ChangeLog

   1 2005-08-30  Richard B. Kreckel  <kreckel@ginac.de>
   2
   3         * include/cln/modules.h (CL_OUTPUT_LABEL): Work around redundant
   4         duplication of basic blocks on m68k.
   5
   6 2005-08-30  Richard B. Kreckel  <kreckel@ginac.de>
   7
   8         * include/cln/modules.h (CL_JUMP_TO): Fix mips* brokenness.
   9
  10 2005-08-27  Bruno Haible  <bruno@clisp.org>
  11
  12         * src/integer/bitwise/cl_I_ash_I.cc (ash): Avoid shifting a 32-bit
  13         zero value by more than 31 bits.
  14
  15 2005-08-27  Bruno Haible  <bruno@clisp.org>
  16
  17         Make the long-float overflow check work on 64-bit platforms.
  18         * src/float/lfloat/cl_LF.h (LF_exp_mid, LF_exp_high): Define as
  19         'unsigned int', not 'unsigned long'.
  20
  21 2005-08-27  Bruno Haible  <bruno@clisp.org>
  22
  23         * modules.h (CL_OUTPUT_LABEL): Work around redundant duplication
  24         of basic blocks by g++ 4.0.
  25         See <http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23345>.
  26
  27 2005-08-27  Bruno Haible  <bruno@clisp.org>
  28
  29         Make it possible to cross-compile CLN.
  30         * m4/intparam.m4: New file.
  31         * m4/floatparam.m4: New file.
  32         * autoconf/aclocal.m4: Include both.
  33         (CL_MACHINE): Add an additional CROSS_MACRO parameter.
  34         * configure.ac (CL_MACHINE): When cross-compiling, use
  35         CL_INTPARAM_CROSS and CL_FLOATPARAM_CROSS.
  36
  37 2005-08-27  Bruno Haible  <bruno@clisp.org>
  38
  39         Define HAVE_LONGLONG and HAVE_LONGDOUBLE when cross-compiling.
  40         * autoconf/aclocal.m4 (CL_LONGLONG, CL_LONGDOUBLE): When cross-
  41         compiling, use the test code from gnulib.
  42
  43 2005-08-27  Bruno Haible  <bruno@clisp.org>
  44
  45         * autoconf/aclocal.m4 (CL_RUSAGE): Fix error when cross-compiling.
  46
  47 2005-08-16  Richard B. Kreckel  <kreckel@ginac.de>
  48
  49         The patch of 2005-05-01 made it impossible to test the type of a cl_UP
  50         by comparing with &cl_class_univpoly_ring. We need an alternative:
  51         * include/cln/object.h (cl_class_flags_modint_ring): New #define...
  52         * src/polynomial/elem/cl_UP.cc (cl_class_univpoly_ring): ...used here.
  53         * src/polynomial/elem/cl_UP_GF2.h: Likewise.
  54         * src/polynomial/elem/cl_UP_MI.h: Likewise.
  55         * src/polynomial/elem/cl_UP_gen.h: Likewise.
  56         * src/polynomial/elem/cl_UP_number.h: Likewise.
  57
  58 2005-08-15  Richard B. Kreckel  <kreckel@ginac.de>
  59
  60         * m4/cc.m4 (CL_AS_NOEXECSTACK): New macro...
  61         * configure.ac: ...used here for setting ASMFLAGS...
  62         * src/Makefile.in: ...which are used here.
  63
  64 2005-08-02  Andreas Jochens  <aj@andaco.de>
  65
  66         * include/cln/config.h.in: Add support for PowerPC 64 CPU.
  67         * include/cln/modules.h: Likewise.
  68         * include/cln/object.h: Likewise.
  69         * include/cln/types.h: Likewise.
  70
  71 2005-07-24  Richard B. Kreckel  <kreckel@ginac.de>
  72
  73         Make out of the box build on x86_64 system with complete 32 bit
  74         userland possible
  75         * include/cln/config.h.in: Don't #define __x86_64__ when
  76         __i386__ is defined.
  77         * src/base/digitseq/cl_asm_x86_64_.cc: New file.
  78         * doc/cln.tex: Revert workaround description introduced 2005-05-02.
  79
  80 2005-06-10  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
  81
  82         * Makefile.in: Don't enter nonexisting directories.
  83
  84 2005-05-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
  85
  86         Speed up the linking step
  87         * src/Makefile.in: Use -objectlist for linking libcln.la.
  88
  89 2005-05-15  Richard B. Kreckel  <kreckel@ginac.de>
  90
  91         * autoconf/ltmain.sh: Upgrade to libtool-1.5.16.
  92         * m4/libtool.m4: Upgrade to libtool-1.5.16 (without Comeau patch).
  93
  94 2005-05-02  Richard B. Kreckel  <kreckel@ginac.de>
  95
  96         * doc/cln.tex: Document what to do on a x86_64 machine with 32-bit
  97         userland.
  98
  99 2005-05-01  Richard B. Kreckel  <kreckel@ginac.de>
 100
 101         Fix crashes in find_univpoly_ring and related functions
 102         * include/cln/modinteger.h: Remove vptr from cl_heap_modint_ring;
 103         remove declaration of cl_class cl_class_modint_ring.
 104         * include/cln/univpoly.h: Remove vptr from cl_heap_univpoly_ring;
 105         remove declaration of cl_class_univpoly_ring.
 106         * include/cln/object.h: cl_class_flags_modint_ring: New #define...
 107         * src/modinteger/cl_MI.cc: ...used in cl_class_modint_ring.
 108         * src/modinteger/cl_MI_fix16.h: No vptr, but static dtor and type flag.
 109         * src/modinteger/cl_MI_fix29.h: Likewise.
 110         * src/modinteger/cl_MI_fix32.h: Likewise.
 111         * src/modinteger/cl_MI_int32.h: Likewise.
 112         * src/modinteger/cl_MI_montgom.h: Likewise.
 113         * src/modinteger/cl_MI_pow2: Likewise.
 114         * src/modinteger/cl_MI_pow2m1.h: Likewise.
 115         * src/modinteger/cl_MI_pow2p1.h: Likewise.
 116         * src/modinteger/cl_MI_std.h: Likewise.
 117         * src/polynomial/elem/cl_UP.cc (cl_make_univpoly_ring): Compare with
 118         cl_class_flags_modint_ring, not with cl_class_modint_ring.
 119         * src/polynomial/elem/cl_UP_GF2.h (cl_class_num_univpoly_ring): New.
 120         * src/polynomial/elem/cl_UP_MI.h (cl_class_modint_univpoly_ring): New.
 121         * src/polynomial/elem/cl_UP_gen.h (cl_class_gen_univpoly_ring): New.
 122         * src/polynomial/elem/cl_UP_number.h (cl_class_num_univpoly_ring): New.
 123         Reported by Ralf Goertz <R_Goertz@web.de>.
 124
 125 2005-04-29  Richard B. Kreckel  <kreckel@ginac.de>
 126             Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 127
 128         * m4/cc.m4: Emit a warning if g++ is used and optimization turned off.
 129
 130 2005-04-24  Richard B. Kreckel  <kreckel@ginac.de>
 131
 132         Make GCC compiler flags default to -O
 133         * m4/cc.m4: New file...
 134         * configure.ac: ...used here.
 135         * autoconf/aclocal.m4: Regenerate.
 136
 137 2005-04-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 138
 139         * include/cln/string.h: Declare cl_string.
 140
 141 2005-03-17  Richard B. Kreckel  <kreckel@ginac.de>
 142
 143         * autoconf/ltmain.sh: Upgrade to libtool-1.5.14.
 144         * m4/libtool.m4: Upgrade to libtool-1.5.14 with Comeau patch.
 145         * autoconf/aclocal.m4: Regenerate.
 146
 147 2005-03-15  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 148
 149         * src/Makefile.in: Use $CXX instead of $CC when linking.
 150
 151 2005-02-27  Richard B. Kreckel  <kreckel@ginac.de>
 152
 153         * examples/perfnum.cc: update to presumed 42st Mersenne prime.
 154
 155 2004-11-28  Richard B. Kreckel  <kreckel@ginac.de>
 156
 157         Disambiguate binary operators of CLN types with float/double
 158         * include/cln/dfloat.h: Add binary operator overloads for arguments of
 159         type double.
 160         * include/cln/ffloat.h: Likewise, for arguments of type float.
 161         * include/cln/float.h: Likewise, both for arguments of types double and
 162         float.
 163         * include/cln/real.h: Likewise.
 164         Reported by Isidro Cachadiña Gutiérrez <icacha@unex.es>.
 165
 166 2004-11-03  Richard B. Kreckel  <kreckel@ginac.de>
 167
 168         * Version 1.1.9 released.
 169
 170 2004-10-28  Ralf Wildenhues <Ralf.Wildenhues@gmx.de>
 171
 172         * src/Makefile.in: Let config.status set LDFLAGS.
