1 2008-01-19, version 1.2.0
2 =========================
7 * Save big amounts of memory in computation of some functions and constants by:
8 - Avoiding pre-computation of series terms and instead computing them in a
10 - Avoiding computation with excess precision in binary splitting algorithm
11 by coercion of intermediate integer values to floating-point as early as
14 Implementation changes
15 ----------------------
17 * Added support for huge numbers:
18 - intC used for all counter loops.
19 - intE is now a 64-bit type on all 64-bit platforms and even on selected
22 * CLN now uses C++ exceptions for error handling. The cl_abort() hook is not
23 supported any more. Please refer to the documentation to learn about existing
26 * Fixed a bug on i386 where comparing Z/2Z ring zeros returnd random results.
28 * Removed cl_boolean. Use built-in C++ bool instead.
33 * Dropped the cln-config script. Please use pkg-config instead.
35 * Updated infrastructure to that of libtool-1.5.24.
37 * Changed encoding of all files to UTF-8.
39 * Fix compilation issues with GCC-4.3.
41 * Fix linking issues on platforms that don't feature weak symbols (like win32).
44 2006-08-08, version 1.1.13
45 ==========================
47 * Compilation fixes for 64-bit brokenness introduced in last release.
50 2006-08-06, version 1.1.12
51 ==========================
53 Implementation changes
54 ----------------------
56 * Fix rare assertion when printing quite large floats.
61 * Compilation fixes for several platforms: *BSD, Intel Mac, and MinGW.
64 2005-11-23, version 1.1.11
65 ==========================
70 * Considerably improved performance of number input.
73 2005-10-22, version 1.1.10
74 ==========================
76 Implementation changes
77 ----------------------
79 * Removed the vptr of modular integer and univariate polynomial
80 classes in order to fix some crashes in that sector of CLN. Code
81 using this needs to be recompiled.
83 * Many more operator signatures make the integration of CLN types with
89 * Several compilation fixes.
91 * Made it possible to cross-compile CLN.
94 2004-11-03, version 1.1.9
95 =========================
100 * Input of numbers in bases 2, 4, 8, 16 and 32 is now done in linear bit
101 complexity as opposed to O(N^2). Useful for all kinds of persistency.
103 Implementation changes
104 ----------------------
106 * Fixed several bugs in the integer input and output routines that could be
107 blamed for random crashes in the following cases: output in base 32 for
108 quite large numbers, input in base 2 for fixnums and input in base 3 for
109 fixnums on a 64 bit target.
111 * Fixed crash when radix specifiers were used in input streams.
113 * Speed up on x86_64 and ia64 by adding some inline assembly.
118 * Fixes for compilation on MacOS X and little endian Mips.
121 2004-07-01, version 1.1.8
122 =========================
124 Implementation changes
125 ----------------------
127 * Fix bug in log(a,b) when b is 1/n.
129 * Fix crash in shared library on AMD64.
134 * Fixes for some language conformance issues.
137 2004-05-02, version 1.1.7
138 =========================
140 Implementation changes
141 ----------------------
143 * Fixed two corner case bugs in cl_LF multiplication and sqrt.
145 * Workaround GCC 3.[123] compiler bug.
147 * Parallel build should work reliably now.
150 2004-01-01, version 1.1.6
151 =========================
157 `sintL ldegree(const cl_UP& x)'.
158 (Suggested by Munagala Ramanath.)
160 Implementation changes
161 ----------------------
163 * Fixed a bug in polynomial subtraction.
164 (Reported by Munagala Ramanath.)
169 * Provide a package control file cln.pc.
171 * Updated infrastructure to autoconf-2.57.
173 * Fixed compilation issues with GCC-3.4 prereleases and on Mac OSX.
176 2002-05-28, version 1.1.5
177 =========================
182 * Fix compilation issues with GCC-3.1.
185 2002-01-04, version 1.1.4
186 =========================
191 * Fix compilation issues with GMP-4.0 and on IBM s390.
