1 2001-07-25, version 1.1.2
6 * Minor cleanups and portability issues.
7 (No need to upgrade if you have been happy so far.)
10 2001-05-31, version 1.1.1
11 =========================
13 Implementation changes
14 ----------------------
16 * Minor cleanups for GCC 3.0 and libstdc++-v3.
18 * Fixes several little bugs and portability issues.
21 2000-12-14, version 1.1.0
22 =========================
27 * ISO/IEC 14882 fashion adjustments:
28 Put everything into namespace cln. All fundamental data types still
29 carry their old names. Other non-macro identifiers are now written as
30 cln::foo instead of cl_foo, except where the cl_ comes from a data type
31 (as in cl_I_to_int()). Headers are installed into a separate
32 directory, #include <cln/foo.h> instead of <cl_foo.h>. Applications
33 must be manually ported to the new style. We apologize for the
36 Implementation changes
37 ----------------------
39 * Removed internal copy of mpn subdirectory tree from libgmp-2.0.2.
40 Configuring with --with-gmp now relies on an installed libgmp library
41 version 3 or above. We recommend libgmp-3.1 or newer.
43 * Adjusted several break-even points to make better use of a faster libgmp and
44 better match present-day CPUs.
46 * Fix several errors for architectures with 64-bit wide words.
47 (Others than Alpha, which worked already.)
49 * Fix compilation errors with current GCC-3.0 snapshots.
54 * Added package tools: script cln-config and autoconf macro AC_PATH_CLN
58 2000-01-13, version 1.0.3
59 =========================
65 `cl_I doublefactorial (uintL n)'.
67 Implementation changes
68 ----------------------
70 * Fix several little configuration errors.
72 * Fix some compilation errors with gcc-2.96 prereleases.
75 1999-09-07, version 1.0.2
76 =========================
82 `cl_I numerator (const cl_RA&)'
83 `cl_I denominator (const cl_RA&)'.
84 Suggested by Richard Kreckel and Sylvain Pion.
86 * New function `cl_equal_hashcode' for the classes cl_N, cl_R, cl_RA, cl_I,
87 cl_F, cl_SF, cl_FF, cl_DF, cl_LF.
88 Suggested by Richard Kreckel.
90 Implementation changes
91 ----------------------
93 * Fix an endless loop when either of the functions
94 `cl_boolean rootp (const cl_RA&, uintL, cl_RA*)',
95 `cl_boolean rootp (const cl_RA&, const cl_I&, cl_RA*)'
96 was called with an integer argument. Thanks to Richard Kreckel.
98 * Fix a bug in the addition and subtraction of rational numbers which could
99 cause unnormalized rational numbers like 3/1 to be created.
102 1999-06-09, version 1.0.1
103 =========================
108 * Speed up the functions `logand', `lognand', `logandc2'.
110 Implementation changes
111 ----------------------
113 * Fix compilation errors with gcc-2.7.2, egcs-1.1.2 and gcc-2.95.
115 * Fix compilation errors on HPPA, MIPS, some versions of DEC Alpha, OpenBSD,
116 and SPARC with LiDIA.
118 * Fix compilation errors on Win32. Thanks to Mumit Khan.
121 1999-01-12, version 1.0
122 =======================
127 * New include files, containing I/O functions:
131 <cl_sfloat_io.h> <cl_ffloat_io.h> <cl_dfloat_io.h> <cl_lfloat_io.h>
134 The file <cl_io.h> now contains only I/O functions for characters and
137 * To access the base ring of a univariate polynomial ring, now use
138 `R->basering()' instead of `R->basering'.
140 * Implement `plusp', `max', `min' for the classes cl_F, cl_SF, cl_FF, cl_DF,
143 * Implement `abs' for the class cl_RA.
145 * Replaced `read_number' with specialized functions `read_complex',
146 `read_real', `read_float', `read_rational', `read_integer'.
148 * Replaced the functions
149 `void fprint (cl_ostream stream, int x)'
150 `void fprint (cl_ostream stream, unsigned int x)'
151 `void fprint (cl_ostream stream, long x)'
152 `void fprint (cl_ostream stream, unsigned long x)'
154 `void fprintdecimal (cl_ostream stream, int x)'
155 `void fprintdecimal (cl_ostream stream, unsigned int x)'
156 `void fprintdecimal (cl_ostream stream, long x)'
157 `void fprintdecimal (cl_ostream stream, unsigned long x)'
162 * The function `cl_I xgcd (const cl_I& a, const cl_I& b, cl_I* u, cl_I* v)'
163 now normalizes the resulting u and v to be of minimum absolute value.
164 (Suggested by the LiDIA group.)
166 * The conversion from string to cl_F, cl_R, cl_N now chooses the format of
167 the resulting float depending on the number of significant digits present
170 * Speed up the functions
171 `cl_R operator* (const cl_RA&, const cl_LF&)'
172 `cl_R operator* (const cl_LF&, const cl_RA&)'
173 `cl_R operator/ (const cl_RA&, const cl_LF&)'
174 `cl_LF operator/ (const cl_LF&, const cl_RA&)'
176 Implementation changes
177 ----------------------
179 * Fix for `cl_I ash (const cl_I& x, const cl_I& n)' when |n| is between
181 (Reported by Thomas Papanikolaou.)
