1 2014-10-16, version 1.3.4
2 =========================
7 * Fix input of huge floats.
12 * Support the AArch64 and the Mips64 ABIs.
15 2013-07-21, version 1.3.3
16 =========================
18 Implementation changes
19 ----------------------
21 * Fix integer input of leading zeros in power-of-two base.
23 * Fix several floating-poing conversion bugs involving huge numbers.
25 * Fix bug that would set input stream fail state when reading a number at EOF.
30 * Support the x32 ABI.
33 2011-05-08, version 1.3.2
34 =========================
36 Implementation changes
37 ----------------------
39 * Fixed a bug in scale_float(cl_LF, cl_I) when the scale factor exceeded 2^31
45 * Improved portability to some non-GCC compilers.
48 2009-09-24, version 1.3.1
49 =========================
51 Implementation changes
52 ----------------------
54 * Fixed a crash in double_approx(cl_RA) on 64-bit platforms.
56 * Add basic support for Renesas SH (sh4).
59 2009-06-30, version 1.3.0
60 =========================
65 * Use the GNU autotools as build system.
67 * Implemented a more portable module dependency mechanism.
70 2008-04-05, version 1.2.2
71 =========================
73 Implementation changes
74 ----------------------
76 * Re-establish CLN-1.2.0 ABI and fix ARM build, both inadvertently broken in
80 2008-03-24, version 1.2.1
81 =========================
83 Implementation changes
84 ----------------------
86 * Fixed some bugs in the output of numbers larger than 2^32 decimal digits.
91 * Modifying C/C++ operators like +=, ++, etc. are now enabled by default.
94 2008-01-19, version 1.2.0
95 =========================
100 * Save big amounts of memory in computation of some functions and constants by:
101 - Avoiding pre-computation of series terms and instead computing them in a
103 - Avoiding computation with excess precision in binary splitting algorithm
104 by coercion of intermediate integer values to floating-point as early as
107 Implementation changes
108 ----------------------
110 * Added support for huge numbers:
111 - intC used for all counter loops.
112 - intE is now a 64-bit type on all 64-bit platforms and even on selected
115 * CLN now uses C++ exceptions for error handling. The cl_abort() hook is not
116 supported any more. Please refer to the documentation to learn about existing
119 * Fixed a bug on i386 where comparing Z/2Z ring zeros returnd random results.
121 * Removed cl_boolean. Use built-in C++ bool instead.
126 * Dropped the cln-config script. Please use pkg-config instead.
128 * Updated infrastructure to that of libtool-1.5.24.
130 * Changed encoding of all files to UTF-8.
132 * Fix compilation issues with GCC-4.3.
134 * Fix linking issues on platforms that don't feature weak symbols (like win32).
137 2006-08-08, version 1.1.13
138 ==========================
140 * Compilation fixes for 64-bit brokenness introduced in last release.
143 2006-08-06, version 1.1.12
144 ==========================
146 Implementation changes
147 ----------------------
149 * Fix rare assertion when printing quite large floats.
154 * Compilation fixes for several platforms: *BSD, Intel Mac, and MinGW.
157 2005-11-23, version 1.1.11
158 ==========================
163 * Considerably improved performance of number input.
166 2005-10-22, version 1.1.10
167 ==========================
169 Implementation changes
170 ----------------------
172 * Removed the vptr of modular integer and univariate polynomial
173 classes in order to fix some crashes in that sector of CLN. Code
174 using this needs to be recompiled.
176 * Many more operator signatures make the integration of CLN types with
182 * Several compilation fixes.
184 * Made it possible to cross-compile CLN.
187 2004-11-03, version 1.1.9
188 =========================
193 * Input of numbers in bases 2, 4, 8, 16 and 32 is now done in linear bit
194 complexity as opposed to O(N^2). Useful for all kinds of persistency.
196 Implementation changes
197 ----------------------
199 * Fixed several bugs in the integer input and output routines that could be
200 blamed for random crashes in the following cases: output in base 32 for
201 quite large numbers, input in base 2 for fixnums and input in base 3 for
202 fixnums on a 64 bit target.
204 * Fixed crash when radix specifiers were used in input streams.
206 * Speed up on x86_64 and ia64 by adding some inline assembly.
