1 2004-01-01, version 1.1.6
7 `sintL ldegree(const cl_UP& x)'.
8 (Suggested by Munagala Ramanath.)
10 Implementation changes
11 ----------------------
13 * Fixed a bug in polynomial subtraction.
14 (Reported by Munagala Ramanath.)
19 * Provide a package control file cln.pc.
21 * Updated infrastructure to autoconf-2.57.
23 * Fixed compilation issues with GCC-3.4 prereleases and on Mac OSX.
26 2002-05-28, version 1.1.5
27 =========================
32 * Fix compilation issues with GCC-3.1.
35 2002-01-04, version 1.1.4
36 =========================
41 * Fix compilation issues with GMP-4.0 and on IBM s390.
43 * Updated infrastructure to autoconf-2.52.
46 2001-11-05, version 1.1.3
47 =========================
49 Implementation changes
50 ----------------------
52 * Fix a computation error of sin and sinh for small arguments.
53 (Reported by Christopher Kennedy.)
55 * Module ordering works with gcc-3.0 -fuse-cxa-atexit now.
57 * -DNO_ASM works properly on Sparc.
60 2001-07-25, version 1.1.2
62 Implementation changes
63 ----------------------
65 * Minor cleanups and portability issues.
66 (No need to upgrade if you have been happy so far.)
69 2001-05-31, version 1.1.1
70 =========================
72 Implementation changes
73 ----------------------
75 * Minor cleanups for GCC 3.0 and libstdc++-v3.
77 * Fixes several little bugs and portability issues.
80 2000-12-14, version 1.1.0
81 =========================
86 * ISO/IEC 14882 fashion adjustments:
87 Put everything into namespace cln. All fundamental data types still
88 carry their old names. Other non-macro identifiers are now written as
89 cln::foo instead of cl_foo, except where the cl_ comes from a data type
90 (as in cl_I_to_int()). Headers are installed into a separate
91 directory, #include <cln/foo.h> instead of <cl_foo.h>. Applications
92 must be manually ported to the new style. We apologize for the
95 Implementation changes
96 ----------------------
98 * Removed internal copy of mpn subdirectory tree from libgmp-2.0.2.
99 Configuring with --with-gmp now relies on an installed libgmp library
100 version 3 or above. We recommend libgmp-3.1 or newer.
102 * Adjusted several break-even points to make better use of a faster libgmp and
103 better match present-day CPUs.
105 * Fix several errors for architectures with 64-bit wide words.
106 (Others than Alpha, which worked already.)
108 * Fix compilation errors with current GCC-3.0 snapshots.
113 * Added package tools: script cln-config and autoconf macro AC_PATH_CLN
117 2000-01-13, version 1.0.3
118 =========================
124 `cl_I doublefactorial (uintL n)'.
126 Implementation changes
127 ----------------------
129 * Fix several little configuration errors.
131 * Fix some compilation errors with gcc-2.96 prereleases.
134 1999-09-07, version 1.0.2
135 =========================
141 `cl_I numerator (const cl_RA&)'
142 `cl_I denominator (const cl_RA&)'.
143 Suggested by Richard Kreckel and Sylvain Pion.
145 * New function `cl_equal_hashcode' for the classes cl_N, cl_R, cl_RA, cl_I,
146 cl_F, cl_SF, cl_FF, cl_DF, cl_LF.
147 Suggested by Richard Kreckel.
149 Implementation changes
150 ----------------------
152 * Fix an endless loop when either of the functions
153 `cl_boolean rootp (const cl_RA&, uintL, cl_RA*)',
154 `cl_boolean rootp (const cl_RA&, const cl_I&, cl_RA*)'
155 was called with an integer argument. Thanks to Richard Kreckel.
157 * Fix a bug in the addition and subtraction of rational numbers which could
158 cause unnormalized rational numbers like 3/1 to be created.
161 1999-06-09, version 1.0.1
162 =========================
167 * Speed up the functions `logand', `lognand', `logandc2'.
169 Implementation changes
170 ----------------------
172 * Fix compilation errors with gcc-2.7.2, egcs-1.1.2 and gcc-2.95.
