1 2007-mm-dd, version 1.2.0
2 =========================
7 * Added support for huge numbers...
8 * CLN now uses C++ exceptions for error handling.
9 * Removed cl_boolean. Use built-in C++ bool instead.
10 * Dropped the cln-config script. Please use pkg-config instead.
15 * Save a lot of memory during computation of some constants.
18 2006-08-08, version 1.1.13
19 ==========================
21 * Compilation fixes for 64-bit brokenness introduced in last release.
24 2006-08-06, version 1.1.12
25 ==========================
27 Implementation changes
28 ----------------------
30 * Fix rare assertion when printing quite large floats.
35 * Compilation fixes for several platforms: *BSD, Intel Mac, and MinGW.
38 2005-11-23, version 1.1.11
39 ==========================
44 * Considerably improved performance of number input.
47 2005-10-22, version 1.1.10
48 ==========================
50 Implementation changes
51 ----------------------
53 * Removed the vptr of modular integer and univariate polynomial
54 classes in order to fix some crashes in that sector of CLN. Code
55 using this needs to be recompiled.
57 * Many more operator signatures make the integration of CLN types with
63 * Several compilation fixes.
65 * Made it possible to cross-compile CLN.
68 2004-11-03, version 1.1.9
69 =========================
74 * Input of numbers in bases 2, 4, 8, 16 and 32 is now done in linear bit
75 complexity as opposed to O(N^2). Useful for all kinds of persistency.
77 Implementation changes
78 ----------------------
80 * Fixed several bugs in the integer input and output routines that could be
81 blamed for random crashes in the following cases: output in base 32 for
82 quite large numbers, input in base 2 for fixnums and input in base 3 for
83 fixnums on a 64 bit target.
85 * Fixed crash when radix specifiers were used in input streams.
87 * Speed up on x86_64 and ia64 by adding some inline assembly.
92 * Fixes for compilation on MacOS X and little endian Mips.
95 2004-07-01, version 1.1.8
96 =========================
98 Implementation changes
99 ----------------------
101 * Fix bug in log(a,b) when b is 1/n.
103 * Fix crash in shared library on AMD64.
108 * Fixes for some language conformance issues.
111 2004-05-02, version 1.1.7
112 =========================
114 Implementation changes
115 ----------------------
117 * Fixed two corner case bugs in cl_LF multiplication and sqrt.
119 * Workaround GCC 3.[123] compiler bug.
121 * Parallel build should work reliably now.
124 2004-01-01, version 1.1.6
125 =========================
131 `sintL ldegree(const cl_UP& x)'.
132 (Suggested by Munagala Ramanath.)
134 Implementation changes
135 ----------------------
137 * Fixed a bug in polynomial subtraction.
138 (Reported by Munagala Ramanath.)
143 * Provide a package control file cln.pc.
145 * Updated infrastructure to autoconf-2.57.
147 * Fixed compilation issues with GCC-3.4 prereleases and on Mac OSX.
150 2002-05-28, version 1.1.5
151 =========================
156 * Fix compilation issues with GCC-3.1.
159 2002-01-04, version 1.1.4
160 =========================
165 * Fix compilation issues with GMP-4.0 and on IBM s390.
167 * Updated infrastructure to autoconf-2.52.
170 2001-11-05, version 1.1.3
171 =========================
173 Implementation changes
174 ----------------------
176 * Fix a computation error of sin and sinh for small arguments.
177 (Reported by Christopher Kennedy.)
179 * Module ordering works with gcc-3.0 -fuse-cxa-atexit now.
181 * -DNO_ASM works properly on Sparc.
184 2001-07-25, version 1.1.2
186 Implementation changes
187 ----------------------
189 * Minor cleanups and portability issues.
190 (No need to upgrade if you have been happy so far.)
193 2001-05-31, version 1.1.1
194 =========================
196 Implementation changes
197 ----------------------
199 * Minor cleanups for GCC 3.0 and libstdc++-v3.
201 * Fixes several little bugs and portability issues.
204 2000-12-14, version 1.1.0
205 =========================
210 * ISO/IEC 14882 fashion adjustments:
211 Put everything into namespace cln. All fundamental data types still
212 carry their old names. Other non-macro identifiers are now written as
213 cln::foo instead of cl_foo, except where the cl_ comes from a data type
214 (as in cl_I_to_int()). Headers are installed into a separate
215 directory, #include <cln/foo.h> instead of <cl_foo.h>. Applications
216 must be manually ported to the new style. We apologize for the
219 Implementation changes
220 ----------------------
222 * Removed internal copy of mpn subdirectory tree from libgmp-2.0.2.