 173
 174 2004-10-27  Peter Breitenlohner  <peb@mppmu.mpg.de>
 175
 176         * cln.pc.in: Fix typo.
 177         * Makefile.in (INSTALL_SCRIPT): Added, to be used for scripts like
 178         cln-config.  Allows us to do special things to binaries, like
 179         installing them with install -s.
 180         * doc/Makefile.in: add/remove the cln.info from the installed
 181         $(infodir)/dir unless this is debian install-info (code copied from
 182         what GNU automake would produce).
 183
 184 2004-10-26  Richard B. Kreckel  <kreckel@ginac.de>
 185
 186         * src/integer/input/cl_I_read_stream.cc (read_integer): Fix a bug
 187         that caused radix specifiers to not work when reading from a stream.
 188         * src/rational/input/cl_RA_read_stream.cc (read_rational): Likewise.
 189         * src/real/input/cl_R_read_stream.cc (read_real): Likewise.
 190         * src/float/input/cl_F_read_stream.cc (read_float): Likewise.
 191         * src/complex/input/cl_N_read_stream.cc (read_complex): Likewise.
 192
 193 2004-10-25  Richard B. Kreckel  <kreckel@ginac.de>
 194
 195         * src/base/cl_low.h: Add mulu64 assembler macro for ia64.
 196
 197 2004-10-24  Richard B. Kreckel  <kreckel@ginac.de>
 198
 199         * src/base/cl_low.h: Add mul and div macros for x86_64.
 200
 201 2004-10-23  Richard B. Kreckel  <kreckel@ginac.de>
 202
 203         * src/integer/conv/cl_I_from_digits.cc (digits_to_I): Fix thinko in
 204         new code for base power of two.
 205
 206 2004-10-22  Richard B. Kreckel  <kreckel@ginac.de>
 207
 208         * src/integer/conv/cl_I_to_digits (I_to_digits): Fix an elusive stack
 209         overwriting problem. That was the real cause for Debian bug#246319.
 210         * src/integer/output/cl_I_print.cc (print_integer): Revert workaround
 211         for the bug fixed above.
 212
 213 2004-10-20  Richard B. Kreckel  <kreckel@ginac.de>
 214
 215         * include/cln/types.h: Use 64 bit digits on x86_64 CPU.
 216
 217 2004-10-12  Richard B. Kreckel  <kreckel@ginac.de>
 218
 219         * src/integer/conv/cl_I_from_digits.cc (digits_to_I): Speedup when
 220         the base is a power of two.
 221
 222 2004-10-05  Richard B. Kreckel  <kreckel@ginac.de>
 223
 224         * src/integer/conv/cl_I_to_digits.cc (I_to_digits): Fix bug in base 32.
 225
 226 2004-09-27  Richard B. Kreckel  <kreckel@ginac.de>
 227
 228         Support for little-endian Mips, second shot
 229         * src/base/digitseq/cl_asm_mipsel_.cc: New file...
 230         * src/base/digitseq/cl_asm_cc: ...used here.
 231         * src/base/digitseq/cl_asm.h: Include cl_asm_mips.h for any endianness.
 232         * include/cln/object.h: Set alignment for mipsel explicitly.
 233
 234 2004-09-05  Richard B. Kreckel  <kreckel@ginac.de>
 235
 236         Support for little-endian Mips
 237         * include/cln/config.h.in: Add __mipsel__.
 238         * include/cln/modules.h: For Mips, this is endianness-agnostic.
 239         * src/base/digitseq/cl_asm_.cc, src/base/digitseq/cl_asm.h:
 240         Mask out assembler for little-endian Mips.
 241
 242 2004-08-30  Bruno Haible  <bruno@clisp.org>
 243
 244         * benchmarks/timebench2.sh: Multiply all repeat counts by 100.
 245         * benchmarks/timebench2.results: Add recent PowerPC G4 results.
 246
 247 2004-08-26  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 248
 249         * examples/e.cc: remove extra semicolon.
 250
 251 2004-08-25  Bruno Haible  <bruno@clisp.org>
 252
 253         * autoconf/ltmain.sh: Upgrade to libtool-1.5.6.
 254         * m4/libtool.m4: New file, from libtool-1.5.6 with modifications:
 255         2004-08-22  Bruno Haible  <bruno@clisp.org>
 256                 * m4/libtool.m4: Add support for Comeau C++ on Linux.
 257                 Reported by Prof. Roberto Bagnara <bagnara@cs.unipr.it>.
 258         * autoconf/aclocal.m4: Regenerate.
 259
 260 2004-08-19  Bruno Haible  <bruno@clisp.org>
 261
 262         * include/cln/modules.h (CL_GLOBALIZE_JUMP_LABEL, CL_JUMP_TO): When
 263         converting a label to a string, use ASM_UNDERSCORE_PREFIX. Needed on
 264         MacOS X.
 265         Reported by Darren Bane <darren.bane@ul.ie>.
 266
 267 2004-07-01  Richard B. Kreckel  <kreckel@ginac.de>
 268
 269         * Version 1.1.8 released.
 270
 271 2004-06-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 272
 273         * src/base/cl_macros.h: alloca(3) has size_t argument type.
 274
 275 2004-06-30  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 276
 277         * include/cln/floatformat.h: Do define a type here.
 278
 279 2004-06-27  Richard B. Kreckel  <kreckel@ginac.de>
 280
 281         * include/cln/modules.h (CL_JUMP_TO): Fix AMD64 brokenness.
 282
 283 2004-06-23  Bruno Haible  <bruno@clisp.org>
 284
 285         * configure.ac: Pretend ftime() is not available. Needed by
 286         CL_TIMES_CLOCK.
 287
 288 2004-06-21  Ralf Stephan  <ralf@ark.in-berlin.de>
 289
 290         * doc/cln.tex: Document jacobi, isprobprime and nextprobprime.
 291
 292 2004-06-18  Richard B. Kreckel  <kreckel@ginac.de>
 293
 294         * rational/transcendental/cl_RA_logp.cc: fix bug where base is
 295         reciprocal of an integer.
 296         Reported by Niklas Knutsson <nq@altern.org>.
 297
 298 2004-06-18  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 299
 300         * src/complex/ring/cl_C_ring.cc, src/integer/ring/cl_I_ring.cc,
 301         src/rational/ring/cl_RA_ring.cc, src/real/ring/cl_R_ring.cc:
 302         Make template specializations explicit.
 303
 304 2004-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 305
 306         * src/complex/input/cl_N_read.cc, src/float/input/cl_F_read.cc,
 307         src/integer/input/cl_I_read.cc, src/rational/input/cl_RA_read.cc,
 308         src/real/input/cl_R_read.cc:  Remove unused labels.
 309
 310 2004-06-17  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 311
 312         * benchmarks/timebench1.cc, benchmarks/timebench2a.LiDIA.cc,
 313         benchmarks/timebench2a.cc, benchmarks/timebench2ap.cc,
 314         benchmarks/timebench2b.LiDIA.cc, benchmarks/timebench2b.cc,
 315         examples/atan_recip.cc, examples/atanh_recip.cc,
 316         examples/contfrac.cc, examples/e.cc, examples/legendre.cc,
 317         examples/lucaslehmer.cc, examples/pi.cc, include/cln/GV.h,
 318         include/cln/SV.h, include/cln/malloc.h, include/cln/modules.h,
 319         include/cln/object.h, include/cln/string.h, src/base/cl_abort.cc,
 320         src/base/cl_alloca.h, src/base/cl_malloc.cc,
 321         src/base/random/cl_random_from.cc,
 322         src/base/string/cl_spushstring_append.cc,
 323         src/base/string/cl_spushstring_push.cc,
 324         src/base/string/cl_st_debug.cc,
 325         src/base/string/input/cl_st_gettoken.cc,
 326         src/complex/input/cl_N_read.cc, src/float/input/cl_F_read.cc,
 327         src/float/output/cl_F_dprint.cc, src/integer/input/cl_I_read.cc,
 328         src/rational/input/cl_RA_read.cc,
 329         src/real/format-output/cl_fmt_integer.cc,
 330         src/real/format-output/cl_fmt_paddedstring.cc,
 331         src/real/input/cl_R_read.cc, src/timing/cl_t_current.cc,
 332         src/timing/cl_t_current2.cc, tests/exam.cc, tests/tests.cc,
 333         tests/timeLFRAmul.cc, tests/timeLFatan-compare.cc,
 334         tests/timeLFatan.cc, tests/timeLFatanh-compare.cc,
 335         tests/timeLFatanh.cc, tests/timeLFcos-compare.cc, tests/timeLFcos.cc,
 336         tests/timeLFcosh.cc, tests/timeLFexp-compare.cc, tests/timeLFexp.cc,
 337         tests/timeLFln-compare.cc, tests/timeLFln.cc,
 338         tests/timeLFsin-compare.cc, tests/timeLFsin.cc, tests/timeLFsinh.cc,
 339         tests/timeLFsqrt.cc, tests/timeMImisc5.cc, tests/timeMIpow2div.cc,
 340         tests/timeMIpow2recip.cc, tests/timeRALFdiv.cc, tests/timeRAtoLF.cc,
 341         tests/timeUPMImul.cc, tests/timecatalan.cc, tests/timediv.cc,
 342         tests/timediv2adic-compare.cc, tests/timediv2adic.cc,
 343         tests/timeeuler.cc, tests/timeexp1.cc, tests/timefact.cc,
 344         tests/timegcd.cc, tests/timemul-compare.cc, tests/timemul.cc,
 345         tests/timepi.cc, tests/timeprint-compare.cc, tests/timeprint.cc,
 346         tests/timerecip2adic-compare.cc, tests/timerecip2adic.cc,
 347         tests/timesqrt.cc, tests/timesqrtmodp.cc, tests/timesquare.cc,
 348         tests/timezeta3.cc: Change all C include headers to ISO style
 349         within C++ code.