174 * Fix compilation errors on HPPA, MIPS, some versions of DEC Alpha, OpenBSD,
175 and SPARC with LiDIA.
177 * Fix compilation errors on Win32. Thanks to Mumit Khan.
180 1999-01-12, version 1.0
181 =======================
186 * New include files, containing I/O functions:
190 <cl_sfloat_io.h> <cl_ffloat_io.h> <cl_dfloat_io.h> <cl_lfloat_io.h>
193 The file <cl_io.h> now contains only I/O functions for characters and
196 * To access the base ring of a univariate polynomial ring, now use
197 `R->basering()' instead of `R->basering'.
199 * Implement `plusp', `max', `min' for the classes cl_F, cl_SF, cl_FF, cl_DF,
202 * Implement `abs' for the class cl_RA.
204 * Replaced `read_number' with specialized functions `read_complex',
205 `read_real', `read_float', `read_rational', `read_integer'.
207 * Replaced the functions
208 `void fprint (cl_ostream stream, int x)'
209 `void fprint (cl_ostream stream, unsigned int x)'
210 `void fprint (cl_ostream stream, long x)'
211 `void fprint (cl_ostream stream, unsigned long x)'
213 `void fprintdecimal (cl_ostream stream, int x)'
214 `void fprintdecimal (cl_ostream stream, unsigned int x)'
215 `void fprintdecimal (cl_ostream stream, long x)'
216 `void fprintdecimal (cl_ostream stream, unsigned long x)'
221 * The function `cl_I xgcd (const cl_I& a, const cl_I& b, cl_I* u, cl_I* v)'
222 now normalizes the resulting u and v to be of minimum absolute value.
223 (Suggested by the LiDIA group.)
225 * The conversion from string to cl_F, cl_R, cl_N now chooses the format of
226 the resulting float depending on the number of significant digits present
229 * Speed up the functions
230 `cl_R operator* (const cl_RA&, const cl_LF&)'
231 `cl_R operator* (const cl_LF&, const cl_RA&)'
232 `cl_R operator/ (const cl_RA&, const cl_LF&)'
233 `cl_LF operator/ (const cl_LF&, const cl_RA&)'
235 Implementation changes
236 ----------------------
238 * Fix for `cl_I ash (const cl_I& x, const cl_I& n)' when |n| is between
240 (Reported by Thomas Papanikolaou.)
242 * Fixed a memory leak in the long-float truncate function.
243 (Reported by Keith Briggs.)
245 * Fixed a memory leak in the modular integer exponentiation.
247 * Reduced the stack space consumption. The default stack size (typically
248 8 MB on Unix) should now be sufficient in order to run all CLN programs.
250 * Portability fixes for m68k platforms.
251 (Reported and fixed by Roman Hodek.)
256 * Restructured the sources. The subpackages
259 - base + integer + rational
260 - base + integer + rational + float
261 - base + integer + rational + float + real
262 - base + integer + rational + float + real + complex
263 - base + integer + modinteger
264 are now self-contained.
266 * The configuration script can be called with option `--enable-shared',
267 to build CLN as a shared library.
269 * Updated the LiDIA interface.
272 1998-05-07, version 0.98
273 ========================
275 Implementation changes
276 ----------------------
278 * Portability fixes for 64-bit platforms like DEC Alpha.
279 (Reported by John Cremona and Thomas Papanikolaou.)
288 * Portability fixes for new ISO/ANSI C++ compilers.
290 * Updated the LiDIA interface for LiDIA version 1.3 and fixed a couple of
297 Implementation changes
298 ----------------------
300 * Portability fix for i686 platforms.
301 (Reported by Michael Pfeifer.)
311 <cl_ring.h>, declaring general rings,
312 <cl_GV.h>, <cl_SV.h>, defining vectors,
313 <cl_univpoly.h>, defining univariate polynomial rings.
318 * Speed up the internal computation of ln(2) and ln(10) by about 20%.
320 * Speed up the computation of exp (for 1000 digits or more) by about 10%.
