1 2001-05-31, version 1.1.1
2 =========================
7 * Minor cleanups for GCC 3.0 and libstdc++-v3.
9 * Fixes several little bugs and portability issues.
12 2000-12-14, version 1.1.0
13 =========================
18 * ISO/IEC 14882 fashion adjustments:
19 Put everything into namespace cln. All fundamental data types still
20 carry their old names. Other non-macro identifiers are now written as
21 cln::foo instead of cl_foo, except where the cl_ comes from a data type
22 (as in cl_I_to_int()). Headers are installed into a separate
23 directory, #include <cln/foo.h> instead of <cl_foo.h>. Applications
24 must be manually ported to the new style. We apologize for the
27 Implementation changes
28 ----------------------
30 * Removed internal copy of mpn subdirectory tree from libgmp-2.0.2.
31 Configuring with --with-gmp now relies on an installed libgmp library
32 version 3 or above. We recommend libgmp-3.1 or newer.
34 * Adjusted several break-even points to make better use of a faster libgmp and
35 better match present-day CPUs.
37 * Fix several errors for architectures with 64-bit wide words.
38 (Others than Alpha, which worked already.)
40 * Fix compilation errors with current GCC-3.0 snapshots.
45 * Added package tools: script cln-config and autoconf macro AC_PATH_CLN
49 2000-01-13, version 1.0.3
50 =========================
56 `cl_I doublefactorial (uintL n)'.
58 Implementation changes
59 ----------------------
61 * Fix several little configuration errors.
63 * Fix some compilation errors with gcc-2.96 prereleases.
66 1999-09-07, version 1.0.2
67 =========================
73 `cl_I numerator (const cl_RA&)'
74 `cl_I denominator (const cl_RA&)'.
75 Suggested by Richard Kreckel and Sylvain Pion.
77 * New function `cl_equal_hashcode' for the classes cl_N, cl_R, cl_RA, cl_I,
78 cl_F, cl_SF, cl_FF, cl_DF, cl_LF.
79 Suggested by Richard Kreckel.
81 Implementation changes
82 ----------------------
84 * Fix an endless loop when either of the functions
85 `cl_boolean rootp (const cl_RA&, uintL, cl_RA*)',
86 `cl_boolean rootp (const cl_RA&, const cl_I&, cl_RA*)'
87 was called with an integer argument. Thanks to Richard Kreckel.
89 * Fix a bug in the addition and subtraction of rational numbers which could
90 cause unnormalized rational numbers like 3/1 to be created.
93 1999-06-09, version 1.0.1
94 =========================
99 * Speed up the functions `logand', `lognand', `logandc2'.
101 Implementation changes
102 ----------------------
104 * Fix compilation errors with gcc-2.7.2, egcs-1.1.2 and gcc-2.95.
106 * Fix compilation errors on HPPA, MIPS, some versions of DEC Alpha, OpenBSD,
107 and SPARC with LiDIA.
109 * Fix compilation errors on Win32. Thanks to Mumit Khan.
112 1999-01-12, version 1.0
113 =======================
118 * New include files, containing I/O functions:
122 <cl_sfloat_io.h> <cl_ffloat_io.h> <cl_dfloat_io.h> <cl_lfloat_io.h>
125 The file <cl_io.h> now contains only I/O functions for characters and
128 * To access the base ring of a univariate polynomial ring, now use
129 `R->basering()' instead of `R->basering'.
131 * Implement `plusp', `max', `min' for the classes cl_F, cl_SF, cl_FF, cl_DF,
134 * Implement `abs' for the class cl_RA.
136 * Replaced `read_number' with specialized functions `read_complex',
137 `read_real', `read_float', `read_rational', `read_integer'.
139 * Replaced the functions
140 `void fprint (cl_ostream stream, int x)'
141 `void fprint (cl_ostream stream, unsigned int x)'
142 `void fprint (cl_ostream stream, long x)'
143 `void fprint (cl_ostream stream, unsigned long x)'
145 `void fprintdecimal (cl_ostream stream, int x)'
146 `void fprintdecimal (cl_ostream stream, unsigned int x)'
147 `void fprintdecimal (cl_ostream stream, long x)'
148 `void fprintdecimal (cl_ostream stream, unsigned long x)'
153 * The function `cl_I xgcd (const cl_I& a, const cl_I& b, cl_I* u, cl_I* v)'
154 now normalizes the resulting u and v to be of minimum absolute value.
155 (Suggested by the LiDIA group.)
