2000-11-17 Richard Kreckel * cln-config.1: added manpage, as required by a couple of distros. * Makefile.in, doc/Makefile.in: target install depends on installdirs. 2000-11-16 Richard Kreckel * autoconf/aclocal.m4 (CL_GMP_SET_UINTD): New macro... * configure.in: ...used here. * include/cln/config.h.in: Put in macros defined by CL_GMP_SET_UINTD... * include/cln/types.h: ...used here. * autoconf/acgeneral.m4, autoconf/aclocal.m4: Adhere to FHS. 2000-11-13 Richard Kreckel * src/base/digitseq/cl_asm.h: Test if (intDsize==32) for MIPS and HPPA, in order to guard against an accidented configuration. * src/integer/conv/cl_I_to_digits.cc (table): member b_hoch_k of struct power_table_entry initialized as ULL instead of as UL, if intDsize==64 (caused misprinting on MIPS w/ GMP). * src/base/cl_macros.h (minus_bitm, bitc): implemented ULL, if HAVE_FAST_LONGLONG. * src/integer/bitwise/cl_I_mkfx.cc (mkf_extract): Likewise. * src/integer/conv/cl_I_from_L.cc (cl_I_from_L): Added trival generation of Bignum for intDsize==64 and a notreached-check at end. * autoconf/config.guess, autoconf/config.sub: updated from FSF. * include/cln/config.h.in: Prepared support for IA64. * include/cln/types.h: Likewise. * include/cln/object.h: Likewise. * include/cln/modules.h: Likewise. * src/base/cl_macros.h (nonreturning_function): Likewise (NUE's compiler claims __GNUC_MINOR__==9). 2000-11-03 Richard Kreckel * src/base/cl_macros.h (bit, minus_bit): changed criterion for ULL from HAVE_DD to HAVE_FAST_LONGLONG. * src/base/cl_macros.h (bitm): implemented ULL, if HAVE_FAST_LONGLONG. * src/base/cl_low.h: actually no need to include "cln/types.h" here. * src/base/cl_low.h (logcount_64): always ULL, independent of HAVE_DD. * src/base/random/cl_UL_random.cc (random32): a is always ULL. 2000-11-01 Richard Kreckel * include/cln/object.h (cl_combine): define additional signatures, if HAVE_LONGLONG is defined, in order to keep the compiler happy. * src/base/cl_macros.h: include "cln/types.h", since we need HAVE_DD... * src/base/cl_macros.h (bit): ...for this macro... * src/base/cl_macros.h (minus_bit): ...and this one. * src/base/cl_low.h: include "cln/types.h", since we need HAVE_DD... * src/base/cl_low.h (logcount_64): ...for this macro. * src/base/random/cl_UL_random.cc (random32): if HAVE_DD a is an ULL. * src/integer/gcd/cl_I_gcd_aux2.cc (floorDD): fixed algorithmic bug that turned up when intDsize==32 and cl_word_size==64. * src/float/dfloat/elem/cl_DF_div.cc (operator/): fixed a missing cast to uint64 that turned up when intDsize==32 and cl_word_size==64. 2000-10-29 Richard Kreckel * src/real/input/cl_R_read.cc, src/complex/input/cl_N_read.cc: #undef stderr, if it's defined so cln::stderr isn't confused. * src/base/input/cl_read_globals.cc: stdin should not be extern. 2000-09-05 Richard Kreckel * include/cln/number.h (As): Fix it in namespace by suffixing `_As' to the appropiate method instead of prefixing `as_'. * src/complex/misc/cl_N_as.cc (cl_N_As): Likewise. * src/real/misc/cl_R_as.cc (cl_R_As): Likewise. * src/rational/misc/cl_RA_as.cc (cl_RA_As): Likewise. * src/integer/misc/cl_I_as.cc (cl_I_As): Likewise. * src/float/misc/cl_F_as.cc (cl_F_As): Likewise. * src/float/sfloat/misc/cl_SF_as.cc (cl_SF_As): Likewise. * src/float/lfloat/misc/cl_LF_as.cc (cl_LF_As): Likewise. * src/float/ffloat/misc/cl_FF_as.cc (cl_FF_As): Likewise. * src/float/dfloat/misc/cl_DF_as.cc (cl_DF_As): Likewise. 