183 * Fixed a memory leak in the long-float truncate function.
184 (Reported by Keith Briggs.)
186 * Fixed a memory leak in the modular integer exponentiation.
188 * Reduced the stack space consumption. The default stack size (typically
189 8 MB on Unix) should now be sufficient in order to run all CLN programs.
191 * Portability fixes for m68k platforms.
192 (Reported and fixed by Roman Hodek.)
197 * Restructured the sources. The subpackages
200 - base + integer + rational
201 - base + integer + rational + float
202 - base + integer + rational + float + real
203 - base + integer + rational + float + real + complex
204 - base + integer + modinteger
205 are now self-contained.
207 * The configuration script can be called with option `--enable-shared',
208 to build CLN as a shared library.
210 * Updated the LiDIA interface.
213 1998-05-07, version 0.98
214 ========================
216 Implementation changes
217 ----------------------
219 * Portability fixes for 64-bit platforms like DEC Alpha.
220 (Reported by John Cremona and Thomas Papanikolaou.)
229 * Portability fixes for new ISO/ANSI C++ compilers.
231 * Updated the LiDIA interface for LiDIA version 1.3 and fixed a couple of
238 Implementation changes
239 ----------------------
241 * Portability fix for i686 platforms.
242 (Reported by Michael Pfeifer.)
252 <cl_ring.h>, declaring general rings,
253 <cl_GV.h>, <cl_SV.h>, defining vectors,
254 <cl_univpoly.h>, defining univariate polynomial rings.
259 * Speed up the internal computation of ln(2) and ln(10) by about 20%.
261 * Speed up the computation of exp (for 1000 digits or more) by about 10%.
263 Implementation changes
264 ----------------------
266 * Portability fix for 64-bit platforms like DEC Alpha:
267 Fixed a bug: endless loop in `sqrt' for large precision.
268 (Reported by Blake Jones.)
270 * Portability fix for RS/6000 platforms.
279 * Speed up the computation of the elementary transcendental functions:
280 exp, sinh, cosh, asinh, atanh (for 100 digits) by about 30%,
281 log (for 100 digits) by about 25%,
282 sin, cos (for 1000 or 10000 digits) by about 20%,
283 asin, acos, atan (for 1000 digits) by about 25%.
285 Implementation changes
286 ----------------------
288 * Portability fix for SunOS 4 and other BSD platforms.
297 * Fixed a bug: endless recursion in `cl_boolean sqrtp (const cl_RA&, cl_RA*)'.
299 * Fixed a bug: Only the first 85% of the digits returned by `cl_eulerconst'
300 for moderate precision (< 11370 digits) were correct.
301 (Reported by Thomas Papanikolaou.)
303 Implementation changes
304 ----------------------
306 * Fix static initialization order bug. Requires g++ 2.7.0 or newer
307 (g++ 2.8.0 or newer on HPPA).
309 * New method `debug_print()', for pretty printing of CLN objects, intended
310 to be called from the debugger.
319 * New functions `float cl_float_approx (const cl_R&)',
320 `double cl_double_approx (const cl_R&)'
321 for converting a real number to a C `float' or `double'.
326 * Speed up `cl_zeta' using Cohen-Villegas-Zagier convergence acceleration.
328 Implementation changes
329 ----------------------
331 * New function `cl_print', for pretty printing of CLN objects, intended to be
332 called from the debugger.
334 * Portability fixes for 64-bit platforms like DEC Alpha.
344 <cl_string.h>, defining strings,
345 <cl_symbol.h>, defining symbols,
346 <cl_proplist.h>, defining property lists,
347 <cl_condition.h>, defining conditions/exceptions.
349 * New functions `cl_F cl_catalanconst ()',
350 `cl_F cl_catalanconst (const cl_F&)',
351 `cl_F cl_catalanconst (cl_float_format_t)'
352 which return Catalan's constant.
353 (Suggested by Thomas Papanikolaou.)
355 * New functions `cl_F cl_zeta (int s)',
356 `cl_F cl_zeta (int s, const cl_F&)',
357 `cl_F cl_zeta (int s, cl_float_format_t)'
358 which return the Riemann zeta function at an integral point s>1.
360 * New functions `cl_F cl_exp1 ()',
361 `cl_F cl_exp1 (const cl_F&)',
362 `cl_F cl_exp1 (cl_float_format_t)'
363 which return e = exp(1).
365 * New function `cl_I binomial (uintL n, uintL k)' which returns the
366 binomial coefficient (n choose k).
368 * New functions `int cl_I_to_int (const cl_I&)',
369 `unsigned int cl_I_to_uint (const cl_I&)',
370 `long cl_I_to_long (const cl_I&)',
371 `unsigned long cl_I_to_ulong (const cl_I&)'
372 for converting an integer to a C `int' or `long'.