 350
 351 2004-06-10  Richard B. Kreckel  <kreckel@ginac.de>
 352
 353         * examples/perfnum.cc: update to presumed 41st Mersenne prime.
 354
 355 2004-05-02  Richard B. Kreckel  <kreckel@ginac.de>
 356
 357         * Version 1.1.7 released.
 358
 359 2004-05-02  Richard B. Kreckel  <kreckel@ginac.de>
 360
 361         * examples/pi.cc and examples/pi.1: New files.
 362         * examples/Makefile.in: Build the pi executable.
 363
 364 2004-05-01  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 365
 366         * src/Makefile.in: Fix for parallel build: wait for subdir objects to
 367         be finished before creating the library.
 368
 369 2004-04-30  Richard B. Kreckel  <kreckel@ginac.de>
 370
 371         * src/integer/output/cl_I_print.cc (print_integer): workaround
 372         GCC compiler bug (cf. Debian bug#246319).
 373
 374 2004-03-20  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 375
 376         * m4/gmp.m4 (CL_GMP_CHECK): Do not lose LIBS setting with config.cache
 377         enabled.
 378
 379 2004-03-08  Bruno Haible  <bruno@clisp.org>
 380
 381         * src/float/lfloat/elem/cl_LF_mul.cc (operator*): Fix the second
 382         underflow condition.
 383         * src/float/lfloat/algebraic/cl_LF_sqrt.cc (sqrt): Fix a bug with large
 384         uexp whereby SQRT of MOST-POSITIVE-LONG-FLOAT was less than 1.
 385
 386 2004-03-04  Richard B. Kreckel  <kreckel@ginac.de>
 387
 388         * Makefile.in (install): Add ${srcdir} for cln.m4.
 389         * m4/gmp.m4: quote macro names.
 390         Reported by Ralf Wildenhues <Ralf.Wildenhues@gmx.de>.
 391
 392 2004-01-01  Richard B. Kreckel  <kreckel@ginac.de>
 393
 394         * Version 1.1.6 released.
 395
 396 2004-01-01  Richard B. Kreckel  <kreckel@ginac.de>
 397
 398         * include/cln/univpoly.h, include/cln/univpoly_complex.h,
 399           include/cln//univpoly_integer.h, include/cln/univpoly_modint.h,
 400           include/cln/univpoly_rational.h, include/cln/univpoly_real.h,
 401           src/polynomial/elem/cl_UP_GF2.h, src/polynomial/elem/cl_UP_MI.h,
 402           src/polynomial/elem/cl_UP_gen.h, src/polynomial/elem/cl_UP_no_ring.cc,
 403           src/polynomial/elem/cl_UP_number.h (ldegree): New function.
 404         * doc/cln.tex: Document `ldegree'.
 405
 406 2003-12-29  Richard B. Kreckel  <kreckel@ginac.de>
 407
 408         Rework of autoconfiscation infrastructure
 409         * autoconf/config.{guess,sub}: Update to GNU version 2003-10-07.
 410         * autoconf/ltmain.sh: Update to GNU version 1.4.3.
 411         * autoconf/autoconf: Remove (from now on we assume autoconf is
 412           installed properly on the sytem).
 413         * autoconf/autoconf.m4: Likewise.
 414         * autoconf/autoconf.m4f: Likewise.
 415         * autoconf/acgeneral.m4: Likewise.
 416         * autoconf/acspecific.m4: Likewise.
 417         * autoconf/aclocal.m4: Regenerate.
 418         * autoconf/acinclude.m4: Remove (while moving the macros...)
 419         * m4/gmp.m4: New file (...to here).
 420         * Makefile.devel: Update to new scheme.
 421         * configure.ac: Likewise.
 422         * include/cln/GV_integer.h: Assume template specializations work.
 423         * include/cln/GV_modinteger.h: Likewise.
 424         * include/cln/config.h.in: Likewise, and drop HAVE_BOOL.
 425         * src/base/cl_base_config.h.in: Drop support for obsolete ftime(3).
 426         * src/base/random/cl_random_from.cc: Likewise.
 427         * src/timing/cl_base_config.h.in: Likewise.
 428         * src/timing/cl_t_current.cc: Likewise.
 429
 430 2003-12-27  Richard B. Kreckel  <kreckel@ginac.de>
 431
 432         * src/polynomial/cl_UP_gen.h (gen_minus): Fix case where first
 433           argument is zero.
 434         * src/polynomial/cl_UP_MI.h (modint_minus): Likewise.
 435         * src/polynomial/cl_UP_number.h (num_minus): Likewise.
 436         Reported by Munagala Ramanath <amberarrow@yahoo.com>.
 437
 438 2003-12-02  Richard B. Kreckel  <kreckel@ginac.de>
 439
 440         * examples/perfnum.cc: update to presumed 40th Mersenne prime.
 441
 442 2003-11-20  Christian Bauer  <cbauer@ginac.de>
 443
 444         Added pkg-config support
 445         * cln.pc.in: New file.
 446         * Makefile.in: Take care of cln.pc.
 447         * configure.ac: Likewise.
 448
 449 2003-08-06  Richard B. Kreckel  <kreckel@ginac.de>
 450
 451         * src/numtheory/cl_nt_sqrtmodp.cc: #undef _R.
 452         Reported by Andrew Rechnitzer <A.Rechnitzer@ms.unimelb.edu.au>.
 453
 454 2003-08-01  Richard Kreckel  <kreckel@ginac.de>
 455
 456         More dependent base resolution issues
 457         * src/base/hash/cl_hash.h (cl_heap_hashtable<T>::iterator()):
 458         portable syntactic simplification.
 459         * src/base/hash/cl_hashset.h: Preceed inherited members with this->.
 460         * src/base/hash/cl_hash1.h: Likewise for member functions.
 461         * src/base/hash/cl_hash2.h: Likewise.
 462         * src/base/hash/cl_hashuniq.h: Likewise.
 463         * src/base/hash/cl_hashuniqweak.h: Likewise.
 464         * src/base/hash/cl_hash.h: Revert explicit static member function
 465         lookup since that was GCC's fault.
 466         * src/base/hash/cl_hash2weak.h: Likewise.
 467         * src/base/hash/cl_hashuniqweak.h: Likewise.
 468
 469 2003-06-29  Richard Kreckel  <kreckel@ginac.de>
 470
 471         Dependent base resolution needed for GCC-3.4
 472         * include/cln/GV.h: Preceed inherited members with this->.
 473         * include/cln/SV.h: Likewise.
 474         * include/cln/object.h: Likewise.
 475         * src/base/hash/cl_hash1.h: Likewise.
 476         * src/base/hash/cl_hash1weak.h: Likewise.
 477         * src/base/hash/cl_hash2.h: Likewise.
 478         * src/base/hash/cl_hashuniq.h: Likewise.
 479         * src/base/hash/cl_hash.h: Make lookup of static member function
 480         explicit.
 481         * src/base/hash/cl_hash2weak.h: Likewise.
 482         * src/base/hash/cl_hashuniqweak.h: Likewise.
 483         * src/modinteger/cl_MI.cc: Make shell-comment a C-comment. Geez.
 484
 485 2003-02-24  Bruno Haible  <bruno@clisp.org>
 486
 487         * src/base/random/cl_random_from.cc (random_state::random_state): Add
 488         support for MacOS X.
 489         * src/modinteger/cl_MI.cc: #undef _R.
 490         Reported by Erann Gat <gat@jpl.nasa.gov>.
 491
 492 2002-08-03  Joerg Arndt  <jj@suse.de>
 493
 494         * include/cln/config.h.in: Add support for x86_64 CPU.
 495         * include/cln/modules.h: Likewise.
 496         * include/cln/types.h: Likewise.
 497         * include/cln/object.h: Likewise.