193 * Updated infrastructure to autoconf-2.52.
196 2001-11-05, version 1.1.3
197 =========================
199 Implementation changes
200 ----------------------
202 * Fix a computation error of sin and sinh for small arguments.
203 (Reported by Christopher Kennedy.)
205 * Module ordering works with gcc-3.0 -fuse-cxa-atexit now.
207 * -DNO_ASM works properly on Sparc.
210 2001-07-25, version 1.1.2
212 Implementation changes
213 ----------------------
215 * Minor cleanups and portability issues.
216 (No need to upgrade if you have been happy so far.)
219 2001-05-31, version 1.1.1
220 =========================
222 Implementation changes
223 ----------------------
225 * Minor cleanups for GCC 3.0 and libstdc++-v3.
227 * Fixes several little bugs and portability issues.
230 2000-12-14, version 1.1.0
231 =========================
236 * ISO/IEC 14882 fashion adjustments:
237 Put everything into namespace cln. All fundamental data types still
238 carry their old names. Other non-macro identifiers are now written as
239 cln::foo instead of cl_foo, except where the cl_ comes from a data type
240 (as in cl_I_to_int()). Headers are installed into a separate
241 directory, #include <cln/foo.h> instead of <cl_foo.h>. Applications
242 must be manually ported to the new style. We apologize for the
245 Implementation changes
246 ----------------------
248 * Removed internal copy of mpn subdirectory tree from libgmp-2.0.2.
249 Configuring with --with-gmp now relies on an installed libgmp library
250 version 3 or above. We recommend libgmp-3.1 or newer.
252 * Adjusted several break-even points to make better use of a faster libgmp and
253 better match present-day CPUs.
255 * Fix several errors for architectures with 64-bit wide words.
256 (Others than Alpha, which worked already.)
258 * Fix compilation errors with current GCC-3.0 snapshots.
263 * Added package tools: script cln-config and autoconf macro AC_PATH_CLN
267 2000-01-13, version 1.0.3
268 =========================
274 `cl_I doublefactorial (uintL n)'.
276 Implementation changes
277 ----------------------
279 * Fix several little configuration errors.
281 * Fix some compilation errors with gcc-2.96 prereleases.
284 1999-09-07, version 1.0.2
285 =========================
291 `cl_I numerator (const cl_RA&)'
292 `cl_I denominator (const cl_RA&)'.
293 Suggested by Richard Kreckel and Sylvain Pion.
295 * New function `cl_equal_hashcode' for the classes cl_N, cl_R, cl_RA, cl_I,
296 cl_F, cl_SF, cl_FF, cl_DF, cl_LF.
297 Suggested by Richard Kreckel.
299 Implementation changes
300 ----------------------
302 * Fix an endless loop when either of the functions
303 `cl_boolean rootp (const cl_RA&, uintL, cl_RA*)',
304 `cl_boolean rootp (const cl_RA&, const cl_I&, cl_RA*)'
305 was called with an integer argument. Thanks to Richard Kreckel.
307 * Fix a bug in the addition and subtraction of rational numbers which could
308 cause unnormalized rational numbers like 3/1 to be created.
311 1999-06-09, version 1.0.1
312 =========================
317 * Speed up the functions `logand', `lognand', `logandc2'.
319 Implementation changes
320 ----------------------
322 * Fix compilation errors with gcc-2.7.2, egcs-1.1.2 and gcc-2.95.
324 * Fix compilation errors on HPPA, MIPS, some versions of DEC Alpha, OpenBSD,
325 and SPARC with LiDIA.
327 * Fix compilation errors on Win32. Thanks to Mumit Khan.
330 1999-01-12, version 1.0
331 =======================
336 * New include files, containing I/O functions:
340 <cl_sfloat_io.h> <cl_ffloat_io.h> <cl_dfloat_io.h> <cl_lfloat_io.h>
343 The file <cl_io.h> now contains only I/O functions for characters and
346 * To access the base ring of a univariate polynomial ring, now use
347 `R->basering()' instead of `R->basering'.
