1 2024-01-22, version 1.3.7
2 =========================
7 * Fixed compilation on ARM platforms.
8 * Add support for LoongArch.
11 2019-12-05, version 1.3.6
12 =========================
17 * Fixed cross-compilation (broken in 1.3.5).
20 2019-11-17, version 1.3.5
21 =========================
23 Implementation changes
24 ----------------------
26 * Fix bug in scale_float() throwing underflow_exception.
38 2014-10-16, version 1.3.4
39 =========================
41 Implementation changes
42 ----------------------
44 * Fix input of huge floats.
49 * Support the AArch64 and the Mips64 ABIs.
52 2013-07-21, version 1.3.3
53 =========================
55 Implementation changes
56 ----------------------
58 * Fix integer input of leading zeros in power-of-two base.
60 * Fix several floating-poing conversion bugs involving huge numbers.
62 * Fix bug that would set input stream fail state when reading a number at EOF.
67 * Support the x32 ABI.
70 2011-05-08, version 1.3.2
71 =========================
73 Implementation changes
74 ----------------------
76 * Fixed a bug in scale_float(cl_LF, cl_I) when the scale factor exceeded 2^31
82 * Improved portability to some non-GCC compilers.
85 2009-09-24, version 1.3.1
86 =========================
88 Implementation changes
89 ----------------------
91 * Fixed a crash in double_approx(cl_RA) on 64-bit platforms.
93 * Add basic support for Renesas SH (sh4).
96 2009-06-30, version 1.3.0
97 =========================
102 * Use the GNU autotools as build system.
104 * Implemented a more portable module dependency mechanism.
107 2008-04-05, version 1.2.2
108 =========================
110 Implementation changes
111 ----------------------
113 * Re-establish CLN-1.2.0 ABI and fix ARM build, both inadvertently broken in
114 the previous release.
117 2008-03-24, version 1.2.1
118 =========================
120 Implementation changes
121 ----------------------
123 * Fixed some bugs in the output of numbers larger than 2^32 decimal digits.
128 * Modifying C/C++ operators like +=, ++, etc. are now enabled by default.
131 2008-01-19, version 1.2.0
132 =========================
137 * Save big amounts of memory in computation of some functions and constants by:
138 - Avoiding pre-computation of series terms and instead computing them in a
140 - Avoiding computation with excess precision in binary splitting algorithm
141 by coercion of intermediate integer values to floating-point as early as
144 Implementation changes
145 ----------------------
147 * Added support for huge numbers:
148 - intC used for all counter loops.
149 - intE is now a 64-bit type on all 64-bit platforms and even on selected
152 * CLN now uses C++ exceptions for error handling. The cl_abort() hook is not
153 supported any more. Please refer to the documentation to learn about existing
156 * Fixed a bug on i386 where comparing Z/2Z ring zeros returnd random results.
158 * Removed cl_boolean. Use built-in C++ bool instead.
163 * Dropped the cln-config script. Please use pkg-config instead.
165 * Updated infrastructure to that of libtool-1.5.24.
167 * Changed encoding of all files to UTF-8.
169 * Fix compilation issues with GCC-4.3.
171 * Fix linking issues on platforms that don't feature weak symbols (like win32).
174 2006-08-08, version 1.1.13
175 ==========================
177 * Compilation fixes for 64-bit brokenness introduced in last release.
180 2006-08-06, version 1.1.12
181 ==========================
183 Implementation changes
184 ----------------------
186 * Fix rare assertion when printing quite large floats.
191 * Compilation fixes for several platforms: *BSD, Intel Mac, and MinGW.
194 2005-11-23, version 1.1.11
195 ==========================
200 * Considerably improved performance of number input.
203 2005-10-22, version 1.1.10
204 ==========================
206 Implementation changes
207 ----------------------
209 * Removed the vptr of modular integer and univariate polynomial
210 classes in order to fix some crashes in that sector of CLN. Code
211 using this needs to be recompiled.
213 * Many more operator signatures make the integration of CLN types with
219 * Several compilation fixes.
221 * Made it possible to cross-compile CLN.
224 2004-11-03, version 1.1.9
225 =========================
230 * Input of numbers in bases 2, 4, 8, 16 and 32 is now done in linear bit
231 complexity as opposed to O(N^2). Useful for all kinds of persistency.
