Mar  7 04:58:11 1996  Torbjorn Granlund  <tege@tiny.tmg.se>
                    387:
                    388:        * mpn/mp_bases.c (__mp_bases): 1e100 => 1e39 to avoid overflow warning.
                    389:
                    390: Wed Mar  6 01:10:42 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    391:
                    392:        * mpf/tests/t-sqrt.c: New file.
                    393:        * mpf/tests/Makefile.in: Corresponding changes.
                    394:
                    395:        * mpf/sqrt.c: Special case for square root of zero.
                    396:
                    397:        * mpq/add.c: Clean up variable names.
                    398:        * mpq/sub.c: Update from mpq/add.c.
                    399:
                    400:        * mpz/divexact.c: abs => ABS.
                    401:        * mpz/gcd.c: Likewise.  Rewrite final fixup code, to decrease
                    402:        allocation.  Misc cleanups.
                    403:
                    404: Tue Mar  5 22:24:56 1996  Torbjorn Granlund  <tege@tmg.se>
                    405:
                    406:        * mpn/configure.in: Recognize linuxoldld as a synonym for linuxaout.
                    407:
                    408:        * gmp.h (mpn_add, mpn_add_1, mpn_sub, mpn_sub_1): Add prototypes.
                    409:
                    410:        * mpn/configure.in: Use t-freebsd also for netbsd.
                    411:
                    412: Mon Mar  4 15:13:28 1996  Torbjorn Granlund  <tege@tmg.se>
                    413:
                    414:        * mpq/Makefile.in (cmp.o): Depend on longlong.h.
                    415:
                    416:        * mpq/equal.c: New file.
                    417:        * mpq/Makefile.in: Corresponding changes.
                    418:
                    419:        * mpf/tests/t-add.c: New file.
                    420:        * mpf/tests/t-sub.c: Renamed from t-addsub.c.
                    421:        * mpf/tests/ref.c: New file.
                    422:        * mpf/tests/Makefile.in: Corresponding changes.
                    423:
                    424:        * gmp-impl.h (SIZ, ABSIZ, PTR, EXP, PREC, ALLOC): New #defines.
                    425:
                    426: Sun Mar  3 07:45:46 1996  Torbjorn Granlund  <tege@tmg.se>
                    427:
                    428:        * mpf/set_str.c: In exponentialization code, allocate 3 extra
                    429:        limbs, not just 2.
                    430:
                    431:        * mpf/get_str.c: Allocate sufficient space for tstr.
                    432:        When calculating exp_in_base, round result down.
                    433:
                    434:        * mpf/tests/t-conv.c: New file.
                    435:        * mpf/tests/Makefile.in: Corresponding changes.
                    436:
                    437:        * mp_bpl.c: New file.
                    438:        * gmp.h: Declare it.
                    439:        * Makefile.in: Corresponding changes.
                    440:
                    441: Sat Mar  2 06:27:56 1996  Torbjorn Granlund  <tege@tmg.se>
                    442:
                    443:        * mpf/set_prc_raw.c: New file.
                    444:        * mpf/set_prc.c: Renamed from set_prec.c.
                    445:        * mpf/get_prc.c: New file.
                    446:        * mpf/Makefile.in: Corresponding changes.
                    447:        * gmp.h: Declare new functions.
                    448:
                    449:        * mpn/generic/gcdext.c: Add copyright header.
                    450:
                    451: Fri Mar  1 01:22:24 1996  Torbjorn Granlund  <tege@tmg.se>
                    452:
                    453:        * mpn/configure.in: For ppc601, search "power" before "powerpc32".
                    454:
                    455:        * mp?/Makefile.in (AR_FLAGS): New variable.
                    456:        (libmp?.a): Use it.
                    457:
                    458:        * make.bat: New file.
                    459:        * mpn/msdos: New directory.
                    460:        * mpn/msdos/asm-syntax.h: New file.
                    461:
                    462:        * mpn/Makefile.in (distclean maintainer-clean): Delete asm-syntax.h.
                    463:
                    464:        * config.sub: Recognize [ctj]90-cray.
                    465:
                    466:        * mpn/configure.in: Recognize [ctj]90-cray-unicos*.
                    467:
                    468:        * mpn/generic/gcdext.c: Don't use alloca directly, use TMP_* macros.
                    469:
                    470:        * mpn/generic/gcd.c: Split increment from use of USIZE to avoid
                    471:        undefined behaviour.
                    472:
                    473: Thu Feb 29 04:11:24 1996  Torbjorn Granlund  <tege@tmg.se>
                    474:
                    475:        * Makefile.in (install-info-files): Update for new install-info
                    476:        behaviour.
                    477:
                    478:        * mpn/power/add_n.s: Rewrite.
                    479:        * mpn/power/sub_n.s: Rewrite.
                    480:
                    481: Wed Feb 28 01:34:30 1996  Torbjorn Granlund  <tege@tmg.se>
                    482:
                    483:        * mpz/pow_ui.c: Compute allocation more aggressively for small bases.
                    484:        * mpz/ui_pow_ui.c: Likewise.
                    485:
                    486:        * mpn/mp_bases.c (__mp_bases): Put huge value in 2nd field for index 1.
                    487:
                    488:        * mpn/generic/sqrtrem.c: sizeof (mp_limb_t) => BYTES_PER_MP_LIMB.
                    489:        * mpn/generic/gcd.c: Likewise.
                    490:        (SIGN_BIT): Compute differently.
                    491:
                    492: Mon Feb 26 00:07:36 1996  Torbjorn Granlund  <tege@tmg.se>
                    493:
                    494:        * All files: mp_limb => mp_limb_t, mp_limb_signed => mp_limb_signed_t.
                    495:
                    496:        * Makefile.in (install, install-bsdmp, install-info-files): Depend
                    497:        on installdirs.  chmod all installed files.
                    498:
                    499: Sun Feb 25 01:47:41 1996  Torbjorn Granlund  <tege@tmg.se>
                    500:
                    501:        * mpbsd/configure.in: Delete debugging code.
                    502:
                    503:        * All Makefile.in: Update clean targets.
                    504:
                    505:        * Makefile.in (AR_FLAGS): New variable.
                    506:        (libgmp.a): Use it.
                    507:        (libmp.a): Likewise.
                    508:
                    509:        * VERSION: Delete file.
                    510:
                    511:        * Makefile.in (installdirs): New target.
                    512:        * mkinstalldirs: New file (from the texinfo package).
                    513:
                    514:        * Makefile.in (INSTALL, INSTALL_DATA, INSTALL_PROGRAM): New variables.
                    515:        (MAKEINFO, MAKEINFOFLAGS, TEXI2DVI): New variables.
                    516:        (install-info): New target.
                    517:        (install, install-bsdmp): Depend on install-info.
                    518:        ($(srcdir)/gmp.info): Changed from plain gmp.info; put info files
                    519:        into source directory.
                    520:        (distclean, mostlyclean): New targets.
                    521:        (maintainer-clean): New name for realclean.
                    522:        (uninstall): New target.
                    523:        (TAGS): New target.
                    524:        (info, dvi): New targets.
                    525:        (.PHONY): Assign.
                    526:
                    527:        * Makefile.in (install, install-bsdmp): Use INSTALL_DATA.
                    528:
                    529:        * mp{n,z,f,bsd}/move-if-change: Delete.
                    530:
                    531:        * mpbsd/Makefile.in (stamp-stddefh): Delete target.
                    532:
                    533:        * Makefile.in (.c.o): Pass CFLAGS last.
                    534:        * mpbsd/Makefile.in (.c.o): Likewise.
                    535:        * mpf/Makefile.in (.c.o): Likewise.
                    536:        * mpq/Makefile.in (.c.o): Likewise.
                    537:        * mpz/Makefile.in (.c.o): Likewise.
                    538:        * mpn/Makefile.in (.c.o): Likewise.
                    539:        (.S.o): Likewise.
                    540:
                    541:        * memory.c: Change allocation error message.
                    542:
                    543:        * Makefile.in (install): Prefix gmp.h with $(srcdir).
                    544:        (install-bsdmp): Prefix mp.h with $(srcdir).
                    545:
                    546:        * mp{n,z,f,bsd}/{configure,config.sub}: Delete.
                    547:
                    548:        * Makefile.in (gmp.dvi): Set TEXINPUTS also for 2nd tex invocation
                    549:        (install targets): Install gmp.info-N.
                    550:
                    551: Sat Feb 24 03:36:52 1996  Torbjorn Granlund  <tege@tmg.se>
                    552:
                    553:        * mpf/get_str.c: Fix typo.
                    554:
                    555:        * mpz/legendre.c: Clarify expression with extra parens.
                    556:
                    557:        * version.c (gmp_version): Not static.
                    558:
                    559:        * mpf/iset_str.c: Properly return error code.
                    560:
                    561:        * mpf/add.c: Delete unused variables.
                    562:        * mpf/inp_str.c: Likewise.
                    563:        * mpq/get_d.c: Likewise.
                    564:
                    565:        * mpn/generic/dump.c: #include <stdio.h>.
                    566:        * mpf/dump.c: Likewise.
                    567:        * mpf/set_str.c: #include <ctype.h>.
                    568:        (strtol): Declare.
                    569:
                    570:        * gmp.h: mpn_sqrt => mpn_sqrtrem.
                    571:
                    572:        * Makefile.in (clean, realclean): Clean in mpbsd.
                    573:        (check): Test in mpf.
                    574:
                    575:        * mpf/Makefile.in (clean): Clean in tests.
                    576:        * mpq/Makefile.in (clean): Clean in tests.
                    577:
                    578:        * mpf/tests/Makefile.in: New file.
                    579:        * mpf/tests/configure.in: New file.
                    580:        * mpf/tests/t-addsub.c: New file.
                    581:
                    582:        * mpf/sub_ui.c: Simply call mpf_sub for now.
                    583:
                    584:        * mpf/sub.c: Increase prec by 1.
                    585:        * mpf/ui_sub.c: Likewise.
                    586:
                    587: Fri Feb 23 00:59:54 1996  Torbjorn Granlund  <tege@tmg.se>
                    588:
                    589:        * mpf/ui_sub.c: Fix typos.
                    590:
                    591:        * mpf/get_str.c: When allocating space for tmp, allow for an extra
                    592:        limb.  In code for fraction conversion, add special case for bases
                    593:        that are a power of 2.
                    594:
                    595:        * mpf/out_str.c: Output leading "0.".
                    596:        Default base to 10, before computing string allocation.
                    597:
                    598:        * mpf/get_str.c: Make variables for string size have type size_t.
                    599:        * gmp.h: Corresponding change.
                    600:
                    601:        * mpf/random2.c: Allow creation of prec+1 large mantissas.
                    602:
                    603:        * mpf/add_ui.c: Don't abort if u < 0; special case for u <= 0.
                    604:        Fix typo in MPN_COPY offset.
                    605:        * mpf/sub_ui.c: Analogous changes.
                    606:
                    607:        * mpf/set_prec.c: Rewrite.
                    608:
                    609:        * mpf/init2.c: Compute precision as in set_prec.c.
                    610:
                    611:        * mpf/div_2exp.c: Special case for u == 0.
                    612:        * mpf/mul_2exp.c: Likewise.  Write r->_mp_size always.
                    613:
                    614:        * mpf/sqrt_ui.c: mpn_sqrt => mpn_sqrtrem.
                    615:        * mpf/sqrt.c: Likewise.  When computing new exponent, round quotient
                    616:        towards -infinity.
                    617:
                    618:        * mpf/add.c: Fix typos.
                    619:        * mpf/sub.c: Fix typos.
                    620:
                    621: Thu Feb 22 00:24:48 1996  Torbjorn Granlund  <tege@tmg.se>
                    622:
                    623:        * mpz/Makefile.in (stamp-stddefh): Delete target.
                    624:        (test): Delete target.
                    625:        * Makefile.in (stamp-stddefh): Delete target.
                    626:        (cre-stddefh.o): Delete target.
                    627:        (gmp.dvi): Set TEXINPUTS before invoking tex.
                    628:
                    629:        * cre-stddefh.c: Delete.
                    630:
                    631:        * mpz/sqrt.c: Fix typo.
                    632:
                    633:        * mpz/powm.c: Special case for mod == 0.
                    634:        * mpz/powm_ui.c: Likewise.
                    635:
                    636:        * mpz/get_si.c: Handle -0x80000000 correctly.
                    637:
                    638:        * mpz/inp_str.c: Now retutns size_t.
                    639:        Make it return number of bytes read or error indication.
                    640:        * mpf/inp_str.c: Likewise.
                    641:
                    642:        * mpz/out_raw.c: Replace by mpz/out_binary.c, with modifications.
                    643:        * mpz/inp_raw.c: Rewrite, using mpz/inp_binary as a base.
                    644:        * mpz/inp_binary.c: Delete.
                    645:
                    646:        * mpn/Makefile.in (XCFLAGS): Remove variable.
                    647:        (.c.o): Don't pass XCFLAGS.
                    648:        (SFLAGS): Set to nothing.
                    649:        (.S.o): Pass SFLAGS, not XCFLAGS.
                    650:
                    651:        * mpn/config/t-freebsd (SFLAGS): New name for XCFLAGS.
                    652:
                    653:        * mpf/out_str.c: Make return number of bytes written or error
                    654:        indication.
                    655:        * mpz/out_str.c: Likewise.
                    656:        * gmp.h: Corresponding changes.
                    657:
                    658:        * gmp.h (__mpz_struct): mp_size_t => int.
                    659:        (__mpq_struct): Likewise.
                    660:        (__mpf_struct): Likewise.
                    661:        (mp_size_t): int => long int.
                    662:
                    663:        * mpn/cray: New directory.
                    664:        * mpn/cray/gmp-mparam.h: New file.
                    665:        * mpn/configure.in: Recognize cray variants.
                    666:
                    667:        * Makefile.in: Set defaults for prefix, libdir, etc.
                    668:        (install): New target.
                    669:        (install-bsdmp): New target.
                    670:        (gmp.html): New target.
                    671:
                    672:        * stack-alloc.c (__tmp_alloc): Cast void ptrs to char * in comparison.
                    673:
                    674: Wed Feb 21 04:35:02 1996  Torbjorn Granlund  <tege@tmg.se>
                    675:
                    676:        * gmp.h: Sort mpn declarations.
                    677:        (mpn_gcdext): Add declaration.
                    678:
                    679:        * mpn/generic/divrem_1.c: New file.
                    680:        * mpn/Makefile.in (divrem_1.o): New rule.
                    681:        * configure.in (functions): Add divrem_1.
                    682:
                    683:        * mpn/generic/divmod.c: Delete file.
                    684:        * mpn/configure.in (functions): Delete divmod.
                    685:        * Makefile.in (divmod.o): Delete rule.
                    686:        * gmp.h (mpn_divmod): New #define.
                    687:
                    688:        * gmp.h (mpn_next_bit_set): Delete spurious declaration.
                    689:
                    690:        * mpn/generic/divrem.c (default case): In code assigning
                    691:        most_significant_q_limb, move reassignment of n0 into if statement.
                    692:
                    693:        * gmp.h (mpf_inp_str): Fix typo.
                    694:        (mpf_out_str): Make prototype match reality.
                    695:        * mpf/inp_str.c: New file.
                    696:        * mpf/out_str.c: New file.
                    697:        * mpf/Makefile.in: Compile new files.
                    698:
                    699:        * mpn/Makefile.in (dump.o): Fix dependency path.
                    700:        (inlines.o): Likewise.
                    701:
                    702:        * mpn/configure.in: Make m68060 be the same as m68000.  Clean up
                    703:        m68k configs.
                    704:
                    705: Tue Feb 20 01:35:11 1996  Torbjorn Granlund  <tege@tmg.se>
                    706:
                    707:        * mpn/generic/sqrtrem.c: Renamed from sqrt.
                    708:        * mpn/configure.in (functions): Corresponding change.
                    709:        * mpn/Makefile.in: Likewise.
                    710:        * mpz/sqrtrem.c: Likewise.
                    711:        * mpz/sqrt.c: Likewise.
                    712:        * mpn/generic/perfsqr.c: Likewise.
                    713:
                    714:        * Makefile.in (clean): Also remove libmp.a.
                    715:        Don't compile cre-conv-tab.c or mp_bases.c.
                    716:        cre-conv-tab.c: Delete file.
                    717:        (gmp.ps): New rule.
                    718:
                    719:        * mpn/mp_bases.c: New file.
                    720:        * mpn/Makefile.in: Compile mp_bases.c.
                    721:
                    722:        * mpz/set_str.c: Skip initial whitespace.
                    723:        * mpf/set_str.c: Likewise.
                    724:        * mpbsd/xtom.c: Likewise.
                    725:
                    726:        * gmp.h: Add missing mpz declarations.
                    727:        Delete all formal parameter names from declarations.
                    728:
                    729:        * mpn/Makefile.in: Add dependencies for .c files.
                    730:
                    731:        * Makefile.in (check): Write recursive make calls separately, not as
                    732:        a loop.
                    733:        (FLAGS_TO_PASS): New variable.  Use it for most recursive makes.
                    734:
                    735: Mon Feb 19 01:02:20 1996  Torbjorn Granlund  <tege@tmg.se>
                    736:
                    737:        * mpn/Makefile.in (.S.o): Pipe cpp output to grep in order to delete
                    738:        lines starting with #.
                    739:        (CPP): Set to $(CC) -E to avoid gcc dependency.
                    740:
                    741:        * mpn/m68k/syntax.h (moveql): Define to moveq for MIT_SYNTAX.
                    742:
                    743:        * mpn/hppa/hppa1_1/pa7100/addmul_1.S: Fix typo in s1_ptr alignment
                    744:        code.
