976a13157c
Prepare for autoconf-2.60.
20 years ago
6fdd87f5e7
* Cumulative patch including Bruno's work on large fixnums on 64 bit machines.
20 years ago
84bb8ee640
* Updated for main branch.
20 years ago
497f359571
* Add support for OpenBSD.
20 years ago
eaff3d906d
* Release date corrected.
20 years ago
37a73f2097
* Prepare for release 1.1.11.
20 years ago
515cd347c4
* src/integer/conv/cl_I_cached_power.h: New file.
* src/integer/conv/cl_I_cached_power.cc: New file.
Contains power_table and cached_power_table previously...
* src/integer/conv/cl_I_to_digits.cc: ...here.
* src/integer/conv/cl_I_from_digits.cc: Use cached powers.
20 years ago
3480230ed3
* src/integer/conv/cl_I_from_digits.cc: Made input of all numbers in
non-power-of-two base much faster.
* tests/test_I_io.cc: New file...
* tests/Makefile.in, tests/test_I.cc: ...used here.
20 years ago
a0c8cec755
* Release 1.1.10.
20 years ago
80af318f28
* src/Makefile.in: Accept CPPFLAGS from environment.
* examples/Makefile.in: Likewise.
* benchmarks/Makefile.in: Likewise.
* tests/Makefile.in: Likewise.
20 years ago
42a192cf59
Use full pathname.
20 years ago
ceddc7b6d5
Have split aclocal.m4 into individual files.
20 years ago
8d0e9dadb8
* include/cln/modules.h (CL_OUTPUT_LABEL): Work around redundant
duplication of basic blocks on m68k.
20 years ago
8dd3e6cf4b
* include/cln/modules.h (CL_JUMP_TO): Fix mips* brokenness.
20 years ago
1e7aede9bd
Avoid shifting a 32-bit zero value by more than 31 bits.
20 years ago
c4e47eb4f1
Make the long-float overflow check work on 64-bit platforms.
20 years ago
134b80ebc1
Work around g++ 4.0 "duplicated labels" problem.
20 years ago
c437fd0700
Make it possible to cross-compile CLN.
20 years ago
0e41626478
Define HAVE_LONGLONG and HAVE_LONGDOUBLE when cross-compiling.
20 years ago
0eff524c4d
Fix for an error message when cross-compiling:
checking whether getrusage works... guessing no
./configure: line 20174: test: too many arguments
20 years ago
e269351d54
The patch of 2005-05-01 made it impossible to test the type of a cl_UP
by comparing with &cl_class_univpoly_ring. We need an alternative:
* include/cln/object.h (cl_class_flags_modint_ring): New #define...
* src/polynomial/elem/cl_UP.cc (cl_class_univpoly_ring): ...used here.
* src/polynomial/elem/cl_UP_GF2.h: Likewise.
* src/polynomial/elem/cl_UP_MI.h: Likewise.
* src/polynomial/elem/cl_UP_gen.h: Likewise.
* src/polynomial/elem/cl_UP_number.h: Likewise.
21 years ago
126552d49e
* m4/cc.m4 (CL_AS_NOEXECSTACK): New macro...
* configure.ac: ...used here for setting ASMFLAGS...
* src/Makefile.in: ...which are used here.
21 years ago
77bcf72d79
* include/cln/config.h.in: Add support for PowerPC 64 CPU.
* include/cln/modules.h: Likewise.
* include/cln/object.h: Likewise.
* include/cln/types.h: Likewise.
21 years ago
2ebe834823
Make out of the box build on x86_64 system with complete 32 bit
userland possible
* include/cln/config.h.in: Don't #define __x86_64__ when
__i386__ is defined.
* src/base/digitseq/cl_asm_x86_64_.cc: New file.
* doc/cln.tex: Revert workaround description introduced 2005-05-02.
21 years ago
afdca11f60
If a directory doesn't exist, fail instead of doing infinite recursion.
21 years ago
48bd67094d
Speed up the linking step
* src/Makefile.in: Use -objectlist for linking libcln.la.
21 years ago
bf3efa7255
* autoconf/ltmain.sh: Upgrade to libtool-1.5.16.
* m4/libtool.m4: Upgrade to libtool-1.5.16 (without Comeau patch).
21 years ago
5395692316
* doc/cln.tex: Document what to do on a x86_64 machine with 32-bit userland.
21 years ago
6d4cabf35e
Fix crashes in find_univpoly_ring and related functions
* include/cln/modinteger.h: Remove vptr from cl_heap_modint_ring;
remove declaration of cl_class cl_class_modint_ring.
* include/cln/univpoly.h: Remove vptr from cl_heap_univpoly_ring;
remove declaration of cl_class_univpoly_ring.
* include/cln/object.h: cl_class_flags_modint_ring: New #define...
* src/modinteger/cl_MI.cc: ...used in cl_class_modint_ring.
* src/modinteger/cl_MI_fix16.h: No vptr, but static dtor and type flag.
* src/modinteger/cl_MI_fix29.h: Likewise.
* src/modinteger/cl_MI_fix32.h: Likewise.
* src/modinteger/cl_MI_int32.h: Likewise.
* src/modinteger/cl_MI_montgom.h: Likewise.