211 * Fixes for compilation on MacOS X and little endian Mips.
214 2004-07-01, version 1.1.8
215 =========================
217 Implementation changes
218 ----------------------
220 * Fix bug in log(a,b) when b is 1/n.
222 * Fix crash in shared library on AMD64.
227 * Fixes for some language conformance issues.
230 2004-05-02, version 1.1.7
231 =========================
233 Implementation changes
234 ----------------------
236 * Fixed two corner case bugs in cl_LF multiplication and sqrt.
238 * Workaround GCC 3.[123] compiler bug.
240 * Parallel build should work reliably now.
243 2004-01-01, version 1.1.6
244 =========================
250 `sintL ldegree(const cl_UP& x)'.
251 (Suggested by Munagala Ramanath.)
253 Implementation changes
254 ----------------------
256 * Fixed a bug in polynomial subtraction.
257 (Reported by Munagala Ramanath.)
262 * Provide a package control file cln.pc.
264 * Updated infrastructure to autoconf-2.57.
266 * Fixed compilation issues with GCC-3.4 prereleases and on Mac OSX.
269 2002-05-28, version 1.1.5
270 =========================
275 * Fix compilation issues with GCC-3.1.
278 2002-01-04, version 1.1.4
279 =========================
284 * Fix compilation issues with GMP-4.0 and on IBM s390.
286 * Updated infrastructure to autoconf-2.52.
289 2001-11-05, version 1.1.3
290 =========================
292 Implementation changes
293 ----------------------
295 * Fix a computation error of sin and sinh for small arguments.
296 (Reported by Christopher Kennedy.)
298 * Module ordering works with gcc-3.0 -fuse-cxa-atexit now.
300 * -DNO_ASM works properly on Sparc.
303 2001-07-25, version 1.1.2
305 Implementation changes
306 ----------------------
308 * Minor cleanups and portability issues.
309 (No need to upgrade if you have been happy so far.)
312 2001-05-31, version 1.1.1
313 =========================
315 Implementation changes
316 ----------------------
318 * Minor cleanups for GCC 3.0 and libstdc++-v3.
320 * Fixes several little bugs and portability issues.
323 2000-12-14, version 1.1.0
324 =========================
329 * ISO/IEC 14882 fashion adjustments:
330 Put everything into namespace cln. All fundamental data types still
331 carry their old names. Other non-macro identifiers are now written as
332 cln::foo instead of cl_foo, except where the cl_ comes from a data type
333 (as in cl_I_to_int()). Headers are installed into a separate
334 directory, #include <cln/foo.h> instead of <cl_foo.h>. Applications
335 must be manually ported to the new style. We apologize for the
338 Implementation changes
339 ----------------------
341 * Removed internal copy of mpn subdirectory tree from libgmp-2.0.2.
342 Configuring with --with-gmp now relies on an installed libgmp library
343 version 3 or above. We recommend libgmp-3.1 or newer.
345 * Adjusted several break-even points to make better use of a faster libgmp and
346 better match present-day CPUs.
348 * Fix several errors for architectures with 64-bit wide words.
349 (Others than Alpha, which worked already.)
351 * Fix compilation errors with current GCC-3.0 snapshots.
356 * Added package tools: script cln-config and autoconf macro AC_PATH_CLN
360 2000-01-13, version 1.0.3
361 =========================
367 `cl_I doublefactorial (uintL n)'.
369 Implementation changes
370 ----------------------
372 * Fix several little configuration errors.
374 * Fix some compilation errors with gcc-2.96 prereleases.
377 1999-09-07, version 1.0.2
378 =========================
384 `cl_I numerator (const cl_RA&)'
385 `cl_I denominator (const cl_RA&)'.
386 Suggested by Richard Kreckel and Sylvain Pion.
388 * New function `cl_equal_hashcode' for the classes cl_N, cl_R, cl_RA, cl_I,
389 cl_F, cl_SF, cl_FF, cl_DF, cl_LF.
390 Suggested by Richard Kreckel.
392 Implementation changes
393 ----------------------
395 * Fix an endless loop when either of the functions
396 `cl_boolean rootp (const cl_RA&, uintL, cl_RA*)',
397 `cl_boolean rootp (const cl_RA&, const cl_I&, cl_RA*)'
398 was called with an integer argument. Thanks to Richard Kreckel.