 498
 499 2002-06-08  Richard Kreckel  <kreckel@ginac.de>
 500
 501         * src/base/digitseq/cl_asm.h: ensure intCsize==16 when including
 502         m68k Assembler routines.
 503         * src/base/digitseq/cl_asm_.cc: Likewise.
 504
 505 2002-05-28  Richard Kreckel  <kreckel@ginac.de>
 506
 507         * Version 1.1.5 released.
 508
 509 2002-05-27  Richard Kreckel  <kreckel@ginac.de>
 510
 511         * include/cln/modules.h (CL_CONCATENATE): New macro.
 512
 513 2002-05-10  Richard Kreckel  <kreckel@ginac.de>
 514
 515         * doc/cln.tex (Building the library): Update recommendations for
 516         compiling on Tru64 using g++ 3.0 and 3.1.
 517         * README: Update homepage.
 518
 519 2002-05-05  Bruno Haible  <bruno@clisp.org>
 520
 521         * doc/cln.tex (Building the library): Give some recommendations for
 522         g++ 3.0 and 3.1.
 523
 524         Force link-time references despite optimizations done by g++ 2.95
 525         and newer.
 526         * include/cln/modules.h (CL_FORCE_LINK): New macro.
 527         * Use CL_FORCE_LINK.
 528         * include/cln/GV_integer.h (cl_GV_I_debug_dummy): Likewise.
 529         * include/cln/GV_number.h (cl_GV_number_debug_dummy): Likewise.
 530         * include/cln/SV_number.h (cl_SV_number_debug_dummy): Likewise.
 531         * include/cln/SV_ringelt.h (cl_SV_ringelt_debug_dummy): Likewise.
 532         * include/cln/dfloat.h (cl_DF_debug_dummy): Likewise.
 533         * include/cln/ffloat.h (cl_FF_classes_dummy, cl_FF_debug_dummy):
 534         Likewise.
 535         * include/cln/integer.h (cl_I_classes_dummy, cl_I_debug_dummy):
 536         Likewise.
 537         * include/cln/lfloat.h (cl_LF_debug_dummy): Likewise.
 538         * include/cln/modinteger.h (cl_MI_debug_dummy): Likewise.
 539         * include/cln/rational.h (cl_RA_debug_dummy): Likewise.
 540         * include/cln/real.h (cl_R_debug_dummy): Likewise.
 541         * include/cln/ring.h (cl_ring_debug_dummy): Likewise.
 542         * include/cln/sfloat.h (cl_SF_classes_dummy, cl_SF_classes_dummy):
 543         Likewise.
 544         * include/cln/string.h (cl_string_debug_dummy): Likewise.
 545         * include/cln/univpoly.h (cl_UP_debug_dummy): Likewise.
 546         * src/float/base/cl_ieee.h (cl_ieee_dummy_NNN): Likewise.
 547
 548         Avoid g++ 3.1 warnings.
 549         * src/base/cl_offsetof.h (offsetof): Redefine each time.
 550         * src/base/digitseq/cl_DS.h: Include "cl_offsetof.h" after <gmp.h>.
 551
 552         Avoid g++ 3.1 warnings.
 553         * src/base/hash/cl_hash.h (struct cl_heap_hashtable): Use typename
 554         where needed.
 555         * src/base/hash/cl_hash1.h (struct cl_heap_hashtable_1): typedef
 556         htxentry as a shortcut.
 557         * src/base/hash/cl_hash2.h (struct cl_heap_hashtable_2): Likewise.
 558         * src/base/hash/cl_hashset.h (struct cl_heap_hashtable_set): Likewise.
 559         * src/base/hash/cl_hashuniq.h (struct cl_heap_hashtable_uniq):
 560         Likewise.
 561
 562 2002-03-15  Bruno Haible  <bruno@clisp.org>
 563
 564         * cln.tex: Document problem with GNU make 3.77.
 565         Reported by Michael Somos <somos@grail.cba.csuohio.edu>.
 566
 567 2002-02-16  Richard Kreckel  <kreckel@ginac.de>
 568
 569         * cln.m4: quote macro name.
 570           Pointed out by Roberto Bagnara.
 571
 572 2002-01-20  Richard Kreckel  <kreckel@ginac.de>
 573
 574         * autoconf/config.{guess,sub}: Update to GNU version 2002-01-02.
 575           (the old one was broken on Linux/Mips.)
 576
 577 2002-01-04  Richard Kreckel  <kreckel@ginac.de>
 578
 579         * autoconf/autoconf.m4f: get brutal in order to adhere to FHS.
 580         * Version 1.1.4 released.
 581
 582 2002-01-03  Richard Kreckel  <kreckel@ginac.de>
 583
 584         * autoconf/acinclude.m4: revamp MPN-matcher.
 585         * autoconf/aclocal.m4: upgrade to autoconf-2.52 infrastructure,
 586           sync with CLisp from CVS.
 587         * autoconf/autoconf: Likewise.
 588         * autoconf/autoconf.m4f: Likewise (new file).
 589         * configure.ac: Likewise (new file, replaces configure.in).
 590         * configure.in: Likewise (deleted, replaced by configure.ac).
 591         * autoconf/config.{guess,sub}: Update to GNU version 2001-12-13.
 592         * src/Makefile.in: made VPATH safe for autoconf-2.52.
 593         * include/cln/config.h.in: Add __s390__.
 594
 595 2001-12-31  Richard Kreckel  <kreckel@ginac.de>
 596
 597         * src/base/digitseq/cl_DS.h: <gmp.h> is not included extern "C"
 598           any more since GMP4 has some C++ support in it.
 599
 600 2001-12-14  Richard Kreckel  <kreckel@ginac.de>
 601
 602         * include/cln/modules.h, include/cln/object.h: add support for
 603           s390.
 604         * src/numtheory/cl_nt_sqrtmodp.cc: workaround for GCC2.x compiler-bug
 605           on s390, provided by Gerhard Tonn.
 606
 607 2001-11-05  Richard Kreckel  <kreckel@ginac.de>
 608
 609         * autoconf/ltmain.sh: Upgrade to libtool-1.4.2.
 610         * autoconf/config.{guess,sub}: Update to GNU version 2001-09-07.
 611         * Version 1.1.3 released.
 612
 613 2001-11-04  Bruno Haible  <haible@clisp.cons.org>
 614
 615         Interoperability with gcc-3.0 -fuse-cxa-atexit.
 616         * autoconf/aclocal.m4 (CL_GLOBAL_CONSTRUCTORS): Add test whether
 617           global destructors actually exist.
 618         * include/cln/modules.h (CL_PROVIDE, CL_PROVIDE_END, CL_PROVIDE_END):
 619           Don't hack the global destructors if there is no global destructors
 620           function.
 621
 622 2001-11-03  Richard Kreckel  <kreckel@ginac.de>
 623
 624         * src/float/transcendental/cl_F_sinx.cc (sinx_naive): For small
 625           values of x, return square(x) instead of x.
 626         * src/float/transcendental/cl_F_sinhx.cc (sinhx_naive): Likewise.
 627
 628 2001-07-25  Richard Kreckel  <kreckel@ginac.de>
 629
 630         * Version 1.1.2 released.
 631
 632 2001-07-24  Richard Kreckel  <kreckel@ginac.de>
 633
 634         * src/base/hash/cl_hash.h: declare _cl_hashtable_iterator<htentry> a
 635           friend of cl_heap_hashtable<htentry>.
 636
 637 2001-07-22  Richard Kreckel  <kreckel@ginac.de>
 638
 639         * src/float/base/cl_ieee.cc: try to do magic to the FPU only if
 640           _FPU_IEEE is really defined.
 641         * include/cln/modules.h: change assembler labels from `label' to
 642           `label:' on hppa, needed by Linux (see comment).
 643         * autoconf/acinclude.m4: new file (for storing CLN-specific macros).
 644         * Makefile.devel: adjusted.
 645         * autoconf/aclocal.m4: regenerate.
 646         * src/base/low/cl_low_mul.cc: moved POD variables that are declared
 647           extern "C" elsewhere out of the namespace.
 648         * src/base/low/cl_low_div.cc: Likewise.
 649
 650 2001-06-08  Bruno Haible  <haible@clisp.cons.org>
 651
 652         * autoconf/config.{guess,sub}: Update to GNU version 2001-05-11.
 653         * autoconf/aclocal.m4: Upgrade to libtool-1.4.
 654         * autoconf/ltmain.sh: Likewise.
 655         * autoconf/ltconfig: Remove file.
 656         * autoconf/install-sh: New file.
 657         * configure.in: Add AC_CONFIG_AUX_DIR call.
 658
 659 2001-06-05  Richard Kreckel  <kreckel@ginac.de>
 660
 661         * tests/tests.cc: resolve namespace ambiguity about strcmp().
 662
 663 2001-05-31  Richard Kreckel  <kreckel@ginac.de>
 664
 665         * Version 1.1.1 released.