2000-09-05 Richard Kreckel * src/complex/transcendental/cl_C_expt_C.cc (expt): fix logic for the 0^y cases. 2000-08-30 Richard Kreckel * include/cln/number.h, cl_number::_as_cl_private_thing(): removed. Rearranged for a clearer distinction between macros and remaining identifiers, so Cint can parse it smoothly. 2000-08-29 Richard Kreckel * include/cln/number.h, the(const cl_number& x): New template function. 2000-08-29 Richard Kreckel * */*: Pushed CLN into a namespace `cln'. While doing so, the following identifiers got their poor-man's namespace (i.e. the prefix `cl_') stripped off: cl_catalanconst() -> catalanconst() cl_compare() -> compare() cl_cos_sin() -> cos_sin() cl_cos_sin_t -> cos_sin_t cl_cosh_sinh() -> cosh_sinh() cl_cosh_sinh_t -> cosh_sinh_t cl_decoded_dfloat -> decoded_dfloat cl_decoded_ffloat -> decoded_ffloat cl_decoded_float -> decoded_float cl_decoded_lfloat -> decoded_lfloat cl_decoded_sfloat -> decoded_sfloat cl_default_float_format -> default_float_format cl_default_print_flags -> default_print_flags cl_default_random_state -> default_random_state cl_double_approx() -> double_approx() cl_equal() -> equal() cl_equal_hashcode() -> equal_hashcode() cl_eulerconst() -> eulerconst() cl_find_modint_ring() -> find_modint_ring() cl_find_univpoly_ring() -> find_univ_poly_ring() cl_float_approx() -> float_approx cl_float_format() -> float_format() cl_float_format_t -> float_format_t cl_free_hook() -> free_hook() cl_hermite() -> hermite() cl_laguerre() -> laguerre() cl_legendre() -> legandre() cl_malloc_hook() -> malloc_hook() cl_pi() -> pi() cl_tschebychev() -> tschebychev() cl_zeta() -> zeta() NB: For functions these changes includes all signatures. * include/*: moved to include/cln/*, stripped `cl_' off filenames. * cln-config.in: new file. 2000-08-26 Bruno Haible * autoconf/acgeneral.m4 (AC_OUTPUT): Use braces in exec_prefix default value, not parens. 2000-08-18 Bruno Haible * include/cl_univpoly_modint.h: Fix typo. 2000-07-13 Bruno Haible * src/float/input/cl_F_read_stream.cc (number_char_p): Accept '_', used as precision marker for floats. Reported by Keith Briggs (in 1998) and Thomas Roessler. * src/integer/input/cl_I_read_stream.cc (number_char_p): Likewise. * src/rational/input/cl_RA_read_stream.cc (number_char_p): Likewise. * src/real/input/cl_R_read_stream.cc (number_char_p): Likewise. * src/complex/input/cl_N_read_stream.cc (number_char_p): Likewise. 2000-06-22 Bruno Haible * include/cl_object.h: Rename cl_word_size to cl_pointer_size. * include/cl_types.h (HAVE_FAST_LONGLONG): Also define on Irix6 with N32 ABI. (cl_word_size): New macro. * src/float/sfloat/cl_SF.h: Use cl_pointer_size instead of cl_word_size. 2000-05-31 Bruno Haible * tests/exam_I_floor.cc (integer_floor_tests): Add one more entry. From a sample that fails with gcc-2.95.2 on Sparc. * tests/exam_I_gcd.cc (integer_gcd_tests): Likewise. 