                    745:        * mpn/hppa/hppa1_1/pa7100/submul_1.S: Likewise.
                    746:
                    747:        * gmp.h: Fix typos in #defines of recently added mpn functions.
                    748:
                    749:        * mpz/inp_str.c: Skip all whitespace, not just plain space.
                    750:        * mpbsd/min.c: Likewise.
                    751:
                    752:        * mpn/configure.in (functions): Add gcdext.
                    753:        * mpn/generic/gcdext.c: New file.
                    754:
                    755:        * mpz/legendre.c: mpz_div_2exp => mpz_tdiv_q_2exp.
                    756:
                    757:        * gmp.h: Surround mpn declarations with extern "C" { ... }.
                    758:
                    759:        * Makefile.in (check): New target.
                    760:
                    761:        * mpq/get_d.c: Update comments.  Use rsize instead of dsize + N_QLIMBS
                    762:        when possible.  Add special case for nsize == 0.
                    763:
                    764:        * gmp.h (mpq_get_d): Add declaration.
                    765:        (mpq_canonicalize): Likewise.
                    766:        (mpq_cmp_ui): Likewise.
                    767:        (mpf_diff): Likewise.
                    768:        (mpf_ui_sub): Likewise.
                    769:        (mpf_set_prec): Likewise.
                    770:        (mpf_random2): Likewise.
                    771:
                    772:        * gmp.h (mpz_cmp_ui): New #define.
                    773:        (mpz_cmp_si): New #define.
                    774:        (mpq_cmp_ui): New #define.
                    775:        (mpz_sign): New #define.
                    776:        (mpq_sign): New #define.
                    777:        (mpf_sign): New #define.
                    778:        (mpq_numref): New #define.
                    779:        (mpq_denref): New #define.
                    780:
                    781:        * mpq/set_z.c: File deleted.
                    782:        * mpq/Makefile.in: Corresponding changes.
                    783:
                    784: Sun Feb 18 01:34:47 1996  Torbjorn Granlund  <tege@tmg.se>
                    785:
                    786:        * mpbsd/sdiv.c: Use _mp_realloc, not _mpz_realloc.
                    787:
                    788:        * mpz/inp_binary.c: Default stream to stdin.
                    789:        * mpz/inp_str.c: Likewise.
                    790:        * mpz/inp_raw.c: Likewise.
                    791:        * mpz/out_binary.c: Default stream to stdout.
                    792:        * mpz/out_raw.c: Likewise.
                    793:        * mpz/out_str.c: Likewise.
                    794:
                    795:        * mpbsd/realloc.c: New file.
                    796:        * mpbsd/Makefile.in: Corresponding changes.
                    797:
                    798:        * mpbsd/min.c: Rewrite (base on mpz/inp_str.c).
                    799:        * mpbsd/mtox.c: Rewrite (base on mpz/get_str.c).
                    800:
                    801:        * mpbsd/mout.c: Rewrite (base on mpz/out_str) but make it output
                    802:        spaces in each 10th position.
                    803:        * mpbsd/xtom.c: Rewrite (base on mpz/set_str).
                    804:
                    805:        * mpq/tests/Makefile.in (st-cmp): New file.
                    806:        * mpq/tests/configure.in (srcname): New file.
                    807:
                    808:        * mpz/tests/configure.in (srcname): Fix typo.
                    809:
                    810:        * mpq/cmp.c: Add check using number of significant bits, to avoid
                    811:        general multiplication.
                    812:
                    813: Sat Feb 17 11:58:30 1996  Torbjorn Granlund  <tege@tmg.se>
                    814:
                    815:        * mpq/cmp_ui.c: Store cy_limb after the mpn_mul_1 calls.
                    816:
                    817:        * mpq/tests: New directory.
                    818:        * mpq/tests/t-cmp.c: New file.
                    819:        * mpq/tests/t-cmp_ui.c: New file.
                    820:
                    821:        * mpz/tests/dive.c (main): Generate zero numerator.
                    822:        (get_random_size) : Delete.
                    823:
                    824:        * mpz/divexact.c: Add special case for 0/x.
                    825:
                    826:        * gmp.h (mpz_mod): Add declaration.
                    827:
                    828: Fri Feb 16 18:18:39 1996  Andreas Schwab  <schwab@informatik.uni-dortmund.de>
                    829:
                    830:        * mpn/m68k/*: Rewrite code not to use the INSN macros.
                    831:        (L): New macro to properly prefix local labels for ELF.
                    832:
                    833: Fri Feb 16 00:20:56 1996  Torbjorn Granlund  <tege@tmg.se>
                    834:
                    835:        * gmp-impl.h (ieee_double_extract): Use plain `unsigned int' for
                    836:        fields.
                    837:        * mpn/generic/inlines.c (_FORCE_INLINES): New #define.  Delete
                    838:        conditional __GNUC__.
                    839:        * gmp.h (mpn_add, mpn_sub, mpn_add_1, mpn_sub_1):
                    840:        Only define these if __GNUC__ || _FORCE_INLINES.
                    841:        * mpf/random2.c: Add missing parameter in non-ANSI header.
                    842:        * mpn/generic/gcd.c (SIGN_BIT): Do as #define to work around bug
                    843:        in AIX compilers.
                    844:        * mpq/get_d.c: #define N_QLIMBS.
                    845:        * mpz/divexact.c: Obscure division by 0 to silent compiler warnings.
                    846:        * stack-alloc.c: Cast void* pointer to char* before doing arithmetic
                    847:        on it.
                    848:
                    849:        * Makefile.in (mpbsd/libmpbsd.a): New rule.
                    850:        * configure.in (configdirs): Add mpbsd.
                    851:
                    852:        * gmp.h: Add declarations for a few missing mpn functions.
                    853:
                    854:        * Makefile.in (libmp.a): New rule.
                    855:
                    856:        * mpbsd/mdiv.c: #include "dmincl.c", not "mpz_dmincl.c"
                    857:        * gmp.h: Move #define of __GNU_MP__ into the `#if __GNU_MP__' block.
                    858:        * mp.h: Likewise.  Update typedefs from gmp.h.
                    859:        * mpbsd/configure.in: New file.
                    860:        * mpbsd/Makefile.in: New file.
                    861:        * mpbsd/configure: Link to master configure.
                    862:        * mpbsd/config.sub: Link to master config.sub.
                    863:
                    864:        * Makefile.in: Set RANLIB_TEST.
                    865:        * (libgmp.a): Use it.
                    866:        * (libgmp.a): Do ranlib before moving the libgmp.a to the build
                    867:        directory.
                    868:        * mp?/Makefile.in: Don't use or set RANLIB.
                    869:
                    870: Thu Feb 15 16:38:41 1996  Torbjorn Granlund  <tege@tmg.se>
                    871:
                    872:        * mpz/add_ui.c: MP_INT => mpz_t.
                    873:        * mpz/cmp_ui.c: Likewise.
                    874:        * mpz/fac_ui.c: Likewise.
                    875:        * mpz/inp_binary.c: Likewise.
                    876:        * mpz/inp_raw.c: Likewise.
                    877:        * mpz/legendre.c: Likewise.
                    878:        * mpz/jacobi.c: Likewise.
                    879:        * mpz/out_binary.c: Likewise.
                    880:        * mpz/out_raw.c: Likewise.
                    881:        * mpz/random2.c: Likewise.
                    882:        * mpz/random.c: Likewise.
                    883:        * mpz/realloc.c: Likewise.
                    884:
                    885:        * mpz/legendre.c: __mpz_2factor(X) => mpz_scan1(X,0),
                    886:        __mpz_odd_less1_2factor => mpz_scan1(X,1).
                    887:        * mpz/ntsup.c: File deleted.
                    888:        * mpz/Makefile.in: Corresponding changes.
                    889:
                    890:        * mpz/pprime_p: Use mpz_scan1 to avoid looping.
                    891:
                    892:        * mpz/fac_ui.c: Type of `k' and `p' is `unsigned long'.
                    893:        * mpz/pprime_p.c: Pass long to *_ui functions.
                    894:        * mpz/gcdext.c: Likewise.
                    895:        * mpz/fdiv_r_2exp.c: Likewise.
                    896:        * mpz/fac_ui.c: Likewise.
                    897:
                    898:        * mpz/powm.c: Don't use mpn_rshift when mod_shift_cnt is 0.
                    899:
                    900:        * mpz/tests/Makefile.in (st-sqrtrem): Fix typo.
                    901:
                    902:        * mpz/cmp_ui.c: #undef mpz_cmp_ui.
                    903:        * mpz/cmp_si.c: #undef mpz_cmp_si.
                    904:        * gmp.h (mpz_cmp_ui): New #define.
                    905:        (mpz_cmp_si): New #define.
                    906:
                    907: Wed Feb 14 22:11:24 1996  Torbjorn Granlund  <tege@tmg.se>
                    908:
                    909:        * gmp.h: Test __cplusplus in addition to __STDC__.
                    910:        * gmp-impl.h: Likewise.
                    911:
                    912:        * gmp.h: Surround declarations with extern "C" { ... }.
                    913:
                    914: Tue Feb 13 15:20:45 1996  Torbjorn Granlund  <tege@tmg.se>
                    915:
                    916:        * mpz/fdiv_r_2exp.c: Use MPN_NORMALIZE.
                    917:        * mpz/tdiv_r_2exp.c: Likewise.
                    918:
                    919:        * mpz/fdiv_r_2exp.c: New file.
                    920:        * mpz/fdiv_q_2exp.c: New file.
                    921:        * mpz/tdiv_r_2exp.c: Renamed from mpz/mod_2exp.c.
                    922:        * mpz/tdiv_q_2exp.c: Renamed from mpz/div_2exp.c
                    923:        * mpz/Makefile.in: Corresponding changes.
                    924:
                    925:        * mpz/scan0.c,scan1.c: New files.
                    926:        * mpz/Makefile.in: Compile them.
                    927:
                    928:        * gmp.h (mpn_normal_size): Delete.
                    929:
                    930:        * config.guess: Update from Cygnus version.
                    931:
                    932:        * mpn/m68k/rshift.S: Use INSN2 macro for lea instructions.
                    933:        * mpn/m68k/lshift.S: Likewise.
                    934:
                    935:        * mpn/configure.in: Fix configuration for plain 68000.
                    936:
                    937: Mon Feb 12 01:06:06 1996  Torbjorn Granlund  <tege@tmg.se>
                    938:
                    939:        * mpz/tests/t-powm.c: Generate negative BASE operand.
                    940:
                    941:        * mpz/powm.c: Make result always positive.
                    942:
                    943: Sun Feb 11 01:44:56 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    944:
                    945:        * mpz/tests/*.c: Add t- prefix.
                    946:        * mpz/tests/Makefile.in: Corresponding changes.
                    947:        * mpz/tests/configure.in: Update srctrigger.
                    948:
                    949:        * mpz/tests/gcd.c: Generate negative operands.
                    950:        * mpz/tests/gcd2.c: Likewise.
                    951:
                    952:        * mpz/gcdext.c: At end, if G is negative, negate all G, S, and T.
                    953:
                    954: Thu Feb  8 17:16:12 UTC 1996 Ken Weber <kweber@mat.ufrgs.br>
                    955:
                    956:        * mp{z,n}/gcd.c: Change mpn_gcd interface.
                    957:        * gmp.h: Ditto.
                    958:        * gmp.texi: update documentation.
                    959:
                    960: Mon Feb  7 23:58:43 1996  Andreas Schwab  <schwab@informatik.uni-dortmund.de>
                    961:
                    962:        * mpn/m68k/{lshift,rshift}.S: New files.
                    963:        * mpn/m68k/syntax.h: New ELF_SYNTAX macros.
                    964:        (MEM_INDX, R, PROLOG, EPILOG): New macros.
                    965:        * mpn/m68k/*.S: Use R macro with register name.  Use PROLOG and EPILOG
                    966:        macros.  Rename `size' to `s_size' or s1_size to avoid clash with ELF
                    967:        .size directive.
                    968:        * mpn/configure.in: New target m68k-*-linux*.
                    969:
                    970: Wed Feb  7 07:41:31 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    971:
                    972:        * Makefile.in (cre-conv-tab): Workaround for SunOS make.
                    973:
                    974:        * mpz/tests/reuse.c: New file.
                    975:        * mpz/tests/Makefile.in: Handle reuse.c.
                    976:
                    977: Tue Feb  6 11:56:24 UTC 1996 Ken Weber <kweber@mat.ufrgs.br>
                    978:
                    979:        * mpz/gcd.c: Fix g->size when one op is 0 and g == other op.
                    980:
                    981: Tue Feb  6 01:36:39 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    982:
                    983:        * gmp.h (mpz_divexact): Delete parameter names.
                    984:        (mpz_lcm): Delete spurious declaration.
                    985:
                    986:        * mpz/dmincl.c: Fix typo.
                    987:
                    988: Mon Feb  5 01:11:56 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                    989:
                    990:        * mpn/generic/gcd.c (gcd_2): Declare consistently.
                    991:
                    992:        * mpz/tdiv_q.c: Optimize division by a single-limb divisor.
                    993:        * mpz/dmincl.c: Likewise.
                    994:
                    995:        * mpz/add.c: Use MPN_NORMALIZE instead of mpn_normal_size.
                    996:        * mpz/sub.c: Likewise.
                    997:        * mpn/generic/sqrt.c: Likewise.
                    998:
                    999:        * mpn/tests/{add_n,sub_n,lshift,rshift}.c: Put garbage in the
                   1000:        destination arrays.
                   1001:
                   1002: Fri Feb  2 02:21:27 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1003:
                   1004:        * mpz/{jacobi.c,legendre.c,ntsup.c,invert.c}: New files.
                   1005:        * mpz/Makefile.in: Compile them.
                   1006:
                   1007:        * mpn/Makefile.in (INCLUDES): Don't search in `generic'.
                   1008:
                   1009: Thu Feb  1 02:15:11 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1010:
                   1011:        Change from Ken Weber:
                   1012:        * mpz/divexact.c: Make it work when quot is identical to either input.
                   1013:
                   1014:        * mpf/ui_sub.c: New file.
                   1015:        * mpf/Makefile.in: Compile it.
                   1016:
                   1017:        * gmp-impl.h (MPZ_TMP_INIT): alloca -> TMP_ALLOC.
                   1018:        * mpz/{c,f}div_{q,qr,r}.c: Use TMP_DECL/TMP_MARK/TMP_FREE since
                   1019:        these use MPZ_TMP_INIT.
                   1020:        * mpz/mod.c: Likewise.
                   1021:        * mpq/{add,sub}.c: Likewise.
                   1022:        * mpq/canonicalize: Likewise.
                   1023:
                   1024:        * mpq/{add,sub,mul,div}.c: Use mpz_divexact. MP_INT -> mpz_t.
                   1025:        * mpq/canonicalize.c: Likewise.
                   1026:
                   1027: Wed Jan 31 01:45:00 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1028:
                   1029:        * mpn/generic/gcd.c: Misc changes from Ken.
                   1030:
                   1031:        * mpz/tests/gcd2.c: New file.
                   1032:        * mpz/tests/Makefile.in: Handle gcd2.c.
                   1033:
                   1034:        * mpn/generic/gcd.c (mpn_gcd): When GCD == ORIG_V, return vsize,
                   1035:        not orig_vsize.  Fix parameter declaration.
                   1036:
                   1037:        * mpz/mod_ui.c: Delete file.
                   1038:        * mpz/Makefile.in: Don't try to compile mod_ui.
                   1039:
                   1040:        * mpz/cdiv_*_ui.c): Make them work right.
                   1041:        * gmp.h: Declare cdiv*.
                   1042:
                   1043: Tue Jan 30 02:22:56 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1044:
                   1045:        * mpz/{cdiv_q.c,cdiv_q_ui.c,cdiv_qr.c,cdiv_qr_ui.c,cdiv_r.c,
                   1046:        cdiv_r_ui.c,cdiv_ui.c}: New files.
                   1047:        * mpz/Makefile.in: Compile them.
                   1048:
                   1049:        * All files: Make file permissions right.
                   1050:
                   1051:        Changes from Ken Weber:
                   1052:        * mpn/generic/accelgcd.c: Delete.
                   1053:        * mpn/generic/bingcd.c: Delete.
                   1054:        * mpn/generic/numbits.c: Delete.
                   1055:        * mpn/generic/gcd.c: New file.
                   1056:        * mpn/configure.in (functions): Update accordingly.
                   1057:        * mpz/divexact.c: New file.
                   1058:        * mpz/Makefile.in: Compile divexact.c.
                   1059:        * mpz/gcd.c: Rewrite to accommodate for gcd changes in mpn.
                   1060:        * gmp.h: declare new functions, delete obsolete declarations.
                   1061:        * mpz/tests/dive.c: New file.
                   1062:        * mpz/tests/Makefile.in: Handle dive.c.
                   1063:
                   1064: Mon Jan 29 03:53:24 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1065:
                   1066:        * mpz/random.c: Handle negative SIZE parameter.
                   1067:
                   1068:        * mpz/tests/tdiv(_ui).c: New name for tst-dm(_ui).c.
                   1069:        * mpz/tests/tst-mdm(_ui).c: Delete.
                   1070:        * mpz/tests/fdiv(_ui).c: New test based in tst-mdm(_ui).
                   1071:        * mpz/tests/*.c: Get rid of tst- prefix for DOS 8+3 naming.
                   1072:        * mpz/tests/Makefile.in: Corresponding changes.
                   1073:        * mpz/tests/configure.in: Update srctrigger.
                   1074:
                   1075:        * mpn/generic/divmod.c: Update from divrem.