 666
 667 2001-05-28  Richard Kreckel  <kreckel@ginac.de>
 668
 669         * cln/cln.tex: documented problems with shared library on Sparc
 670           using gcc older than 2.95.3.
 671         * configure.in: Fixed typos in versioning docu.
 672
 673 2001-05-25  Bruno Haible  <haible@clisp.cons.org>
 674
 675         * src/base/digitseq/cl_asm_arm_.cc: Use #0x instead of #& to designate
 676           hexadecimal constants.
 677
 678 2001-05-25  Richard Kreckel  <kreckel@ginac.de>
 679
 680         * autoconf/floatparam.c (double_wordorder_bigendian_p): new symbol.
 681         * src/float/dfloat/cl_DF.h: Check for double_wordorder_bigendian_p.
 682         * Removed LiDIA interface since that is now outdated (namespace cln)
 683           and maintained elsewhere.
 684         * Adjusted dates and final touches for 1.1.1.
 685
 686 2001-05-19  Richard Kreckel  <kreckel@ginac.de>
 687
 688         * INSTALL: Update toolchain info: no egcs, some more platforms.
 689         * doc/cln.tex: Likewise.
 690
 691 2001-05-18  Richard Kreckel  <kreckel@ginac.de>
 692
 693         * src/base/cl_low.h: prepended variables declared inside macros
 694           with underscore.  Fixes equal_hashcode() on various platforms.
 695
 696 2001-04-25  Richard Kreckel  <kreckel@ginac.de>
 697
 698         * src/base/cl_low.h: Added several checks if NO_ASM is defined, so
 699           this definition becomes actually useful.  This is needed for
 700           compilation on Arm until somebody fixes the assembler files for Arm.
 701         * src/base/digitseq/cl_asm.h: Likewise.
 702         * src/base/digitseq/cl_asm_.cc: Likewise.
 703         * */Makefile.in: Added `override' in front of `CPPFLAGS +=' so
 704            one can say `make CPPFLAGS=-DFOOBAR'.
 705
 706 2001-03-26  Arvid Norberg <c99ang@cs.umu.se>
 707
 708         * src/base/random/cl_random_from.cc: ported to beos.
 709
 710 2001-03-05  Richard Kreckel  <kreckel@ginac.de>
 711
 712         * include/cln/modules.h (CL_JUMP_TO): Fix IA64 brokenness.
 713
 714 2001-01-28  Richard Kreckel  <kreckel@ginac.de>
 715
 716         * include/cln/number.h (cl_as_N): Remove bogus comment.
 717
 718 2001-01-22  Richard Kreckel  <kreckel@ginac.de>
 719
 720         * configure.in: Make build in separate builddir possible (again).
 721
 722 2001-01-22  Richard Kreckel  <kreckel@ginac.de>
 723
 724         * include/cln/*.h: Change signatures of all classes' methods
 725           cln::cl_foo::operator new(size_t, cl_foo*) to
 726           cln::cl_foo::operator new(size_t, void*) so one can declare
 727           std::vector<cln::cl_foo>, std::list<cln::cl_foo> etc. for
 728           certain STL implementations (like libstdc++-v3).
 729
 730 2000-12-14  Richard Kreckel  <kreckel@ginac.de>
 731
 732         * Version 1.1 released.
 733
 734 2000-12-13  Richard Kreckel  <kreckel@ginac.de>
 735
 736         * */*: cl_istream -> std::istream, cl_ostream -> std::ostream.
 737
 738 2000-12-05  Richard Kreckel  <kreckel@ginac.de>
 739
 740         * Makefile.in, src/Makefile.in, doc/Makefile.in: Use mkdir -p.
 741         * include/cln/version.h.in, src/base/verion.cc: New files.
 742         * configure.in: Generate include/cln/version.h.
 743         * cln.m4: Rewrote it. Check result of cln-config without compiling.
 744           Do cross-check library version and header version information.
 745
 746 2000-12-02  Christian Bauer  <cbauer@ginac.de>
 747
 748         * Makefile.in, src/Makefile.in, doc/Makefile.in: Added $DESTDIR.
 749         * cln.m4, cln.spec.in: some minor fixes with respect to RPM package
 750           building.
 751
 752 2000-11-24  Richard Kreckel  <kreckel@ginac.de>
 753
 754         * */*: Removed problematic stdin, stdout and stderr definitions.
 755           Use std::cin, std::cout, std::cerr instead (obsoletes 2000-10-29).
 756
 757 2000-11-20  Bruno Haible
 758
 759         * cln-config.1: change title.
 760
 761 2000-11-18  Richard Kreckel  <kreckel@ginac.de>
 762
 763         * cln.m4: New file.
 764         * doc/cln.tex: Document package tools cln-config and cln.m4.
 765         * Makefile.in: Care about cln.m4.
 766
 767 2000-11-17  Richard Kreckel  <kreckel@ginac.de>
 768
 769         * cln-config.1: added manpage, as required by a couple of distros.
 770         * Makefile.in, doc/Makefile.in: target install depends on installdirs.
 771
 772 2000-11-16  Richard Kreckel  <kreckel@ginac.de>
 773
 774         * autoconf/aclocal.m4 (CL_GMP_SET_UINTD): New macro...
 775         * configure.in: ...used here.
 776         * include/cln/config.h.in: Put in macros defined by CL_GMP_SET_UINTD...
 777         * include/cln/types.h: ...used here.
 778         * autoconf/acgeneral.m4, autoconf/aclocal.m4: Adhere to FHS.
 779
 780 2000-11-13  Richard Kreckel  <kreckel@ginac.de>
 781
 782         * src/base/digitseq/cl_asm.h: Test if (intDsize==32) for MIPS and HPPA,
 783           in order to guard against an accidented configuration.
 784         * src/integer/conv/cl_I_to_digits.cc (table): member b_hoch_k of
 785           struct power_table_entry initialized as ULL instead of as UL, if
 786           intDsize==64 (caused misprinting on MIPS w/ GMP).
 787         * src/base/cl_macros.h (minus_bitm, bitc): implemented ULL, if
 788           HAVE_FAST_LONGLONG.
 789         * src/integer/bitwise/cl_I_mkfx.cc (mkf_extract): Likewise.
 790         * src/integer/conv/cl_I_from_L.cc (cl_I_from_L): Added trival
 791           generation of Bignum for intDsize==64 and a notreached-check at end.
 792         * autoconf/config.guess, autoconf/config.sub: updated from FSF.
 793         * include/cln/config.h.in: Prepared support for IA64.
 794         * include/cln/types.h: Likewise.
 795         * include/cln/object.h: Likewise.
 796         * include/cln/modules.h: Likewise.
 797         * src/base/cl_macros.h (nonreturning_function): Likewise (NUE's
 798           compiler claims __GNUC_MINOR__==9).
 799
 800 2000-11-03  Richard Kreckel  <kreckel@ginac.de>
 801
 802         * src/base/cl_macros.h (bit, minus_bit): changed criterion for ULL from
 803           HAVE_DD to HAVE_FAST_LONGLONG.
 804         * src/base/cl_macros.h (bitm): implemented ULL, if HAVE_FAST_LONGLONG.
 805         * src/base/cl_low.h: actually no need to include "cln/types.h" here.
 806         * src/base/cl_low.h (logcount_64): always ULL, independent of HAVE_DD.
 807         * src/base/random/cl_UL_random.cc (random32): a is always ULL.
 808
 809 2000-11-01  Richard Kreckel  <kreckel@ginac.de>
 810
 811         * include/cln/object.h (cl_combine): define additional signatures, if
 812           HAVE_LONGLONG is defined, in order to keep the compiler happy.
 813         * src/base/cl_macros.h: include "cln/types.h", since we need HAVE_DD...
 814         * src/base/cl_macros.h (bit): ...for this macro...
 815         * src/base/cl_macros.h (minus_bit): ...and this one.
 816         * src/base/cl_low.h: include "cln/types.h", since we need HAVE_DD...
 817         * src/base/cl_low.h (logcount_64): ...for this macro.
 818         * src/base/random/cl_UL_random.cc (random32): if HAVE_DD a is an ULL.
 819         * src/integer/gcd/cl_I_gcd_aux2.cc (floorDD): fixed algorithmic bug.
 820           that turned up when intDsize==32 and cl_word_size==64.
 821         * src/float/dfloat/elem/cl_DF_div.cc (operator/): fixed a missing cast
 822           to uint64 that turned up when intDsize==32 and cl_word_size==64.
 823
 824 2000-10-29  Richard Kreckel  <kreckel@ginac.de>
 825
 826         * src/real/input/cl_R_read.cc, src/complex/input/cl_N_read.cc:
 827           #undef stderr, if it's defined so cln::stderr isn't confused.
 828         * src/base/input/cl_read_globals.cc: stdin should not be extern.
 829
 830 2000-09-05  Richard Kreckel  <kreckel@ginac.de>
 831
 832         * include/cln/number.h (As): Fix it in namespace by suffixing `_As'
 833           to the appropiate method instead of prefixing `as_'.