2000-05-30 Richard Kreckel * configure.in, autoconf/aclocal.m4 (CL_GMP_H_VERSION, CL_GMP_CHECK): New macros. * configure.in, include/cl_config.h.in (CL_VERSION, CL_VERSION_MINOR, CL_VERSION_PATCHLEVEL): New definitions. * autoconf/config.guess, autoconf/config.sub, autoconf/ltconfig, autoconf/ltmain.sh: updated from FSF (libtool 1.3.5, etc). * src/Makefile.in, configure.in: release-variables renamed from CLN_* to CL_*. * configure.in: default to build both shared and static library (i.e. default to the most common behaviour). 2000-05-29 Richard Kreckel * autoconf/aclocal.m4 (CL_CANONICAL_HOST): Added missing changequote environment around the patch of 2000-05-23. 2000-05-29 Bruno Haible * autoconf/aclocal.m4 (CL_PROG_INSTALL): Fix typo. Reported by Thomas Klausner . 2000-05-27 Richard Kreckel * src/float/lfloat/algebraic/cl_LF_sqrt.cc, src/base/digitseq/cl_DS_sqrt.cc: Readjusted break-even points. 2000-05-24 Richard Kreckel * autoconf/config.*: Updated to new version from FSF (the new libtool wants this). * src/Makefile.in: added $(LDFLAGS) to link step. * src/base/digitseq/cl_2DS_div.cc, cl_2DS_recip.cc: Readjusted break-even points. 2000-05-23 Bruno Haible * autoconf/aclocal.m4 (CL_CANONICAL_HOST): Determine host_cpu, host_vendor, host_os correctly if $host has more than two hyphens. 2000-05-19 Richard Kreckel * src/base/digitseq/cl_DS_mul.cc: Rearranged break-even points to better match present-day CPUs whenever GMP3 is used. * src/base/digitseq/cl_DS_div.cc: dto. * src/TUNING: Added comment about order of tuning. * configure, configure.in: Safer GMP3-detection. * INSTALL.generic: Clarified behaviour of --with-gmp. * autoconf/config.guess: updated from Clisp-2000-03-06 sources. 2000-05-04 Richard Kreckel * gmp/: removed completely. From now on we are going to link externally against libgmp3.0 or above! * configure, configure.in, Makefile.in, */Makeflags: removed support of internal gmp 2.0.2, like $GMP_INCLUDES, which should be done by setting $CPPFLAGS instead. * Makefile.in: Added libtool inter-library dependency for -lgmp and conforming interface versioning (-version-info). * autoconf/ltconfig, autoconf/ltmain.sh: Updated to newer versions from libtool 1.3.4. 2000-02-22 Bruno Haible * src/base/digitseq/cl_asm_mips_.cc: Starting at argument 5 the parameter passing was changed for the MIPS n32 ABI. 2000-01-24 Richard Kreckel * gmp/*: Replaced the complete mpn sources with the ones from Debian since they are maintained while the ones from FSF aren't and there were problems on some architectures, PowerPC in particular. See the file gmp/README.CLN. This way the hard links in this directory have vanished, they were causing trouble for people working in AFS. This became necessary for Debian, because there it woudn't compile on PPC. 2000-01-13 Richard Kreckel * Version 1.0.3 released. 