157 * The conversion from string to cl_F, cl_R, cl_N now chooses the format of
158 the resulting float depending on the number of significant digits present
161 * Speed up the functions
162 `cl_R operator* (const cl_RA&, const cl_LF&)'
163 `cl_R operator* (const cl_LF&, const cl_RA&)'
164 `cl_R operator/ (const cl_RA&, const cl_LF&)'
165 `cl_LF operator/ (const cl_LF&, const cl_RA&)'
167 Implementation changes
168 ----------------------
170 * Fix for `cl_I ash (const cl_I& x, const cl_I& n)' when |n| is between
172 (Reported by Thomas Papanikolaou.)
174 * Fixed a memory leak in the long-float truncate function.
175 (Reported by Keith Briggs.)
177 * Fixed a memory leak in the modular integer exponentiation.
179 * Reduced the stack space consumption. The default stack size (typically
180 8 MB on Unix) should now be sufficient in order to run all CLN programs.
182 * Portability fixes for m68k platforms.
183 (Reported and fixed by Roman Hodek.)
188 * Restructured the sources. The subpackages
191 - base + integer + rational
192 - base + integer + rational + float
193 - base + integer + rational + float + real
194 - base + integer + rational + float + real + complex
195 - base + integer + modinteger
196 are now self-contained.
198 * The configuration script can be called with option `--enable-shared',
199 to build CLN as a shared library.
201 * Updated the LiDIA interface.
204 1998-05-07, version 0.98
205 ========================
207 Implementation changes
208 ----------------------
210 * Portability fixes for 64-bit platforms like DEC Alpha.
211 (Reported by John Cremona and Thomas Papanikolaou.)
220 * Portability fixes for new ISO/ANSI C++ compilers.
222 * Updated the LiDIA interface for LiDIA version 1.3 and fixed a couple of
229 Implementation changes
230 ----------------------
232 * Portability fix for i686 platforms.
233 (Reported by Michael Pfeifer.)
243 <cl_ring.h>, declaring general rings,
244 <cl_GV.h>, <cl_SV.h>, defining vectors,
245 <cl_univpoly.h>, defining univariate polynomial rings.
250 * Speed up the internal computation of ln(2) and ln(10) by about 20%.
252 * Speed up the computation of exp (for 1000 digits or more) by about 10%.
254 Implementation changes
255 ----------------------
257 * Portability fix for 64-bit platforms like DEC Alpha:
258 Fixed a bug: endless loop in `sqrt' for large precision.
259 (Reported by Blake Jones.)
261 * Portability fix for RS/6000 platforms.
270 * Speed up the computation of the elementary transcendental functions:
271 exp, sinh, cosh, asinh, atanh (for 100 digits) by about 30%,
272 log (for 100 digits) by about 25%,
273 sin, cos (for 1000 or 10000 digits) by about 20%,
274 asin, acos, atan (for 1000 digits) by about 25%.
276 Implementation changes
277 ----------------------
279 * Portability fix for SunOS 4 and other BSD platforms.
288 * Fixed a bug: endless recursion in `cl_boolean sqrtp (const cl_RA&, cl_RA*)'.
290 * Fixed a bug: Only the first 85% of the digits returned by `cl_eulerconst'
291 for moderate precision (< 11370 digits) were correct.
292 (Reported by Thomas Papanikolaou.)
294 Implementation changes
295 ----------------------
297 * Fix static initialization order bug. Requires g++ 2.7.0 or newer
298 (g++ 2.8.0 or newer on HPPA).
300 * New method `debug_print()', for pretty printing of CLN objects, intended
301 to be called from the debugger.
310 * New functions `float cl_float_approx (const cl_R&)',
311 `double cl_double_approx (const cl_R&)'
312 for converting a real number to a C `float' or `double'.
317 * Speed up `cl_zeta' using Cohen-Villegas-Zagier convergence acceleration.
319 Implementation changes
320 ----------------------
322 * New function `cl_print', for pretty printing of CLN objects, intended to be
323 called from the debugger.
325 * Portability fixes for 64-bit platforms like DEC Alpha.
335 <cl_string.h>, defining strings,
336 <cl_symbol.h>, defining symbols,
337 <cl_proplist.h>, defining property lists,
338 <cl_condition.h>, defining conditions/exceptions.
340 * New functions `cl_F cl_catalanconst ()',
341 `cl_F cl_catalanconst (const cl_F&)',
342 `cl_F cl_catalanconst (cl_float_format_t)'
343 which return Catalan's constant.