233 Implementation changes
234 ----------------------
236 * Fixed several bugs in the integer input and output routines that could be
237 blamed for random crashes in the following cases: output in base 32 for
238 quite large numbers, input in base 2 for fixnums and input in base 3 for
239 fixnums on a 64 bit target.
241 * Fixed crash when radix specifiers were used in input streams.
243 * Speed up on x86_64 and ia64 by adding some inline assembly.
248 * Fixes for compilation on MacOS X and little endian Mips.
251 2004-07-01, version 1.1.8
252 =========================
254 Implementation changes
255 ----------------------
257 * Fix bug in log(a,b) when b is 1/n.
259 * Fix crash in shared library on AMD64.
264 * Fixes for some language conformance issues.
267 2004-05-02, version 1.1.7
268 =========================
270 Implementation changes
271 ----------------------
273 * Fixed two corner case bugs in cl_LF multiplication and sqrt.
275 * Workaround GCC 3.[123] compiler bug.
277 * Parallel build should work reliably now.
280 2004-01-01, version 1.1.6
281 =========================
287 `sintL ldegree(const cl_UP& x)'.
288 (Suggested by Munagala Ramanath.)
290 Implementation changes
291 ----------------------
293 * Fixed a bug in polynomial subtraction.
294 (Reported by Munagala Ramanath.)
299 * Provide a package control file cln.pc.
301 * Updated infrastructure to autoconf-2.57.
303 * Fixed compilation issues with GCC-3.4 prereleases and on Mac OSX.
306 2002-05-28, version 1.1.5
307 =========================
312 * Fix compilation issues with GCC-3.1.
315 2002-01-04, version 1.1.4
316 =========================
321 * Fix compilation issues with GMP-4.0 and on IBM s390.
323 * Updated infrastructure to autoconf-2.52.
326 2001-11-05, version 1.1.3
327 =========================
329 Implementation changes
330 ----------------------
332 * Fix a computation error of sin and sinh for small arguments.
333 (Reported by Christopher Kennedy.)
335 * Module ordering works with gcc-3.0 -fuse-cxa-atexit now.
337 * -DNO_ASM works properly on Sparc.
340 2001-07-25, version 1.1.2
342 Implementation changes
343 ----------------------
345 * Minor cleanups and portability issues.
346 (No need to upgrade if you have been happy so far.)
349 2001-05-31, version 1.1.1
350 =========================
352 Implementation changes
353 ----------------------
355 * Minor cleanups for GCC 3.0 and libstdc++-v3.
357 * Fixes several little bugs and portability issues.
360 2000-12-14, version 1.1.0
361 =========================
366 * ISO/IEC 14882 fashion adjustments:
367 Put everything into namespace cln. All fundamental data types still
368 carry their old names. Other non-macro identifiers are now written as
369 cln::foo instead of cl_foo, except where the cl_ comes from a data type
370 (as in cl_I_to_int()). Headers are installed into a separate
371 directory, #include <cln/foo.h> instead of <cl_foo.h>. Applications
372 must be manually ported to the new style. We apologize for the
375 Implementation changes
376 ----------------------
378 * Removed internal copy of mpn subdirectory tree from libgmp-2.0.2.
379 Configuring with --with-gmp now relies on an installed libgmp library
380 version 3 or above. We recommend libgmp-3.1 or newer.
382 * Adjusted several break-even points to make better use of a faster libgmp and
383 better match present-day CPUs.
385 * Fix several errors for architectures with 64-bit wide words.
386 (Others than Alpha, which worked already.)
388 * Fix compilation errors with current GCC-3.0 snapshots.
393 * Added package tools: script cln-config and autoconf macro AC_PATH_CLN
397 2000-01-13, version 1.0.3
398 =========================
404 `cl_I doublefactorial (uintL n)'.
406 Implementation changes
407 ----------------------
409 * Fix several little configuration errors.
411 * Fix some compilation errors with gcc-2.96 prereleases.
414 1999-09-07, version 1.0.2
415 =========================
421 `cl_I numerator (const cl_RA&)'
422 `cl_I denominator (const cl_RA&)'.
423 Suggested by Richard Kreckel and Sylvain Pion.
425 * New function `cl_equal_hashcode' for the classes cl_N, cl_R, cl_RA, cl_I,
426 cl_F, cl_SF, cl_FF, cl_DF, cl_LF.
