1 2009-09-24, version 1.3.1
2 =========================
7 * Fixed a crash in double_approx(cl_RA) on 64-bit platforms.
9 * Add basic support for Renesas SH (sh4).
12 2009-06-30, version 1.3.0
13 =========================
18 * Use the GNU autotools as build system.
20 * Implemented a more portable module dependency mechanism.
23 2008-04-05, version 1.2.2
24 =========================
26 Implementation changes
27 ----------------------
29 * Re-establish CLN-1.2.0 ABI and fix ARM build, both inadvertently broken in
33 2008-03-24, version 1.2.1
34 =========================
36 Implementation changes
37 ----------------------
39 * Fixed some bugs in the output of numbers larger than 2^32 decimal digits.
44 * Modifying C/C++ operators like +=, ++, etc. are now enabled by default.
47 2008-01-19, version 1.2.0
48 =========================
53 * Save big amounts of memory in computation of some functions and constants by:
54 - Avoiding pre-computation of series terms and instead computing them in a
56 - Avoiding computation with excess precision in binary splitting algorithm
57 by coercion of intermediate integer values to floating-point as early as
60 Implementation changes
61 ----------------------
63 * Added support for huge numbers:
64 - intC used for all counter loops.
65 - intE is now a 64-bit type on all 64-bit platforms and even on selected
68 * CLN now uses C++ exceptions for error handling. The cl_abort() hook is not
69 supported any more. Please refer to the documentation to learn about existing
72 * Fixed a bug on i386 where comparing Z/2Z ring zeros returnd random results.
74 * Removed cl_boolean. Use built-in C++ bool instead.
79 * Dropped the cln-config script. Please use pkg-config instead.
81 * Updated infrastructure to that of libtool-1.5.24.
83 * Changed encoding of all files to UTF-8.
85 * Fix compilation issues with GCC-4.3.
87 * Fix linking issues on platforms that don't feature weak symbols (like win32).
90 2006-08-08, version 1.1.13
91 ==========================
93 * Compilation fixes for 64-bit brokenness introduced in last release.
96 2006-08-06, version 1.1.12
97 ==========================
99 Implementation changes
100 ----------------------
102 * Fix rare assertion when printing quite large floats.
107 * Compilation fixes for several platforms: *BSD, Intel Mac, and MinGW.
110 2005-11-23, version 1.1.11
111 ==========================
116 * Considerably improved performance of number input.
119 2005-10-22, version 1.1.10
120 ==========================
122 Implementation changes
123 ----------------------
125 * Removed the vptr of modular integer and univariate polynomial
126 classes in order to fix some crashes in that sector of CLN. Code
127 using this needs to be recompiled.
129 * Many more operator signatures make the integration of CLN types with
135 * Several compilation fixes.
137 * Made it possible to cross-compile CLN.
140 2004-11-03, version 1.1.9
141 =========================
146 * Input of numbers in bases 2, 4, 8, 16 and 32 is now done in linear bit
147 complexity as opposed to O(N^2). Useful for all kinds of persistency.
149 Implementation changes
150 ----------------------
152 * Fixed several bugs in the integer input and output routines that could be
153 blamed for random crashes in the following cases: output in base 32 for
154 quite large numbers, input in base 2 for fixnums and input in base 3 for
155 fixnums on a 64 bit target.
157 * Fixed crash when radix specifiers were used in input streams.
159 * Speed up on x86_64 and ia64 by adding some inline assembly.
164 * Fixes for compilation on MacOS X and little endian Mips.
167 2004-07-01, version 1.1.8
168 =========================
170 Implementation changes
171 ----------------------
173 * Fix bug in log(a,b) when b is 1/n.
175 * Fix crash in shared library on AMD64.
180 * Fixes for some language conformance issues.
183 2004-05-02, version 1.1.7
184 =========================
186 Implementation changes
187 ----------------------
189 * Fixed two corner case bugs in cl_LF multiplication and sqrt.
191 * Workaround GCC 3.[123] compiler bug.
193 * Parallel build should work reliably now.