400 * Fix a bug in the addition and subtraction of rational numbers which could
401 cause unnormalized rational numbers like 3/1 to be created.
404 1999-06-09, version 1.0.1
405 =========================
410 * Speed up the functions `logand', `lognand', `logandc2'.
412 Implementation changes
413 ----------------------
415 * Fix compilation errors with gcc-2.7.2, egcs-1.1.2 and gcc-2.95.
417 * Fix compilation errors on HPPA, MIPS, some versions of DEC Alpha, OpenBSD,
418 and SPARC with LiDIA.
420 * Fix compilation errors on Win32. Thanks to Mumit Khan.
423 1999-01-12, version 1.0
424 =======================
429 * New include files, containing I/O functions:
433 <cl_sfloat_io.h> <cl_ffloat_io.h> <cl_dfloat_io.h> <cl_lfloat_io.h>
436 The file <cl_io.h> now contains only I/O functions for characters and
439 * To access the base ring of a univariate polynomial ring, now use
440 `R->basering()' instead of `R->basering'.
442 * Implement `plusp', `max', `min' for the classes cl_F, cl_SF, cl_FF, cl_DF,
445 * Implement `abs' for the class cl_RA.
447 * Replaced `read_number' with specialized functions `read_complex',
448 `read_real', `read_float', `read_rational', `read_integer'.
450 * Replaced the functions
451 `void fprint (cl_ostream stream, int x)'
452 `void fprint (cl_ostream stream, unsigned int x)'
453 `void fprint (cl_ostream stream, long x)'
454 `void fprint (cl_ostream stream, unsigned long x)'
456 `void fprintdecimal (cl_ostream stream, int x)'
457 `void fprintdecimal (cl_ostream stream, unsigned int x)'
458 `void fprintdecimal (cl_ostream stream, long x)'
459 `void fprintdecimal (cl_ostream stream, unsigned long x)'
464 * The function `cl_I xgcd (const cl_I& a, const cl_I& b, cl_I* u, cl_I* v)'
465 now normalizes the resulting u and v to be of minimum absolute value.
466 (Suggested by the LiDIA group.)
468 * The conversion from string to cl_F, cl_R, cl_N now chooses the format of
469 the resulting float depending on the number of significant digits present
472 * Speed up the functions
473 `cl_R operator* (const cl_RA&, const cl_LF&)'
474 `cl_R operator* (const cl_LF&, const cl_RA&)'
475 `cl_R operator/ (const cl_RA&, const cl_LF&)'
476 `cl_LF operator/ (const cl_LF&, const cl_RA&)'
478 Implementation changes
479 ----------------------
481 * Fix for `cl_I ash (const cl_I& x, const cl_I& n)' when |n| is between
483 (Reported by Thomas Papanikolaou.)
485 * Fixed a memory leak in the long-float truncate function.
486 (Reported by Keith Briggs.)
488 * Fixed a memory leak in the modular integer exponentiation.
490 * Reduced the stack space consumption. The default stack size (typically
491 8 MB on Unix) should now be sufficient in order to run all CLN programs.
493 * Portability fixes for m68k platforms.
494 (Reported and fixed by Roman Hodek.)
499 * Restructured the sources. The subpackages
502 - base + integer + rational
503 - base + integer + rational + float
504 - base + integer + rational + float + real
505 - base + integer + rational + float + real + complex
506 - base + integer + modinteger
507 are now self-contained.
509 * The configuration script can be called with option `--enable-shared',
510 to build CLN as a shared library.
512 * Updated the LiDIA interface.
515 1998-05-07, version 0.98
516 ========================
518 Implementation changes
519 ----------------------
521 * Portability fixes for 64-bit platforms like DEC Alpha.
522 (Reported by John Cremona and Thomas Papanikolaou.)
531 * Portability fixes for new ISO/ANSI C++ compilers.
533 * Updated the LiDIA interface for LiDIA version 1.3 and fixed a couple of
540 Implementation changes
541 ----------------------
543 * Portability fix for i686 platforms.
544 (Reported by Michael Pfeifer.)