2000-01-13 Richard Kreckel * src/base/cl_macros.h (nonreturning_function): For egcs and newer use __attribute__ ((__noreturn__)) instead of the __volatile__ storage class. (nonreturning): Remove macro. * include/*: Minor fixes to stop -ansi -pedantic from complaining. * include/cl_integer.h: (doublefactorial): New declaration. * src/integer/misc/combin/cl_I_doublefactorial.cc: New file. 1999-12-18 Bruno Haible * autoconf/acgeneral.m4 (AC_ARG_PROGRAM): Create conftestsed using "cat", not "echo". * autoconf/ltconfig: Improve support for recent FreeBSD 3. * include/cl_GV.h (cl_GV_vectorops): Change return type from 'T' to 'const T', to match definition in src/vector/cl_GV_number.cc. Reported by Duncan Simpson . * gmp/mpn/Makefile.in (.S.lo): Use 'if', not '&&', because '&&' may yield return code 1. 1999-09-07 Bruno Haible * Version 1.0.2 released. 1999-09-06 Bruno Haible * src/rational/cl_RA.h (integerp, ratiop): Instead of returning a boolean expression, write alternatives ending with either "return cl_true;" or "return cl_false;". This way, g++ does a better job inlining it. * src/float/cl_F.h (longfloatp): Likewise. * src/real/cl_R.h (rationalp, integerp, floatp): Likewise. * src/complex/cl_C.h (realp, complexp): Likewise. 1999-09-05 Bruno Haible * include/cl_integer.h (cl_equal_hashcode): New declaration. * include/cl_rational.h (cl_equal_hashcode): New declaration. * include/cl_sfloat.h (cl_equal_hashcode): New declaration. * include/cl_ffloat.h (cl_equal_hashcode): New declaration. * include/cl_dfloat.h (cl_equal_hashcode): New declaration. * include/cl_lfloat.h (cl_equal_hashcode): New declaration. * include/cl_float.h (cl_equal_hashcode): New declaration. * include/cl_real.h (cl_equal_hashcode): New declaration. * include/cl_complex.h (cl_equal_hashcode): New declaration. * src/base/cl_N.h (equal_hashcode_low, equal_hashcode_one): New macros. * src/integer/misc/cl_I_eqhashcode.cc: New file. * src/rational/misc/cl_RA_eqhashcode.cc: New file. * src/float/sfloat/misc/cl_SF_eqhashcode.cc: New file. * src/float/ffloat/misc/cl_FF_eqhashcode.cc: New file. * src/float/dfloat/misc/cl_DF_eqhashcode.cc: New file. * src/float/lfloat/misc/cl_LF_eqhashcode.cc: New file. * src/float/misc/cl_F_eqhashcode.cc: New file. * src/real/misc/cl_R_eqhashcode.cc: New file. * src/complex/misc/cl_C_eqhashcode.cc: New file. * doc/cln.tex: Document `cl_equal_hashcode'. 1999-09-05 Bruno Haible * include/cl_ring.h (cl_number_ring_ops): Add `contains' member. (cl_number_ring): New class. (cl_specialized_number_ring): Inherit from cl_number_ring. (instanceof): New function. * src/integer/ring/cl_I_ring.cc (cl_I_p): New function. * src/integer/misc/cl_I_as.cc (cl_I_p): Add comment. * src/rational/ring/cl_RA_ring.cc (cl_RA_p): New function. * src/rational/misc/cl_RA_as.cc (cl_RA_p): Add comment. * src/real/ring/cl_R_ring.cc (cl_R_p): New function. * src/real/misc/cl_R_as.cc (cl_R_p): Add comment. * src/complex/ring/cl_C_ring.cc (cl_N_p): New function. * src/complex/misc/cl_N_as.cc (cl_N_p): Add comment. * doc/cln.tex: Document `instanceof'. 1999-09-05 Bruno Haible * include/cl_rational.h (numerator, denominator): New declarations. * src/rational/elem/cl_RA_numerator.cc: New file. * src/rational/elem/cl_RA_denominator.cc: New file. * include/cl_integer.h (numerator, denominator): New inline functions. * doc/cln.tex: Document `numerator' and `denominator'. 1999-09-05 Bruno Haible * src/rational/algebraic/cl_RA_rootp.cc (rootp): Fix endless loop in the integer case. * src/rational/algebraic/cl_RA_rootp_I.cc (rootp): Likewise. 