196 2004-01-01, version 1.1.6
197 =========================
203 `sintL ldegree(const cl_UP& x)'.
204 (Suggested by Munagala Ramanath.)
206 Implementation changes
207 ----------------------
209 * Fixed a bug in polynomial subtraction.
210 (Reported by Munagala Ramanath.)
215 * Provide a package control file cln.pc.
217 * Updated infrastructure to autoconf-2.57.
219 * Fixed compilation issues with GCC-3.4 prereleases and on Mac OSX.
222 2002-05-28, version 1.1.5
223 =========================
228 * Fix compilation issues with GCC-3.1.
231 2002-01-04, version 1.1.4
232 =========================
237 * Fix compilation issues with GMP-4.0 and on IBM s390.
239 * Updated infrastructure to autoconf-2.52.
242 2001-11-05, version 1.1.3
243 =========================
245 Implementation changes
246 ----------------------
248 * Fix a computation error of sin and sinh for small arguments.
249 (Reported by Christopher Kennedy.)
251 * Module ordering works with gcc-3.0 -fuse-cxa-atexit now.
253 * -DNO_ASM works properly on Sparc.
256 2001-07-25, version 1.1.2
258 Implementation changes
259 ----------------------
261 * Minor cleanups and portability issues.
262 (No need to upgrade if you have been happy so far.)
265 2001-05-31, version 1.1.1
266 =========================
268 Implementation changes
269 ----------------------
271 * Minor cleanups for GCC 3.0 and libstdc++-v3.
273 * Fixes several little bugs and portability issues.
276 2000-12-14, version 1.1.0
277 =========================
282 * ISO/IEC 14882 fashion adjustments:
283 Put everything into namespace cln. All fundamental data types still
284 carry their old names. Other non-macro identifiers are now written as
285 cln::foo instead of cl_foo, except where the cl_ comes from a data type
286 (as in cl_I_to_int()). Headers are installed into a separate
287 directory, #include <cln/foo.h> instead of <cl_foo.h>. Applications
288 must be manually ported to the new style. We apologize for the
291 Implementation changes
292 ----------------------
294 * Removed internal copy of mpn subdirectory tree from libgmp-2.0.2.
295 Configuring with --with-gmp now relies on an installed libgmp library
296 version 3 or above. We recommend libgmp-3.1 or newer.
298 * Adjusted several break-even points to make better use of a faster libgmp and
299 better match present-day CPUs.
301 * Fix several errors for architectures with 64-bit wide words.
302 (Others than Alpha, which worked already.)
304 * Fix compilation errors with current GCC-3.0 snapshots.
309 * Added package tools: script cln-config and autoconf macro AC_PATH_CLN
313 2000-01-13, version 1.0.3
314 =========================
320 `cl_I doublefactorial (uintL n)'.
322 Implementation changes
323 ----------------------
325 * Fix several little configuration errors.
327 * Fix some compilation errors with gcc-2.96 prereleases.
330 1999-09-07, version 1.0.2
331 =========================
337 `cl_I numerator (const cl_RA&)'
338 `cl_I denominator (const cl_RA&)'.
339 Suggested by Richard Kreckel and Sylvain Pion.
341 * New function `cl_equal_hashcode' for the classes cl_N, cl_R, cl_RA, cl_I,
342 cl_F, cl_SF, cl_FF, cl_DF, cl_LF.
343 Suggested by Richard Kreckel.
345 Implementation changes
346 ----------------------
348 * Fix an endless loop when either of the functions
349 `cl_boolean rootp (const cl_RA&, uintL, cl_RA*)',
350 `cl_boolean rootp (const cl_RA&, const cl_I&, cl_RA*)'
351 was called with an integer argument. Thanks to Richard Kreckel.
353 * Fix a bug in the addition and subtraction of rational numbers which could
354 cause unnormalized rational numbers like 3/1 to be created.
357 1999-06-09, version 1.0.1
358 =========================
363 * Speed up the functions `logand', `lognand', `logandc2'.
365 Implementation changes
366 ----------------------
368 * Fix compilation errors with gcc-2.7.2, egcs-1.1.2 and gcc-2.95.