427 Suggested by Richard Kreckel.
429 Implementation changes
430 ----------------------
432 * Fix an endless loop when either of the functions
433 `cl_boolean rootp (const cl_RA&, uintL, cl_RA*)',
434 `cl_boolean rootp (const cl_RA&, const cl_I&, cl_RA*)'
435 was called with an integer argument. Thanks to Richard Kreckel.
437 * Fix a bug in the addition and subtraction of rational numbers which could
438 cause unnormalized rational numbers like 3/1 to be created.
441 1999-06-09, version 1.0.1
442 =========================
447 * Speed up the functions `logand', `lognand', `logandc2'.
449 Implementation changes
450 ----------------------
452 * Fix compilation errors with gcc-2.7.2, egcs-1.1.2 and gcc-2.95.
454 * Fix compilation errors on HPPA, MIPS, some versions of DEC Alpha, OpenBSD,
455 and SPARC with LiDIA.
457 * Fix compilation errors on Win32. Thanks to Mumit Khan.
460 1999-01-12, version 1.0
461 =======================
466 * New include files, containing I/O functions:
470 <cl_sfloat_io.h> <cl_ffloat_io.h> <cl_dfloat_io.h> <cl_lfloat_io.h>
473 The file <cl_io.h> now contains only I/O functions for characters and
476 * To access the base ring of a univariate polynomial ring, now use
477 `R->basering()' instead of `R->basering'.
479 * Implement `plusp', `max', `min' for the classes cl_F, cl_SF, cl_FF, cl_DF,
482 * Implement `abs' for the class cl_RA.
484 * Replaced `read_number' with specialized functions `read_complex',
485 `read_real', `read_float', `read_rational', `read_integer'.
487 * Replaced the functions
488 `void fprint (cl_ostream stream, int x)'
489 `void fprint (cl_ostream stream, unsigned int x)'
490 `void fprint (cl_ostream stream, long x)'
491 `void fprint (cl_ostream stream, unsigned long x)'
493 `void fprintdecimal (cl_ostream stream, int x)'
494 `void fprintdecimal (cl_ostream stream, unsigned int x)'
495 `void fprintdecimal (cl_ostream stream, long x)'
496 `void fprintdecimal (cl_ostream stream, unsigned long x)'
501 * The function `cl_I xgcd (const cl_I& a, const cl_I& b, cl_I* u, cl_I* v)'
502 now normalizes the resulting u and v to be of minimum absolute value.
503 (Suggested by the LiDIA group.)
505 * The conversion from string to cl_F, cl_R, cl_N now chooses the format of
506 the resulting float depending on the number of significant digits present
509 * Speed up the functions
510 `cl_R operator* (const cl_RA&, const cl_LF&)'
511 `cl_R operator* (const cl_LF&, const cl_RA&)'
512 `cl_R operator/ (const cl_RA&, const cl_LF&)'
513 `cl_LF operator/ (const cl_LF&, const cl_RA&)'
515 Implementation changes
516 ----------------------
518 * Fix for `cl_I ash (const cl_I& x, const cl_I& n)' when |n| is between
520 (Reported by Thomas Papanikolaou.)
522 * Fixed a memory leak in the long-float truncate function.
523 (Reported by Keith Briggs.)
525 * Fixed a memory leak in the modular integer exponentiation.
527 * Reduced the stack space consumption. The default stack size (typically
528 8 MB on Unix) should now be sufficient in order to run all CLN programs.
530 * Portability fixes for m68k platforms.
531 (Reported and fixed by Roman Hodek.)
536 * Restructured the sources. The subpackages
539 - base + integer + rational
540 - base + integer + rational + float
541 - base + integer + rational + float + real
542 - base + integer + rational + float + real + complex
543 - base + integer + modinteger
544 are now self-contained.
546 * The configuration script can be called with option `--enable-shared',
547 to build CLN as a shared library.
549 * Updated the LiDIA interface.
552 1998-05-07, version 0.98
553 ========================
555 Implementation changes
556 ----------------------
558 * Portability fixes for 64-bit platforms like DEC Alpha.
559 (Reported by John Cremona and Thomas Papanikolaou.)
568 * Portability fixes for new ISO/ANSI C++ compilers.
570 * Updated the LiDIA interface for LiDIA version 1.3 and fixed a couple of
577 Implementation changes
578 ----------------------
580 * Portability fix for i686 platforms.