322 Implementation changes
323 ----------------------
325 * Portability fix for 64-bit platforms like DEC Alpha:
326 Fixed a bug: endless loop in `sqrt' for large precision.
327 (Reported by Blake Jones.)
329 * Portability fix for RS/6000 platforms.
338 * Speed up the computation of the elementary transcendental functions:
339 exp, sinh, cosh, asinh, atanh (for 100 digits) by about 30%,
340 log (for 100 digits) by about 25%,
341 sin, cos (for 1000 or 10000 digits) by about 20%,
342 asin, acos, atan (for 1000 digits) by about 25%.
344 Implementation changes
345 ----------------------
347 * Portability fix for SunOS 4 and other BSD platforms.
356 * Fixed a bug: endless recursion in `cl_boolean sqrtp (const cl_RA&, cl_RA*)'.
358 * Fixed a bug: Only the first 85% of the digits returned by `cl_eulerconst'
359 for moderate precision (< 11370 digits) were correct.
360 (Reported by Thomas Papanikolaou.)
362 Implementation changes
363 ----------------------
365 * Fix static initialization order bug. Requires g++ 2.7.0 or newer
366 (g++ 2.8.0 or newer on HPPA).
368 * New method `debug_print()', for pretty printing of CLN objects, intended
369 to be called from the debugger.
378 * New functions `float cl_float_approx (const cl_R&)',
379 `double cl_double_approx (const cl_R&)'
380 for converting a real number to a C `float' or `double'.
385 * Speed up `cl_zeta' using Cohen-Villegas-Zagier convergence acceleration.
387 Implementation changes
388 ----------------------
390 * New function `cl_print', for pretty printing of CLN objects, intended to be
391 called from the debugger.
393 * Portability fixes for 64-bit platforms like DEC Alpha.
403 <cl_string.h>, defining strings,
404 <cl_symbol.h>, defining symbols,
405 <cl_proplist.h>, defining property lists,
406 <cl_condition.h>, defining conditions/exceptions.
408 * New functions `cl_F cl_catalanconst ()',
409 `cl_F cl_catalanconst (const cl_F&)',
410 `cl_F cl_catalanconst (cl_float_format_t)'
411 which return Catalan's constant.
412 (Suggested by Thomas Papanikolaou.)
414 * New functions `cl_F cl_zeta (int s)',
415 `cl_F cl_zeta (int s, const cl_F&)',
416 `cl_F cl_zeta (int s, cl_float_format_t)'
417 which return the Riemann zeta function at an integral point s>1.
419 * New functions `cl_F cl_exp1 ()',
420 `cl_F cl_exp1 (const cl_F&)',
421 `cl_F cl_exp1 (cl_float_format_t)'
422 which return e = exp(1).
424 * New function `cl_I binomial (uintL n, uintL k)' which returns the
425 binomial coefficient (n choose k).
427 * New functions `int cl_I_to_int (const cl_I&)',
428 `unsigned int cl_I_to_uint (const cl_I&)',
429 `long cl_I_to_long (const cl_I&)',
430 `unsigned long cl_I_to_ulong (const cl_I&)'
431 for converting an integer to a C `int' or `long'.
433 * New functions `float cl_float_approx (const cl_I&)',
434 `float cl_float_approx (const cl_RA&)',
435 `double cl_double_approx (const cl_I&)',
436 `double cl_double_approx (const cl_RA&)'
437 for converting a rational number to a C `float' or `double'.
439 Implementation changes
440 ----------------------
442 * Moved the sources to subdirectories.
443 (Suggested by Jörg Arndt.)
448 * New benchmark for better comparing LiDIA, Pari and CLN.
450 * Added an interface to LiDIA, allows using CLN as a replacement of libI.
451 (Suggested by Thomas Papanikolaou.)
453 * Added an ILOG Talk interface, for interactive use of the library.
462 * New include file <cl_modinteger.h>, defining modular integer rings.
464 * New functions `cl_F cl_eulerconst ()', `cl_F cl_eulerconst (const cl_F&)',
465 `cl_F cl_eulerconst (cl_float_format_t)' which return Euler's constant.
