1 2019-12-05, version 1.3.6
2 =========================
7 * Fixed cross-compilation (broken in 1.3.5).
10 2019-11-17, version 1.3.5
11 =========================
13 Implementation changes
14 ----------------------
16 * Fix bug in scale_float() throwing underflow_exception.
28 2014-10-16, version 1.3.4
29 =========================
31 Implementation changes
32 ----------------------
34 * Fix input of huge floats.
39 * Support the AArch64 and the Mips64 ABIs.
42 2013-07-21, version 1.3.3
43 =========================
45 Implementation changes
46 ----------------------
48 * Fix integer input of leading zeros in power-of-two base.
50 * Fix several floating-poing conversion bugs involving huge numbers.
52 * Fix bug that would set input stream fail state when reading a number at EOF.
57 * Support the x32 ABI.
60 2011-05-08, version 1.3.2
61 =========================
63 Implementation changes
64 ----------------------
66 * Fixed a bug in scale_float(cl_LF, cl_I) when the scale factor exceeded 2^31
72 * Improved portability to some non-GCC compilers.
75 2009-09-24, version 1.3.1
76 =========================
78 Implementation changes
79 ----------------------
81 * Fixed a crash in double_approx(cl_RA) on 64-bit platforms.
83 * Add basic support for Renesas SH (sh4).
86 2009-06-30, version 1.3.0
87 =========================
92 * Use the GNU autotools as build system.
94 * Implemented a more portable module dependency mechanism.
97 2008-04-05, version 1.2.2
98 =========================
100 Implementation changes
101 ----------------------
103 * Re-establish CLN-1.2.0 ABI and fix ARM build, both inadvertently broken in
104 the previous release.
107 2008-03-24, version 1.2.1
108 =========================
110 Implementation changes
111 ----------------------
113 * Fixed some bugs in the output of numbers larger than 2^32 decimal digits.
118 * Modifying C/C++ operators like +=, ++, etc. are now enabled by default.
121 2008-01-19, version 1.2.0
122 =========================
127 * Save big amounts of memory in computation of some functions and constants by:
128 - Avoiding pre-computation of series terms and instead computing them in a
130 - Avoiding computation with excess precision in binary splitting algorithm
131 by coercion of intermediate integer values to floating-point as early as
134 Implementation changes
135 ----------------------
137 * Added support for huge numbers:
138 - intC used for all counter loops.
139 - intE is now a 64-bit type on all 64-bit platforms and even on selected
142 * CLN now uses C++ exceptions for error handling. The cl_abort() hook is not
143 supported any more. Please refer to the documentation to learn about existing
146 * Fixed a bug on i386 where comparing Z/2Z ring zeros returnd random results.
148 * Removed cl_boolean. Use built-in C++ bool instead.
153 * Dropped the cln-config script. Please use pkg-config instead.
155 * Updated infrastructure to that of libtool-1.5.24.
157 * Changed encoding of all files to UTF-8.
159 * Fix compilation issues with GCC-4.3.
161 * Fix linking issues on platforms that don't feature weak symbols (like win32).
164 2006-08-08, version 1.1.13
165 ==========================
167 * Compilation fixes for 64-bit brokenness introduced in last release.
170 2006-08-06, version 1.1.12
171 ==========================
173 Implementation changes
174 ----------------------
176 * Fix rare assertion when printing quite large floats.
181 * Compilation fixes for several platforms: *BSD, Intel Mac, and MinGW.
184 2005-11-23, version 1.1.11
185 ==========================
190 * Considerably improved performance of number input.
193 2005-10-22, version 1.1.10
194 ==========================
196 Implementation changes
197 ----------------------
199 * Removed the vptr of modular integer and univariate polynomial
200 classes in order to fix some crashes in that sector of CLN. Code
201 using this needs to be recompiled.
203 * Many more operator signatures make the integration of CLN types with
209 * Several compilation fixes.
211 * Made it possible to cross-compile CLN.
214 2004-11-03, version 1.1.9
215 =========================
220 * Input of numbers in bases 2, 4, 8, 16 and 32 is now done in linear bit
221 complexity as opposed to O(N^2). Useful for all kinds of persistency.