* src/modinteger/cl_MI_pow2: Likewise.
* src/modinteger/cl_MI_pow2m1.h: Likewise.
* src/modinteger/cl_MI_pow2p1.h: Likewise.
* src/modinteger/cl_MI_std.h: Likewise.
* src/polynomial/elem/cl_UP.cc (cl_make_univpoly_ring): Compare with
cl_class_flags_modint_ring, not with cl_class_modint_ring.
* src/polynomial/elem/cl_UP_GF2.h (cl_class_num_univpoly_ring): New.
* src/polynomial/elem/cl_UP_MI.h (cl_class_modint_univpoly_ring): New.
* src/polynomial/elem/cl_UP_gen.h (cl_class_gen_univpoly_ring): New.
* src/polynomial/elem/cl_UP_number.h (cl_class_num_univpoly_ring): New.
Reported by Ralf Goertz <R_Goertz@web.de>.
21 years ago
d27593cce6
* m4/cc.m4: Emit a warning if g++ is used and optimization turned off.
21 years ago
95464ffca5
Make GCC compiler flags default to -O
* m4/cc.m4: New file...
* configure.ac: ...used here.
* autoconf/aclocal.m4: Regenerate.
21 years ago
63aedca00b
* include/cln/string.h: Declare cl_string.
21 years ago
c14014b041
* autoconf/ltmain.sh: Upgrade to libtool-1.5.14.
* m4/libtool.m4: Upgrade to libtool-1.5.14 with Comeau patch.
* autoconf/aclocal.m4: Regenerate.
21 years ago
b4bf50e16a
* src/Makefile.in: Use $CXX instead of $CC when linking.
21 years ago
5332b309a5
* examples/perfnum.cc: update to presumed 42st Mersenne prime.
21 years ago
1f0efb29d4
Disambiguate binary operators of CLN types with float/double
* include/cln/dfloat.h: Add binary operator overloads for arguments of
type double.
* include/cln/ffloat.h: Likewise, for arguments of type float.
* include/cln/float.h: Likewise, both for arguments of types double and
float.
* include/cln/real.h: Likewise.
Reported by Isidro Cachadiña Gutiérrez <icacha@unex.es>.
----------------------------------------------------------------------
include/cln/dfloat.h include/cln/ffloat.h CVS: include/cln/float.h
include/cln/real.h CVS:
----------------------------------------------------------------------
21 years ago
1dbaa44490
* Version 1.1.9 released.
21 years ago
57023a0ee9
* src/Makefile.in: Let config.status set LDFLAGS.
21 years ago
337a21d2e0
* cln.pc.in: Fix typo.
* Makefile.in (INSTALL_SCRIPT): Added, to be used for scripts like
cln-config. Allows us to do special things to binaries, like
installing them with install -s.
* doc/Makefile.in: add/remove the cln.info from the installed
$(infodir)/dir unless this is debian install-info (code copied from
what GNU automake would produce).
21 years ago
7792c43d12
* src/integer/input/cl_I_read_stream.cc (read_integer): Fix a bug
that caused radix specifiers to not work when reading from a stream.
* src/rational/input/cl_RA_read_stream.cc (read_rational): Likewise.
* src/real/input/cl_R_read_stream.cc (read_real): Likewise.
* src/float/input/cl_F_read_stream.cc (read_float): Likewise.
* src/complex/input/cl_N_read_stream.cc (read_complex): Likewise.
21 years ago
ccc9b05698
* src/base/cl_low.h: Add mulu64 assembler macro for ia64.
21 years ago
1bf866178f
* src/base/cl_low.h: Add mul and div macros for x86_64.
21 years ago
b31a3a904d
* src/integer/conv/cl_I_from_digits.cc (digits_to_I): Fix thinko in
new code for base power of two.
21 years ago
abe559d7d8
* src/integer/conv/cl_I_to_digits (I_to_digits): Fix an elusive stack
overwriting problem. That was the real cause for Debian bug#246319.
* src/integer/output/cl_I_print.cc (print_integer): Revert workaround
for the bug fixed above.
21 years ago
b8a5b7150b
* include/cln/types.h: Use 64 bit digits on x86_64 CPU.
21 years ago
ac82f8ef8b
* src/integer/conv/cl_I_from_digits.cc (digits_to_I): Speedup when
the base is a power of two.
21 years ago
8548d5d0a5
* src/integer/conv/cl_I_to_digits.cc (I_to_digits): Fix bug in base 32.
21 years ago
089003015a
Support for little-endian Mips, second shot
* src/base/digitseq/cl_asm_mipsel_.cc: New file...
* src/base/digitseq/cl_asm_cc: ...used here.
* src/base/digitseq/cl_asm.h: Include cl_asm_mips.h for any endianness.
* include/cln/object.h: Set alignment for mipsel explicitly.
21 years ago
abbb561b2c
Support for little-endian Mips
* include/cln/config.h.in: Add __mipsel__.
* include/cln/modules.h: For Mips, this is endianness-agnostic.
* src/base/digitseq/cl_asm_.cc, src/base/digitseq/cl_asm.h:
Mask out assembler for little-endian Mips.
21 years ago
6562d864cb
Update benchmark results.
21 years ago
Bruno Haible
Richard Kreckel