349 * Implement `plusp', `max', `min' for the classes cl_F, cl_SF, cl_FF, cl_DF,
352 * Implement `abs' for the class cl_RA.
354 * Replaced `read_number' with specialized functions `read_complex',
355 `read_real', `read_float', `read_rational', `read_integer'.
357 * Replaced the functions
358 `void fprint (cl_ostream stream, int x)'
359 `void fprint (cl_ostream stream, unsigned int x)'
360 `void fprint (cl_ostream stream, long x)'
361 `void fprint (cl_ostream stream, unsigned long x)'
363 `void fprintdecimal (cl_ostream stream, int x)'
364 `void fprintdecimal (cl_ostream stream, unsigned int x)'
365 `void fprintdecimal (cl_ostream stream, long x)'
366 `void fprintdecimal (cl_ostream stream, unsigned long x)'
371 * The function `cl_I xgcd (const cl_I& a, const cl_I& b, cl_I* u, cl_I* v)'
372 now normalizes the resulting u and v to be of minimum absolute value.
373 (Suggested by the LiDIA group.)
375 * The conversion from string to cl_F, cl_R, cl_N now chooses the format of
376 the resulting float depending on the number of significant digits present
379 * Speed up the functions
380 `cl_R operator* (const cl_RA&, const cl_LF&)'
381 `cl_R operator* (const cl_LF&, const cl_RA&)'
382 `cl_R operator/ (const cl_RA&, const cl_LF&)'
383 `cl_LF operator/ (const cl_LF&, const cl_RA&)'
385 Implementation changes
386 ----------------------
388 * Fix for `cl_I ash (const cl_I& x, const cl_I& n)' when |n| is between
390 (Reported by Thomas Papanikolaou.)
392 * Fixed a memory leak in the long-float truncate function.
393 (Reported by Keith Briggs.)
395 * Fixed a memory leak in the modular integer exponentiation.
397 * Reduced the stack space consumption. The default stack size (typically
398 8 MB on Unix) should now be sufficient in order to run all CLN programs.
400 * Portability fixes for m68k platforms.
401 (Reported and fixed by Roman Hodek.)
406 * Restructured the sources. The subpackages
409 - base + integer + rational
410 - base + integer + rational + float
411 - base + integer + rational + float + real
412 - base + integer + rational + float + real + complex
413 - base + integer + modinteger
414 are now self-contained.
416 * The configuration script can be called with option `--enable-shared',
417 to build CLN as a shared library.
419 * Updated the LiDIA interface.
422 1998-05-07, version 0.98
423 ========================
425 Implementation changes
426 ----------------------
428 * Portability fixes for 64-bit platforms like DEC Alpha.
429 (Reported by John Cremona and Thomas Papanikolaou.)
438 * Portability fixes for new ISO/ANSI C++ compilers.
440 * Updated the LiDIA interface for LiDIA version 1.3 and fixed a couple of
447 Implementation changes
448 ----------------------
450 * Portability fix for i686 platforms.
451 (Reported by Michael Pfeifer.)
461 <cl_ring.h>, declaring general rings,
462 <cl_GV.h>, <cl_SV.h>, defining vectors,
463 <cl_univpoly.h>, defining univariate polynomial rings.
468 * Speed up the internal computation of ln(2) and ln(10) by about 20%.
470 * Speed up the computation of exp (for 1000 digits or more) by about 10%.
472 Implementation changes
473 ----------------------
475 * Portability fix for 64-bit platforms like DEC Alpha:
476 Fixed a bug: endless loop in `sqrt' for large precision.
477 (Reported by Blake Jones.)
479 * Portability fix for RS/6000 platforms.