                   1076:        * mpn/generic/divrem.c: Misc cleanups.
                   1077:
                   1078: Sun Jan 28 03:25:08 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1079:
                   1080:        * All files: Use new TMP_ALLOC interface.
                   1081:
                   1082:        * mpz/powm_ui.c: Make Jan 25 changes to powm.c also here.
                   1083:
                   1084:        * mpz/tests/powm_ui.c: New file.
                   1085:        * mpz/tests/Makefile.in: Add rules for tst-powm and tst-powm_ui.
                   1086:
                   1087:        * Makefile.in: Update dependency list.
                   1088:        * mpf/Makefile.in: Likewise.
                   1089:        * mpz/Makefile.in: Likewise.
                   1090:        * mpq/Makefile.in: Likewise.
                   1091:        * Makefile.in: Set RANLIB simply to ranlib, and allow configure
                   1092:        to override it.
                   1093:
                   1094:        * mpz/Makefile.in (conf): Delete spurious target.
                   1095:        (mp_bases.c): Delete.
                   1096:        (cre-conv-tab rules): Delete.
                   1097:
                   1098:        * Makefile.in (cre-conv-tab): Greatly simplify.
                   1099:
                   1100: Sat Jan 27 13:38:15 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1101:
                   1102:        * stack-alloc.c: New file.
                   1103:        * stack-alloc.h: New file.
                   1104:
                   1105:        * gmp.h (__gmp_inline): Define using __inline__.
                   1106:
                   1107: Thu Jan 25 00:28:37 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1108:
                   1109:        * mpn/generic/scan0.c: New file.
                   1110:        * mpn/generic/scan1.c: Renamed from next_bit.c.
                   1111:        * mpn/configure.in (functions): Include scan0 and scan1.
                   1112:
                   1113:        * mpn/m68k/*: #include sysdep.h.  Use C_GLOBAL_NAME.
                   1114:
                   1115:        * configure: Update from Cygnus version.
                   1116:        * config.guess: Likewise.
                   1117:        * config.sub: Likewise.
                   1118:        * configure: Pass --nfp to recursive configures.
                   1119:
                   1120:        * mpz/tests/tst-*.c: Adjust SIZE and reps.
                   1121:
                   1122:        * mpz/powm.c: Move esize==0 test earlier.
                   1123:        In final reduction of rp,rsize, don't call mpn_divmod unless
                   1124:        reduction is really needed.
                   1125:
                   1126:        * mpz/tests/tst-powm.c: Fix thinko in checking code.
                   1127:
                   1128:        * All files: Get rid of `__' prefix from mpn_* calls and declarations.
                   1129:        * gmp.h: #define __MPN.
                   1130:        * gmp.h: Use __MPN in #defines for mpn calls.
                   1131:
                   1132:        * mpn/generic/mul_n.c: Prepend `i' to internal routines.
                   1133:        * gmp-impl.h: Add #defines using __MPN for those internal routines.
                   1134:
                   1135:        * mpn/generic/sqrt.c: Change call to mpn_mul to mpn_mul_n.
                   1136:
                   1137: Wed Jan 24 13:28:19 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1138:
                   1139:        * mpn/sparc32/udiv_fp.S: New name for udiv_qrnnd.S.
                   1140:        * mpn/sparc32/udiv_nfp.S: New name for v8/udiv_qrnnd.S.
                   1141:        * mpn/sparc32/v8/supersparc: New directory.
                   1142:        * mpn/sparc32/v8/supersparc/udiv.S: New file.
                   1143:
                   1144: Tue Jan 23 01:10:11 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1145:
                   1146:        This major contribution is from Ken Weber:
                   1147:        * mpn/generic/accelgcd.c: New file.
                   1148:        * mpn/generic/bdivmod.c: New file.
                   1149:        * mpn/generic/bingcd.c: New file.
                   1150:        * mpn/generic/gcd_1.c: Rewrite.
                   1151:        * mpn/generic/numbits.c: New file (to go away soon).
                   1152:        * mpz/gcd.c: Rewrite.
                   1153:        * mpz/tests/tst-gcd.c (SIZE): Now 128.
                   1154:        * gmp.h: Declare new functions.
                   1155:        * mpn/configure.in (functions): List new files.
                   1156:        * gmp-impl.h (MPN_SWAP): Delete.
                   1157:        (MPN_LESS_BITS_LIMB, MPN_LESS_BITS, MPN_MORE_BITS): Delete.
                   1158:        (MPN_COMPL_INCR, MPN_COMPL): Delete.
                   1159:
                   1160: Mon Jan 22 02:04:59 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1161:
                   1162:        * gmp.h (mpn_name): New #define.
                   1163:
                   1164:        * mpn/m88k/mc88110/addmul_1.s: New file.
                   1165:        * mpn/m88k/mc88110/add_n.S: New file.
                   1166:        * mpn/m88k/mc88110/sub_n.S: New file.
                   1167:
                   1168:        * mpn/m88k/sub_n.s: Correctly initialize carry.
                   1169:
                   1170:        * mpn/sparc32/{add_n.S,sub_n.S,lshift.S,rshift.S): `beq' => `be'.
                   1171:
                   1172: Sun Jan 21 00:04:35 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1173:
                   1174:        * mpn/sparc64/addmul_1.s: New file.
                   1175:        * mpn/sparc64/submul_1.s: New file.
                   1176:        * mpn/sparc64/rshift.s: New file.
                   1177:
                   1178: Sat Jan 20 00:32:54 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1179:
                   1180:        * mpz/iset.c: Fix typo introduced Dec 25.
                   1181:
                   1182: Wed Jan 17 13:16:44 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1183:
                   1184:        * config/mt-sprc8-gcc: New name for mt-sparc8-gcc.
                   1185:        * config/mt-sparcv8-gcc: Delete.
                   1186:        * configure.in: Corresponding changes.
                   1187:
                   1188: Tue Jan 16 16:31:01 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1189:
                   1190:        * gmp-impl.h: #include alloca.h when necessary.
                   1191:
                   1192:        * longlong.h: Test __alpha instead of __alpha__, since the former
                   1193:        is the standard symbol.
                   1194:
                   1195: Mon Jan 15 18:06:57 1996  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1196:
                   1197:        * mpn/sparc64/mul_1.s: Swap operands of mulx instructions.
                   1198:        * mpn/sparc64/lshift.s: New file.
                   1199:
                   1200: Fri Dec 29 17:34:03 1995  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1201:
                   1202:        * mpn/x86/pentium/add_n.S: Get rid of #defines for register names.
                   1203:        * mpn/x86/pentium/sub_n.S: Likewise.
                   1204:
                   1205: Thu Dec 28 03:16:57 1995  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1206:
                   1207:        * mpn/x86/pentium/mul_1.S: Rework loop to avoid AGI between update
                   1208:        of loop induction variable and load insn at beginning of loop.
                   1209:        * mpn/x86/pentium/addmul_1.S: Likewise.
                   1210:        * mpn/x86/pentium/submul_1.S: Likewise.
                   1211:
                   1212: Mon Dec 25 23:22:55 1995  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1213:
                   1214:        * All files: Prefix user-visible structure fields with _mp_.
                   1215:
                   1216: Fri Dec 22 20:42:17 1995  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1217:
                   1218:        * mpn/configure.in (m68k configs): Terminate path variable with
                   1219:        plain "m68k".
                   1220:
                   1221: Fri Dec 22 03:29:33 1995  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1222:
                   1223:        * mpn/sparc32/add_n.S: Update from sub_n.S to fix bugs, and to
                   1224:        clean things up.
                   1225:
                   1226:        * mpn/configure.in (m68k configs): Update #include path for new
                   1227:        mpn directory organization.
                   1228:
                   1229: Tue Dec 12 02:53:02 1995  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1230:
                   1231:        * gmp.h: Prefix all structure field with _mp_.
                   1232:        * gmp-impl.h: Define access macros for these fields.
                   1233:
                   1234: Sun Dec 10 00:47:17 1995  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1235:
                   1236:        * mpn/alpha/addmul_1.s: Prefix labels with `.'.
                   1237:        * mpn/alpha/submul_1.s: Likewise.
                   1238:        * mpn/alpha/[lr]shift.s: Likewise.
                   1239:        * mpn/alpha/udiv_qrnnd.S: Likewise.
                   1240:        * mpn/alpha/ev5/[lr]shift.s: Likewise.
                   1241:
                   1242:        * mpn/alpha/ev5/lshift.s: Fix typos.
                   1243:
                   1244: Fri Dec  1 14:28:20 1995  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1245:
                   1246:        * mpn/Makefile.in (.SUFFIXES): Define.
                   1247:
                   1248: Wed Nov 29 23:11:57 1995  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1249:
                   1250:        * mpn/sparc64/{add_n.s, sub_n.s}: New files.
                   1251:
                   1252: Tue Nov 28 06:03:13 1995  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1253:
                   1254:        * mpn/x86/syntax.h: Handle ELF_SYNTAX.
                   1255:        Rename GAS_SYNTAX => BSD_SYNTAX.
                   1256:
                   1257:        * mpn/configure.in: Handle linuxelf and SysV for x86 variants.
                   1258:
                   1259: Mon Nov 27 01:32:12 1995  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1260:
                   1261:        * mpn/hppa/hppa1_1/pa7100/submul_1.S: New file.
                   1262:
                   1263: Sun Nov 26 04:30:47 1995  Torbjorn Granlund  <tege@noisy.tmg.se>
                   1264:
                   1265:        * mpn/hppa/hppa1_1/pa7100/addmul_1.S: New file.
                   1266:
                   1267:        * mpn/sparc32/add_n.S: Rewrite to use 64 bit loads/stores.
                   1268:        * mpn/sparc32/sub_n.S: Likewise.
                   1269:
                   1270: Fri Nov 17 00:18:46 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1271:
                   1272:        * mpn/configure.in: Handle m68k on NextStep.
                   1273:
                   1274: Thu Nov 16 02:30:26 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1275:
                   1276:        * mpn: Reorganize machine-specific directories.
                   1277:        * mpn/configure.in: Corresponding changes.
                   1278:        (sh, sh2): Handle these.
                   1279:        (m68k targets): Create asm-syntax.h.
                   1280:
                   1281: Thu Nov  9 02:20:50 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1282:
                   1283:        * mpn/generic/mul_n.c (____mpn_sqr_n): Delete code that calls abort.
                   1284:        (____mpn_mul_n): Likewise.
                   1285:
                   1286: Tue Nov  7 03:25:12 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1287:
                   1288:        * mpf/get_str.c: In exponentiation code (two places), don't swap
                   1289:        input and output areas when calling mpn_mul_1.
                   1290:        * mpf/set_str.c: Likewise.
                   1291:
                   1292: Fri Nov  3 02:35:58 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1293:
                   1294:        * mpf/Makefile.in: Make sure all objects are listed in dependency list;
                   1295:        delete spurious entries.
                   1296:
                   1297:        * mpf/mul.c: Handle U or V being 0.  Allow prec+1 for result precision.
                   1298:
                   1299:        * mpf/set_prec.c: New computation of limb precision.
                   1300:        * mpf/set_dfl_prec.c: Likewise.
                   1301:
                   1302:        * mpf/random2.c: Fix typo computing exp.
                   1303:        * mpf/get_str.c: In (uexp > usize) case, set n_limbs as a function of
                   1304:        the user-requested number of digits, n_digits.
                   1305:
                   1306: Thu Nov  2 16:25:07 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1307:
                   1308:        * mpn/generic/divrem.c (case 2): Don't move np vector back, it is
                   1309:        never read.
                   1310:        (default case): Put most significant limb from np in new variable n2;
                   1311:        decrease size argument for MPN_COPY_DECR; use n2 instead of np[dsize].
                   1312:
                   1313: Wed Nov  1 02:59:53 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1314:
                   1315:        * mpn/sparc/[lr]shift.S: New files.
                   1316:
                   1317: Tue Oct 31 00:08:12 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1318:
                   1319:        * mpz/gcd_ui.c: Set w->size unconditionally when v is zero.
                   1320:
                   1321:        * gmp-impl.h (assert): Delete definition.
                   1322:
                   1323:        * mpf/sub.c: Delete all assert calls.  Delete variable `cy'.
                   1324:
                   1325:        * mpf/neg.c: Use prec+1 as precision.  Optimize for when arguments
                   1326:        are the same.
                   1327:        * mpf/abs.c: Likewise.
                   1328:        * mpf/{set,neg,abs}.c: Make structure and variable names similar.
                   1329:
                   1330: Mon Oct 30 12:45:26 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1331:
                   1332:        * mpf/random2.c (random): Test __SVR4 in addition to __svr4__.
                   1333:        * mpn/generic/random2.c (random): Likewise.
                   1334:
                   1335: Sun Oct 29 01:54:28 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1336:
                   1337:        * mpf/div.c: Special handle U or V being 0.
                   1338:
                   1339:        * mpf/random2.c: New file.
                   1340:
                   1341:        * longlong.h (i860 rshift_rhlc): Define.
                   1342:        (i960 udiv_qrnnd): Define.
                   1343:        (i960 count_leading_zeros): Define.
                   1344:        (i960 add_ssaaaa): Define.
                   1345:        (i960 sub_ddmmss): Define.
                   1346:        (i960 rshift_rhlc): Define.
                   1347:
                   1348: Sat Oct 28 19:09:15 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1349:
                   1350:        * mpn/pentium/rshift.S: Fix and generalize condition for when to use
                   1351:        special code for shift by 1.
                   1352:        * mpn/pentium/lshift.S: Likewise.
                   1353:
                   1354: Thu Oct 26 00:02:56 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1355:
                   1356:        * gmp.h: #undef __need_size_t.
                   1357:        * mp.h: Update from gmp.h.
                   1358:
                   1359: Wed Oct 25 00:17:27 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1360:
                   1361:        * mpf/Makefile.in: Compile set_prec.c.
                   1362:        * mpf/realloc.c: Delete this file.
                   1363:        * mpf/Makefile.in: Delete mentions of realloc.c.
                   1364:
                   1365:        * gmp.h (__mpf_struct): Get rid of `alloc' field.
                   1366:        * mpf/clear.c: Likewise.
                   1367:        * mpf/init*.c: Likewise.
                   1368:        * mpf/set_prec.c: Likewise.
                   1369:        * mpf/iset*.c: Likewise.
                   1370:
                   1371:        * mpf/iset_str.c: New file.
                   1372:
                   1373:        * mpn/configure.in: Handle pyramid.
                   1374:
                   1375:        * mpf/set.c: Use prec+1 as precision.
                   1376:
                   1377:        * mpf/set_prec.c: New file.
                   1378:
                   1379: Tue Oct 24 00:56:41 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1380:
                   1381:        * mpn/generic/divrem.c: New file.  Will replace mpn/generic/divmod.c
                   1382:        when rest of source is converted.
                   1383:        * mpn/configure.in (functions): Add `divrem'
                   1384:        * mpn/generic/set_str.c: Never call __mpn_mul_1 with zero size.
                   1385:
                   1386:        * mpf/get_str.c: Completely rewritten.
                   1387:        * mpf/add.c: Fix several problems.
                   1388:        * mpf/sub.c: Compare operands from most significant end until
                   1389:        first difference, exclude skipped limbs from computation.
                   1390:        Accordingly simplify normalization code.
                   1391:        * mpf/set_str.c: Fix several problems.
                   1392:        * mpf/dump.c: New file.
                   1393:        * mpf/Makefile.in: Compile dump.c.
                   1394:        * mpf/init2.c: Set prec field correctly.
                   1395:
                   1396: Sun Oct 22 03:02:09 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1397:
                   1398:        * cre-conv-tab.c: #include math.h; don't declare log and floor.
                   1399:
                   1400: Sat Oct 21 23:04:10 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1401:
                   1402:        * mpf/mul_ui.c: Handle U being 0.
                   1403:
                   1404: Wed Oct 18 19:39:27 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1405:
                   1406:        * mpn/generic/set_str.c: Correctly handle input like "000000000000".
                   1407:        Misc cleanups.
                   1408:
                   1409: Tue Oct 17 15:14:13 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1410:
                   1411:        * longlong.h: Define COUNT_LEADING_ZEROS_0 for machines where
                   1412:        appropriate.
                   1413:
                   1414: Mon Oct 16 19:14:43 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1415:
                   1416:        * mpf/add.c: Rewrite.
                   1417:        * mpf/set_str.c: New file.  Needs more work.
                   1418:
                   1419: Sat Oct 14 00:14:04 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1420:
                   1421:        * mpf/div_2exp.c: Vastly simplify.
                   1422:        * mpf/mul_2exp.c: Likewise.
                   1423:
                   1424:        * mpf/sub.c: Rewrite.
                   1425:
                   1426:        * gmp-impl.h (udiv_qrnnd_preinv2gen): Terminate comment.
                   1427:
                   1428:        * mpf/dump.c: Free allocated memory.
                   1429:
                   1430:        * gmp-impl.h (assert): Define.
                   1431:
                   1432: Wed Oct 11 13:31:00 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1433:
                   1434:        * mpn/pentium/rshift.S: Install new code to optimize shift-by-1.
                   1435:
                   1436: Tue Oct 10 00:37:21 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1437:
                   1438:        * mpn/pentium/lshift.S: Install new code to optimize shift-by-1.
                   1439:
                   1440:        * mpn/powerpc32/{lshift.s,rshift.s}: New files.
                   1441:
                   1442:        * configure.in: Fix typo.
                   1443:
                   1444: Sat Oct  7 08:17:09 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1445:
                   1446:        * longlong.h (smul_ppmm): Correct type of __m0 and __m1.