370 * Fix compilation errors on HPPA, MIPS, some versions of DEC Alpha, OpenBSD,
371 and SPARC with LiDIA.
373 * Fix compilation errors on Win32. Thanks to Mumit Khan.
376 1999-01-12, version 1.0
377 =======================
382 * New include files, containing I/O functions:
386 <cl_sfloat_io.h> <cl_ffloat_io.h> <cl_dfloat_io.h> <cl_lfloat_io.h>
389 The file <cl_io.h> now contains only I/O functions for characters and
392 * To access the base ring of a univariate polynomial ring, now use
393 `R->basering()' instead of `R->basering'.
395 * Implement `plusp', `max', `min' for the classes cl_F, cl_SF, cl_FF, cl_DF,
398 * Implement `abs' for the class cl_RA.
400 * Replaced `read_number' with specialized functions `read_complex',
401 `read_real', `read_float', `read_rational', `read_integer'.
403 * Replaced the functions
404 `void fprint (cl_ostream stream, int x)'
405 `void fprint (cl_ostream stream, unsigned int x)'
406 `void fprint (cl_ostream stream, long x)'
407 `void fprint (cl_ostream stream, unsigned long x)'
409 `void fprintdecimal (cl_ostream stream, int x)'
410 `void fprintdecimal (cl_ostream stream, unsigned int x)'
411 `void fprintdecimal (cl_ostream stream, long x)'
412 `void fprintdecimal (cl_ostream stream, unsigned long x)'
417 * The function `cl_I xgcd (const cl_I& a, const cl_I& b, cl_I* u, cl_I* v)'
418 now normalizes the resulting u and v to be of minimum absolute value.
419 (Suggested by the LiDIA group.)
421 * The conversion from string to cl_F, cl_R, cl_N now chooses the format of
422 the resulting float depending on the number of significant digits present
425 * Speed up the functions
426 `cl_R operator* (const cl_RA&, const cl_LF&)'
427 `cl_R operator* (const cl_LF&, const cl_RA&)'
428 `cl_R operator/ (const cl_RA&, const cl_LF&)'
429 `cl_LF operator/ (const cl_LF&, const cl_RA&)'
431 Implementation changes
432 ----------------------
434 * Fix for `cl_I ash (const cl_I& x, const cl_I& n)' when |n| is between
436 (Reported by Thomas Papanikolaou.)
438 * Fixed a memory leak in the long-float truncate function.
439 (Reported by Keith Briggs.)
441 * Fixed a memory leak in the modular integer exponentiation.
443 * Reduced the stack space consumption. The default stack size (typically
444 8 MB on Unix) should now be sufficient in order to run all CLN programs.
446 * Portability fixes for m68k platforms.
447 (Reported and fixed by Roman Hodek.)
452 * Restructured the sources. The subpackages
455 - base + integer + rational
456 - base + integer + rational + float
457 - base + integer + rational + float + real
458 - base + integer + rational + float + real + complex
459 - base + integer + modinteger
460 are now self-contained.
462 * The configuration script can be called with option `--enable-shared',
463 to build CLN as a shared library.
465 * Updated the LiDIA interface.
468 1998-05-07, version 0.98
469 ========================
471 Implementation changes
472 ----------------------
474 * Portability fixes for 64-bit platforms like DEC Alpha.
475 (Reported by John Cremona and Thomas Papanikolaou.)
484 * Portability fixes for new ISO/ANSI C++ compilers.
486 * Updated the LiDIA interface for LiDIA version 1.3 and fixed a couple of
493 Implementation changes
494 ----------------------
496 * Portability fix for i686 platforms.
497 (Reported by Michael Pfeifer.)
507 <cl_ring.h>, declaring general rings,
508 <cl_GV.h>, <cl_SV.h>, defining vectors,
509 <cl_univpoly.h>, defining univariate polynomial rings.
514 * Speed up the internal computation of ln(2) and ln(10) by about 20%.
516 * Speed up the computation of exp (for 1000 digits or more) by about 10%.