488 * Speed up the computation of the elementary transcendental functions:
489 exp, sinh, cosh, asinh, atanh (for 100 digits) by about 30%,
490 log (for 100 digits) by about 25%,
491 sin, cos (for 1000 or 10000 digits) by about 20%,
492 asin, acos, atan (for 1000 digits) by about 25%.
494 Implementation changes
495 ----------------------
497 * Portability fix for SunOS 4 and other BSD platforms.
506 * Fixed a bug: endless recursion in `cl_boolean sqrtp (const cl_RA&, cl_RA*)'.
508 * Fixed a bug: Only the first 85% of the digits returned by `cl_eulerconst'
509 for moderate precision (< 11370 digits) were correct.
510 (Reported by Thomas Papanikolaou.)
512 Implementation changes
513 ----------------------
515 * Fix static initialization order bug. Requires g++ 2.7.0 or newer
516 (g++ 2.8.0 or newer on HPPA).
518 * New method `debug_print()', for pretty printing of CLN objects, intended
519 to be called from the debugger.
528 * New functions `float cl_float_approx (const cl_R&)',
529 `double cl_double_approx (const cl_R&)'
530 for converting a real number to a C `float' or `double'.
535 * Speed up `cl_zeta' using Cohen-Villegas-Zagier convergence acceleration.
537 Implementation changes
538 ----------------------
540 * New function `cl_print', for pretty printing of CLN objects, intended to be
541 called from the debugger.
543 * Portability fixes for 64-bit platforms like DEC Alpha.
553 <cl_string.h>, defining strings,
554 <cl_symbol.h>, defining symbols,
555 <cl_proplist.h>, defining property lists,
556 <cl_condition.h>, defining conditions/exceptions.
558 * New functions `cl_F cl_catalanconst ()',
559 `cl_F cl_catalanconst (const cl_F&)',
560 `cl_F cl_catalanconst (cl_float_format_t)'
561 which return Catalan's constant.
562 (Suggested by Thomas Papanikolaou.)
564 * New functions `cl_F cl_zeta (int s)',
565 `cl_F cl_zeta (int s, const cl_F&)',
566 `cl_F cl_zeta (int s, cl_float_format_t)'
567 which return the Riemann zeta function at an integral point s>1.
569 * New functions `cl_F cl_exp1 ()',
570 `cl_F cl_exp1 (const cl_F&)',
571 `cl_F cl_exp1 (cl_float_format_t)'
572 which return e = exp(1).
574 * New function `cl_I binomial (uintL n, uintL k)' which returns the
575 binomial coefficient (n choose k).
577 * New functions `int cl_I_to_int (const cl_I&)',
578 `unsigned int cl_I_to_uint (const cl_I&)',
579 `long cl_I_to_long (const cl_I&)',
580 `unsigned long cl_I_to_ulong (const cl_I&)'
581 for converting an integer to a C `int' or `long'.
583 * New functions `float cl_float_approx (const cl_I&)',
584 `float cl_float_approx (const cl_RA&)',
585 `double cl_double_approx (const cl_I&)',
586 `double cl_double_approx (const cl_RA&)'
587 for converting a rational number to a C `float' or `double'.
589 Implementation changes
590 ----------------------
592 * Moved the sources to subdirectories.
593 (Suggested by Jörg Arndt.)
598 * New benchmark for better comparing LiDIA, Pari and CLN.
600 * Added an interface to LiDIA, allows using CLN as a replacement of libI.
601 (Suggested by Thomas Papanikolaou.)
603 * Added an ILOG Talk interface, for interactive use of the library.
612 * New include file <cl_modinteger.h>, defining modular integer rings.
614 * New functions `cl_F cl_eulerconst ()', `cl_F cl_eulerconst (const cl_F&)',
615 `cl_F cl_eulerconst (cl_float_format_t)' which return Euler's constant.
620 * Speed up square root of large numbers by use of Newton's algorithm.
622 * Speed up multiplication and division of large long-floats by small integers.