                   1447:
                   1448: Wed Oct  4 16:31:28 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1449:
                   1450:        * mpn/configure.in: Handle alphaev5.
                   1451:        * mpn/ev4: New name for alpha subdir.
                   1452:        * mpn/ev5: New subdir.
                   1453:        * mpn/ev5/lshift.s: New file.
                   1454:
                   1455: Tue Oct  3 15:06:45 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1456:
                   1457:        * mpn/alpha/mul_1.s: Avoid static increments of pointers; use
                   1458:        corresponding offsets in ldq and stq instructions instead.
                   1459:        (Loop): Swap cmpult and stq to save one cycle on EV5.
                   1460:
                   1461:        * mpn/tests/{add_n.s,sub_n.s,lshift.s,rshift.s,mul_1.s,addmul_1.s,
                   1462:        submul_1.s}: Don't check results if NOCHECK is defined.
                   1463:
                   1464: Mon Oct  2 11:40:18 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1465:
                   1466:        * longlong.h (mips umul_ppmm [32 and 64 bit versions]):
                   1467:        Make new variants, based on GCC version number, that use `l' and `h'
                   1468:        constraints instead of explicit mflo and mfhi instructions
                   1469:
                   1470: Sun Oct  1 00:17:47 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1471:
                   1472:        * mpn/mc88100/add_n.s: Decrease unrolling factor from 16 to 8.
                   1473:        * mpn/mc88100/sub_n.s: Likewise.
                   1474:
                   1475:        * config/mt-m88110: New file.
                   1476:        * configure.in: Use it.
                   1477:
                   1478:        * mpn/mc88110/mul_1.s: Fix thinko.
                   1479:
                   1480: Sat Sep 30 21:28:19 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1481:
                   1482:        * mpz/set_d.c: Declare `size' at function start.
                   1483:
                   1484:        * experimental: New directory for mpx and mpz2.
                   1485:
                   1486:        * mpz/tdiv_q.c: Clarify comments.
                   1487:        * mpz/{mod.c,mod_ui.c}: New file, for math mod function.
                   1488:
                   1489:        * mpn/sh2/{mul_1.s,addmul_1.s,submul_1.s}: New files.
                   1490:
                   1491:        * mpn/sh/{add_n.s,sub_n.s}: New files.
                   1492:
                   1493:        * mpn/pyr/{add_n.s,sub_n.s,mul_1.s,addmul_1.s}: New files.
                   1494:
                   1495:        * mpn/i960/{add_n.s,sub_n.s}: New files.
                   1496:
                   1497:        * mpn/alpha/addmul_1.s (Loop): Move decrement of r18 to before umulh,
                   1498:        to save cycles on EV5.
                   1499:        * mpn/alpha/submul_1.s: Ditto.
                   1500:        * mpn/alpha/mul_1.s: Ditto.
                   1501:
                   1502: Thu Sep 28 02:48:59 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1503:
                   1504:        * gmp.h (mp_limb, mp_limb_signed): Define as `long long' if
                   1505:        _LONG_LONG_LIMB is defined.
                   1506:
                   1507:        * longlong.h (m88110): Test __m88110__, not __mc88110__
                   1508:
                   1509:        * mpn/mc88110/mul_1.s: Rewrite.
                   1510:
                   1511: Tue Sep 26 23:29:05 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1512:
                   1513:        * config.sub: Update from current Cygnus version.
                   1514:
                   1515:        * mpn/configure.in: Recognize canonical m88*, not mc88*.
                   1516:
                   1517: Fri Sep 22 14:58:05 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1518:
                   1519:        * mpz/set_d.c: New file.
                   1520:        * mpz/Makefile.in: Build new files.
                   1521:
                   1522:        * mpq/get_d.c: Replace usage of scalbn with ldexp.
                   1523:
                   1524:        * mpn/{vax,i386}/gmp-mparam.h: New files.
                   1525:        * gmp-impl.h (ieee_double_extract): Define here.
                   1526:        * mpf/set_d.c (ieee_double_extract): Not here.
                   1527:
                   1528: Thu Sep 21 00:56:36 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1529:
                   1530:        * longlong.h (C umul_ppmm): Use UWtype, not USItype for temps.
                   1531:        (udiv_qrnnd): For cases implemented with call to __udiv_qrnnd,
                   1532:        protect with new symbol LONGLONG_STANDALONE.
                   1533:        (68000 umul_ppmm): Use %# prefix for immediate constants.
                   1534:
                   1535: Wed Sep 20 15:36:23 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1536:
                   1537:        * mpn/generic/divmod_1.c: Handle
                   1538:        divisor_limb == 1 << (BITS_PER_MP_LIMB - 1)
                   1539:        specifically also when normalization_steps != 0.
                   1540:
                   1541: Mon Sep 18 15:42:30 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1542:
                   1543:        * mpq/get_d.c: New file.
                   1544:
                   1545: Sun Sep 17 02:04:36 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1546:
                   1547:        * longlong.h (pyr): Botch up for now.
                   1548:
                   1549: Sat Sep 16 00:11:50 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1550:
                   1551:        * mpn/clipper/mul_1.s: New file.
                   1552:        * mpn/clipper/add_n.s: New file.
                   1553:        * mpn/clipper/sub_n.s: New file.
                   1554:        * mpn/configure.in: Handle clipper*-*-*.
                   1555:
                   1556:        * mpn/configure.in: Recognize rs6000-*-*.
                   1557:
                   1558: Fri Sep 15 00:41:34 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1559:
                   1560:        * mpn/alpha/add_n.s: New file.
                   1561:        * mpn/alpha/sub_n.s: New file.
                   1562:
                   1563:        * mpn/mips3: New name for mpn/r4000.
                   1564:        * mpn/mips2: New name for mpn/r3000.
                   1565:        * mpn/configure.in: Corresponding changes.
                   1566:
                   1567:        * mpn/generic/perfsqr.c (primes): Delete.
                   1568:        (residue_map): Delete.
                   1569:
                   1570: Thu Sep 14 00:07:58 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1571:
                   1572:        * mpn/r3000/sub_n.s: Fix typo.
                   1573:
                   1574:        * dm_trunc.c: Delete spurious file.
                   1575:
                   1576:        * mpz/out_binary.c: Fix typo.
                   1577:
                   1578:        * mpn/configure.in (per-target): Make mips*-*-irix6* imply r4000.
                   1579:
                   1580:        * gmp-impl.h: For sparc and sgi, include alloca.h.
                   1581:
                   1582:        * mpn/z8000/mul_1.s: Replace `test r' with `and r,r'.  Replace
                   1583:        `ldk r,#0' with `xor r,r'.
                   1584:
                   1585: Wed Sep  6 00:58:38 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1586:
                   1587:        * mpz/inp_binary.c: New file.
                   1588:        * mpz/out_binary.c: New file.
                   1589:        * mpz/Makefile.in: Build new files.
                   1590:
                   1591: Tue Sep  5 22:53:51 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1592:
                   1593:        * gmp.h (__mpz_struct): Change `long int' => `mp_size_t' for alloc
                   1594:        and size fields.
                   1595:
                   1596: Sat Sep  2 17:47:59 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1597:
                   1598:        * mpn/r4000/{add_n.s,sub_n.s}: Optimize away some pointer arithmetic.
                   1599:        * mpn/r3000/{add_n.s,sub_n.s,lshift.s,rshift.s}: New files,
                   1600:        derived from r4000 code.
                   1601:
                   1602: Fri Sep  1 05:35:52 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1603:
                   1604:        * mpn/r3000/mul_1.s: Fix typo.
                   1605:
                   1606:        * mpn/powerpc32: Fix some old vs new mnemonic issues.
                   1607:
                   1608:        * mpn/powerpc32/{add_n.s,sub_n.s}: New files.
                   1609:        * mpn/r4000/{add_n.s,sub_n.s,lshift.s,rshift.s}: New files.
                   1610:
                   1611: Wed Aug 30 10:43:47 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1612:
                   1613:        * mpn/r3000/mul_1.s ($LC1): Use addiu for immediate add.
                   1614:        * mpn/r4000/{mul_1.s,addmul_1.s,submul_1.s}: New files.
                   1615:
                   1616:        * config.guess: Update to latest FSF revision.
                   1617:
                   1618: Mon Aug 28 02:18:13 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1619:
                   1620:        * mpz/out_str.c: Cast str to char * in fputs call.
                   1621:
                   1622:        * gmp-impl.h: Define UQItype, SItype, and USItype also
                   1623:        when not __GNUC__.
                   1624:
                   1625: Fri Aug 25 01:45:04 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1626:
                   1627:        * mpn/i386/syntax.h: Renamed from asm-syntax.h.
                   1628:        * mpn/mc68020/syntax.h: Renamed from asm-syntax.h.
                   1629:        * mpn/configure.in: Corresponding changes.
                   1630:
                   1631: Sun Aug 13 19:20:04 1995  Torbjorn Granlund  <tege@bozo.tmg.se>
                   1632:
                   1633:        * mpn/generic/random2.c: Test __hpux, not hpux.
                   1634:
                   1635: Sat Apr 15 20:50:33 1995  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1636:
                   1637:        * mpn/sparc/add_n.S: Make it work for PIC.
                   1638:        * mpn/sparc/sub_n.s: Likewise.
                   1639:        * mpn/sparc8/addmul_1.S: Likewise.
                   1640:        * mpn/sparc8/mul_1.S: Likewise.
                   1641:        * mpn/i386/add_n.S: Likewise.
                   1642:        * mpn/i386/sub_n.S: Likewise.
                   1643:
                   1644: Thu Apr 13 23:15:03 1995  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1645:
                   1646:        * mpn/configure.in: Don't search power subdir for generic ppc configs.
                   1647:        Add some ppc cpu-specific configs.  Misc clean up.
                   1648:
                   1649: Mon Apr 10 00:16:35 1995  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1650:
                   1651:        * mpz/ui_pow_ui.c: Delete spurious code to handle negative results.
                   1652:
                   1653: Sun Apr  9 12:38:11 1995  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1654:
                   1655:        * longlong.h (SPARC v8 udiv_qrnnd): Generate remainder in C,
                   1656:        not in asm.
                   1657:
                   1658:        * mpn/generic/sqrt.c (SQRT): Test for __SOFT_FLOAT.
                   1659:
                   1660: Tue Mar 28 00:19:52 1995  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1661:
                   1662:        * mpn/generic/hamdist.c (popc_limb): Make Mar 16 change here too.
                   1663:
                   1664: Fri Mar 17 23:29:22 1995  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1665:
                   1666:        * longlong.h (SH umul_ppmm): Define.
                   1667:
                   1668: Thu Mar 16 16:40:44 1995  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1669:
                   1670:        * mpn/generic/popcount.c (popc_limb): Rearrange 32 bit case
                   1671:        to help CSE.
                   1672:
                   1673: Fri Mar 10 20:03:49 1995  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1674:
                   1675:        * mpn/powerpc32/mul_1.s: Clear cy before entering loop.
                   1676:        Rearrange loop to save a cycle.
                   1677:        * mpn/powerpc32/addmul_1.s: New file.
                   1678:        * mpn/powerpc32/submul_1.s: New file.
                   1679:
                   1680: Fri Feb 17 22:44:45 1995  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1681:
                   1682:        * mpn/configure.in: Set target_makefile_frag for freebsd
                   1683:        in new case stmt.
                   1684:        * mpn/config/t-freebsd: New file.
                   1685:        * mpn/Makefile.in: Add #### for frag insertion.
                   1686:        (XCFLAGS): Clear by default.
                   1687:        (.c.o, .S.o rules): Pass XCFLAGS.
                   1688:
                   1689: Tue Feb  7 16:27:50 1995  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1690:
                   1691:        * longlong.h (68000 umul_ppmm): Merge improvements from henderson.
                   1692:
                   1693: Tue Jan 24 04:23:20 1995  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1694:
                   1695:        * longlong.h (default umul_ppmm): Store input parameters in temporaries
                   1696:        to avoid reading them twice.
                   1697:        (default smul_ppmm): New definition.
                   1698:
                   1699: Thu Dec 29 04:20:07 1994  Jim Meyering  (meyering@comco.com)
                   1700:
                   1701:        * generic/perfsqr.c (__mpn_perfect_square_p): Remove declaration
                   1702:        of unused variable.
                   1703:        * generic/pre_mod_1.c (__mpn_preinv_mod_1): Likewise.
                   1704:        * mpz/powm.c (pow): Likewise.
                   1705:
                   1706:        * mpz/and.c (mpz_and): Use {} instead of `;' for empty else clause
                   1707:        to placate `gcc -Wall'.
                   1708:        * mpz/ior.c (mpz_ior): Likewise.
                   1709:
                   1710: Wed Dec 28 13:31:40 1994  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1711:
                   1712:        * mpn/m*68*/*.S: #include asm-syntax.h, not asm.h.
                   1713:
                   1714: Mon Dec 26 17:15:36 1994  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1715:
                   1716:        * longlong.h: Test for more symbols, in __mc68000__ case.
                   1717:
                   1718:        * mpn/mpn/config.sub: Recognize m68060.
                   1719:        * mpn/configure.in: Change mc* to m* for 68k targets.
                   1720:        * mpn/Makefile.in (.S.o): Delete spurious creation of temp .c file.
                   1721:
                   1722: Mon Dec 19 01:56:30 1994  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1723:
                   1724:        * config.sub: Recognize pentium as a valid CPU.
                   1725:        * mpn/configure.in: Handle pentium specifically, to use new assembly
                   1726:        code.
                   1727:
                   1728: Mon Dec 19 00:13:01 1994  Jim Meyering  (meyering@comco.com)
                   1729:
                   1730:        * gmp.h: Define _GMP_H_HAVE_FILE if FILE, __STDIO_H__, or H_STDIO
                   1731:        is defined.
                   1732:        * gmp.h: test _GMP_H_HAVE_FILE instead of FILE everywhere else.
                   1733:
                   1734: Mon Dec 19 00:04:54 1994  Kent Boortz  (boortz@sics.se)
                   1735:
                   1736:        * Makefile.in (recursive makes): Pass CFLAGS.
                   1737:
                   1738: Sun Dec 18 22:34:49 1994  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1739:
                   1740:        * mpn/pentium: New directory.
                   1741:
                   1742:        * mpz/pprime.c: Make sure to mpz_clear all temporaries.
                   1743:
                   1744:        * longlong.h: Don't use udiv instruction when SUPERSPARC is defined.
                   1745:        * configure.in: Handle supersparc*-.
                   1746:        * config/mt-supspc-gcc: New file.
                   1747:        * config/mt-sparc8-gcc: New name for mt-sparcv8-gcc.
                   1748:
                   1749: Mon Dec 12 22:22:10 1994  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1750:
                   1751:        * mpn/i386/*.S: #include "asm-syntax.h", not "asm.h".
                   1752:        #include sysdep.h before asm-syntax.h.
                   1753:
                   1754:        * mpn/mc68020/asm-syntax.h: #undef ALIGN before defining it.
                   1755:        * mpn/i386/asm-syntax.h: Likewise.
                   1756:
                   1757:        * mpn/mc68020/asm-syntax.h: New name for asm.h.
                   1758:        * mpn/i386/asm-syntax.h: New name for asm.h.
                   1759:
                   1760: Tue Dec  6 21:55:25 1994  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1761:
                   1762:        * mpz/array_init.c: Fix typo in declaration.
                   1763:
                   1764: Fri Nov 18 19:50:52 1994  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1765:
                   1766:        * mpn/Makefile.in (.S.o): Pass CFLAGS and INCLUDES.
                   1767:
                   1768: Mon Nov 14 00:34:12 1994  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1769:
                   1770:        * mpn/generic/random2.c (random): Test for __svr4__.
                   1771:
                   1772: Wed Oct 12 23:28:16 1994  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1773:
                   1774:        * cre-conv-tab.c (main): Avoid upper-case X in printf format string.
                   1775:
                   1776: Tue Aug 23 17:16:35 1994  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1777:
                   1778:        * mpz/perfsqr.c: Use mpn_perfect_square_p.
                   1779:        * mpn/generic/perfsqr.c: New file.
                   1780:
                   1781: Wed Jul  6 13:46:51 1994  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1782:
                   1783:        * mpz/array_init.c: New file.
                   1784:        * mpz/Makefile.in: Compile array_init.
                   1785:        * gmp.h: Declare mpz_array_init.
                   1786:
                   1787: Mon Jul  4 01:10:03 1994  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1788:
                   1789:        * mpz/add.c: Fix bogus comment.
                   1790:        * mpz/sub.c: Likewise.
                   1791:
                   1792: Sat Jul  2 02:14:56 1994  Torbjorn Granlund  (tege@adder.cygnus.com)
                   1793:
                   1794:        * mpn/generic/pre_mod_1.c: New file.
                   1795:        * mpz/perfsqr.c: Use __mpn_preinv_mod_1 when faster.
                   1796:
                   1797: Fri Jul 01 22:10:19 1994  Richard Earnshaw (rwe11@cl.cam.ac.uk)
                   1798:
                   1799:        * longlong.h (arm umul_ppmm): Fix typos in last change.  Mark
                   1800:        hard-coded registers with "%|"
                   1801:
                   1802: Thu Jun 30 03:59:33 1994  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1803:
                   1804:        * mpz/perfsqr.c: Define PP, etc, for machines with 64 bit limbs.
                   1805:        Use __mpn_mod_1.
                   1806:        * mpz/perfsqr.c: Don't clobber REM in quadratic residue check loop.