223 Configuring with --with-gmp now relies on an installed libgmp library
224 version 3 or above. We recommend libgmp-3.1 or newer.
226 * Adjusted several break-even points to make better use of a faster libgmp and
227 better match present-day CPUs.
229 * Fix several errors for architectures with 64-bit wide words.
230 (Others than Alpha, which worked already.)
232 * Fix compilation errors with current GCC-3.0 snapshots.
237 * Added package tools: script cln-config and autoconf macro AC_PATH_CLN
241 2000-01-13, version 1.0.3
242 =========================
248 `cl_I doublefactorial (uintL n)'.
250 Implementation changes
251 ----------------------
253 * Fix several little configuration errors.
255 * Fix some compilation errors with gcc-2.96 prereleases.
258 1999-09-07, version 1.0.2
259 =========================
265 `cl_I numerator (const cl_RA&)'
266 `cl_I denominator (const cl_RA&)'.
267 Suggested by Richard Kreckel and Sylvain Pion.
269 * New function `cl_equal_hashcode' for the classes cl_N, cl_R, cl_RA, cl_I,
270 cl_F, cl_SF, cl_FF, cl_DF, cl_LF.
271 Suggested by Richard Kreckel.
273 Implementation changes
274 ----------------------
276 * Fix an endless loop when either of the functions
277 `cl_boolean rootp (const cl_RA&, uintL, cl_RA*)',
278 `cl_boolean rootp (const cl_RA&, const cl_I&, cl_RA*)'
279 was called with an integer argument. Thanks to Richard Kreckel.
281 * Fix a bug in the addition and subtraction of rational numbers which could
282 cause unnormalized rational numbers like 3/1 to be created.
285 1999-06-09, version 1.0.1
286 =========================
291 * Speed up the functions `logand', `lognand', `logandc2'.
293 Implementation changes
294 ----------------------
296 * Fix compilation errors with gcc-2.7.2, egcs-1.1.2 and gcc-2.95.
298 * Fix compilation errors on HPPA, MIPS, some versions of DEC Alpha, OpenBSD,
299 and SPARC with LiDIA.
301 * Fix compilation errors on Win32. Thanks to Mumit Khan.
304 1999-01-12, version 1.0
305 =======================
310 * New include files, containing I/O functions:
314 <cl_sfloat_io.h> <cl_ffloat_io.h> <cl_dfloat_io.h> <cl_lfloat_io.h>
317 The file <cl_io.h> now contains only I/O functions for characters and
320 * To access the base ring of a univariate polynomial ring, now use
321 `R->basering()' instead of `R->basering'.
323 * Implement `plusp', `max', `min' for the classes cl_F, cl_SF, cl_FF, cl_DF,
326 * Implement `abs' for the class cl_RA.
328 * Replaced `read_number' with specialized functions `read_complex',
329 `read_real', `read_float', `read_rational', `read_integer'.
331 * Replaced the functions
332 `void fprint (cl_ostream stream, int x)'
333 `void fprint (cl_ostream stream, unsigned int x)'
334 `void fprint (cl_ostream stream, long x)'
335 `void fprint (cl_ostream stream, unsigned long x)'
337 `void fprintdecimal (cl_ostream stream, int x)'
338 `void fprintdecimal (cl_ostream stream, unsigned int x)'
339 `void fprintdecimal (cl_ostream stream, long x)'
340 `void fprintdecimal (cl_ostream stream, unsigned long x)'
345 * The function `cl_I xgcd (const cl_I& a, const cl_I& b, cl_I* u, cl_I* v)'
346 now normalizes the resulting u and v to be of minimum absolute value.
347 (Suggested by the LiDIA group.)
349 * The conversion from string to cl_F, cl_R, cl_N now chooses the format of
350 the resulting float depending on the number of significant digits present
353 * Speed up the functions
354 `cl_R operator* (const cl_RA&, const cl_LF&)'
355 `cl_R operator* (const cl_LF&, const cl_RA&)'
356 `cl_R operator/ (const cl_RA&, const cl_LF&)'
357 `cl_LF operator/ (const cl_LF&, const cl_RA&)'
359 Implementation changes
360 ----------------------
362 * Fix for `cl_I ash (const cl_I& x, const cl_I& n)' when |n| is between
364 (Reported by Thomas Papanikolaou.)
366 * Fixed a memory leak in the long-float truncate function.
367 (Reported by Keith Briggs.)
369 * Fixed a memory leak in the modular integer exponentiation.
371 * Reduced the stack space consumption. The default stack size (typically
372 8 MB on Unix) should now be sufficient in order to run all CLN programs.
374 * Portability fixes for m68k platforms.
375 (Reported and fixed by Roman Hodek.)