624 * Speed up the computation of pi, e, exp, log, sin, cos, tan, sinh, cosh, tanh,
625 asin, acos, atan, asinh, acosh, atanh.
626 All of them are now computed with asymptotically fast algorithms, of
627 bit complexity O(log(N)^2*M(N)) = O(N^(1+epsilon)) for N digits.
629 * Fixed several bugs in the transcendental functions routines. In particular,
630 the `log' function went into an endless loop.
632 * Fixed a bug: The cl_I -> cl_DF and cl_I -> cl_LF conversions didn't work
633 correctly for negative integers.
642 * New include file <cl_numtheory.h>, containing number theoretic functions,
643 for now only the Jacobi symbol.
648 * Speed up squaring of large numbers by 30%.
650 * Speed up division of large numbers by use of Newton's algorithm. The
651 improvement is noticeable for numbers with at least about 1000 words =
652 32000 bits = 10000 decimal digits.
654 * Speed up the binary-to-decimal conversion of large numbers by use of
655 a divide-and-conquer method. The improvement is noticeable for numbers
656 of at least 1250 words = 40000 bits = 12000 decimal digits.
658 * Fixed a bug: The functions `cl_boolean minusp (const type&)' and
659 `cl_boolean zerop (const type&)' were missing for type=cl_I and type=cl_RA.
661 * Fixed a bug: The function `cl_boolean logtest (const cl_I&, const cl_I&)'
662 didn't work correctly if both arguments were bignums.
664 * Fixed a bug: The function `cl_I mod (const cl_I&, const cl_I&)', when called
665 as `mod(-n*m,m)' (n>0), returned m instead of 0.
667 * Fixed a bug: The function `uintL power2p (const cl_I&)' returned wrong
668 values for arguments x = 2^n, n >= 29.
670 Implementation changes
671 ----------------------
673 * Speed up CLN by using the low-level addition/subtraction, shift and
674 multiplication loops from GNU gmp-2.0.2.
676 * Fixed a severe reference counting bug in the constructors
677 `type::type (const long)' and `type::type (const unsigned long)'
678 for type=cl_number, type=cl_N, type=cl_RA, type=cl_I that most often
681 * CLN's word sequences can be chosen to be stored big-endian in memory
682 (like it was up to now) or little-endian (needed for interfacing to GMP).
691 * New include file <cl_timing.h>, contains timing facilities.
696 * Speed up the multiplication of large numbers by use of an FFT based
697 algorithm (Schönhage-Strassen). The improvement is noticeable when both
698 factors have at least 1250 words = 40000 bits = 12000 decimal digits.
700 * Speed up the functions `cl_I gcd (const cl_I&, const cl_I&)' and
701 `cl_I xgcd (const cl_I&, const cl_I&, cl_I*, cl_I*)' by a factor
702 of 1.5. Simple trick.
704 * Speed up the function `cl_boolean sqrtp (const cl_I&, cl_I*)' using a
705 trick from Henri Cohen.
707 * Fixed an endless loop in the function `uint32 gcd (uint32, uint32)'
708 which occurred when one of the arguments is zero.
710 * Fixed an endless loop in the function `cl_I factorial (uintL)'.
712 Implementation changes
713 ----------------------
715 * <cl_io.h> now uses iostreams by default instead of stdio FILE pointers.
716 (Reported by Tito Flagella.)
718 * Fixed linking error when compiling without -O.
719 (Reported by Tito Flagella.)
721 * Fixed syntax error in <cl_input.h>.
722 (Reported by Tito Flagella.)
724 * Fixed syntax errors in src/cl_DS.h and src/cl_FF_plus.h.
725 (Reported by Marcus Daniels.)
727 * More portability fixes.
729 * Configuration now supports the --srcdir option.
730 (Reported by Marcus Daniels.)
736 * Compiles well on Sparc, using gcc.
738 * Mail to beta testers.
744 * Compiles well on Linux, using gcc.