                   1807:
                   1808: Wed Jun 29 18:45:41 1994  Torbjorn Granlund  (tege@adder.cygnus.com)
                   1809:
                   1810:        * mpn/generic/sqrt.c (SQRT): New asm for IBM POWER2.
                   1811:
                   1812:        * mpz/gcd_ui.c: Return 0 if result does not fit an unsigned long.
                   1813:
                   1814:        * gmp.h: Use "defined (__STDC__)" consistently.
                   1815:
                   1816: Tue Jun 28 18:44:58 1994  Torbjorn Granlund  (tege@adder.cygnus.com)
                   1817:
                   1818:        * gmp.h (mpz_get_si): Don't use "signed" keyword for return type.
                   1819:
                   1820:        * mpz/tests/Makefile.in: Use CFLAGS for linking.
                   1821:
                   1822:        * Makefile.in (CFLAGS): Use -O2 here.
                   1823:        * mpn/Makefile (CFLAGS): Not here.
                   1824:
                   1825:        * mpq/cmp_ui.c: Fix typo.
                   1826:        * mpq/canonicalize.c: Fix typo.
                   1827:        * mpz/gcd_ui.c: Handle gcd(0,v) and gcd(u,0) correctly.
                   1828:        * mpn/generic/gcd_1.c: Fix braino in last change.
                   1829:
                   1830: Mon Jun 27 16:10:27 1994  Torbjorn Granlund  (tege@rtl.cygnus.com)
                   1831:
                   1832:        * mpz/gcd_ui.c: Change return type and return result.
                   1833:        Allow destination param to be NULL.
                   1834:        * gmp.h: Corresponding change.
                   1835:        * mpn/generic/gcd_1.c: Handle zero return from mpn_mod_1.
                   1836:
                   1837: Tue Jun 14 02:17:43 1994  Torbjorn Granlund  (tege@tiny.cygnus.com)
                   1838:
                   1839:        * mpn/i386/asm.h (ALIGN): Make it take a parameter.
                   1840:        * mpn/i386/*.S: Use ALIGN to align all loops.
                   1841:
                   1842:        * mpn/i386/*.S: Move colon inside C_GLOBAL_NAME expression.
                   1843:        (Makes old versions of GAS happy.)
                   1844:
                   1845: Sat May 28 01:43:54 1994  Torbjorn Granlund  (tege@adder.cygnus.com)
                   1846:
                   1847:        * Many files: Delete unused variables and labels.
                   1848:        * mpn/generic/dump.c: cast printf width argument to int.
                   1849:
                   1850: Wed May 25 00:42:37 1994  Torbjorn Granlund  (tege@thepub.cygnus.com)
                   1851:
                   1852:        * mpz/gcd.c (mpz_gcd): Normalize after __mpn_sub calls.
                   1853:        (xmod): Ignore return value of __mpn_divmod.
                   1854:        (xmod): Improve normalization code.
                   1855:
                   1856: Sat May 21 01:30:09 1994  Torbjorn Granlund  (tege@adder.cygnus.com)
                   1857:
                   1858:        * mpz/gcdext.c: Cosmetic changes.
                   1859:
                   1860:        * mpz/fdiv_ui.c: New file.
                   1861:
                   1862: Fri May 20 00:24:53 1994  Torbjorn Granlund  (tege@adder.cygnus.com)
                   1863:
                   1864:        * mpz/tests/Makefile.in: Use explicit rules for running tests,
                   1865:        not a shell loop.
                   1866:        (clean): Delete stmp-*.
                   1867:
                   1868:        * mpz/Makefile.in: Update.
                   1869:
                   1870:        * mpz/div_ui.c: Don't include longlong.h.
                   1871:        * mpz/dm_ui.c: Likewise.
                   1872:
                   1873:        * mpz/fdiv_q.c, mpz/fdiv_q_ui.c, mpz/fdiv_qr.c, mpz/fdiv_qr_ui.c,
                   1874:        mpz/fdiv_r.c, mpz/fdiv_r_ui.c: New files.  Code partly from deleted
                   1875:        mdm.c, mdm_ui.c, etc, partly rewritten.
                   1876:        * mpz/dm_floor_ui.c, mpz/dm_floor.c: Delete.
                   1877:        * mpz/mdm.c, mpz/mdm_ui.c, mpz/mdiv.c, mpz/mdiv_ui.c, mpz/mmod.c,
                   1878:        mpz/mmod_ui.c: Delete.
                   1879:
                   1880:        * mpz/tdiv_q.c, mpz/tdiv_q_ui.c, mpz/tdiv_qr.c, mpz/tdiv_qr_ui.c,
                   1881:        mpz/tdiv_r.c, mpz/tdiv_r_ui.c:
                   1882:        New names for files implementing truncating division.
                   1883:        * mpz/div_ui.c, mpz/dm_ui.c, mpz/mod_ui.c: Simplify.
                   1884:
                   1885:        * mpn/Makefile.in (.S.o): Don't rely on CPP being defined, use CC
                   1886:        instead.
                   1887:        (clean): Delete tmp-*.
                   1888:
                   1889: Thu May 19 01:37:44 1994  Torbjorn Granlund  (tege@adder.cygnus.com)
                   1890:
                   1891:        * mpz/cmp.c: Call __mpn_cmp.
                   1892:
                   1893:        * mpz/popcount.c: Fix typo.
                   1894:
                   1895:        * mpz/powm_ui.c: Simplify main loop.  Keep principal operand size
                   1896:        smaller than MSIZE when possible.
                   1897:        * mpz/powm.c: Likewise.
                   1898:
                   1899:        * mpn/generic/sqrt.c: Move alloca calls into where the memory is
                   1900:        needed.  Simplify.
                   1901:
                   1902:        * gmp.h: (_PROTO): New macro.
                   1903:        Add many function declarations; use _PROTO macro in all declarations.
                   1904:
                   1905:        * mpf/*.c: Prepend mpn calls with __.
                   1906:
                   1907: Wed May 18 20:57:06 1994  Torbjorn Granlund  (tege@adder.cygnus.com)
                   1908:
                   1909:        * mpf/*ui*.c: Make ui argument `long' for consistency with mpz
                   1910:        functions.
                   1911:
                   1912:        * mpf/div_ui.c: Simplify.
                   1913:
                   1914: Tue May 17 01:05:14 1994  Torbjorn Granlund  (tege@adder.cygnus.com)
                   1915:
                   1916:        * mpz/*.c: Prepend mpn calls with __.
                   1917:
                   1918:        * mpz/mul_ui.c: Use mpn_mul_1.
                   1919:
                   1920: Mon May 16 17:19:41 1994  Torbjorn Granlund  (tege@adder.cygnus.com)
                   1921:
                   1922:        * mpn/i386/mul_1.S: Use C_GLOBAL_NAME.
                   1923:        * mpn/i386/mul_1.S, mpn/i386/addmul_1.S, mpn/i386/submul_1.S:
                   1924:        Nuke use of LAB.
                   1925:
                   1926: Sat May 14 14:21:02 1994  Torbjorn Granlund  (tege@adder.cygnus.com)
                   1927:
                   1928:        * gmp-impl.h: Don't define abort here.
                   1929:
                   1930:        * mpz/pow_ui.c: Increase temporary allocation.
                   1931:        * mpz/ui_pow_ui.c: Likewise.
                   1932:
                   1933:        * gmp.h (mpz_add_1, mpz_sub_1): Don't call memcpy.
                   1934:
                   1935:        * All Makefile.in: Delete spurious -I arguments.
                   1936:        Update dependencies.
                   1937:
                   1938:        * mpz/popcount.c: New file.
                   1939:        * mpz/hamdist.c: New file.
                   1940:
                   1941:        * All configure: Latest version from Cygnus.
                   1942:
                   1943:        * mpq/Makefile.in: New file.
                   1944:        * mpq/configure.in: New file.
                   1945:        * Makefile.in, configure.in: Enable compilation of mpq.
                   1946:
                   1947:        * mpq/set_z.c: Fix typos.
                   1948:        * mpq/canonicalize.c: Fix typos.
                   1949:        * mpq/cmp_ui.c: Fix typos.
                   1950:
                   1951:        * mpf/add_ui.c: Read U->D into UP always.  Delete spurious MPN_COPY.
                   1952:        * mpf/sub_ui.c: Likewise.
                   1953:
                   1954:        * gmp-impl.h: Don't redefine alloca.
                   1955:
                   1956:        * COPYING.LIB: Renamed from COPYING.
                   1957:
                   1958: Wed May 11 01:45:44 1994  Torbjorn Granlund  (tege@adder.cygnus.com)
                   1959:
                   1960:        * mpz/powm_ui.c: When shifting E left by C+1, handle out-of-range
                   1961:        shift counts.  Fix typo when testing negative_result.
                   1962:        * mpz/powm.c: Likewise.
                   1963:
                   1964:        * mpz/ui_pow_ui.c: New file.
                   1965:        * mpz/Makefile.in: Update.
                   1966:
                   1967:        * mpz/pow_ui.c: Call __mpn_mul_n instead of __mpn_mul when possible.
                   1968:
                   1969:        * mpz/div.c, mpz/div_ui.c, mpz/gcd.c: Prefix external mpn calls.
                   1970:        * mpz/gcd.c: Declare mpn_xmod.
                   1971:
                   1972:        * mpz/powm.c: Major changes to accommodate changed mpn semantics.
                   1973:        * mpz/powm_ui.c: Update from mpz/powm.c.
                   1974:
                   1975:        * mpz/tests/tst-io.c: New file.
                   1976:        * mpz/tests/tst-logic: New file.
                   1977:        * mpz/tests/Makefile.in: Update.
                   1978:
                   1979:        * mpz/inp_str.c: Get base right when checking for first digit.
                   1980:        * mpz/inp_str.c: Allocate more space for DEST when needed.
                   1981:
                   1982:        * mpz/com.c: Use mpn_add_1 and mpn_sub_1.
                   1983:        * mpz/and.c, mpz/ior.c: Likewise.  Simplify somewhat.
                   1984:
                   1985:        * mpz/add_ui.c: Use mpn_add_1 and mpn_sub_1.
                   1986:        Rename parameters to be consistent with mpz/sub_ui.
                   1987:        General simplifications.
                   1988:        * mpz/sub_ui.x: Likewise.
                   1989:
                   1990: Tue Aug 10 19:41:16 1993  Torbjorn Granlund  (tege@prudens.matematik.su.se)
                   1991:
                   1992:        * mpf: New directory.
                   1993:        * mpf/*.c: Merge basic set of mpf functions.
                   1994:
                   1995:        * Many logs missing...
                   1996:
                   1997: Sun Apr 25 18:40:26 1993  Torbjorn Granlund  (tege@pde.nada.kth.se)
                   1998:
                   1999:        * memory.c: Use #if instead of #ifdef for __STDC__ for consistency.
                   2000:        * bsd/xtom.c: Likewise.
                   2001:
                   2002:        * mpz/div.c: Remove free_me and free_me_size and their usage.
                   2003:        Use mpn_divmod for division; corresponding changes in return value
                   2004:        convention.
                   2005:        * mpz/powm.c: `carry_digit' => `carry_limb'.
                   2006:        * bsd/sdiv.c: Clearify comment.
                   2007:
                   2008: Sun Apr 25 00:31:28 1993  Torbjorn Granlund  (tege@pde.nada.kth.se)
                   2009:
                   2010:        * longlong.h (__udiv_qrnnd_c): Make all variables `unsigned long int'.
                   2011:
                   2012: Sat Apr 24 16:23:33 1993  Torbjorn Granlund  (tege@pde.nada.kth.se)
                   2013:
                   2014:        * longlong.h (__udiv_qrnnd_c): Make all variables `unsigned long int'.
                   2015:
                   2016:        * gmp-impl.h: #define ABS.
                   2017:        * (Many files): Use ABS instead of abs.
                   2018:
                   2019:        * mpn/generic/sqrt.c, mpz/clrbit.c, mpz/get_si.c, mpz/mod_2exp.c,
                   2020:        mpz/pow_ui.c: Cast 1 to mp_limb before shifting.
                   2021:
                   2022:        * mpz/perfsqr.c: Use #if, not plain if for exclusion of code for
                   2023:        non-32-bit machines.
                   2024:
                   2025: Tue Apr 20 13:13:58 1993  Torbjorn Granlund  (tege@du.nada.kth.se)
                   2026:
                   2027:        * mpn/generic/sqrt.c: Handle overflow for intermediate quotients by
                   2028:        rounding them down to fit.
                   2029:
                   2030:        * mpz/perfsqr.c (PP): Define in hexadecimal to avoid GCC warnings.
                   2031:
                   2032:        * mpz/inp_str.c (char_ok_for_base): New function.
                   2033:        (mpz_inp_str): Use it.
                   2034:
                   2035: Sun Mar 28 21:54:06 1993  Torbjorn Granlund  (tege@cyklop.nada.kth.se)
                   2036:
                   2037:        * mpz/inp_raw.c: Allocate x_index, not xsize limbs.
                   2038:
                   2039: Mon Mar 15 11:44:06 1993  Torbjorn Granlund  (tege@pde.nada.kth.se)
                   2040:
                   2041:        * mpz/pprime.c: Declare param `const'.
                   2042:        * gmp.h: Add declarations for mpz_com.
                   2043:
                   2044: Thu Feb 18 14:10:34 1993  Torbjorn Granlund  (tege@pde.nada.kth.se)
                   2045:
                   2046:        * mpq/add.c, mpq/sub.c: Call mpz_clear for t.
                   2047:
                   2048: Fri Feb 12 20:27:34 1993  Torbjorn Granlund  (tege@cyklop.nada.kth.se)
                   2049:
                   2050:        * mpz/inp_str.c: Recog minus sign as first character.
                   2051:
                   2052: Wed Feb  3 01:36:02 1993  Torbjorn Granlund  (tege@cyklop.nada.kth.se)
                   2053:
                   2054:        * mpz/iset.c: Handle 0 size.
                   2055:
                   2056: Tue Feb  2 13:03:33 1993  Torbjorn Granlund  (tege@cyklop.nada.kth.se)
                   2057:
                   2058:        * mpz/mod_ui.c: Initialize dividend_size before it's used.
                   2059:
                   2060: Mon Jan  4 09:11:15 1993  Torbjorn Granlund  (tege@sics.se)
                   2061:
                   2062:        * bsd/itom.c: Declare param explicitly 'signed'.
                   2063:        * bsd/sdiv.c: Likewise.
                   2064:
                   2065:        * mpq/cmp.c: Remove unused variable tmp_size.
                   2066:        * mpz/powm_ui.c: Fix typo in esize==0 if stmt.
                   2067:        * mpz/powm.c: Likewise.
                   2068:
                   2069: Sun Nov 29 01:16:11 1992  Torbjorn Granlund  (tege@sics.se)
                   2070:
                   2071:        * mpn/generic/divmod_1.c (mpn_divmod_1): Handle
                   2072:        divisor_limb == 1 << (BITS_PER_MP_LIMB - 1)
                   2073:        specifically.
                   2074:
                   2075:        * Reorganize sources.  New directories mpn, mpn/MACH, mpn/generic,
                   2076:        mpz, mpq, bsd.  Use full file name for change logs hereafter.
                   2077:
                   2078: Wed Oct 28 17:40:04 1992  Torbjorn Granlund  (tege@jupiter.sics.se)
                   2079:
                   2080:        * longlong.h (__hppa umul_ppmm): Fix typos.
                   2081:        (__hppa sub_ddmmss): Swap input arguments.
                   2082:
                   2083:        * mpz_perfsqr.c (mpz_perfect_square_p): Avoid , before } in
                   2084:        initializator.
                   2085:
                   2086: Sun Oct 25 20:30:06 1992  Torbjorn Granlund  (tege@jupiter.sics.se)
                   2087:
                   2088:        * mpz_pprime.c (mpz_probab_prime_p): Handle numbers <= 3
                   2089:        specifically (used to consider all negative numbers prime).
                   2090:
                   2091:        * mpz_powm_ui: `carry_digit' => `carry_limb'.
                   2092:
                   2093:        * sdiv: Handle zero dividend specifically.  Replace most code in
                   2094:        this function with a call to mpn_divmod_1.
                   2095:
                   2096: Fri Sep 11 22:15:55 1992  Torbjorn Granlund  (tege@tarrega.sics.se)
                   2097:
                   2098:        * mpq_clear: Don't free the MP_RAT!
                   2099:
                   2100:        * mpn_lshift, mpn_rshift, mpn_rshiftci: Remove `long' from 4:th arg.
                   2101:
                   2102: Thu Sep  3 01:47:07 1992  Torbjorn Granlund  (tege@jupiter.sics.se)
                   2103:
                   2104:        * All files: Remove leading _ from mpn function names.
                   2105:
                   2106: Wed Sep  2 22:21:16 1992  Torbjorn Granlund  (tege@jupiter.sics.se)
                   2107:
                   2108:        Fix from Jan-Hein Buhrman:
                   2109:        * mpz_mdiv.c, mpz_mmod.c, mpz_mdm.c: Make them work as documented.
                   2110:
                   2111:        * mpz_mmod.c, mpz_mdm.c: Move decl of TEMP_DIVISOR to reflect its
                   2112:        life.
                   2113:
                   2114: Sun Aug 30 18:37:15 1992  Torbjorn Granlund  (tege@jupiter.sics.se)
                   2115:
                   2116:        * _mpz_get_str: Use mpz_sizeinbase for computing out_len.
                   2117:        * _mpz_get_str: Don't remove leading zeros.  Abort if there are some.
                   2118:
                   2119: Wed Mar  4 17:56:56 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2120:
                   2121:        * gmp.h: Change definition of MP_INT to make the & before params
                   2122:        optional.  Use typedef to define it.