 834         * src/complex/misc/cl_N_as.cc (cl_N_As): Likewise.
 835         * src/real/misc/cl_R_as.cc (cl_R_As): Likewise.
 836         * src/rational/misc/cl_RA_as.cc (cl_RA_As): Likewise.
 837         * src/integer/misc/cl_I_as.cc (cl_I_As): Likewise.
 838         * src/float/misc/cl_F_as.cc (cl_F_As): Likewise.
 839         * src/float/sfloat/misc/cl_SF_as.cc (cl_SF_As): Likewise.
 840         * src/float/lfloat/misc/cl_LF_as.cc (cl_LF_As): Likewise.
 841         * src/float/ffloat/misc/cl_FF_as.cc (cl_FF_As): Likewise.
 842         * src/float/dfloat/misc/cl_DF_as.cc (cl_DF_As): Likewise.
 843
 844 2000-09-05  Richard Kreckel  <kreckel@ginac.de>
 845
 846         * src/complex/transcendental/cl_C_expt_C.cc (expt): fix logic for
 847           the 0^y cases.
 848
 849 2000-08-30  Richard Kreckel  <kreckel@ginac.de>
 850
 851         * include/cln/number.h, cl_number::_as_cl_private_thing(): removed.
 852           Rearranged for a clearer distinction between macros and remaining
 853           identifiers, so Cint can parse it smoothly.
 854
 855 2000-08-29  Richard Kreckel  <kreckel@ginac.de>
 856
 857         * include/cln/number.h, the(const cl_number& x): New template
 858           function.
 859
 860 2000-08-29  Richard Kreckel  <kreckel@ginac.de>
 861
 862         * */*: Pushed CLN into a namespace `cln'.  While doing so, the
 863           following identifiers got their poor-man's namespace (i.e. the
 864           prefix `cl_') stripped off:
 865           cl_catalanconst()        ->  catalanconst()
 866           cl_compare()             ->  compare()
 867           cl_cos_sin()             ->  cos_sin()
 868           cl_cos_sin_t             ->  cos_sin_t
 869           cl_cosh_sinh()           ->  cosh_sinh()
 870           cl_cosh_sinh_t           ->  cosh_sinh_t
 871           cl_decoded_dfloat        ->  decoded_dfloat
 872           cl_decoded_ffloat        ->  decoded_ffloat
 873           cl_decoded_float         ->  decoded_float
 874           cl_decoded_lfloat        ->  decoded_lfloat
 875           cl_decoded_sfloat        ->  decoded_sfloat
 876           cl_default_float_format  ->  default_float_format
 877           cl_default_print_flags   ->  default_print_flags
 878           cl_default_random_state  ->  default_random_state
 879           cl_double_approx()       ->  double_approx()
 880           cl_equal()               ->  equal()
 881           cl_equal_hashcode()      ->  equal_hashcode()
 882           cl_eulerconst()          ->  eulerconst()
 883           cl_find_modint_ring()    ->  find_modint_ring()
 884           cl_find_univpoly_ring()  ->  find_univ_poly_ring()
 885           cl_float_approx()        ->  float_approx
 886           cl_float_format()        ->  float_format()
 887           cl_float_format_t        ->  float_format_t
 888           cl_free_hook()           ->  free_hook()
 889           cl_hermite()             ->  hermite()
 890           cl_laguerre()            ->  laguerre()
 891           cl_legendre()            ->  legandre()
 892           cl_malloc_hook()         ->  malloc_hook()
 893           cl_pi()                  ->  pi()
 894           cl_tschebychev()         ->  tschebychev()
 895           cl_zeta()                ->  zeta()
 896           NB: For functions these changes includes all signatures.
 897         * include/*: moved to include/cln/*, stripped `cl_' off filenames.
 898         * cln-config.in: new file.
 899
 900 2000-08-26  Bruno Haible  <haible@clisp.cons.org>
 901
 902         * autoconf/acgeneral.m4 (AC_OUTPUT): Use braces in exec_prefix default
 903           value, not parens.
 904
 905 2000-08-18  Bruno Haible  <haible@clisp.cons.org>
 906
 907         * include/cl_univpoly_modint.h: Fix typo.
 908
 909 2000-07-13  Bruno Haible  <haible@clisp.cons.org>
 910
 911         * src/float/input/cl_F_read_stream.cc (number_char_p): Accept '_',
 912           used as precision marker for floats.
 913           Reported by Keith Briggs (in 1998) and Thomas Roessler.
 914         * src/integer/input/cl_I_read_stream.cc (number_char_p): Likewise.
 915         * src/rational/input/cl_RA_read_stream.cc (number_char_p): Likewise.
 916         * src/real/input/cl_R_read_stream.cc (number_char_p): Likewise.
 917         * src/complex/input/cl_N_read_stream.cc (number_char_p): Likewise.
 918
 919 2000-06-22  Bruno Haible  <haible@clisp.cons.org>
 920
 921         * include/cl_object.h: Rename cl_word_size to cl_pointer_size.
 922         * include/cl_types.h (HAVE_FAST_LONGLONG): Also define on Irix6
 923           with N32 ABI.
 924           (cl_word_size): New macro.
 925         * src/float/sfloat/cl_SF.h: Use cl_pointer_size instead of
 926           cl_word_size.
 927
 928 2000-05-31  Bruno Haible  <haible@clisp.cons.org>
 929
 930         * tests/exam_I_floor.cc (integer_floor_tests): Add one more entry.
 931           From a sample that fails with gcc-2.95.2 on Sparc.
 932         * tests/exam_I_gcd.cc (integer_gcd_tests): Likewise.
 933
 934 2000-05-30  Richard Kreckel  <kreckel@ginac.de>
 935
 936         * configure.in, autoconf/aclocal.m4 (CL_GMP_H_VERSION, CL_GMP_CHECK):
 937           New macros.
 938         * configure.in, include/cl_config.h.in (CL_VERSION, CL_VERSION_MINOR,
 939           CL_VERSION_PATCHLEVEL): New definitions.
 940         * autoconf/config.guess, autoconf/config.sub, autoconf/ltconfig,
 941           autoconf/ltmain.sh: updated from FSF (libtool 1.3.5, etc).
 942         * src/Makefile.in, configure.in: release-variables renamed from
 943           CLN_* to CL_*.
 944         * configure.in: default to build both shared and static library
 945           (i.e. default to the most common behaviour).
 946
 947 2000-05-29  Richard Kreckel  <kreckel@ginac.de>
 948
 949         * autoconf/aclocal.m4 (CL_CANONICAL_HOST): Added missing changequote
 950           environment around the patch of 2000-05-23.
 951
 952 2000-05-29  Bruno Haible  <haible@clisp.cons.org>
 953
 954         * autoconf/aclocal.m4 (CL_PROG_INSTALL): Fix typo.
 955           Reported by Thomas Klausner <wiz@danbala.ifoer.tuwien.ac.at>.
 956
 957 2000-05-27  Richard Kreckel  <kreckel@ginac.de>
 958
 959         * src/float/lfloat/algebraic/cl_LF_sqrt.cc,
 960           src/base/digitseq/cl_DS_sqrt.cc: Readjusted break-even points.
 961
 962 2000-05-24  Richard Kreckel  <kreckel@ginac.de>
 963
 964         * autoconf/config.*: Updated to new version from FSF
 965           (the new libtool wants this).
 966         * src/Makefile.in: added $(LDFLAGS) to link step.
 967         * src/base/digitseq/cl_2DS_div.cc, cl_2DS_recip.cc: Readjusted
 968           break-even points.
 969
 970 2000-05-23  Bruno Haible  <haible@clisp.cons.org>
 971
 972         * autoconf/aclocal.m4 (CL_CANONICAL_HOST): Determine host_cpu,
 973           host_vendor, host_os correctly if $host has more than two hyphens.
 974
 975 2000-05-19  Richard Kreckel  <kreckel@ginac.de>
 976
 977         * src/base/digitseq/cl_DS_mul.cc: Rearranged break-even points to
 978           better match present-day CPUs whenever GMP3 is used.
 979         * src/base/digitseq/cl_DS_div.cc: dto.
 980         * src/TUNING: Added comment about order of tuning.
 981         * configure, configure.in: Safer GMP3-detection.
 982         * INSTALL.generic: Clarified behaviour of --with-gmp.
 983         * autoconf/config.guess: updated from Clisp-2000-03-06 sources.
 984
 985 2000-05-04  Richard Kreckel  <kreckel@ginac.de>
 986
 987         * gmp/: removed completely.  From now on we are going to link
 988           externally against libgmp3.0 or above!
 989         * configure, configure.in, Makefile.in, */Makeflags: removed support
 990           of internal gmp 2.0.2, like $GMP_INCLUDES, which should be done by
 991           setting $CPPFLAGS instead.