380 * Restructured the sources. The subpackages
383 - base + integer + rational
384 - base + integer + rational + float
385 - base + integer + rational + float + real
386 - base + integer + rational + float + real + complex
387 - base + integer + modinteger
388 are now self-contained.
390 * The configuration script can be called with option `--enable-shared',
391 to build CLN as a shared library.
393 * Updated the LiDIA interface.
396 1998-05-07, version 0.98
397 ========================
399 Implementation changes
400 ----------------------
402 * Portability fixes for 64-bit platforms like DEC Alpha.
403 (Reported by John Cremona and Thomas Papanikolaou.)
412 * Portability fixes for new ISO/ANSI C++ compilers.
414 * Updated the LiDIA interface for LiDIA version 1.3 and fixed a couple of
421 Implementation changes
422 ----------------------
424 * Portability fix for i686 platforms.
425 (Reported by Michael Pfeifer.)
435 <cl_ring.h>, declaring general rings,
436 <cl_GV.h>, <cl_SV.h>, defining vectors,
437 <cl_univpoly.h>, defining univariate polynomial rings.
442 * Speed up the internal computation of ln(2) and ln(10) by about 20%.
444 * Speed up the computation of exp (for 1000 digits or more) by about 10%.
446 Implementation changes
447 ----------------------
449 * Portability fix for 64-bit platforms like DEC Alpha:
450 Fixed a bug: endless loop in `sqrt' for large precision.
451 (Reported by Blake Jones.)
453 * Portability fix for RS/6000 platforms.
462 * Speed up the computation of the elementary transcendental functions:
463 exp, sinh, cosh, asinh, atanh (for 100 digits) by about 30%,
464 log (for 100 digits) by about 25%,
465 sin, cos (for 1000 or 10000 digits) by about 20%,
466 asin, acos, atan (for 1000 digits) by about 25%.
468 Implementation changes
469 ----------------------
471 * Portability fix for SunOS 4 and other BSD platforms.
480 * Fixed a bug: endless recursion in `cl_boolean sqrtp (const cl_RA&, cl_RA*)'.
482 * Fixed a bug: Only the first 85% of the digits returned by `cl_eulerconst'
483 for moderate precision (< 11370 digits) were correct.
484 (Reported by Thomas Papanikolaou.)
486 Implementation changes
487 ----------------------
489 * Fix static initialization order bug. Requires g++ 2.7.0 or newer
490 (g++ 2.8.0 or newer on HPPA).
492 * New method `debug_print()', for pretty printing of CLN objects, intended
493 to be called from the debugger.
502 * New functions `float cl_float_approx (const cl_R&)',
503 `double cl_double_approx (const cl_R&)'
504 for converting a real number to a C `float' or `double'.
509 * Speed up `cl_zeta' using Cohen-Villegas-Zagier convergence acceleration.
511 Implementation changes
512 ----------------------
514 * New function `cl_print', for pretty printing of CLN objects, intended to be
515 called from the debugger.
517 * Portability fixes for 64-bit platforms like DEC Alpha.
527 <cl_string.h>, defining strings,
528 <cl_symbol.h>, defining symbols,
529 <cl_proplist.h>, defining property lists,
530 <cl_condition.h>, defining conditions/exceptions.
532 * New functions `cl_F cl_catalanconst ()',
533 `cl_F cl_catalanconst (const cl_F&)',
534 `cl_F cl_catalanconst (cl_float_format_t)'
535 which return Catalan's constant.
536 (Suggested by Thomas Papanikolaou.)
538 * New functions `cl_F cl_zeta (int s)',
539 `cl_F cl_zeta (int s, const cl_F&)',
540 `cl_F cl_zeta (int s, cl_float_format_t)'
541 which return the Riemann zeta function at an integral point s>1.
543 * New functions `cl_F cl_exp1 ()',
544 `cl_F cl_exp1 (const cl_F&)',
545 `cl_F cl_exp1 (cl_float_format_t)'
546 which return e = exp(1).
548 * New function `cl_I binomial (uintL n, uintL k)' which returns the
549 binomial coefficient (n choose k).
551 * New functions `int cl_I_to_int (const cl_I&)',
552 `unsigned int cl_I_to_uint (const cl_I&)',
553 `long cl_I_to_long (const cl_I&)',
554 `unsigned long cl_I_to_ulong (const cl_I&)'
555 for converting an integer to a C `int' or `long'.
557 * New functions `float cl_float_approx (const cl_I&)',
558 `float cl_float_approx (const cl_RA&)',
559 `double cl_double_approx (const cl_I&)',
560 `double cl_double_approx (const cl_RA&)'
561 for converting a rational number to a C `float' or `double'.