                   2123:        * mp.h: Use typedef to define MINT.
                   2124:
                   2125: Tue Feb 18 14:38:39 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2126:
                   2127:        longlong.h (hppa umul_ppmm): Add missing semicolon.  Declare type
                   2128:        of __w1 and __w0.
                   2129:
                   2130: Fri Feb 14 21:33:21 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2131:
                   2132:        * longlong.h: Make default count_leading_zeros work for machines >
                   2133:        32 bits.  Prepend `__' before local variables to avoid conflicts
                   2134:        with users' variables.
                   2135:
                   2136:        * mpn_dm_1.c: Remove udiv_qrnnd_preinv ...
                   2137:        * gmp-impl.h: ... and put it here.
                   2138:        * mpn_mod_1: Use udiv_qrnnd_preinv if it is faster than udiv_qrnnd.
                   2139:
                   2140: Tue Feb 11 17:20:12 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2141:
                   2142:        * mpn_mul: Enhance base case by handling small multiplicands.
                   2143:        * mpn_dm_1.c: Revert last change.
                   2144:
                   2145: Mon Feb 10 11:55:15 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2146:
                   2147:        * mpn_dm_1.c: Don't define udiv_qrnnd_preinv unless needed.
                   2148:
                   2149: Fri Feb  7 16:26:16 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2150:
                   2151:        * mpn_mul: Replace code for base case.
                   2152:
                   2153: Thu Feb  6 15:10:42 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2154:
                   2155:        * mpn_dm_1.c (_mpn_divmod_1): Add code for avoiding division by
                   2156:        pre-inverting divisor.
                   2157:
                   2158: Sun Feb  2 11:10:25 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2159:
                   2160:        * longlong.h: Make __LLDEBUG__ work differently.
                   2161:        (_IBMR2): Reinsert old code.
                   2162:
                   2163: Sat Feb  1 16:43:00 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2164:
                   2165:        * longlong.h (#ifdef _IBMR2): Replace udiv_qrnnd with new code
                   2166:        using floating point operations.  Don't define
                   2167:        UDIV_NEEDS_NORMALIZATION any longer.
                   2168:
                   2169: Fri Jan 31 15:09:13 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2170:
                   2171:        * longlong.h: Define UMUL_TIME and UDIV_TIME for most machines.
                   2172:        * longlong.h (#ifdef __hppa): Define umul_ppmm.
                   2173:
                   2174: Wed Jan 29 16:41:36 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2175:
                   2176:        * mpn_cmp: Only one length parameter, assume operand lengths are
                   2177:        the same.  Don't require normalization.
                   2178:        * mpq_cmp, mpz_add, mpz_sub, mpz_gcd, mpn_mul, mpn_sqrt: Change for
                   2179:        new mpn_cmp definition.
                   2180:
                   2181: Tue Jan 28 11:18:55 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2182:
                   2183:        * _mpz_get_str: Fix typo in comment.
                   2184:
                   2185: Mon Jan 27 09:44:16 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2186:
                   2187:        * Makefile.in: Add new files.
                   2188:
                   2189:        * mpn_dm_1.c: New file with function _mpn_divmod_1.
                   2190:        * mpz_dm_ui.c (mpz_divmod_ui): Use _mpn_divmod_1.
                   2191:        * mpz_div_ui: Likewise.
                   2192:
                   2193:        * mpn_mod_1.c: New file with function _mpn_mod_1.
                   2194:        * mpz_mod_ui: Use _mpn_mod_1.
                   2195:
                   2196: Thu Jan 23 18:54:09 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2197:
                   2198:        Bug found by Paul Zimmermann (zimmermann@inria.inria.fr):
                   2199:        * mpz_div_ui.c (mpz_div_ui), mpz_dm_ui.c (mpz_divmod_ui):
                   2200:        Handle dividend == 0.
                   2201:
                   2202: Wed Jan 22 12:02:26 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2203:
                   2204:        * mpz_pprime.c: Use "" for #include.
                   2205:
                   2206: Sun Jan 19 13:36:55 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2207:
                   2208:        * mpn_rshiftci.c (header): Correct comment.
                   2209:
                   2210: Wed Jan 15 18:56:04 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2211:
                   2212:        * mpz_powm, mpz_powm_ui (if (bsize > msize)): Do alloca (bsize + 1)
                   2213:        to make space for ignored quotient at the end.  (The quotient might
                   2214:        always be an extra limb.)
                   2215:
                   2216: Tue Jan 14 21:28:48 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2217:
                   2218:        * mpz_powm_ui: Fix comment.
                   2219:        * mpz_powm: Likewise.
                   2220:
                   2221: Mon Jan 13 18:16:25 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2222:
                   2223:        * tests/Makefile.in: Prepend $(TEST_PREFIX) to Makefile target.
                   2224:
                   2225: Sun Jan 12 13:54:28 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2226:
                   2227:        Fixes from Kazumaro Aoki:
                   2228:        * mpz_out_raw: Take abs of size to handle negative values.
                   2229:        * mpz_inp_raw: Reallocate before reading ptr from X.
                   2230:        * mpz_inp_raw: Store, don't read, size to x->size.
                   2231:
                   2232: Tue Jan  7 17:50:25 1992  Torbjorn Granlund  (tege@zevs.sics.se)
                   2233:
                   2234:        * gmp.h, mp.h: Remove parameter names from prototypes.
                   2235:
                   2236: Sun Dec 15 00:09:36 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2237:
                   2238:        * tests/Makefile.in: Prepend "./" to file names when executing
                   2239:        tests.
                   2240:
                   2241:        * Makefile.in: Fix many problems.
                   2242:
                   2243: Sat Dec 14 01:00:02 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2244:
                   2245:        * mpn_sqrt.c: New file with _mpn_sqrt.
                   2246:        * mpz_sqrt, mpz_sqrtrem, mpz_perfect_square_p: Use _mpn_sqrt.
                   2247:        * msqrt.c: Delete.  Create from mpz_sqrtrem.c in Makefile.in.
                   2248:        * mpz_do_sqrt.c: Delete.
                   2249:        * Makefile.in: Update to reflect these changes.
                   2250:
                   2251:        * Makefile.in, configure, configure.subr: New files
                   2252:        (from bothner@cygnus.com).
                   2253:        * dist-Makefile: Delete.
                   2254:
                   2255:        * mpz_fac_ui: Fix comment.
                   2256:
                   2257:        * mpz_random2: Rewrite a bit to make it possible for the most
                   2258:        significant limb to be == 1.
                   2259:
                   2260:        * mpz_pprime.c (mpz_probab_prime_p): Remove \t\n.
                   2261:
                   2262: Fri Dec 13 23:10:02 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2263:
                   2264:        * mpz_do_sqrt: Simplify special case for U == 0.
                   2265:        * m*sqrt*.c, mpz_perfsqr.c (mpz_perfect_square_p):
                   2266:          Rename _mpz_impl_sqrt to _mpz_do_sqrt.
                   2267:
                   2268: Fri Dec 13 12:52:28 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2269:
                   2270:        * gmp-impl.h (MPZ_TMP_INIT): Cast to the right type.
                   2271:
                   2272: Thu Dec 12 22:17:29 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2273:
                   2274:        * mpn_add, mpn_sub, mpn_mul, mpn_div: Change type of several
                   2275:        variables to mp_size.
                   2276:
                   2277: Wed Dec 11 22:00:34 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2278:
                   2279:        * mpn_rshift.c: Fix header comments.
                   2280:
                   2281: Mon Dec  9 17:46:10 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2282:
                   2283:        Released 1.2.
                   2284:
                   2285:        * gmp-impl.h (MPZ_TMP_INIT): Cast alloca return value.
                   2286:
                   2287:        * dist-Makefile: Add missing dependency for cre-mparam.
                   2288:
                   2289:        * mpz_mdiv.c, mpz_mmod.c, mpz_mdm.c, mpz_mdiv_ui.c,
                   2290:          mpz_mmod_ui.c, mpz_mdm_ui.c: Remove obsolete comment.
                   2291:
                   2292:        * dist-Makefile (clean): clean in tests subdir too.
                   2293:        * tests/Makefile: Define default values for ROOT and SUB.
                   2294:
                   2295:        * longlong.h (__a29k__ udiv_qrnnd): Change "q" to "1" for operand
                   2296:        2 constraint.
                   2297:
                   2298: Mon Nov 11 00:06:05 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2299:
                   2300:        * mpz_sizeinb.c (mpz_sizeinbase): Special code for size == 0.
                   2301:
                   2302: Sat Nov  9 23:47:38 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2303:
                   2304:        Released 1.1.94.
                   2305:
                   2306:        * dist-Makefile, Makefile, tests/Makefile: Merge tests into
                   2307:        distribution.
                   2308:
                   2309: Fri Nov  8 22:57:19 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2310:
                   2311:        * gmp.h: Don't use keyword `signed' for non-ANSI compilers.
                   2312:
                   2313: Thu Nov  7 22:06:46 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2314:
                   2315:        * longlong.h: Cosmetic changes to keep it identical to gcc2 version
                   2316:        of longlong.h.
                   2317:        * longlong.h (__ibm032__): Fix operand order for add_ssaaaa and
                   2318:        sub_ddmmss.
                   2319:
                   2320: Mon Nov  4 00:36:46 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2321:
                   2322:        * mpn_mul: Fix indentation.
                   2323:
                   2324:        * mpz_do_sqrt: Don't assume 32 bit limbs (had constant
                   2325:        4294967296.0).
                   2326:        * mpz_do_sqrt: Handle overflow in conversion from double returned
                   2327:        by SQRT to mp_limb.
                   2328:
                   2329:        * gmp.h: Add missing function definitions.
                   2330:
                   2331: Sun Nov  3 18:25:25 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2332:
                   2333:        * mpz_pow_ui: Change type of `i' to int.
                   2334:
                   2335:        * ChangeLog: Add change log entry.
                   2336:        * ChangeLog: Add change log entry.
                   2337:        * ChangeLog: Add change log entry.
                   2338:        * ChangeLog: Add change log entry.
                   2339:        * ChangeLog: Add change log entry.
                   2340:        * ChangeLog: Add change log entry.
                   2341:        * ChangeLog: Add change log entry.
                   2342:        * ChangeLog: Add change log entry.
                   2343: Stack overflow.
                   2344:
                   2345:        * mpz_pow_ui.c: Fix typo in comment.
                   2346:
                   2347:        * dist-Makefile: Create rpow.c from mpz_powm_ui.c.
                   2348:        * mpz_powm_ui.c: Add code for rpow.
                   2349:        * rpow.c: Delete this file.  The rpow function is now implemented
                   2350:        in mpz_powm_ui.c.
                   2351:
                   2352:        * mpz_fac_ui.c: New file.
                   2353:        * gmp.h, dist-Makefile: Add stuff for mpz_fac_ui.
                   2354:
                   2355:        Bug found by John Amanatides (amana@sasquatch.cs.yorku.ca):
                   2356:        * mpz_powm_ui, mpz_powm: Call _mpn_mul in the right way, with
                   2357:        the first argument not smaller than the second.
                   2358:
                   2359: Tue Oct 29 13:56:55 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2360:
                   2361:        * cre-conv-tab.c (main), cre-mparam.c (main): Fix typo in output
                   2362:        header text.
                   2363:
                   2364: Mon Oct 28 00:35:29 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2365:
                   2366:        * mpz_random2: Handle size == 0.
                   2367:
                   2368:        * gmp-impl.h (struct __mp_bases): Rename chars_per_limb_exactly to
                   2369:        chars_per_bit_exactly, and change its definition.
                   2370:        * cre-conv-tab.c (main): Output field according to its new
                   2371:        definition.
                   2372:        * mpz_out_str, _mpz_get_str, mpz_sizeinb, mout:
                   2373:        Use chars_per_bit_exactly.
                   2374:
                   2375:        * mpz_random2: Change the loop termination condition in order to
                   2376:        get a large most significant limb with higher probability.
                   2377:
                   2378:        * gmp.h: Add declaration of new mpz_random2 and mpz_get_si.
                   2379:        * mpz_get_si.c: New file.
                   2380:        * dist-Makefile: Add mpz_random2 and mpz_get_si.
                   2381:
                   2382:        * mpz_sizeinb.c (mpz_sizeinbase): Special code for base being a
                   2383:        power of 2, giving exact result.
                   2384:
                   2385:        * mpn_mul: Fix MPN_MUL_VERIFY in various ways.
                   2386:        * mpn_mul: New macro KARATSUBA_THRESHOLD.
                   2387:        * mpn_mul (karatsuba's algorithm): Don't write intermediate results
                   2388:        to prodp, use temporary pp instead.  (Intermediate results can be
                   2389:        larger than the final result, possibly writing into hyperspace.)
                   2390:        * mpn_mul: Make smarter choice between Karatsuba's algorithm and the
                   2391:        shortcut algorithm.
                   2392:        * mpn_mul: Fix typo, cy instead of xcy.  Unify carry handling code.
                   2393:
                   2394: Sun Oct 27 19:57:32 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2395:
                   2396:        * mpn_mul: In non-classical case, choose Karatsuba's algorithm only
                   2397:        when usize > 1.5 vsize.
                   2398:
                   2399:        * mpn_mul: Break between classical and Karatsuba's algorithm at
                   2400:        KARATSUBA_THRESHOLD, if defined.  Default to 8.
                   2401:
                   2402:        * mpn_div: Kludge to fix stray memory read.
                   2403:
                   2404: Sat Oct 26 20:06:14 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2405:
                   2406:        * mpz_gcdext: Handle a = b = 0.  Remove memory leakage by calling
                   2407:        mpz_clear for all temporary variables.
                   2408:
                   2409:        * mpz_gcd: Reduce w_bcnt in _mpn_lshift call to hold that
                   2410:        function's argument constraints.  Compute wsize correctly.
                   2411:
                   2412:        * mpz_gcd: Fix typo in comment.
                   2413:
                   2414:        * memory.c (_mp_default_allocate, _mp_default_reallocate): Call
                   2415:        abort if allocation fails, don't just exit.
                   2416:
                   2417: Fri Oct 25 22:17:20 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2418:
                   2419:        * mpz_random2.c: New file.
                   2420:
                   2421: Thu Oct 17 18:06:42 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2422:
                   2423:        Bugs found by Pierre-Joseph Gailly (pjg@sunbim.be):
                   2424:        * mpq_cmp: Take sign into account, don't just compare the
                   2425:        magnitudes.
                   2426:        * mpq_cmp: Call _mpn_mul in the right way, with the first argument
                   2427:        not smaller than the second.
                   2428:
                   2429: Wed Oct 16 19:27:32 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2430:
                   2431:        * mpz_random: Ensure the result is normalized.
                   2432:
                   2433: Tue Oct 15 14:55:13 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2434:
                   2435:        * mpz_clrbit: Support non-ANSI compilers.
                   2436:
                   2437: Wed Oct  9 18:03:28 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2438:
                   2439:        * longlong.h (68k add_ssaaaa, sub_ddmmss): Generalize constraints.
                   2440:
                   2441: Tue Oct  8 17:42:59 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2442:
                   2443:        * mpz_mdm_ui: Add comments.
                   2444:
                   2445:        * mpz_mdiv: Use MPZ_TMP_INIT instead of mpz_init.
                   2446:        * mpz_init_ui: Change spacing and header comment.
                   2447:
                   2448: Thu Oct  3 18:36:13 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2449:
                   2450:        * dist-Makefile: Prepend `./' before some filenames.
                   2451:
                   2452: Sun Sep 29 14:02:11 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2453:
                   2454:        Released 1.1 (public).
                   2455:
                   2456:        * mpz_com: New name of mpz_not.
                   2457:        * dist-Makefile: Change mpz_not to mpz_com.
                   2458:
                   2459: Tue Sep 24 12:44:11 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2460:
                   2461:        * longlong.h: Fix header comment.
                   2462:
                   2463: Mon Sep  9 15:16:24 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2464:
                   2465:        Released 1.0.92.
                   2466:
                   2467:        * mpn_mul.c (_mpn_mul): Handle leading zero limbs in non-Karatsuba
                   2468:        case.
                   2469:
                   2470:        * longlong.h (m68000 umul_ppmm): Clobber one register less by
                   2471:        slightly rearranging the code.
                   2472:
                   2473: Sun Sep  1 18:53:25 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2474:
                   2475:        * dist-Makefile (stamp-stddefh): Fix typo.
                   2476:
                   2477: Sat Aug 31 20:41:31 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2478:
                   2479:        Released 1.0.91.
                   2480:
                   2481:        * mpz_mdiv.c, mpz_mmod.c, mpz_mdm.c, mpz_mdiv_ui.c,
                   2482:          mpz_mmod_ui.c, mpz_mdm_ui.c: New files and functions.
                   2483:        * gmp.h, gmp.texi: Define the new functions.
                   2484:
                   2485: Fri Aug 30 08:32:56 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2486:
                   2487:        * mpz_gcdext: Compute t argument from the other quantities at the
                   2488:          end, of the function, not in the loop.  New feature: Allow t to be
                   2489:          NULL.
                   2490:
                   2491:        * mpz_add.c, mpz_sub.c, mpz_mul.c, mpz_powm.c, mpz_gcd.c: Don't
                   2492:          include "mp.h".  Use type name `MP_INT' always.
                   2493:
                   2494:        * dist-Makefile, mpz_cmp.c: Merge mcmp.c from mpz_cmp.c.
                   2495:
                   2496: Wed Aug 28 00:45:11 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2497:
                   2498:        * dist-Makefile (documentation): Go via tmp.texi to avoid the
                   2499:          creation of gmp.dvi if any errors occur.  Make tex read input
                   2500:          from /dev/null.