374 * New functions `float cl_float_approx (const cl_I&)',
375 `float cl_float_approx (const cl_RA&)',
376 `double cl_double_approx (const cl_I&)',
377 `double cl_double_approx (const cl_RA&)'
378 for converting a rational number to a C `float' or `double'.
380 Implementation changes
381 ----------------------
383 * Moved the sources to subdirectories.
384 (Suggested by Jörg Arndt.)
389 * New benchmark for better comparing LiDIA, Pari and CLN.
391 * Added an interface to LiDIA, allows using CLN as a replacement of libI.
392 (Suggested by Thomas Papanikolaou.)
394 * Added an ILOG Talk interface, for interactive use of the library.
403 * New include file <cl_modinteger.h>, defining modular integer rings.
405 * New functions `cl_F cl_eulerconst ()', `cl_F cl_eulerconst (const cl_F&)',
406 `cl_F cl_eulerconst (cl_float_format_t)' which return Euler's constant.
411 * Speed up square root of large numbers by use of Newton's algorithm.
413 * Speed up multiplication and division of large long-floats by small integers.
415 * Speed up the computation of pi, e, exp, log, sin, cos, tan, sinh, cosh, tanh,
416 asin, acos, atan, asinh, acosh, atanh.
417 All of them are now computed with asymptotically fast algorithms, of
418 bit complexity O(log(N)^2*M(N)) = O(N^(1+epsilon)) for N digits.
420 * Fixed several bugs in the transcendental functions routines. In particular,
421 the `log' function went into an endless loop.
423 * Fixed a bug: The cl_I -> cl_DF and cl_I -> cl_LF conversions didn't work
424 correctly for negative integers.
433 * New include file <cl_numtheory.h>, containing number theoretic functions,
434 for now only the Jacobi symbol.
439 * Speed up squaring of large numbers by 30%.
441 * Speed up division of large numbers by use of Newton's algorithm. The
442 improvement is noticeable for numbers with at least about 1000 words =
443 32000 bits = 10000 decimal digits.
445 * Speed up the binary-to-decimal conversion of large numbers by use of
446 a divide-and-conquer method. The improvement is noticeable for numbers
447 of at least 1250 words = 40000 bits = 12000 decimal digits.
449 * Fixed a bug: The functions `cl_boolean minusp (const type&)' and
450 `cl_boolean zerop (const type&)' were missing for type=cl_I and type=cl_RA.
452 * Fixed a bug: The function `cl_boolean logtest (const cl_I&, const cl_I&)'
453 didn't work correctly if both arguments were bignums.
455 * Fixed a bug: The function `cl_I mod (const cl_I&, const cl_I&)', when called
456 as `mod(-n*m,m)' (n>0), returned m instead of 0.
458 * Fixed a bug: The function `uintL power2p (const cl_I&)' returned wrong
459 values for arguments x = 2^n, n >= 29.
461 Implementation changes
462 ----------------------
464 * Speed up CLN by using the low-level addition/subtraction, shift and
465 multiplication loops from GNU gmp-2.0.2.
467 * Fixed a severe reference counting bug in the constructors
468 `type::type (const long)' and `type::type (const unsigned long)'
469 for type=cl_number, type=cl_N, type=cl_RA, type=cl_I that most often
472 * CLN's word sequences can be chosen to be stored big-endian in memory
473 (like it was up to now) or little-endian (needed for interfacing to GMP).
482 * New include file <cl_timing.h>, contains timing facilities.
487 * Speed up the multiplication of large numbers by use of an FFT based
488 algorithm (Schönhage-Strassen). The improvement is noticeable when both
489 factors have at least 1250 words = 40000 bits = 12000 decimal digits.
491 * Speed up the functions `cl_I gcd (const cl_I&, const cl_I&)' and
492 `cl_I xgcd (const cl_I&, const cl_I&, cl_I*, cl_I*)' by a factor
493 of 1.5. Simple trick.
495 * Speed up the function `cl_boolean sqrtp (const cl_I&, cl_I*)' using a
496 trick from Henri Cohen.
498 * Fixed an endless loop in the function `uint32 gcd (uint32, uint32)'
499 which occurred when one of the arguments is zero.
501 * Fixed an endless loop in the function `cl_I factorial (uintL)'.
503 Implementation changes
504 ----------------------
506 * <cl_io.h> now uses iostreams by default instead of stdio FILE pointers.
507 (Reported by Tito Flagella.)
509 * Fixed linking error when compiling without -O.
510 (Reported by Tito Flagella.)
512 * Fixed syntax error in <cl_input.h>.
513 (Reported by Tito Flagella.)
515 * Fixed syntax errors in src/cl_DS.h and src/cl_FF_plus.h.
516 (Reported by Marcus Daniels.)
518 * More portability fixes.
520 * Configuration now supports the --srcdir option.
521 (Reported by Marcus Daniels.)
527 * Compiles well on Sparc, using gcc.
529 * Mail to beta testers.
535 * Compiles well on Linux, using gcc.