1999-09-05 Bruno Haible * include/cl_config.h.in: Support for sparc64 CPU. * include/cl_modules.h: Likewise. * include/cl_types.h: Likewise. * include/cl_object.h: Likewise. * include/cl_GV.h: Likewise. * src/Makefile.in: Likewise. * src/base/cl_alloca.h: Likewise. * src/base/cl_macros.h: Likewise. * src/base/cl_sysdep.h: Likewise. * src/base/cl_low.h: Likewise. * src/base/digitseq/cl_asm.h: Likewise. * src/base/digitseq/cl_asm_.cc: Likewise. * src/base/digitseq/cl_asm_sparc64.h: New file. * src/base/digitseq/cl_asm_sparc64_.cc: New file. * src/modinteger/cl_MI_int.h: Support for sparc64 CPU. * src/polynomial/elem/cl_UP_no_ring.cc: Likewise. * src/polynomial/elem/cl_UP_GF2.h: Likewise. * src/polynomial/elem/cl_asm_GF2.cc: Likewise. 1999-09-04 Bruno Haible * src/base/digitseq/cl_asm_sparc_.cc (orc2_loop_up, orc2_loop_down): Use the `orn' instruction. 1999-08-14 Bruno Haible Assume all platforms have and clock_t. * configure.in: Don't call CL_STDLIB_H and CL_CLOCK_T. * src/base/cl_base_config.h.in (CLOCK_T): Remove definition. * src/base/random/cl_random_from.cc: Use clock_t instead of CLOCK_T. * src/timing/cl_t_config.h.in (CLOCK_T): Remove definition. * src/timing/cl_t_current2.cc: Use clock_t instead of CLOCK_T. 1999-07-18 Bruno Haible * gmp/config.guess: Link to autoconf/config.guess. gmp/config.sub: Link to autoconf/config.sub. Needed for Win32 platforms. 1999-07-17 Bruno Haible * autoconf/aclocal.m4 (CL_CANONICAL_HOST_CPU): Distinguish "sparc" and "sparc64" according to the C compiler, not the uname result. 1999-06-17 Bruno Haible * src/base/digitseq/cl_asm_sparc_.cc (compare_loop_up): Fix COUNTER_LOOPS version, fortunately not used yet. * include/cl_modules.h: Prepare for gcc version 3. 1999-06-12 Bruno Haible * src/rational/elem/cl_RA_plus.cc, src/rational/elem/cl_RA_minus.cc: Change the last call from I_I_to_RT to I_I_to_RA. 1999-06-09 Bruno Haible * Version 1.0.1 released. 1999-06-09 Bruno Haible * src/integer/cl_I.h (pFN_maxlength_digits_at): Define also when intDsize==64. 1999-06-08 Bruno Haible * autoconf/intparam.c (printf_underscored): Change argument type to `const char*'. * include/cl_modules.h (CL_OUTPUT_LABEL, CL_JUMP_TO): New macros. (CL_PROVIDE, CL_PROVIDE_END): Use them. * include/cl_string.h (cl_heap_string::operator new): Return 1, not 0. * include/cl_GV.h (cl_GV_inner::operator new): Likewise. * src/base/ring/cl_no_ring.cc (dummy_canonhom, dummy_expt_pos): Don't cast a cl_I to void here. Works around a bug in g++-2.95. * src/complex/misc/cl_C_class.cc: Include "cl_C.h". * src/polynomial/elem/cl_UP_no_ring.cc (dummy_canonhom, dummy_expt_pos): Don't cast a cl_I to void here. Works around a bug in g++-2.95. * src/polynomial/elem/cl_asm_sparc_GF2.cc (DECLARE_FUNCTION): New macro. * src/rational/misc/cl_RA_class.cc: Include "cl_RA.h". * src/vector/cl_GV_I.cc (cl_heap_GV_I_general::operator new, DEFINE_cl_heap_GV_I_bits): Return 1, not 0. * src/vector/cl_GV_number.cc (cl_heap_GV_number_general::operator new): Likewise. 1999-06-01 Bruno Haible * autoconf/aclocal.m4 (CL_CANONICAL_HOST_CPU): Canonicalize alpha variants to alpha. 