563 Implementation changes
564 ----------------------
566 * Moved the sources to subdirectories.
567 (Suggested by Jörg Arndt.)
572 * New benchmark for better comparing LiDIA, Pari and CLN.
574 * Added an interface to LiDIA, allows using CLN as a replacement of libI.
575 (Suggested by Thomas Papanikolaou.)
577 * Added an ILOG Talk interface, for interactive use of the library.
586 * New include file <cl_modinteger.h>, defining modular integer rings.
588 * New functions `cl_F cl_eulerconst ()', `cl_F cl_eulerconst (const cl_F&)',
589 `cl_F cl_eulerconst (cl_float_format_t)' which return Euler's constant.
594 * Speed up square root of large numbers by use of Newton's algorithm.
596 * Speed up multiplication and division of large long-floats by small integers.
598 * Speed up the computation of pi, e, exp, log, sin, cos, tan, sinh, cosh, tanh,
599 asin, acos, atan, asinh, acosh, atanh.
600 All of them are now computed with asymptotically fast algorithms, of
601 bit complexity O(log(N)^2*M(N)) = O(N^(1+epsilon)) for N digits.
603 * Fixed several bugs in the transcendental functions routines. In particular,
604 the `log' function went into an endless loop.
606 * Fixed a bug: The cl_I -> cl_DF and cl_I -> cl_LF conversions didn't work
607 correctly for negative integers.
616 * New include file <cl_numtheory.h>, containing number theoretic functions,
617 for now only the Jacobi symbol.
622 * Speed up squaring of large numbers by 30%.
624 * Speed up division of large numbers by use of Newton's algorithm. The
625 improvement is noticeable for numbers with at least about 1000 words =
626 32000 bits = 10000 decimal digits.
628 * Speed up the binary-to-decimal conversion of large numbers by use of
629 a divide-and-conquer method. The improvement is noticeable for numbers
630 of at least 1250 words = 40000 bits = 12000 decimal digits.
632 * Fixed a bug: The functions `cl_boolean minusp (const type&)' and
633 `cl_boolean zerop (const type&)' were missing for type=cl_I and type=cl_RA.
635 * Fixed a bug: The function `cl_boolean logtest (const cl_I&, const cl_I&)'
636 didn't work correctly if both arguments were bignums.
638 * Fixed a bug: The function `cl_I mod (const cl_I&, const cl_I&)', when called
639 as `mod(-n*m,m)' (n>0), returned m instead of 0.
641 * Fixed a bug: The function `uintL power2p (const cl_I&)' returned wrong
642 values for arguments x = 2^n, n >= 29.
644 Implementation changes
645 ----------------------
647 * Speed up CLN by using the low-level addition/subtraction, shift and
648 multiplication loops from GNU gmp-2.0.2.
650 * Fixed a severe reference counting bug in the constructors
651 `type::type (const long)' and `type::type (const unsigned long)'
652 for type=cl_number, type=cl_N, type=cl_RA, type=cl_I that most often
655 * CLN's word sequences can be chosen to be stored big-endian in memory
656 (like it was up to now) or little-endian (needed for interfacing to GMP).
665 * New include file <cl_timing.h>, contains timing facilities.
670 * Speed up the multiplication of large numbers by use of an FFT based
671 algorithm (Schönhage-Strassen). The improvement is noticeable when both
672 factors have at least 1250 words = 40000 bits = 12000 decimal digits.
674 * Speed up the functions `cl_I gcd (const cl_I&, const cl_I&)' and
675 `cl_I xgcd (const cl_I&, const cl_I&, cl_I*, cl_I*)' by a factor
676 of 1.5. Simple trick.
678 * Speed up the function `cl_boolean sqrtp (const cl_I&, cl_I*)' using a
679 trick from Henri Cohen.
681 * Fixed an endless loop in the function `uint32 gcd (uint32, uint32)'
682 which occurred when one of the arguments is zero.
684 * Fixed an endless loop in the function `cl_I factorial (uintL)'.
686 Implementation changes
687 ----------------------
689 * <cl_io.h> now uses iostreams by default instead of stdio FILE pointers.
690 (Reported by Tito Flagella.)
692 * Fixed linking error when compiling without -O.
693 (Reported by Tito Flagella.)
695 * Fixed syntax error in <cl_input.h>.
696 (Reported by Tito Flagella.)
698 * Fixed syntax errors in src/cl_DS.h and src/cl_FF_plus.h.
699 (Reported by Marcus Daniels.)
701 * More portability fixes.
703 * Configuration now supports the --srcdir option.
704 (Reported by Marcus Daniels.)
710 * Compiles well on Sparc, using gcc.
712 * Mail to beta testers.
718 * Compiles well on Linux, using gcc.