223 Implementation changes
224 ----------------------
226 * Fixed several bugs in the integer input and output routines that could be
227 blamed for random crashes in the following cases: output in base 32 for
228 quite large numbers, input in base 2 for fixnums and input in base 3 for
229 fixnums on a 64 bit target.
231 * Fixed crash when radix specifiers were used in input streams.
233 * Speed up on x86_64 and ia64 by adding some inline assembly.
238 * Fixes for compilation on MacOS X and little endian Mips.
241 2004-07-01, version 1.1.8
242 =========================
244 Implementation changes
245 ----------------------
247 * Fix bug in log(a,b) when b is 1/n.
249 * Fix crash in shared library on AMD64.
254 * Fixes for some language conformance issues.
257 2004-05-02, version 1.1.7
258 =========================
260 Implementation changes
261 ----------------------
263 * Fixed two corner case bugs in cl_LF multiplication and sqrt.
265 * Workaround GCC 3.[123] compiler bug.
267 * Parallel build should work reliably now.
270 2004-01-01, version 1.1.6
271 =========================
277 `sintL ldegree(const cl_UP& x)'.
278 (Suggested by Munagala Ramanath.)
280 Implementation changes
281 ----------------------
283 * Fixed a bug in polynomial subtraction.
284 (Reported by Munagala Ramanath.)
289 * Provide a package control file cln.pc.
291 * Updated infrastructure to autoconf-2.57.
293 * Fixed compilation issues with GCC-3.4 prereleases and on Mac OSX.
296 2002-05-28, version 1.1.5
297 =========================
302 * Fix compilation issues with GCC-3.1.
305 2002-01-04, version 1.1.4
306 =========================
311 * Fix compilation issues with GMP-4.0 and on IBM s390.
313 * Updated infrastructure to autoconf-2.52.
316 2001-11-05, version 1.1.3
317 =========================
319 Implementation changes
320 ----------------------
322 * Fix a computation error of sin and sinh for small arguments.
323 (Reported by Christopher Kennedy.)
325 * Module ordering works with gcc-3.0 -fuse-cxa-atexit now.
327 * -DNO_ASM works properly on Sparc.
330 2001-07-25, version 1.1.2
332 Implementation changes
333 ----------------------
335 * Minor cleanups and portability issues.
336 (No need to upgrade if you have been happy so far.)
339 2001-05-31, version 1.1.1
340 =========================
342 Implementation changes
343 ----------------------
345 * Minor cleanups for GCC 3.0 and libstdc++-v3.
347 * Fixes several little bugs and portability issues.
350 2000-12-14, version 1.1.0
351 =========================
356 * ISO/IEC 14882 fashion adjustments:
357 Put everything into namespace cln. All fundamental data types still
358 carry their old names. Other non-macro identifiers are now written as
359 cln::foo instead of cl_foo, except where the cl_ comes from a data type
360 (as in cl_I_to_int()). Headers are installed into a separate
361 directory, #include <cln/foo.h> instead of <cl_foo.h>. Applications
362 must be manually ported to the new style. We apologize for the
365 Implementation changes
366 ----------------------
368 * Removed internal copy of mpn subdirectory tree from libgmp-2.0.2.
369 Configuring with --with-gmp now relies on an installed libgmp library
370 version 3 or above. We recommend libgmp-3.1 or newer.
372 * Adjusted several break-even points to make better use of a faster libgmp and
373 better match present-day CPUs.
375 * Fix several errors for architectures with 64-bit wide words.
376 (Others than Alpha, which worked already.)
378 * Fix compilation errors with current GCC-3.0 snapshots.
383 * Added package tools: script cln-config and autoconf macro AC_PATH_CLN
387 2000-01-13, version 1.0.3
388 =========================
394 `cl_I doublefactorial (uintL n)'.
396 Implementation changes
397 ----------------------
399 * Fix several little configuration errors.
401 * Fix some compilation errors with gcc-2.96 prereleases.
404 1999-09-07, version 1.0.2
405 =========================
411 `cl_I numerator (const cl_RA&)'
412 `cl_I denominator (const cl_RA&)'.
413 Suggested by Richard Kreckel and Sylvain Pion.
415 * New function `cl_equal_hashcode' for the classes cl_N, cl_R, cl_RA, cl_I,
416 cl_F, cl_SF, cl_FF, cl_DF, cl_LF.