 992         * Makefile.in: Added libtool inter-library dependency for -lgmp and
 993           conforming interface versioning (-version-info).
 994         * autoconf/ltconfig, autoconf/ltmain.sh: Updated to newer versions
 995           from libtool 1.3.4.
 996
 997 2000-02-22  Bruno Haible  <haible@clisp.cons.org>
 998
 999         * src/base/digitseq/cl_asm_mips_.cc: Starting at argument 5 the
1000           parameter passing was changed for the MIPS n32 ABI.
1001
1002 2000-01-24  Richard Kreckel  <kreckel@ginac.de>
1003
1004         * gmp/*: Replaced the complete mpn sources with the ones from
1005           Debian since they are maintained while the ones from FSF
1006           aren't and there were problems on some architectures, PowerPC
1007           in particular.  See the file gmp/README.CLN.  This way the
1008           hard links in this directory have vanished, they were causing
1009           trouble for people working in AFS.  This became necessary for
1010           Debian, because there it woudn't compile on PPC.
1011
1012 2000-01-13  Richard Kreckel  <kreckel@ginac.de>
1013
1014         * Version 1.0.3 released.
1015
1016 2000-01-13  Richard Kreckel  <kreckel@ginac.de>
1017
1018         * src/base/cl_macros.h (nonreturning_function): For egcs and newer
1019           use __attribute__ ((__noreturn__)) instead of the __volatile__
1020           storage class.
1021           (nonreturning): Remove macro.
1022         * include/*: Minor fixes to stop -ansi -pedantic from complaining.
1023         * include/cl_integer.h: (doublefactorial): New declaration.
1024         * src/integer/misc/combin/cl_I_doublefactorial.cc: New file.
1025
1026 1999-12-18  Bruno Haible  <haible@clisp.cons.org>
1027
1028         * autoconf/acgeneral.m4 (AC_ARG_PROGRAM): Create conftestsed using
1029           "cat", not "echo".
1030         * autoconf/ltconfig: Improve support for recent FreeBSD 3.
1031         * include/cl_GV.h (cl_GV_vectorops): Change return type from 'T' to
1032           'const T', to match definition in src/vector/cl_GV_number.cc.
1033           Reported by Duncan Simpson <dps@io.stargate.co.uk>.
1034         * gmp/mpn/Makefile.in (.S.lo): Use 'if', not '&&', because '&&' may
1035           yield return code 1.
1036
1037 1999-09-07  Bruno Haible  <haible@clisp.cons.org>
1038
1039         * Version 1.0.2 released.
1040
1041 1999-09-06  Bruno Haible  <haible@clisp.cons.org>
1042
1043         * src/rational/cl_RA.h (integerp, ratiop): Instead of returning a
1044           boolean expression, write alternatives ending with either
1045           "return cl_true;" or "return cl_false;". This way, g++ does a
1046           better job inlining it.
1047         * src/float/cl_F.h (longfloatp): Likewise.
1048         * src/real/cl_R.h (rationalp, integerp, floatp): Likewise.
1049         * src/complex/cl_C.h (realp, complexp): Likewise.
1050
1051 1999-09-05  Bruno Haible  <haible@clisp.cons.org>
1052
1053         * include/cl_integer.h (cl_equal_hashcode): New declaration.
1054         * include/cl_rational.h (cl_equal_hashcode): New declaration.
1055         * include/cl_sfloat.h (cl_equal_hashcode): New declaration.
1056         * include/cl_ffloat.h (cl_equal_hashcode): New declaration.
1057         * include/cl_dfloat.h (cl_equal_hashcode): New declaration.
1058         * include/cl_lfloat.h (cl_equal_hashcode): New declaration.
1059         * include/cl_float.h (cl_equal_hashcode): New declaration.
1060         * include/cl_real.h (cl_equal_hashcode): New declaration.
1061         * include/cl_complex.h (cl_equal_hashcode): New declaration.
1062         * src/base/cl_N.h (equal_hashcode_low, equal_hashcode_one): New macros.
1063         * src/integer/misc/cl_I_eqhashcode.cc: New file.
1064         * src/rational/misc/cl_RA_eqhashcode.cc: New file.
1065         * src/float/sfloat/misc/cl_SF_eqhashcode.cc: New file.
1066         * src/float/ffloat/misc/cl_FF_eqhashcode.cc: New file.
1067         * src/float/dfloat/misc/cl_DF_eqhashcode.cc: New file.
1068         * src/float/lfloat/misc/cl_LF_eqhashcode.cc: New file.
1069         * src/float/misc/cl_F_eqhashcode.cc: New file.
1070         * src/real/misc/cl_R_eqhashcode.cc: New file.
1071         * src/complex/misc/cl_C_eqhashcode.cc: New file.
1072         * doc/cln.tex: Document `cl_equal_hashcode'.
1073
1074 1999-09-05  Bruno Haible  <haible@clisp.cons.org>
1075
1076         * include/cl_ring.h (cl_number_ring_ops): Add `contains' member.
1077           (cl_number_ring): New class.
1078           (cl_specialized_number_ring<T>): Inherit from cl_number_ring.
1079           (instanceof): New function.
1080         * src/integer/ring/cl_I_ring.cc (cl_I_p): New function.
1081         * src/integer/misc/cl_I_as.cc (cl_I_p): Add comment.
1082         * src/rational/ring/cl_RA_ring.cc (cl_RA_p): New function.
1083         * src/rational/misc/cl_RA_as.cc (cl_RA_p): Add comment.
1084         * src/real/ring/cl_R_ring.cc (cl_R_p): New function.
1085         * src/real/misc/cl_R_as.cc (cl_R_p): Add comment.
1086         * src/complex/ring/cl_C_ring.cc (cl_N_p): New function.
1087         * src/complex/misc/cl_N_as.cc (cl_N_p): Add comment.
1088         * doc/cln.tex: Document `instanceof'.
1089
1090 1999-09-05  Bruno Haible  <haible@clisp.cons.org>
1091
1092         * include/cl_rational.h (numerator, denominator): New declarations.
1093         * src/rational/elem/cl_RA_numerator.cc: New file.
1094         * src/rational/elem/cl_RA_denominator.cc: New file.
1095         * include/cl_integer.h (numerator, denominator): New inline functions.
1096         * doc/cln.tex: Document `numerator' and `denominator'.
1097
1098 1999-09-05  Bruno Haible  <haible@clisp.cons.org>
1099
1100         * src/rational/algebraic/cl_RA_rootp.cc (rootp): Fix endless loop
1101           in the integer case.
1102         * src/rational/algebraic/cl_RA_rootp_I.cc (rootp): Likewise.
1103
1104 1999-09-05  Bruno Haible  <haible@clisp.cons.org>
1105
1106         * include/cl_config.h.in: Support for sparc64 CPU.
1107         * include/cl_modules.h: Likewise.
1108         * include/cl_types.h: Likewise.
1109         * include/cl_object.h: Likewise.
1110         * include/cl_GV.h: Likewise.
1111         * src/Makefile.in: Likewise.
1112         * src/base/cl_alloca.h: Likewise.
1113         * src/base/cl_macros.h: Likewise.
1114         * src/base/cl_sysdep.h: Likewise.
1115         * src/base/cl_low.h: Likewise.
1116         * src/base/digitseq/cl_asm.h: Likewise.
1117         * src/base/digitseq/cl_asm_.cc: Likewise.
1118         * src/base/digitseq/cl_asm_sparc64.h: New file.
1119         * src/base/digitseq/cl_asm_sparc64_.cc: New file.
1120         * src/modinteger/cl_MI_int.h: Support for sparc64 CPU.
1121         * src/polynomial/elem/cl_UP_no_ring.cc: Likewise.
1122         * src/polynomial/elem/cl_UP_GF2.h: Likewise.
1123         * src/polynomial/elem/cl_asm_GF2.cc: Likewise.
1124
1125 1999-09-04  Bruno Haible  <haible@clisp.cons.org>
1126
1127         * src/base/digitseq/cl_asm_sparc_.cc (orc2_loop_up, orc2_loop_down):
1128           Use the `orn' instruction.
1129
1130 1999-08-14  Bruno Haible  <haible@clisp.cons.org>
1131
1132         Assume all platforms have <stdlib.h> and clock_t.
1133         * configure.in: Don't call CL_STDLIB_H and CL_CLOCK_T.
1134         * src/base/cl_base_config.h.in (CLOCK_T): Remove definition.
1135         * src/base/random/cl_random_from.cc: Use clock_t instead of CLOCK_T.
1136         * src/timing/cl_t_config.h.in (CLOCK_T): Remove definition.
1137         * src/timing/cl_t_current2.cc: Use clock_t instead of CLOCK_T.
1138
1139 1999-07-18  Bruno Haible  <haible@clisp.cons.org>
1140
1141         * gmp/config.guess: Link to autoconf/config.guess.