518 Implementation changes
519 ----------------------
521 * Portability fix for 64-bit platforms like DEC Alpha:
522 Fixed a bug: endless loop in `sqrt' for large precision.
523 (Reported by Blake Jones.)
525 * Portability fix for RS/6000 platforms.
534 * Speed up the computation of the elementary transcendental functions:
535 exp, sinh, cosh, asinh, atanh (for 100 digits) by about 30%,
536 log (for 100 digits) by about 25%,
537 sin, cos (for 1000 or 10000 digits) by about 20%,
538 asin, acos, atan (for 1000 digits) by about 25%.
540 Implementation changes
541 ----------------------
543 * Portability fix for SunOS 4 and other BSD platforms.
552 * Fixed a bug: endless recursion in `cl_boolean sqrtp (const cl_RA&, cl_RA*)'.
554 * Fixed a bug: Only the first 85% of the digits returned by `cl_eulerconst'
555 for moderate precision (< 11370 digits) were correct.
556 (Reported by Thomas Papanikolaou.)
558 Implementation changes
559 ----------------------
561 * Fix static initialization order bug. Requires g++ 2.7.0 or newer
562 (g++ 2.8.0 or newer on HPPA).
564 * New method `debug_print()', for pretty printing of CLN objects, intended
565 to be called from the debugger.
574 * New functions `float cl_float_approx (const cl_R&)',
575 `double cl_double_approx (const cl_R&)'
576 for converting a real number to a C `float' or `double'.
581 * Speed up `cl_zeta' using Cohen-Villegas-Zagier convergence acceleration.
583 Implementation changes
584 ----------------------
586 * New function `cl_print', for pretty printing of CLN objects, intended to be
587 called from the debugger.
589 * Portability fixes for 64-bit platforms like DEC Alpha.
599 <cl_string.h>, defining strings,
600 <cl_symbol.h>, defining symbols,
601 <cl_proplist.h>, defining property lists,
602 <cl_condition.h>, defining conditions/exceptions.
604 * New functions `cl_F cl_catalanconst ()',
605 `cl_F cl_catalanconst (const cl_F&)',
606 `cl_F cl_catalanconst (cl_float_format_t)'
607 which return Catalan's constant.
608 (Suggested by Thomas Papanikolaou.)
610 * New functions `cl_F cl_zeta (int s)',
611 `cl_F cl_zeta (int s, const cl_F&)',
612 `cl_F cl_zeta (int s, cl_float_format_t)'
613 which return the Riemann zeta function at an integral point s>1.
615 * New functions `cl_F cl_exp1 ()',
616 `cl_F cl_exp1 (const cl_F&)',
617 `cl_F cl_exp1 (cl_float_format_t)'
618 which return e = exp(1).
620 * New function `cl_I binomial (uintL n, uintL k)' which returns the
621 binomial coefficient (n choose k).
623 * New functions `int cl_I_to_int (const cl_I&)',
624 `unsigned int cl_I_to_uint (const cl_I&)',
625 `long cl_I_to_long (const cl_I&)',
626 `unsigned long cl_I_to_ulong (const cl_I&)'
627 for converting an integer to a C `int' or `long'.
629 * New functions `float cl_float_approx (const cl_I&)',
630 `float cl_float_approx (const cl_RA&)',
631 `double cl_double_approx (const cl_I&)',
632 `double cl_double_approx (const cl_RA&)'
633 for converting a rational number to a C `float' or `double'.
635 Implementation changes
636 ----------------------
638 * Moved the sources to subdirectories.
639 (Suggested by Jörg Arndt.)
644 * New benchmark for better comparing LiDIA, Pari and CLN.
646 * Added an interface to LiDIA, allows using CLN as a replacement of libI.
647 (Suggested by Thomas Papanikolaou.)
649 * Added an ILOG Talk interface, for interactive use of the library.
658 * New include file <cl_modinteger.h>, defining modular integer rings.
660 * New functions `cl_F cl_eulerconst ()', `cl_F cl_eulerconst (const cl_F&)',
661 `cl_F cl_eulerconst (cl_float_format_t)' which return Euler's constant.