417 Suggested by Richard Kreckel.
419 Implementation changes
420 ----------------------
422 * Fix an endless loop when either of the functions
423 `cl_boolean rootp (const cl_RA&, uintL, cl_RA*)',
424 `cl_boolean rootp (const cl_RA&, const cl_I&, cl_RA*)'
425 was called with an integer argument. Thanks to Richard Kreckel.
427 * Fix a bug in the addition and subtraction of rational numbers which could
428 cause unnormalized rational numbers like 3/1 to be created.
431 1999-06-09, version 1.0.1
432 =========================
437 * Speed up the functions `logand', `lognand', `logandc2'.
439 Implementation changes
440 ----------------------
442 * Fix compilation errors with gcc-2.7.2, egcs-1.1.2 and gcc-2.95.
444 * Fix compilation errors on HPPA, MIPS, some versions of DEC Alpha, OpenBSD,
445 and SPARC with LiDIA.
447 * Fix compilation errors on Win32. Thanks to Mumit Khan.
450 1999-01-12, version 1.0
451 =======================
456 * New include files, containing I/O functions:
460 <cl_sfloat_io.h> <cl_ffloat_io.h> <cl_dfloat_io.h> <cl_lfloat_io.h>
463 The file <cl_io.h> now contains only I/O functions for characters and
466 * To access the base ring of a univariate polynomial ring, now use
467 `R->basering()' instead of `R->basering'.
469 * Implement `plusp', `max', `min' for the classes cl_F, cl_SF, cl_FF, cl_DF,
472 * Implement `abs' for the class cl_RA.
474 * Replaced `read_number' with specialized functions `read_complex',
475 `read_real', `read_float', `read_rational', `read_integer'.
477 * Replaced the functions
478 `void fprint (cl_ostream stream, int x)'
479 `void fprint (cl_ostream stream, unsigned int x)'
480 `void fprint (cl_ostream stream, long x)'
481 `void fprint (cl_ostream stream, unsigned long x)'
483 `void fprintdecimal (cl_ostream stream, int x)'
484 `void fprintdecimal (cl_ostream stream, unsigned int x)'
485 `void fprintdecimal (cl_ostream stream, long x)'
486 `void fprintdecimal (cl_ostream stream, unsigned long x)'
491 * The function `cl_I xgcd (const cl_I& a, const cl_I& b, cl_I* u, cl_I* v)'
492 now normalizes the resulting u and v to be of minimum absolute value.
493 (Suggested by the LiDIA group.)
495 * The conversion from string to cl_F, cl_R, cl_N now chooses the format of
496 the resulting float depending on the number of significant digits present
499 * Speed up the functions
500 `cl_R operator* (const cl_RA&, const cl_LF&)'
501 `cl_R operator* (const cl_LF&, const cl_RA&)'
502 `cl_R operator/ (const cl_RA&, const cl_LF&)'
503 `cl_LF operator/ (const cl_LF&, const cl_RA&)'
505 Implementation changes
506 ----------------------
508 * Fix for `cl_I ash (const cl_I& x, const cl_I& n)' when |n| is between
510 (Reported by Thomas Papanikolaou.)
512 * Fixed a memory leak in the long-float truncate function.
513 (Reported by Keith Briggs.)
515 * Fixed a memory leak in the modular integer exponentiation.
517 * Reduced the stack space consumption. The default stack size (typically
518 8 MB on Unix) should now be sufficient in order to run all CLN programs.
520 * Portability fixes for m68k platforms.
521 (Reported and fixed by Roman Hodek.)
526 * Restructured the sources. The subpackages
529 - base + integer + rational
530 - base + integer + rational + float
531 - base + integer + rational + float + real
532 - base + integer + rational + float + real + complex
533 - base + integer + modinteger
534 are now self-contained.
536 * The configuration script can be called with option `--enable-shared',
537 to build CLN as a shared library.
539 * Updated the LiDIA interface.
542 1998-05-07, version 0.98
543 ========================
545 Implementation changes
546 ----------------------
548 * Portability fixes for 64-bit platforms like DEC Alpha.
549 (Reported by John Cremona and Thomas Papanikolaou.)
558 * Portability fixes for new ISO/ANSI C++ compilers.
560 * Updated the LiDIA interface for LiDIA version 1.3 and fixed a couple of
567 Implementation changes
568 ----------------------
570 * Portability fix for i686 platforms.