                   2501:
                   2502: Fri Aug 23 15:58:52 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2503:
                   2504:        * longlong.h (68020, i386): Don't define machine-dependent
                   2505:          __umulsidi3 (so the default definition is used).
                   2506:        * longlong.h (all machines): Cast all operands, sources and
                   2507:          destinations, to `unsigned long int'.
                   2508:        * longlong.h: Add gmicro support.
                   2509:
                   2510: Thu Aug 22 00:28:29 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2511:
                   2512:        * longlong.h: Rename BITS_PER_LONG to LONG_TYPE_SIZE.
                   2513:        * longlong.h (__ibm032__): Define count_leading_zeros and umul_ppmm.
                   2514:        * longlong.h: Define UMUL_TIME and UDIV_TIME for some CPUs.
                   2515:        * _mpz_get_str.c: Add code to do division by big_base using only
                   2516:          umul_qrnnd, if that is faster.  Use UMUL_TIME and UDIV_TIME to
                   2517:          decide which variant to use.
                   2518:
                   2519: Wed Aug 21 15:45:23 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2520:
                   2521:        * longlong.h (__sparc__ umul_ppmm): Move two insn from end to the
                   2522:          nops.  (Saves two insn.)
                   2523:
                   2524:        * longlong.h (__sparc__ umul_ppmm): Rewrite in order to avoid
                   2525:          branch, and to permit input/output register overlap.
                   2526:
                   2527:        * longlong.h (__29k__): Remove duplicated udiv_qrnnd definition.
                   2528:        * longlong.h (__29k__ umul_ppmm): Split asm instructions into two
                   2529:          asm statements (gives better code if either the upper or lower
                   2530:          part of the product is unused.
                   2531:
                   2532: Tue Aug 20 17:57:59 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2533:
                   2534:        * _mpz_get_str.c (outside of functions): Remove
                   2535:          num_to_ascii_lower_case and num_to_ascii_upper_case.  Use string
                   2536:          constants in the function instead.
                   2537:
                   2538: Mon Aug 19 00:37:42 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2539:
                   2540:        * cre-conv-tab.c (main): Output table in hex.  Output 4 fields, not
                   2541:          3, for components 0 and 1.
                   2542:
                   2543:        * gmp.h: Add declaration of mpq_neg.
                   2544:
                   2545:        Released 1.0beta.13.
                   2546:
                   2547:        * _mpz_set_str.c (mpz_set_str): Cast EOF and SPC to char before
                   2548:          comparing to enum literals SPC and EOF.  This makes the code work
                   2549:          for compilers where `char' is unsigned.  (Bug found by Brian
                   2550:          Beuning).
                   2551:
                   2552:        Released 1.0beta.12.
                   2553:
                   2554:        * mpz_mod_ui: Remove references to quot.  Remove quot_ptr, quot_size
                   2555:          declarations and assignment code.
                   2556:
                   2557: Sun Aug 18 14:44:26 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2558:
                   2559:        * mpz_mod_ui: Handle dividend < 0.
                   2560:
                   2561:        Released 1.0beta.11.
                   2562:
                   2563:        * mpz_dm_ui, mpz_div_ui, mpz_mod_ui, sdiv: Make them share the same
                   2564:          general structure, variable names, etc.
                   2565:
                   2566:        * sdiv: Un-normalize the remainder in n1 before it is negated.
                   2567:
                   2568:        * longlong.h: Mention UDIV_NEEDS_NORMALIZATION in description of
                   2569:          udiv_qrnnd.
                   2570:
                   2571:        * mpz_dm_ui.c (mpz_divmod_ui), mpz_div_ui.c (mpz_div_ui): Increment
                   2572:          the quotient size if the dividend size is incremented.  (Bug found
                   2573:          by Brian Beuning.)
                   2574:
                   2575:        * mpz_mod_ui: Shift back the remainder, if UDIV_NEEDS_NORMALIZATION.
                   2576:          (Bug found by Brian Beuning.)
                   2577:
                   2578:        * mpz_mod_ui: Replace "digit" by "limb".
                   2579:
                   2580:        * mpz_perfsqr.c (mpz_perfect_square_p): Disable second test case
                   2581:          for non-32-bit machines (PP is hardwired for such machines).
                   2582:        * mpz_perfsqr.c (outside of functions): Define PP value with an L.
                   2583:
                   2584:        * mpn_mul.c (_mpn_mul): Add verification code that is activated if
                   2585:          DEBUG is defined.  Replace "digit" by "limb".
                   2586:        * mpn_mul.c (_mpn_mul: Karatsuba's algorithm: 4.): Normalize temp
                   2587:          after the addition.
                   2588:        * mpn_mul.c (_mpn_mul: Karatsuba's algorithm: 1.): Compare u0_size
                   2589:          and v0_size, and according to the result, swap arguments in
                   2590:          recursive call.  (Don't violate mpn_mul's own argument
                   2591:          constraints.)
                   2592:
                   2593: Fri Aug 16 13:47:12 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2594:
                   2595:        Released 1.0beta.10.
                   2596:
                   2597:        * longlong.h (IBMR2): Add udiv_qrnnd.
                   2598:
                   2599:        * mpz_perfsqr: Remove unused variables.
                   2600:
                   2601:        * mpz_and (case for different signs): Initialize loop variable i!
                   2602:
                   2603:        * dist-Makefile: Update automatically generated dependencies.
                   2604:        * dist-Makefile (madd.c, msub.c, pow.c, mult.c, gcd.c): Add mp.h,
                   2605:          etc to dependency file lists.
                   2606:
                   2607:        * longlong.h (add_ssaaaa, sub_ddmmss [C default versions]): Make __x
                   2608:          `unsigned long int'.
                   2609:        * longlong.h: Add `int' after `unsigned' and `long' everywhere.
                   2610:
                   2611: Wed Aug 14 18:06:48 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2612:
                   2613:        * longlong.h: Add ARM, i860 support.
                   2614:
                   2615:        * mpn_lshift, mpn_rshift, mpn_rshiftci: Rename *_word with *_limb.
                   2616:
                   2617: Tue Aug 13 21:57:43 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2618:
                   2619:        * _mpz_get_str.c, _mpz_set_str.c, mpz_sizeinb.c (mpz_sizeinbase),
                   2620:          mpz_out_str.c, mout.c: Remove declaration of __mp_bases.
                   2621:        * gmp-impl.h: Put it here, and make it `const'.
                   2622:        * cre-conv-tab.c (main): Make struct __mp_bases `const'.
                   2623:
                   2624: Mon Aug 12 17:11:46 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2625:
                   2626:        * cre-conv-tab.c (main): Use %lu in printf for long ints.
                   2627:
                   2628:        * dist-Makefile: Fix cre-* dependencies.
                   2629:
                   2630:        * cre-conv-tab.c (main): Output field big_base_inverted.
                   2631:
                   2632:        * gmp-impl.h (struct bases): New field big_base_inverted.
                   2633:        * gmp-impl.h (struct bases): Change type of chars_per_limb_exactly
                   2634:          to float (in order to keep the structure smaller).
                   2635:
                   2636:        * mp.h, gmp.h: Change names of macros for avoiding multiple
                   2637:          includes.
                   2638:
                   2639: Fri Aug  9 18:01:36 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2640:
                   2641:        * _mpz_get_str: Only shift limb array if normalization_steps != 0
                   2642:          (optimization).
                   2643:
                   2644:        * longlong.h (sparc umul_ppmm): Use __asm__, not asm.
                   2645:        * longlong.h (IBMR2 umul_ppmm): Refer to __m0 and __m1, not to m0
                   2646:          and m1 (overlap between output and input operands did not work).
                   2647:        * longlong.h: Add VAX, ROMP and HP-PA support.
                   2648:        * longlong.h: Sort the machine dependent code in alphabetical order
                   2649:          on the CPU name.
                   2650:        * longlong.h: Hack comments.
                   2651:
                   2652: Thu Aug  8 14:13:36 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2653:
                   2654:        Released 1.0beta.9.
                   2655:
                   2656:        * longlong.h: Define BITS_PER_LONG to 32 if it's not already
                   2657:          defined.
                   2658:        * Define __BITS4 to BITS_PER_LONG / 4.
                   2659:        * Don't assume 32 bit word size in "count_leading_zeros" C macro.
                   2660:          Use __BITS4 and BITS_PER_LONG instead.
                   2661:
                   2662:        * longlong.h: Don't #undef internal macros (reverse change of Aug 3).
                   2663:
                   2664:        * longlong.h (68k): Define add_ssaaaa sub_ddmmss, and umul_ppmm
                   2665:          even for plain mc68000.
                   2666:
                   2667:        * mpq_div: Flip the sign of the numerator *and* denominator of the
                   2668:          result if the intermediate denominator is negative.
                   2669:
                   2670:        * mpz_and.c, mpz_ior.c: Use MPN_COPY for all copying operations.
                   2671:
                   2672:        * mpz_and.c: Compute the result size more conservatively.
                   2673:        * mpz_ior.c: Likewise.
                   2674:
                   2675:        * mpz_realloc: Never allocate zero space even if NEW_SIZE == 0.
                   2676:
                   2677:        * dist-Makefile: Remove madd.c, msub.c, pow.c, mult.c, gcd.c from
                   2678:          BSDMP_SRCS.
                   2679:
                   2680:        * dist-Makefile: Create mult.c from mpz_mul.c.
                   2681:        * mult.c: Delete this file.
                   2682:
                   2683:        * _mpz_set_str: Normalize the result (for bases 2, 4, 8... it was
                   2684:          not done properly if the input string had many leading zeros).
                   2685:
                   2686: Sun Aug  4 16:54:14 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2687:
                   2688:        * dist-Makefile (gcd.c, pow.c, madd.c, msub.c): Make these targets
                   2689:          work with VPATH and GNU MP.
                   2690:
                   2691:        * mpz_gcd: Don't call mpz_set; inline its functionality.
                   2692:
                   2693:        * mpq_mul, mpq_div: Fix several serious typos.
                   2694:
                   2695:        * mpz_dmincl, mpz_div: Don't normalize the quotient if it's already
                   2696:          zero.
                   2697:
                   2698:        * mpq_neg.c: New file.
                   2699:
                   2700:        * dist-Makefile: Remove obsolete dependencies.
                   2701:
                   2702:        * mpz_sub: Fix typo.
                   2703:
                   2704:        Bugs found by Pierre-Joseph Gailly (pjg@sunbim.be):
                   2705:        * mpq_mul, mpq_div: Initialize tmp[12] variables even when the gcd
                   2706:          is just 1.
                   2707:        * mpz_gcd: Handle gcd(0,v) and gcd(u,0) in special cases.
                   2708:
                   2709: Sat Aug  3 23:45:28 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2710:
                   2711:        * longlong.h: Clean up comments.
                   2712:        * longlong.h: #undef internal macros.
                   2713:
                   2714: Fri Aug  2 18:29:11 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2715:
                   2716:        * mpq_set_si, mpq_set_ui: Canonicalize 0/x to 0/1.
                   2717:        * mpq_set_si, mpq_set_ui: Cosmetic formatting changes.
                   2718:
                   2719:        * mpz_dmincl.c: Normalize the remainder before shifting it back.
                   2720:
                   2721:        * mpz_dm_ui.c (mpz_divmod_ui): Handle rem == dividend.
                   2722:
                   2723:        * mpn_div.c: Fix comment.
                   2724:
                   2725:        * mpz_add.c, mpz_sub.c: Use __MP_INT (not MP_INT) for intermediate
                   2726:          type, in order to work for both GNU and Berkeley functions.
                   2727:
                   2728:        * dist-Makefile: Create gcd.c from mpz_gcd.c, pow.c from mpz_powm,
                   2729:          madd.c from mpz_add.c, msub.c from mpz_sub.c.
                   2730:          respectively.
                   2731:        * pow.c, gcd.c, mpz_powmincl.c, madd.c, msub.c: Remove these.
                   2732:        * mpz_powm.c, mpz_gcd.c, mpz_add.c, mpz_sub.c: #ifdef for GNU and
                   2733:          Berkeley function name variants.
                   2734:        * dist-Makefile: Add created files to "clean" target.
                   2735:
                   2736: Tue Jul 16 15:19:46 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2737:
                   2738:        * mpq_get_den: No need for absolute value of the size, the
                   2739:          denominator is always positive.
                   2740:
                   2741:        * mpz_get_ui: If the operand is zero, return zero.  Don't read the
                   2742:          limb array!
                   2743:
                   2744:        * mpz_dmincl.c: Don't ignore the return value from _mpn_rshift, it
                   2745:          is the size of the remainder.
                   2746:
                   2747: Mon Jul 15 11:08:05 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2748:
                   2749:        * Several files: Remove unused variables and functions.
                   2750:
                   2751:        * gmp-impl.h: Declare _mpz_impl_sqrt.
                   2752:
                   2753:        * mpz_dm_ui (mpz_divmod_ui), sdiv: Shift back the remainder if
                   2754:          UDIV_NEEDS_NORMALIZATION.  (Fix from Brian Beuning.)
                   2755:
                   2756:        * mpz_dm_ui.c, sdiv: Replace *digit with *limb.
                   2757:
                   2758:        * mpz_ior: Add missing else statement in -OP1 | -OP2 case.
                   2759:        * mpz_ior: Add missing else statement in OP1 | -OP2 case.
                   2760:        * mpz_ior: Swap also OP1 and OP2 pointers in -OP1 & OP2 case.
                   2761:        * mpz_ior: Duplicate _mpz_realloc code.
                   2762:
                   2763:        * mpz_and: Add missing else statement in -OP1 & -OP2 case.
                   2764:        * mpz_and: Rewrite OP1 & -OP2 case.
                   2765:        * mpz_and: Swap also OP1 and OP2 pointers in -OP1 & OP2 case.
                   2766:
                   2767:        * mpz_gcdext: Loop in d1.size (not b->size).  (Fix from Brian
                   2768:          Beuning.)
                   2769:
                   2770:        * mpz_perfsqr: Fix argument order in _mpz_impl_sqrt call.  (Fix from
                   2771:          Brian Beuning.)
                   2772:
                   2773: Fri Jul 12 17:10:33 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2774:
                   2775:        * mpq_set.c, mpq_set_ui.c, mpq_set_si.c, mpq_inv.c,
                   2776:          mpq_get_num.c, mpq_get_den.c, mpq_set_num.c, mpq_set_den.c:
                   2777:          New files.
                   2778:
                   2779:        * mpz_dmincl.c: Remove second re-allocation of rem->d.  It
                   2780:          was never executed.
                   2781:
                   2782:        * dist-Makefile: Use `-r' instead of `-x' for test for ranlib (as
                   2783:          some unixes' test doesn't have the -r option).
                   2784:
                   2785:        * *.*: Cast allocated pointers to the appropriate type (makes old C
                   2786:          compilers happier).
                   2787:
                   2788:        * cre-conv-tab.c (main): Divide max_uli by 2 and multiply again
                   2789:          after conversion to double.  (Kludge for broken C compilers.)
                   2790:
                   2791:        * dist-Makefile (stamp-stddefh): New target.  Test if "stddef.h"
                   2792:          exists in the system and creates a minimal one if it does not
                   2793:          exist.
                   2794:        * cre-stddefh.c: New file.
                   2795:        * dist-Makefile: Make libgmp.a and libmp.a depend on stamp-stddefh.
                   2796:        * dist-Makefile (clean): Add some more.
                   2797:        * gmp.h, mp.h: Unconditionally include "stddef.h".
                   2798:
                   2799: Thu Jul 11 10:08:21 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2800:
                   2801:        * min: Do ungetc of last read character.
                   2802:        * min.c: include stdio.h.
                   2803:
                   2804:        * dist-Makefile: Go via tmp- files for cre* redirection.
                   2805:        * dist-Makefile: Add tmp* to "clean" target.
                   2806:
                   2807:        * dist-Makefile: Use LOCAL_CC for cre*, to simplyfy cross
                   2808:          compilation.
                   2809:
                   2810:        * gmp.h, mp.h: Don't define NULL here.
                   2811:        * gmp-impl.h: Define it here.
                   2812:
                   2813: Wed Jul 10 14:13:33 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2814:
                   2815:        * mpz_mod_2exp: Don't copy too much, overwriting most significant
                   2816:          limb.
                   2817:
                   2818:        * mpz_and, mpz_ior: Don't read op[12]_ptr from op[12] when
                   2819:          reallocating res, if op[12]_ptr got their value from alloca.
                   2820:
                   2821:        * mpz_and, mpz_ior: Clear up comments.
                   2822:
                   2823:        * cre-mparam.c: Output parameters for `short int' and `int'.
                   2824:
                   2825:        * mpz_and, mpz_ior: Negate negative op[12]_size in several places.
                   2826:
                   2827: Tue Jul  9 18:40:30 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2828:
                   2829:        * gmp.h, mp.h: Test for _SIZE_T defined before typedef'ing size_t.
                   2830:          (Fix for Sun lossage.)
                   2831:
                   2832:        * gmp.h: Add declaration of mpq_clear.
                   2833:
                   2834:        * dist-Makefile: Chack if "ranlib" exists, before using it.
                   2835:        * dist-Makefile: Add mpz_sqrtrem.c and mpz_size.c.
                   2836:        * mpz_powm: Fix typo, "pow" instead of "mpz_powm".
                   2837:
                   2838: Fri Jul  5 19:08:09 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2839:
                   2840:        * move: Remove incorrect comment.
                   2841:
                   2842:        * mpz_free, mpq_free: Rename to *_clear.
                   2843:        * dist-Makefile: Likewise.