554 <cl_ring.h>, declaring general rings,
555 <cl_GV.h>, <cl_SV.h>, defining vectors,
556 <cl_univpoly.h>, defining univariate polynomial rings.
561 * Speed up the internal computation of ln(2) and ln(10) by about 20%.
563 * Speed up the computation of exp (for 1000 digits or more) by about 10%.
565 Implementation changes
566 ----------------------
568 * Portability fix for 64-bit platforms like DEC Alpha:
569 Fixed a bug: endless loop in `sqrt' for large precision.
570 (Reported by Blake Jones.)
572 * Portability fix for RS/6000 platforms.
581 * Speed up the computation of the elementary transcendental functions:
582 exp, sinh, cosh, asinh, atanh (for 100 digits) by about 30%,
583 log (for 100 digits) by about 25%,
584 sin, cos (for 1000 or 10000 digits) by about 20%,
585 asin, acos, atan (for 1000 digits) by about 25%.
587 Implementation changes
588 ----------------------
590 * Portability fix for SunOS 4 and other BSD platforms.
599 * Fixed a bug: endless recursion in `cl_boolean sqrtp (const cl_RA&, cl_RA*)'.
601 * Fixed a bug: Only the first 85% of the digits returned by `cl_eulerconst'
602 for moderate precision (< 11370 digits) were correct.
603 (Reported by Thomas Papanikolaou.)
605 Implementation changes
606 ----------------------
608 * Fix static initialization order bug. Requires g++ 2.7.0 or newer
609 (g++ 2.8.0 or newer on HPPA).
611 * New method `debug_print()', for pretty printing of CLN objects, intended
612 to be called from the debugger.
621 * New functions `float cl_float_approx (const cl_R&)',
622 `double cl_double_approx (const cl_R&)'
623 for converting a real number to a C `float' or `double'.
628 * Speed up `cl_zeta' using Cohen-Villegas-Zagier convergence acceleration.
630 Implementation changes
631 ----------------------
633 * New function `cl_print', for pretty printing of CLN objects, intended to be
634 called from the debugger.
636 * Portability fixes for 64-bit platforms like DEC Alpha.
646 <cl_string.h>, defining strings,
647 <cl_symbol.h>, defining symbols,
648 <cl_proplist.h>, defining property lists,
649 <cl_condition.h>, defining conditions/exceptions.
651 * New functions `cl_F cl_catalanconst ()',
652 `cl_F cl_catalanconst (const cl_F&)',
653 `cl_F cl_catalanconst (cl_float_format_t)'
654 which return Catalan's constant.
655 (Suggested by Thomas Papanikolaou.)
657 * New functions `cl_F cl_zeta (int s)',
658 `cl_F cl_zeta (int s, const cl_F&)',
659 `cl_F cl_zeta (int s, cl_float_format_t)'
660 which return the Riemann zeta function at an integral point s>1.
662 * New functions `cl_F cl_exp1 ()',
663 `cl_F cl_exp1 (const cl_F&)',
664 `cl_F cl_exp1 (cl_float_format_t)'
665 which return e = exp(1).
667 * New function `cl_I binomial (uintL n, uintL k)' which returns the
668 binomial coefficient (n choose k).
670 * New functions `int cl_I_to_int (const cl_I&)',
671 `unsigned int cl_I_to_uint (const cl_I&)',
672 `long cl_I_to_long (const cl_I&)',
673 `unsigned long cl_I_to_ulong (const cl_I&)'
674 for converting an integer to a C `int' or `long'.
676 * New functions `float cl_float_approx (const cl_I&)',
677 `float cl_float_approx (const cl_RA&)',
678 `double cl_double_approx (const cl_I&)',
679 `double cl_double_approx (const cl_RA&)'
680 for converting a rational number to a C `float' or `double'.
682 Implementation changes
683 ----------------------
685 * Moved the sources to subdirectories.
686 (Suggested by Jörg Arndt.)
691 * New benchmark for better comparing LiDIA, Pari and CLN.
693 * Added an interface to LiDIA, allows using CLN as a replacement of libI.
694 (Suggested by Thomas Papanikolaou.)
696 * Added an ILOG Talk interface, for interactive use of the library.
705 * New include file <cl_modinteger.h>, defining modular integer rings.