581 (Reported by Michael Pfeifer.)
591 <cl_ring.h>, declaring general rings,
592 <cl_GV.h>, <cl_SV.h>, defining vectors,
593 <cl_univpoly.h>, defining univariate polynomial rings.
598 * Speed up the internal computation of ln(2) and ln(10) by about 20%.
600 * Speed up the computation of exp (for 1000 digits or more) by about 10%.
602 Implementation changes
603 ----------------------
605 * Portability fix for 64-bit platforms like DEC Alpha:
606 Fixed a bug: endless loop in `sqrt' for large precision.
607 (Reported by Blake Jones.)
609 * Portability fix for RS/6000 platforms.
618 * Speed up the computation of the elementary transcendental functions:
619 exp, sinh, cosh, asinh, atanh (for 100 digits) by about 30%,
620 log (for 100 digits) by about 25%,
621 sin, cos (for 1000 or 10000 digits) by about 20%,
622 asin, acos, atan (for 1000 digits) by about 25%.
624 Implementation changes
625 ----------------------
627 * Portability fix for SunOS 4 and other BSD platforms.
636 * Fixed a bug: endless recursion in `cl_boolean sqrtp (const cl_RA&, cl_RA*)'.
638 * Fixed a bug: Only the first 85% of the digits returned by `cl_eulerconst'
639 for moderate precision (< 11370 digits) were correct.
640 (Reported by Thomas Papanikolaou.)
642 Implementation changes
643 ----------------------
645 * Fix static initialization order bug. Requires g++ 2.7.0 or newer
646 (g++ 2.8.0 or newer on HPPA).
648 * New method `debug_print()', for pretty printing of CLN objects, intended
649 to be called from the debugger.
658 * New functions `float cl_float_approx (const cl_R&)',
659 `double cl_double_approx (const cl_R&)'
660 for converting a real number to a C `float' or `double'.
665 * Speed up `cl_zeta' using Cohen-Villegas-Zagier convergence acceleration.
667 Implementation changes
668 ----------------------
670 * New function `cl_print', for pretty printing of CLN objects, intended to be
671 called from the debugger.
673 * Portability fixes for 64-bit platforms like DEC Alpha.
683 <cl_string.h>, defining strings,
684 <cl_symbol.h>, defining symbols,
685 <cl_proplist.h>, defining property lists,
686 <cl_condition.h>, defining conditions/exceptions.
688 * New functions `cl_F cl_catalanconst ()',
689 `cl_F cl_catalanconst (const cl_F&)',
690 `cl_F cl_catalanconst (cl_float_format_t)'
691 which return Catalan's constant.
692 (Suggested by Thomas Papanikolaou.)
694 * New functions `cl_F cl_zeta (int s)',
695 `cl_F cl_zeta (int s, const cl_F&)',
696 `cl_F cl_zeta (int s, cl_float_format_t)'
697 which return the Riemann zeta function at an integral point s>1.
699 * New functions `cl_F cl_exp1 ()',
700 `cl_F cl_exp1 (const cl_F&)',
701 `cl_F cl_exp1 (cl_float_format_t)'
702 which return e = exp(1).
704 * New function `cl_I binomial (uintL n, uintL k)' which returns the
705 binomial coefficient (n choose k).
707 * New functions `int cl_I_to_int (const cl_I&)',
708 `unsigned int cl_I_to_uint (const cl_I&)',
709 `long cl_I_to_long (const cl_I&)',
710 `unsigned long cl_I_to_ulong (const cl_I&)'
711 for converting an integer to a C `int' or `long'.
713 * New functions `float cl_float_approx (const cl_I&)',
714 `float cl_float_approx (const cl_RA&)',
715 `double cl_double_approx (const cl_I&)',
716 `double cl_double_approx (const cl_RA&)'
717 for converting a rational number to a C `float' or `double'.
719 Implementation changes
720 ----------------------
722 * Moved the sources to subdirectories.
723 (Suggested by Jörg Arndt.)
728 * New benchmark for better comparing LiDIA, Pari and CLN.
730 * Added an interface to LiDIA, allows using CLN as a replacement of libI.
731 (Suggested by Thomas Papanikolaou.)
733 * Added an ILOG Talk interface, for interactive use of the library.