571 (Reported by Michael Pfeifer.)
581 <cl_ring.h>, declaring general rings,
582 <cl_GV.h>, <cl_SV.h>, defining vectors,
583 <cl_univpoly.h>, defining univariate polynomial rings.
588 * Speed up the internal computation of ln(2) and ln(10) by about 20%.
590 * Speed up the computation of exp (for 1000 digits or more) by about 10%.
592 Implementation changes
593 ----------------------
595 * Portability fix for 64-bit platforms like DEC Alpha:
596 Fixed a bug: endless loop in `sqrt' for large precision.
597 (Reported by Blake Jones.)
599 * Portability fix for RS/6000 platforms.
608 * Speed up the computation of the elementary transcendental functions:
609 exp, sinh, cosh, asinh, atanh (for 100 digits) by about 30%,
610 log (for 100 digits) by about 25%,
611 sin, cos (for 1000 or 10000 digits) by about 20%,
612 asin, acos, atan (for 1000 digits) by about 25%.
614 Implementation changes
615 ----------------------
617 * Portability fix for SunOS 4 and other BSD platforms.
626 * Fixed a bug: endless recursion in `cl_boolean sqrtp (const cl_RA&, cl_RA*)'.
628 * Fixed a bug: Only the first 85% of the digits returned by `cl_eulerconst'
629 for moderate precision (< 11370 digits) were correct.
630 (Reported by Thomas Papanikolaou.)
632 Implementation changes
633 ----------------------
635 * Fix static initialization order bug. Requires g++ 2.7.0 or newer
636 (g++ 2.8.0 or newer on HPPA).
638 * New method `debug_print()', for pretty printing of CLN objects, intended
639 to be called from the debugger.
648 * New functions `float cl_float_approx (const cl_R&)',
649 `double cl_double_approx (const cl_R&)'
650 for converting a real number to a C `float' or `double'.
655 * Speed up `cl_zeta' using Cohen-Villegas-Zagier convergence acceleration.
657 Implementation changes
658 ----------------------
660 * New function `cl_print', for pretty printing of CLN objects, intended to be
661 called from the debugger.
663 * Portability fixes for 64-bit platforms like DEC Alpha.
673 <cl_string.h>, defining strings,
674 <cl_symbol.h>, defining symbols,
675 <cl_proplist.h>, defining property lists,
676 <cl_condition.h>, defining conditions/exceptions.
678 * New functions `cl_F cl_catalanconst ()',
679 `cl_F cl_catalanconst (const cl_F&)',
680 `cl_F cl_catalanconst (cl_float_format_t)'
681 which return Catalan's constant.
682 (Suggested by Thomas Papanikolaou.)
684 * New functions `cl_F cl_zeta (int s)',
685 `cl_F cl_zeta (int s, const cl_F&)',
686 `cl_F cl_zeta (int s, cl_float_format_t)'
687 which return the Riemann zeta function at an integral point s>1.
689 * New functions `cl_F cl_exp1 ()',
690 `cl_F cl_exp1 (const cl_F&)',
691 `cl_F cl_exp1 (cl_float_format_t)'
692 which return e = exp(1).
694 * New function `cl_I binomial (uintL n, uintL k)' which returns the
695 binomial coefficient (n choose k).
697 * New functions `int cl_I_to_int (const cl_I&)',
698 `unsigned int cl_I_to_uint (const cl_I&)',
699 `long cl_I_to_long (const cl_I&)',
700 `unsigned long cl_I_to_ulong (const cl_I&)'
701 for converting an integer to a C `int' or `long'.
703 * New functions `float cl_float_approx (const cl_I&)',
704 `float cl_float_approx (const cl_RA&)',
705 `double cl_double_approx (const cl_I&)',
706 `double cl_double_approx (const cl_RA&)'
707 for converting a rational number to a C `float' or `double'.
709 Implementation changes
710 ----------------------
712 * Moved the sources to subdirectories.
713 (Suggested by Jörg Arndt.)
718 * New benchmark for better comparing LiDIA, Pari and CLN.
720 * Added an interface to LiDIA, allows using CLN as a replacement of libI.
721 (Suggested by Thomas Papanikolaou.)
723 * Added an ILOG Talk interface, for interactive use of the library.