1999-05-29 Bruno Haible * src/base/digitseq/cl_asm_i386_.cc (DECLARE_FUNCTION): Treat OpenBSD like NetBSD. * src/base/digitseq/cl_asm_sparc_.cc (DECLARE_FUNCTION): Likewise. * src/base/digitseq/cl_asm_m68k_.cc (DECLARE_FUNCTION): Treat OpenBSD like NetBSD, and Linux/ELF like SVR4. 1999-05-16 Bruno Haible * src/base/cl_low.h (integerlength32) [__rs6000__]: Use old assembler syntax on AIX systems and new assembler syntax on non-AIX systems. 1999-05-01 Bruno Haible * autoconf/config.guess, autoconf/config.sub: Upgrade to newest version from GNU CVS. 1999-04-24 Bruno Haible * src/integer/bitwise/cl_I_logand.cc (logand): Optimize the case when either operand is a positive fixnum, O(1) instead of O(N). * src/integer/bitwise/cl_I_lognand.cc (lognand): Likewise. * src/integer/bitwise/cl_I_logandc2.cc (logandc2): Likewise for the first operand. 1999-04-14 Bruno Haible * autoconf/aclocal.m4 (CL_GLOBAL_CONSTRUCTORS): Add check whether ctor/dtor needs to be exported. * include/cl_config.h.in (CL_NEED_GLOBALIZE_CTORDTOR): New macro. * include/cl_modules.h (CL_GLOBALIZE_JUMP_LABEL): Renamed from CL_GLOBALIZE_LABEL. (CL_GLOBALIZE_LABEL): New macro. (CL_GLOBALIZE_CTORDTOR_LABEL): Renamed from CL_GLOBALIZE_ASM_LABEL. (CL_PROVIDE): Update. 1999-04-12 Bruno Haible * src/Makefile.in ($(ASMFILES_S)): On HPPA, ignore preprocessing errors ("unterminated string or character constant"). ($(ASMFILES_LO)): On HPPA, try with various settings of COMPILER_PATH, in order to try /usr/ccs/bin/as and /bin/as. 1999-04-11 Bruno Haible * INSTALL: Mention gmp problems on MIPS. * doc/cln.tex: Likewise. 1999-03-24 Mumit Khan * src/Makefile.in (SUBDIRS): Filter out CVS and RCS directories from the source tree. * include/cl_modules.h (CL_GLOBALIZE_LABEL): Define for Win32. (CL_GLOBALIZE_ASM_LABEL): New macro. (CL_PROVIDE): Use it. * src/base/random/cl_random_from.cc: Handle WIN32. * src/timing/cl_t_current.cc: Likewise. 1999-03-15 Bruno Haible * autoconf/intparam.c (main7): Use %lX instead of %X for a `long'. (main8): Adapt for C++. 1999-03-09 Bruno Haible * INSTALL: Mention egcs-1.1 problems on Sparc. * doc/cln.tex: Likewise. 1999-03-08 Bruno Haible * autoconf/aclocal.m4 (CL_FPU_CONTROL): Fix the "checking for" messages. 1999-02-25 Bruno Haible * autoconf/aclocal.m4: In test programs, declare `int main()', not `main()'. * lidia-interface/src/interfaces/integers/cln/bigint.c (bigint_to_string): Fix for negative arguments. * src/base/cl_low.h: Check for `__sparc__', not `SPARC'. * src/base/cl_alloca.h: Likewise. * src/base/cl_low.h: Eliminate CLISP style "# " comments. * src/base/digitseq/cl_asm_arm_.cc, src/base/digitseq/cl_asm_mips_.cc, src/float/dfloat/elem/cl_DF_mul.cc: Likewise. * src/modinteger/cl_MI_pow2.h, src/modinteger/cl_MI_pow2m1.h, src/modinteger/cl_MI_pow2p1.h: Workaround g++-2.7.2 inlining bug. 1999-01-18 Bruno Haible * autoconf/acgeneral.m4, autoconf/acspecific.m4: Upgrade to autoconf-2.13. * autoconf/config.guess, autoconf/config.sub: Likewise. * autoconf/aclocal.m4 (CL_ALLOCA): Test for _MSC_VER and alloca being a macro. Use ${ac_objext}. * src/base/cl_macros.h (alloca): Put _MSC_VER test before the others, conforming with CL_ALLOCA. 1999-01-12 Bruno Haible * Version 1.0 released.