742 * New include file <cl_modinteger.h>, defining modular integer rings.
744 * New functions `cl_F cl_eulerconst ()', `cl_F cl_eulerconst (const cl_F&)',
745 `cl_F cl_eulerconst (cl_float_format_t)' which return Euler's constant.
750 * Speed up square root of large numbers by use of Newton's algorithm.
752 * Speed up multiplication and division of large long-floats by small integers.
754 * Speed up the computation of pi, e, exp, log, sin, cos, tan, sinh, cosh, tanh,
755 asin, acos, atan, asinh, acosh, atanh.
756 All of them are now computed with asymptotically fast algorithms, of
757 bit complexity O(log(N)^2*M(N)) = O(N^(1+epsilon)) for N digits.
759 * Fixed several bugs in the transcendental functions routines. In particular,
760 the `log' function went into an endless loop.
762 * Fixed a bug: The cl_I -> cl_DF and cl_I -> cl_LF conversions didn't work
763 correctly for negative integers.
772 * New include file <cl_numtheory.h>, containing number theoretic functions,
773 for now only the Jacobi symbol.
778 * Speed up squaring of large numbers by 30%.
780 * Speed up division of large numbers by use of Newton's algorithm. The
781 improvement is noticeable for numbers with at least about 1000 words =
782 32000 bits = 10000 decimal digits.
784 * Speed up the binary-to-decimal conversion of large numbers by use of
785 a divide-and-conquer method. The improvement is noticeable for numbers
786 of at least 1250 words = 40000 bits = 12000 decimal digits.
788 * Fixed a bug: The functions `cl_boolean minusp (const type&)' and
789 `cl_boolean zerop (const type&)' were missing for type=cl_I and type=cl_RA.
791 * Fixed a bug: The function `cl_boolean logtest (const cl_I&, const cl_I&)'
792 didn't work correctly if both arguments were bignums.
794 * Fixed a bug: The function `cl_I mod (const cl_I&, const cl_I&)', when called
795 as `mod(-n*m,m)' (n>0), returned m instead of 0.
797 * Fixed a bug: The function `uintL power2p (const cl_I&)' returned wrong
798 values for arguments x = 2^n, n >= 29.
800 Implementation changes
801 ----------------------
803 * Speed up CLN by using the low-level addition/subtraction, shift and
804 multiplication loops from GNU gmp-2.0.2.
806 * Fixed a severe reference counting bug in the constructors
807 `type::type (const long)' and `type::type (const unsigned long)'
808 for type=cl_number, type=cl_N, type=cl_RA, type=cl_I that most often
811 * CLN's word sequences can be chosen to be stored big-endian in memory
812 (like it was up to now) or little-endian (needed for interfacing to GMP).
821 * New include file <cl_timing.h>, contains timing facilities.
826 * Speed up the multiplication of large numbers by use of an FFT based
827 algorithm (Schönhage-Strassen). The improvement is noticeable when both
828 factors have at least 1250 words = 40000 bits = 12000 decimal digits.
830 * Speed up the functions `cl_I gcd (const cl_I&, const cl_I&)' and
831 `cl_I xgcd (const cl_I&, const cl_I&, cl_I*, cl_I*)' by a factor
832 of 1.5. Simple trick.
834 * Speed up the function `cl_boolean sqrtp (const cl_I&, cl_I*)' using a
835 trick from Henri Cohen.
837 * Fixed an endless loop in the function `uint32 gcd (uint32, uint32)'
838 which occurred when one of the arguments is zero.
840 * Fixed an endless loop in the function `cl_I factorial (uintL)'.
842 Implementation changes
843 ----------------------
845 * <cl_io.h> now uses iostreams by default instead of stdio FILE pointers.
846 (Reported by Tito Flagella.)
848 * Fixed linking error when compiling without -O.
849 (Reported by Tito Flagella.)
851 * Fixed syntax error in <cl_input.h>.
852 (Reported by Tito Flagella.)
854 * Fixed syntax errors in src/cl_DS.h and src/cl_FF_plus.h.
855 (Reported by Marcus Daniels.)
857 * More portability fixes.
859 * Configuration now supports the --srcdir option.
860 (Reported by Marcus Daniels.)
866 * Compiles well on Sparc, using gcc.
868 * Mail to beta testers.
874 * Compiles well on Linux, using gcc.