732 * New include file <cl_modinteger.h>, defining modular integer rings.
734 * New functions `cl_F cl_eulerconst ()', `cl_F cl_eulerconst (const cl_F&)',
735 `cl_F cl_eulerconst (cl_float_format_t)' which return Euler's constant.
740 * Speed up square root of large numbers by use of Newton's algorithm.
742 * Speed up multiplication and division of large long-floats by small integers.
744 * Speed up the computation of pi, e, exp, log, sin, cos, tan, sinh, cosh, tanh,
745 asin, acos, atan, asinh, acosh, atanh.
746 All of them are now computed with asymptotically fast algorithms, of
747 bit complexity O(log(N)^2*M(N)) = O(N^(1+epsilon)) for N digits.
749 * Fixed several bugs in the transcendental functions routines. In particular,
750 the `log' function went into an endless loop.
752 * Fixed a bug: The cl_I -> cl_DF and cl_I -> cl_LF conversions didn't work
753 correctly for negative integers.
762 * New include file <cl_numtheory.h>, containing number theoretic functions,
763 for now only the Jacobi symbol.
768 * Speed up squaring of large numbers by 30%.
770 * Speed up division of large numbers by use of Newton's algorithm. The
771 improvement is noticeable for numbers with at least about 1000 words =
772 32000 bits = 10000 decimal digits.
774 * Speed up the binary-to-decimal conversion of large numbers by use of
775 a divide-and-conquer method. The improvement is noticeable for numbers
776 of at least 1250 words = 40000 bits = 12000 decimal digits.
778 * Fixed a bug: The functions `cl_boolean minusp (const type&)' and
779 `cl_boolean zerop (const type&)' were missing for type=cl_I and type=cl_RA.
781 * Fixed a bug: The function `cl_boolean logtest (const cl_I&, const cl_I&)'
782 didn't work correctly if both arguments were bignums.
784 * Fixed a bug: The function `cl_I mod (const cl_I&, const cl_I&)', when called
785 as `mod(-n*m,m)' (n>0), returned m instead of 0.
787 * Fixed a bug: The function `uintL power2p (const cl_I&)' returned wrong
788 values for arguments x = 2^n, n >= 29.
790 Implementation changes
791 ----------------------
793 * Speed up CLN by using the low-level addition/subtraction, shift and
794 multiplication loops from GNU gmp-2.0.2.
796 * Fixed a severe reference counting bug in the constructors
797 `type::type (const long)' and `type::type (const unsigned long)'
798 for type=cl_number, type=cl_N, type=cl_RA, type=cl_I that most often
801 * CLN's word sequences can be chosen to be stored big-endian in memory
802 (like it was up to now) or little-endian (needed for interfacing to GMP).
811 * New include file <cl_timing.h>, contains timing facilities.
816 * Speed up the multiplication of large numbers by use of an FFT based
817 algorithm (Schönhage-Strassen). The improvement is noticeable when both
818 factors have at least 1250 words = 40000 bits = 12000 decimal digits.
820 * Speed up the functions `cl_I gcd (const cl_I&, const cl_I&)' and
821 `cl_I xgcd (const cl_I&, const cl_I&, cl_I*, cl_I*)' by a factor
822 of 1.5. Simple trick.
824 * Speed up the function `cl_boolean sqrtp (const cl_I&, cl_I*)' using a
825 trick from Henri Cohen.
827 * Fixed an endless loop in the function `uint32 gcd (uint32, uint32)'
828 which occurred when one of the arguments is zero.
830 * Fixed an endless loop in the function `cl_I factorial (uintL)'.
832 Implementation changes
833 ----------------------
835 * <cl_io.h> now uses iostreams by default instead of stdio FILE pointers.
836 (Reported by Tito Flagella.)
838 * Fixed linking error when compiling without -O.
839 (Reported by Tito Flagella.)
841 * Fixed syntax error in <cl_input.h>.
842 (Reported by Tito Flagella.)
844 * Fixed syntax errors in src/cl_DS.h and src/cl_FF_plus.h.
845 (Reported by Marcus Daniels.)
847 * More portability fixes.
849 * Configuration now supports the --srcdir option.
850 (Reported by Marcus Daniels.)
856 * Compiles well on Sparc, using gcc.
858 * Mail to beta testers.
864 * Compiles well on Linux, using gcc.