1142           gmp/config.sub: Link to autoconf/config.sub.
1143           Needed for Win32 platforms.
1144
1145 1999-07-17  Bruno Haible  <haible@clisp.cons.org>
1146
1147         * autoconf/aclocal.m4 (CL_CANONICAL_HOST_CPU): Distinguish "sparc" and
1148           "sparc64" according to the C compiler, not the uname result.
1149
1150 1999-06-17  Bruno Haible  <haible@clisp.cons.org>
1151
1152         * src/base/digitseq/cl_asm_sparc_.cc (compare_loop_up): Fix
1153           COUNTER_LOOPS version, fortunately not used yet.
1154
1155         * include/cl_modules.h: Prepare for gcc version 3.
1156
1157 1999-06-12  Bruno Haible  <haible@clisp.cons.org>
1158
1159         * src/rational/elem/cl_RA_plus.cc, src/rational/elem/cl_RA_minus.cc:
1160           Change the last call from I_I_to_RT to I_I_to_RA.
1161
1162 1999-06-09  Bruno Haible  <haible@clisp.cons.org>
1163
1164         * Version 1.0.1 released.
1165
1166 1999-06-09  Bruno Haible  <haible@clisp.cons.org>
1167
1168         * src/integer/cl_I.h (pFN_maxlength_digits_at): Define also when
1169           intDsize==64.
1170
1171 1999-06-08  Bruno Haible  <haible@clisp.cons.org>
1172
1173         * autoconf/intparam.c (printf_underscored): Change argument type to
1174           `const char*'.
1175         * include/cl_modules.h (CL_OUTPUT_LABEL, CL_JUMP_TO): New macros.
1176           (CL_PROVIDE, CL_PROVIDE_END): Use them.
1177         * include/cl_string.h (cl_heap_string::operator new): Return 1, not 0.
1178         * include/cl_GV.h (cl_GV_inner<T>::operator new): Likewise.
1179         * src/base/ring/cl_no_ring.cc (dummy_canonhom, dummy_expt_pos): Don't
1180           cast a cl_I to void here. Works around a bug in g++-2.95.
1181         * src/complex/misc/cl_C_class.cc: Include "cl_C.h".
1182         * src/polynomial/elem/cl_UP_no_ring.cc (dummy_canonhom,
1183            dummy_expt_pos): Don't cast a cl_I to void here. Works around a bug
1184           in g++-2.95.
1185         * src/polynomial/elem/cl_asm_sparc_GF2.cc (DECLARE_FUNCTION): New
1186           macro.
1187         * src/rational/misc/cl_RA_class.cc: Include "cl_RA.h".
1188         * src/vector/cl_GV_I.cc (cl_heap_GV_I_general::operator new,
1189            DEFINE_cl_heap_GV_I_bits): Return 1, not 0.
1190         * src/vector/cl_GV_number.cc (cl_heap_GV_number_general::operator new):
1191           Likewise.
1192
1193 1999-06-01  Bruno Haible  <haible@clisp.cons.org>
1194
1195         * autoconf/aclocal.m4 (CL_CANONICAL_HOST_CPU): Canonicalize alpha
1196           variants to alpha.
1197
1198 1999-05-29  Bruno Haible  <haible@clisp.cons.org>
1199
1200         * src/base/digitseq/cl_asm_i386_.cc (DECLARE_FUNCTION): Treat
1201           OpenBSD like NetBSD.
1202         * src/base/digitseq/cl_asm_sparc_.cc (DECLARE_FUNCTION): Likewise.
1203         * src/base/digitseq/cl_asm_m68k_.cc (DECLARE_FUNCTION): Treat
1204           OpenBSD like NetBSD, and Linux/ELF like SVR4.
1205
1206 1999-05-16  Bruno Haible  <haible@clisp.cons.org>
1207
1208         * src/base/cl_low.h (integerlength32) [__rs6000__]: Use old assembler
1209           syntax on AIX systems and new assembler syntax on non-AIX systems.
1210
1211 1999-05-01  Bruno Haible  <haible@clisp.cons.org>
1212
1213         * autoconf/config.guess, autoconf/config.sub: Upgrade to newest
1214           version from GNU CVS.
1215
1216 1999-04-24  Bruno Haible  <haible@clisp.cons.org>
1217
1218         * src/integer/bitwise/cl_I_logand.cc (logand): Optimize the case when
1219           either operand is a positive fixnum, O(1) instead of O(N).
1220         * src/integer/bitwise/cl_I_lognand.cc (lognand): Likewise.
1221         * src/integer/bitwise/cl_I_logandc2.cc (logandc2): Likewise for the
1222           first operand.
1223
1224 1999-04-14  Bruno Haible  <haible@clisp.cons.org>
1225
1226         * autoconf/aclocal.m4 (CL_GLOBAL_CONSTRUCTORS): Add check whether
1227           ctor/dtor needs to be exported.
1228         * include/cl_config.h.in (CL_NEED_GLOBALIZE_CTORDTOR): New macro.
1229         * include/cl_modules.h (CL_GLOBALIZE_JUMP_LABEL): Renamed from
1230           CL_GLOBALIZE_LABEL.
1231           (CL_GLOBALIZE_LABEL): New macro.
1232           (CL_GLOBALIZE_CTORDTOR_LABEL): Renamed from CL_GLOBALIZE_ASM_LABEL.
1233           (CL_PROVIDE): Update.
1234
1235 1999-04-12  Bruno Haible  <haible@clisp.cons.org>
1236
1237         * src/Makefile.in ($(ASMFILES_S)): On HPPA, ignore preprocessing
1238           errors ("unterminated string or character constant").
1239           ($(ASMFILES_LO)): On HPPA, try with various settings of
1240           COMPILER_PATH, in order to try /usr/ccs/bin/as and /bin/as.
1241
1242 1999-04-11  Bruno Haible  <haible@clisp.cons.org>
1243
1244         * INSTALL: Mention gmp problems on MIPS.
1245         * doc/cln.tex: Likewise.
1246
1247 1999-03-24  Mumit Khan  <khan@xraylith.wisc.edu>
1248
1249         * src/Makefile.in (SUBDIRS): Filter out CVS and RCS directories from
1250           the source tree.
1251         * include/cl_modules.h (CL_GLOBALIZE_LABEL): Define for Win32.
1252           (CL_GLOBALIZE_ASM_LABEL): New macro.
1253           (CL_PROVIDE): Use it.
1254         * src/base/random/cl_random_from.cc: Handle WIN32.
1255         * src/timing/cl_t_current.cc: Likewise.
1256
1257 1999-03-15  Bruno Haible  <haible@clisp.cons.org>
1258
1259         * autoconf/intparam.c (main7): Use %lX instead of %X for a `long'.
1260           (main8): Adapt for C++.
1261
1262 1999-03-09  Bruno Haible  <haible@clisp.cons.org>
1263
1264         * INSTALL: Mention egcs-1.1 problems on Sparc.
1265         * doc/cln.tex: Likewise.
1266
1267 1999-03-08  Bruno Haible  <haible@clisp.cons.org>
1268
1269         * autoconf/aclocal.m4 (CL_FPU_CONTROL): Fix the "checking for"
1270           messages.
1271
1272 1999-02-25  Bruno Haible  <haible@clisp.cons.org>
1273
1274         * autoconf/aclocal.m4: In test programs, declare `int main()', not
1275           `main()'.
1276
1277         * lidia-interface/src/interfaces/integers/cln/bigint.c
1278           (bigint_to_string): Fix for negative arguments.
1279
1280         * src/base/cl_low.h: Check for `__sparc__', not `SPARC'.
1281         * src/base/cl_alloca.h: Likewise.
1282
1283         * src/base/cl_low.h: Eliminate CLISP style "# " comments.
1284         * src/base/digitseq/cl_asm_arm_.cc,
1285           src/base/digitseq/cl_asm_mips_.cc,
1286           src/float/dfloat/elem/cl_DF_mul.cc: Likewise.
1287
1288         * src/modinteger/cl_MI_pow2.h,
1289           src/modinteger/cl_MI_pow2m1.h,
1290           src/modinteger/cl_MI_pow2p1.h: Workaround g++-2.7.2 inlining bug.
1291
1292 1999-01-18  Bruno Haible  <haible@clisp.cons.org>
1293
1294         * autoconf/acgeneral.m4,
1295           autoconf/acspecific.m4: Upgrade to autoconf-2.13.
1296         * autoconf/config.guess, autoconf/config.sub: Likewise.
1297         * autoconf/aclocal.m4 (CL_ALLOCA): Test for _MSC_VER and alloca being
1298           a macro. Use ${ac_objext}.
1299         * src/base/cl_macros.h (alloca): Put _MSC_VER test before the others,
1300           conforming with CL_ALLOCA.
1301
1302 1999-01-12  Bruno Haible  <haible@clisp.cons.org>
1303
1304         * Version 1.0 released.
1305