                   2844:        * mpq_add, mpq_sub, mpq_mul, mpq_div: Likewise.
                   2845:
                   2846:        * mpz_dmincl.c: Don't call "move", inline its functionality.
                   2847:
                   2848: Thu Jul  4 00:06:39 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2849:
                   2850:        * Makefile: Include dist-Makefile.  Fix dist target to include
                   2851:          dist-Makefile (with the name "Makefile" in the archive).
                   2852:
                   2853:        * dist-Makefile: New file made from Makefile.  Add new mpz_...
                   2854:          functions.
                   2855:
                   2856:        * mpz_powincl.c New file for mpz_powm (Berkeley MP pow)
                   2857:          functionality.  Avoids code duplication.
                   2858:        * pow.c, mpz_powm.c: Include mpz_powincl.c
                   2859:
                   2860:        * mpz_dmincl.c: New file containing general division code.  Avoids
                   2861:          code duplication.
                   2862:        * mpz_dm.c (mpz_divmod), mpz_mod.c (mpz_mod), mdiv.c (mdiv): Include
                   2863:          mpz_dmincl.c.
                   2864:
                   2865:        * _mpz_get_str: Don't call memmove, unless HAS_MEMMOVE is defined.
                   2866:          Instead, write the overlapping memory copying inline.
                   2867:
                   2868:        * mpz_dm_ui.c: New name for mpz_divmod_ui.c (SysV file name limit).
                   2869:
                   2870:        * longlong.h: Don't use #elif.
                   2871:        * mpz_do_sqrt.c: Likewise.
                   2872:
                   2873:        * longlong.h: Use __asm__ instead of asm.
                   2874:        * longlong.h (sparc udiv_qrnnd): Make it to one string over several
                   2875:          lines.
                   2876:
                   2877:        * longlong.h: Preend __ll_ to B, highpart, and lowpart.
                   2878:
                   2879:        * longlong.h: Move array t in count_leading_zeros to the new file
                   2880:          mp_clz_tab.c.  Rename the array __clz_tab.
                   2881:        * All files: #ifdef for traditional C compatibility.
                   2882:
                   2883: Wed Jul  3 11:42:14 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2884:
                   2885:        * mpz_and: Initialize res_ptr always (used to be initialized only
                   2886:          when reallocating).
                   2887:
                   2888:        * longlong.h (umul_ppmm [C variant]): Make __ul...__vh
                   2889:          `unsigned int', and cast the multiplications.  This way
                   2890:          compilers more easily can choose cheaper multiplication
                   2891:          instructions.
                   2892:
                   2893:        * mpz_mod_2exp: Handle input argument < modulo argument.
                   2894:        * mpz_many: Make sure mp_size is the type for sizes, not int.
                   2895:
                   2896:        * mpz_init, mpz_init_set*, mpq_init, mpq_add, mpq_sub, mpq_mul,
                   2897:          mpq_div: Change mpz_init* interface.  Structure pointer as first
                   2898:          arg to initialization function, no longer *return* struct.
                   2899:
                   2900: Sun Jun 30 19:21:44 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2901:
                   2902:        * Rename mpz_impl_sqrt.c to mpz_do_sqrt.c to satisfy SysV 14
                   2903:          character file name length limit.
                   2904:
                   2905:        * Most files: Rename MINT to MP_INT.  Rename MRAT to MP_RAT.
                   2906:        * mpz_sizeinb.c: New file with function mpz_sizeinbase.
                   2907:        * mp_bases.c: New file, with array __mp_bases.
                   2908:        * _mpz_get_str, _mpz_set_str: Remove struct bases, use extern
                   2909:          __mp_bases instead.
                   2910:        * mout, mpz_out_str: Use array __mp_bases instead of function
                   2911:          _mpz_get_cvtlen.
                   2912:        * mpz_get_cvtlen.c: Remove.
                   2913:        * Makefile: Update.
                   2914:
                   2915: Sat Jun 29 21:57:28 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2916:
                   2917:        * longlong.h (__sparc8__ umul_ppmm): Insert 3 nop:s for wr delay.
                   2918:        * longlong.h (___IBMR2__): Define umul_ppmm, add_ssaaaa, sub_ddmmss.
                   2919:        * longlong.h (__sparc__): Don't call .umul; expand asm instead.
                   2920:          Don't define __umulsidi3 (i.e. use default definition).
                   2921:
                   2922: Mon Jun 24 17:37:23 1991  Torbjorn Granlund  (tege@amon.sics.se)
                   2923:
                   2924:        * _mpz_get_str.c (num_to_ascii_lower_case, num_to_ascii_upper_case):
                   2925:          Swap 't' and 's'.
                   2926:
                   2927: Sat Jun 22 13:54:01 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2928:
                   2929:        * mpz_gcdext.c: New file.
                   2930:
                   2931:        * mpn_mul: Handle carry and unexpected operand sizes in last
                   2932:          additions/subtractions.  (Bug trigged when v1_size == 1.)
                   2933:
                   2934:        * mp*_alloc*: Rename functions to mp*_init* (files to mp*_iset*.c).
                   2935:        * mpq_*: Call mpz_init*.
                   2936:
                   2937:        * mpz_pow_ui, rpow: Use _mpn_mul instead of mult.  Restructure.
                   2938:
                   2939: Wed May 29 20:32:33 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2940:
                   2941:        * mpz_get_cvtlen: multiply by size.
                   2942:
                   2943: Sun May 26 15:01:15 1991  Torbjorn Granlund  (tege@bella.nada.kth.se)
                   2944:
                   2945:        Alpha-release 0.95.
                   2946:
                   2947:        Fixes from Doug Lea (dl@g.oswego.edu):
                   2948:        * mpz_mul_ui: Loop to MULT_SIZE (not PROD_SIZE).  Adjust PROD_SIZE
                   2949:          correctly.
                   2950:        * mpz_div: Prepend _ to mpz_realloc.
                   2951:        * mpz_set_xs, mpz_set_ds: Fix typos in function name.
                   2952:
                   2953: Sat May 25 22:51:16 1991  Torbjorn Granlund  (tege@bella.nada.kth.se)
                   2954:
                   2955:        * mpz_divmod_ui: New function.
                   2956:
                   2957:        * sdiv: Make the sign of the remainder correct.
                   2958:
                   2959: Thu May 23 15:28:24 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2960:
                   2961:        * Alpha-release 0.94.
                   2962:
                   2963:        * mpz_mul_ui: Include longlong.h.
                   2964:
                   2965:        * mpz_perfsqr.c (mpz_perfect_square_p): Call _mpz_impl_sqrt instead
                   2966:          of msqrt.
                   2967:
                   2968:        * mpz_impl_sqrt: Don't call "move", inline its functionality.
                   2969:
                   2970:        * mdiv: Use MPN_COPY instead of memcpy.
                   2971:        * rpow, mpz_mul, mpz_mod_2exp: Likewise.
                   2972:        * pow.c: Likewise, and fix bug in the size arg.
                   2973:
                   2974:        * xtom: Don't use mpz_alloc, inline needed code instead.  Call
                   2975:          _mpz_set_str instead of mpz_set_str.
                   2976:
                   2977:        * Makefile: Make two libraries, libmp.a and libgmp.a.
                   2978:
                   2979: Thu May 22 20:25:29 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2980:
                   2981:        * Add manual to distribution.
                   2982:        * Fold in many missing routines descibed in the manual.
                   2983:        * Update Makefile.
                   2984:
                   2985: Wed May 22 13:48:46 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2986:
                   2987:        * mpz_set_str: Make it handle 0x prefix OK.
                   2988:
                   2989: Sat May 18 18:31:02 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2990:
                   2991:        * memory.c (_mp_default_reallocate): Swap OLD_SIZE and NEW_SIZE
                   2992:          arguments.
                   2993:        * mpz_realloc (_mpz_realloc): Swap in call to _mp_reallocate_func.
                   2994:        * min: Likewise.
                   2995:
                   2996: Thu May 16 20:43:05 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   2997:
                   2998:        * memory.c: Make the default allocations functions global.
                   2999:        * mp_set_fns (mp_set_memory_functions): Make a NULL pointer mean the
                   3000:          default memory function.
                   3001:
                   3002: Wed May  8 20:02:42 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   3003:
                   3004:        * mpz_div: Handle DEN the same as QUOT correctly by copying DEN->D
                   3005:          even if no normalization is needed.
                   3006:        * mpz_div: Rework reallocation scheme, to avoid excess copying.
                   3007:
                   3008:        * mpz_sub_ui.c, mpz_add_ui.c: New files.
                   3009:
                   3010:        * mpz_cmp.c, mpz_cmp_ui.c: New files.
                   3011:
                   3012:        * mpz_mul_2exp: Handle zero input MINT correctly.
                   3013:
                   3014:        * mpn_rshiftci:  Don't handle shift counts > BITS_PER_MP_DIGIT.
                   3015:
                   3016:        * mpz_out_raw.c, mpz_inp_raw.c: New files for raw I/O.
                   3017:
                   3018: Tue May  7 15:44:58 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   3019:
                   3020:        * mpn_rshift: Don't handle shift counts > BITS_PER_MP_DIGIT.
                   3021:        * mpz_div_2exp: Don't call _mpn_rshift with cnt > BITS_PER_MP_DIGIT.
                   3022:        * gcd, mpz_gcd: Likewise.
                   3023:
                   3024:        * gcd, mpz_gcd: Handle common 2 factors correctly.
                   3025:
                   3026: Mon May  6 20:22:59 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   3027:
                   3028:        * gmp-impl.h (MPN_COPY): Inline a loop instead of calling memcpy.
                   3029:
                   3030:        * gmp-impl.h, mpz_get_str, rpow: Swap DST and SRC in TMPCOPY* macros.
                   3031:
                   3032: Sun May  5 15:16:23 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   3033:
                   3034:        * mpz_div: Remove test for QUOT == 0.
                   3035:
                   3036: Sun Apr 28 20:21:04 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   3037:
                   3038:        * pow: Don't make MOD normalization in place, as it's a bad idea to
                   3039:          write on an input parameter.
                   3040:        * pow: Reduce BASE if it's > MOD.
                   3041:        * pow, mult, mpz_mul: Simplify realloc code.
                   3042:
                   3043: Sat Apr 27 21:03:11 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   3044:
                   3045:        * Install multplication using Karatsuba's algorithm as default.
                   3046:
                   3047: Fri Apr 26 01:03:57 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   3048:
                   3049:        * msqrt: Store in ROOT even for U==0, to make msqrt(0) defined.
                   3050:
                   3051:        * mpz_div_2exp.c, mpz_mul_2exp.c: New files for shifting right and
                   3052:          left, respectively.
                   3053:        * gmp.h: Add definitions for mpz_div_2exp and mpz_mul_2exp.
                   3054:
                   3055:        * mlshift.c, mrshift.c: Remove.
                   3056:
                   3057: Wed Apr 24 21:39:22 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   3058:
                   3059:        * mpn_mul: Check only for m2_size == 0 in function header.
                   3060:
                   3061: Mon Apr 22 01:31:57 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   3062:
                   3063:        * karatsuba.c: New file for Karatsuba's multplication algorithm.
                   3064:
                   3065:        * mpz_random, mpz_init, mpz_mod_2exp: New files and functions.
                   3066:
                   3067:        * mpn_cmp: Fix header comment.
                   3068:
                   3069: Sun Apr 21 00:10:44 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   3070:
                   3071:        * pow: Switch off initial base reduction.
                   3072:
                   3073: Sat Apr 20 22:06:05 1991  Torbjorn Granlund  (tege@echnaton.sics.se)
                   3074:
                   3075:        * mpz_get_str: Don't generate initial zeros for initial word.
                   3076:          Used to write outside of allocated storage.
                   3077:
                   3078: Mon Apr 15 15:48:08 1991  Torbjorn Granlund  (tege@zevs.sics.se)
                   3079:
                   3080:        * _mpz_realloc: Make it accept size in number of mp_digits.
                   3081:        * Most functions: Use new _mpz_realloc definition.
                   3082:
                   3083:        * mpz_set_str: Remove calls _mp_free_func.
                   3084:
                   3085:        * Most functions: Rename mpn_* to _mpn_*.  Rename mpz_realloc to
                   3086:          _mpz_realloc.
                   3087:        * mpn_lshift: Redefine _mpn_lshift to only handle small shifts.
                   3088:        * mdiv, mpz_div, ...: Changes for new definition of _mpn_lshift.
                   3089:        * msqrt, mp*_*shift*: Define cnt as unsigned (for speed).
                   3090:
                   3091: Sat Apr  6 14:05:16 1991  Torbjorn Granlund  (tege@musta.nada.kth.se)
                   3092:
                   3093:        * mpn_mul: Multiply by the first digit in M2 in a special
                   3094:          loop instead of zeroing the product area.
                   3095:
                   3096:        * mpz_abs.c: New file.
                   3097:
                   3098:        * sdiv: Implement as mpz_div_si for speed.
                   3099:
                   3100:        * mpn_add: Make it work for second source operand == 0.
                   3101:
                   3102:        * msub: Negate the correct operand, i.e. V before swapping, not
                   3103:          the smaller of U and V!
                   3104:        * madd, msub: Update abs_* when swapping operands, and not after
                   3105:          (optimization).
                   3106:
                   3107: Fri Apr  5 00:19:36 1991  Torbjorn Granlund  (tege@black.nada.kth.se)
                   3108:
                   3109:        * mpn_sub: Make it work for subtrahend == 0.
                   3110:
                   3111:        * madd, msub: Rewrite to minimize mpn_cmp calls.  Ensure
                   3112:          mpn_cmp is called with positive sizes (used to be called
                   3113:          incorrectly with negative sizes sometimes).
                   3114:
                   3115:        * msqrt: Make it divide by zero if fed with a negative number.
                   3116:        * Remove if statement at end of precision calculation that was
                   3117:          never true.
                   3118:
                   3119:        * itom, mp.h: The argument is of type short, not int.
                   3120:
                   3121:        * mpz_realloc, gmp.h: Make mpz_realloc return the new digit pointer.
                   3122:
                   3123:        * mpz_get_str.c, mpz_set_str.c, mpz_new_str.c: Don't include mp.h.
                   3124:
                   3125:        * Add COPYING to distribution.
                   3126:
                   3127:        * mpz_div_ui.c, mpz_div_si.c, mpz_new_ui.c, mpz_new_si.c: New files.
                   3128:
                   3129: Fri Mar 15 00:26:29 1991  Torbjorn Granlund  (tege@musta.nada.kth.se)
                   3130:
                   3131:        * Add Copyleft headers to all files.
                   3132:
                   3133:        * mpn_mul.c, mpn_div.c: Add header comments.
                   3134:        * mult.c, mdiv.c: Update header comments.
                   3135:
                   3136:        * mpq_add.c, mpq_sub.c, mpq_div.c, mpq_new.c, mpq_new_ui.c,
                   3137:          mpq_free.c: New files for rational arithmetics.
                   3138:
                   3139:        * mpn_lshift.c: Avoid writing the most significant word if it is 0.
                   3140:
                   3141:        * mdiv.c: Call mpn_lshift for the normalization.
                   3142:        * mdiv.c: Remove #ifdefs.
                   3143:
                   3144:        * Makefile: Add ChangeLog to DISTFILES.
                   3145:
                   3146:        * mpn_div.c: Make the add_back code work (by removing abort()).
                   3147:        * mpn_div.c: Make it return if the quotient is size as compared
                   3148:          with the difference NSIZE - DSIZE.  If the stored quotient is
                   3149:          larger than that, return 1, otherwise 0.
                   3150:        * gmp.h: Fix mpn_div declaration.
                   3151:        * mdiv.c: Adopt call to mpn_div.
                   3152:        * mpz_div.c: New file (developed from mdiv.c).
                   3153:
                   3154:        * README: Update routine names.
                   3155:
                   3156: Thu Mar 14 18:45:28 1991  Torbjorn Granlund  (tege@musta.nada.kth.se)
                   3157:
                   3158:        * mpq_mul.c: New file for rational multplication.
                   3159:
                   3160:        * gmp.h: Add definitions for rational arithmetics.
                   3161:
                   3162:        * mpn_div: Kludge the case where the high numerator digit > the
                   3163:          high denominator digit.  (This code is going to be optimized later.)
                   3164:
                   3165:        * New files: gmp.h for GNU specific functions, gmp-common.h for
                   3166:          definitions common for mp.h and gmp.h.
                   3167:
                   3168:        * Ensure mp.h just defines what BSD mp.h defines.
                   3169:
                   3170:        * pow.c: Fix typo for bp allocation.
                   3171:
                   3172:        * Rename natural number functions to mpn_*, integer functions to
                   3173:          mpz_*.
                   3174:
                   3175: Tue Mar  5 18:47:04 1991  Torbjorn Granlund  (tege@musta.nada.kth.se)
                   3176:
                   3177:        * mdiv.c (_mp_divide, case 2): Change test for estimate of Q from
                   3178:          "n0 >= r" to "n0 > r".
                   3179:
                   3180:        * msqrt: Tune the increasing precision scheme, to do fewer steps.
                   3181:
                   3182: Tue Mar  3 18:50:10 1991  Torbjorn Granlund  (tege@musta.nada.kth.se)
                   3183:
                   3184:        * msqrt: Use the low level routines.  Use low precision in the
                   3185:        beginning, and increase the precision as the result converges.
                   3186:        (This optimization gave a 6-fold speedup.)
                   3187: 
                   3188: Local Variables:
                   3189: mode: indented-text
                   3190: left-margin: 8
                   3191: fill-column: 75
                   3192: version-control: never
                   3193: End:

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>