470 * Speed up square root of large numbers by use of Newton's algorithm.
472 * Speed up multiplication and division of large long-floats by small integers.
474 * Speed up the computation of pi, e, exp, log, sin, cos, tan, sinh, cosh, tanh,
475 asin, acos, atan, asinh, acosh, atanh.
476 All of them are now computed with asymptotically fast algorithms, of
477 bit complexity O(log(N)^2*M(N)) = O(N^(1+epsilon)) for N digits.
479 * Fixed several bugs in the transcendental functions routines. In particular,
480 the `log' function went into an endless loop.
482 * Fixed a bug: The cl_I -> cl_DF and cl_I -> cl_LF conversions didn't work
483 correctly for negative integers.
492 * New include file <cl_numtheory.h>, containing number theoretic functions,
493 for now only the Jacobi symbol.
498 * Speed up squaring of large numbers by 30%.
500 * Speed up division of large numbers by use of Newton's algorithm. The
501 improvement is noticeable for numbers with at least about 1000 words =
502 32000 bits = 10000 decimal digits.
504 * Speed up the binary-to-decimal conversion of large numbers by use of
505 a divide-and-conquer method. The improvement is noticeable for numbers
506 of at least 1250 words = 40000 bits = 12000 decimal digits.
508 * Fixed a bug: The functions `cl_boolean minusp (const type&)' and
509 `cl_boolean zerop (const type&)' were missing for type=cl_I and type=cl_RA.
511 * Fixed a bug: The function `cl_boolean logtest (const cl_I&, const cl_I&)'
512 didn't work correctly if both arguments were bignums.
514 * Fixed a bug: The function `cl_I mod (const cl_I&, const cl_I&)', when called
515 as `mod(-n*m,m)' (n>0), returned m instead of 0.
517 * Fixed a bug: The function `uintL power2p (const cl_I&)' returned wrong
518 values for arguments x = 2^n, n >= 29.
520 Implementation changes
521 ----------------------
523 * Speed up CLN by using the low-level addition/subtraction, shift and
524 multiplication loops from GNU gmp-2.0.2.
526 * Fixed a severe reference counting bug in the constructors
527 `type::type (const long)' and `type::type (const unsigned long)'
528 for type=cl_number, type=cl_N, type=cl_RA, type=cl_I that most often
531 * CLN's word sequences can be chosen to be stored big-endian in memory
532 (like it was up to now) or little-endian (needed for interfacing to GMP).
541 * New include file <cl_timing.h>, contains timing facilities.
546 * Speed up the multiplication of large numbers by use of an FFT based
547 algorithm (Schönhage-Strassen). The improvement is noticeable when both
548 factors have at least 1250 words = 40000 bits = 12000 decimal digits.
550 * Speed up the functions `cl_I gcd (const cl_I&, const cl_I&)' and
551 `cl_I xgcd (const cl_I&, const cl_I&, cl_I*, cl_I*)' by a factor
552 of 1.5. Simple trick.
554 * Speed up the function `cl_boolean sqrtp (const cl_I&, cl_I*)' using a
555 trick from Henri Cohen.
557 * Fixed an endless loop in the function `uint32 gcd (uint32, uint32)'
558 which occurred when one of the arguments is zero.
560 * Fixed an endless loop in the function `cl_I factorial (uintL)'.
562 Implementation changes
563 ----------------------
565 * <cl_io.h> now uses iostreams by default instead of stdio FILE pointers.
566 (Reported by Tito Flagella.)
568 * Fixed linking error when compiling without -O.
569 (Reported by Tito Flagella.)
571 * Fixed syntax error in <cl_input.h>.
572 (Reported by Tito Flagella.)
574 * Fixed syntax errors in src/cl_DS.h and src/cl_FF_plus.h.
575 (Reported by Marcus Daniels.)
577 * More portability fixes.
579 * Configuration now supports the --srcdir option.
580 (Reported by Marcus Daniels.)
586 * Compiles well on Sparc, using gcc.
588 * Mail to beta testers.
594 * Compiles well on Linux, using gcc.