707 * New functions `cl_F cl_eulerconst ()', `cl_F cl_eulerconst (const cl_F&)',
708 `cl_F cl_eulerconst (cl_float_format_t)' which return Euler's constant.
713 * Speed up square root of large numbers by use of Newton's algorithm.
715 * Speed up multiplication and division of large long-floats by small integers.
717 * Speed up the computation of pi, e, exp, log, sin, cos, tan, sinh, cosh, tanh,
718 asin, acos, atan, asinh, acosh, atanh.
719 All of them are now computed with asymptotically fast algorithms, of
720 bit complexity O(log(N)^2*M(N)) = O(N^(1+epsilon)) for N digits.
722 * Fixed several bugs in the transcendental functions routines. In particular,
723 the `log' function went into an endless loop.
725 * Fixed a bug: The cl_I -> cl_DF and cl_I -> cl_LF conversions didn't work
726 correctly for negative integers.
735 * New include file <cl_numtheory.h>, containing number theoretic functions,
736 for now only the Jacobi symbol.
741 * Speed up squaring of large numbers by 30%.
743 * Speed up division of large numbers by use of Newton's algorithm. The
744 improvement is noticeable for numbers with at least about 1000 words =
745 32000 bits = 10000 decimal digits.
747 * Speed up the binary-to-decimal conversion of large numbers by use of
748 a divide-and-conquer method. The improvement is noticeable for numbers
749 of at least 1250 words = 40000 bits = 12000 decimal digits.
751 * Fixed a bug: The functions `cl_boolean minusp (const type&)' and
752 `cl_boolean zerop (const type&)' were missing for type=cl_I and type=cl_RA.
754 * Fixed a bug: The function `cl_boolean logtest (const cl_I&, const cl_I&)'
755 didn't work correctly if both arguments were bignums.
757 * Fixed a bug: The function `cl_I mod (const cl_I&, const cl_I&)', when called
758 as `mod(-n*m,m)' (n>0), returned m instead of 0.
760 * Fixed a bug: The function `uintL power2p (const cl_I&)' returned wrong
761 values for arguments x = 2^n, n >= 29.
763 Implementation changes
764 ----------------------
766 * Speed up CLN by using the low-level addition/subtraction, shift and
767 multiplication loops from GNU gmp-2.0.2.
769 * Fixed a severe reference counting bug in the constructors
770 `type::type (const long)' and `type::type (const unsigned long)'
771 for type=cl_number, type=cl_N, type=cl_RA, type=cl_I that most often
774 * CLN's word sequences can be chosen to be stored big-endian in memory
775 (like it was up to now) or little-endian (needed for interfacing to GMP).
784 * New include file <cl_timing.h>, contains timing facilities.
789 * Speed up the multiplication of large numbers by use of an FFT based
790 algorithm (Schönhage-Strassen). The improvement is noticeable when both
791 factors have at least 1250 words = 40000 bits = 12000 decimal digits.
793 * Speed up the functions `cl_I gcd (const cl_I&, const cl_I&)' and
794 `cl_I xgcd (const cl_I&, const cl_I&, cl_I*, cl_I*)' by a factor
795 of 1.5. Simple trick.
797 * Speed up the function `cl_boolean sqrtp (const cl_I&, cl_I*)' using a
798 trick from Henri Cohen.
800 * Fixed an endless loop in the function `uint32 gcd (uint32, uint32)'
801 which occurred when one of the arguments is zero.
803 * Fixed an endless loop in the function `cl_I factorial (uintL)'.
805 Implementation changes
806 ----------------------
808 * <cl_io.h> now uses iostreams by default instead of stdio FILE pointers.
809 (Reported by Tito Flagella.)
811 * Fixed linking error when compiling without -O.
812 (Reported by Tito Flagella.)
814 * Fixed syntax error in <cl_input.h>.
815 (Reported by Tito Flagella.)
817 * Fixed syntax errors in src/cl_DS.h and src/cl_FF_plus.h.
818 (Reported by Marcus Daniels.)
820 * More portability fixes.
822 * Configuration now supports the --srcdir option.
823 (Reported by Marcus Daniels.)
829 * Compiles well on Sparc, using gcc.
831 * Mail to beta testers.
837 * Compiles well on Linux, using gcc.