666 * Speed up square root of large numbers by use of Newton's algorithm.
668 * Speed up multiplication and division of large long-floats by small integers.
670 * Speed up the computation of pi, e, exp, log, sin, cos, tan, sinh, cosh, tanh,
671 asin, acos, atan, asinh, acosh, atanh.
672 All of them are now computed with asymptotically fast algorithms, of
673 bit complexity O(log(N)^2*M(N)) = O(N^(1+epsilon)) for N digits.
675 * Fixed several bugs in the transcendental functions routines. In particular,
676 the `log' function went into an endless loop.
678 * Fixed a bug: The cl_I -> cl_DF and cl_I -> cl_LF conversions didn't work
679 correctly for negative integers.
688 * New include file <cl_numtheory.h>, containing number theoretic functions,
689 for now only the Jacobi symbol.
694 * Speed up squaring of large numbers by 30%.
696 * Speed up division of large numbers by use of Newton's algorithm. The
697 improvement is noticeable for numbers with at least about 1000 words =
698 32000 bits = 10000 decimal digits.
700 * Speed up the binary-to-decimal conversion of large numbers by use of
701 a divide-and-conquer method. The improvement is noticeable for numbers
702 of at least 1250 words = 40000 bits = 12000 decimal digits.
704 * Fixed a bug: The functions `cl_boolean minusp (const type&)' and
705 `cl_boolean zerop (const type&)' were missing for type=cl_I and type=cl_RA.
707 * Fixed a bug: The function `cl_boolean logtest (const cl_I&, const cl_I&)'
708 didn't work correctly if both arguments were bignums.
710 * Fixed a bug: The function `cl_I mod (const cl_I&, const cl_I&)', when called
711 as `mod(-n*m,m)' (n>0), returned m instead of 0.
713 * Fixed a bug: The function `uintL power2p (const cl_I&)' returned wrong
714 values for arguments x = 2^n, n >= 29.
716 Implementation changes
717 ----------------------
719 * Speed up CLN by using the low-level addition/subtraction, shift and
720 multiplication loops from GNU gmp-2.0.2.
722 * Fixed a severe reference counting bug in the constructors
723 `type::type (const long)' and `type::type (const unsigned long)'
724 for type=cl_number, type=cl_N, type=cl_RA, type=cl_I that most often
727 * CLN's word sequences can be chosen to be stored big-endian in memory
728 (like it was up to now) or little-endian (needed for interfacing to GMP).
737 * New include file <cl_timing.h>, contains timing facilities.
742 * Speed up the multiplication of large numbers by use of an FFT based
743 algorithm (Schönhage-Strassen). The improvement is noticeable when both
744 factors have at least 1250 words = 40000 bits = 12000 decimal digits.
746 * Speed up the functions `cl_I gcd (const cl_I&, const cl_I&)' and
747 `cl_I xgcd (const cl_I&, const cl_I&, cl_I*, cl_I*)' by a factor
748 of 1.5. Simple trick.
750 * Speed up the function `cl_boolean sqrtp (const cl_I&, cl_I*)' using a
751 trick from Henri Cohen.
753 * Fixed an endless loop in the function `uint32 gcd (uint32, uint32)'
754 which occurred when one of the arguments is zero.
756 * Fixed an endless loop in the function `cl_I factorial (uintL)'.
758 Implementation changes
759 ----------------------
761 * <cl_io.h> now uses iostreams by default instead of stdio FILE pointers.
762 (Reported by Tito Flagella.)
764 * Fixed linking error when compiling without -O.
765 (Reported by Tito Flagella.)
767 * Fixed syntax error in <cl_input.h>.
768 (Reported by Tito Flagella.)
770 * Fixed syntax errors in src/cl_DS.h and src/cl_FF_plus.h.
771 (Reported by Marcus Daniels.)
773 * More portability fixes.
775 * Configuration now supports the --srcdir option.
776 (Reported by Marcus Daniels.)
782 * Compiles well on Sparc, using gcc.
784 * Mail to beta testers.
790 * Compiles well on Linux, using gcc.