344 (Suggested by Thomas Papanikolaou.)
346 * New functions `cl_F cl_zeta (int s)',
347 `cl_F cl_zeta (int s, const cl_F&)',
348 `cl_F cl_zeta (int s, cl_float_format_t)'
349 which return the Riemann zeta function at an integral point s>1.
351 * New functions `cl_F cl_exp1 ()',
352 `cl_F cl_exp1 (const cl_F&)',
353 `cl_F cl_exp1 (cl_float_format_t)'
354 which return e = exp(1).
356 * New function `cl_I binomial (uintL n, uintL k)' which returns the
357 binomial coefficient (n choose k).
359 * New functions `int cl_I_to_int (const cl_I&)',
360 `unsigned int cl_I_to_uint (const cl_I&)',
361 `long cl_I_to_long (const cl_I&)',
362 `unsigned long cl_I_to_ulong (const cl_I&)'
363 for converting an integer to a C `int' or `long'.
365 * New functions `float cl_float_approx (const cl_I&)',
366 `float cl_float_approx (const cl_RA&)',
367 `double cl_double_approx (const cl_I&)',
368 `double cl_double_approx (const cl_RA&)'
369 for converting a rational number to a C `float' or `double'.
371 Implementation changes
372 ----------------------
374 * Moved the sources to subdirectories.
375 (Suggested by Jörg Arndt.)
380 * New benchmark for better comparing LiDIA, Pari and CLN.
382 * Added an interface to LiDIA, allows using CLN as a replacement of libI.
383 (Suggested by Thomas Papanikolaou.)
385 * Added an ILOG Talk interface, for interactive use of the library.
394 * New include file <cl_modinteger.h>, defining modular integer rings.
396 * New functions `cl_F cl_eulerconst ()', `cl_F cl_eulerconst (const cl_F&)',
397 `cl_F cl_eulerconst (cl_float_format_t)' which return Euler's constant.
402 * Speed up square root of large numbers by use of Newton's algorithm.
404 * Speed up multiplication and division of large long-floats by small integers.
406 * Speed up the computation of pi, e, exp, log, sin, cos, tan, sinh, cosh, tanh,
407 asin, acos, atan, asinh, acosh, atanh.
408 All of them are now computed with asymptotically fast algorithms, of
409 bit complexity O(log(N)^2*M(N)) = O(N^(1+epsilon)) for N digits.
411 * Fixed several bugs in the transcendental functions routines. In particular,
412 the `log' function went into an endless loop.
414 * Fixed a bug: The cl_I -> cl_DF and cl_I -> cl_LF conversions didn't work
415 correctly for negative integers.
424 * New include file <cl_numtheory.h>, containing number theoretic functions,
425 for now only the Jacobi symbol.
430 * Speed up squaring of large numbers by 30%.
432 * Speed up division of large numbers by use of Newton's algorithm. The
433 improvement is noticeable for numbers with at least about 1000 words =
434 32000 bits = 10000 decimal digits.
436 * Speed up the binary-to-decimal conversion of large numbers by use of
437 a divide-and-conquer method. The improvement is noticeable for numbers
438 of at least 1250 words = 40000 bits = 12000 decimal digits.
440 * Fixed a bug: The functions `cl_boolean minusp (const type&)' and
441 `cl_boolean zerop (const type&)' were missing for type=cl_I and type=cl_RA.
443 * Fixed a bug: The function `cl_boolean logtest (const cl_I&, const cl_I&)'
444 didn't work correctly if both arguments were bignums.
446 * Fixed a bug: The function `cl_I mod (const cl_I&, const cl_I&)', when called
447 as `mod(-n*m,m)' (n>0), returned m instead of 0.
449 * Fixed a bug: The function `uintL power2p (const cl_I&)' returned wrong
450 values for arguments x = 2^n, n >= 29.
452 Implementation changes
453 ----------------------
455 * Speed up CLN by using the low-level addition/subtraction, shift and
456 multiplication loops from GNU gmp-2.0.2.
458 * Fixed a severe reference counting bug in the constructors
459 `type::type (const long)' and `type::type (const unsigned long)'
460 for type=cl_number, type=cl_N, type=cl_RA, type=cl_I that most often
463 * CLN's word sequences can be chosen to be stored big-endian in memory
464 (like it was up to now) or little-endian (needed for interfacing to GMP).
473 * New include file <cl_timing.h>, contains timing facilities.
478 * Speed up the multiplication of large numbers by use of an FFT based
479 algorithm (Schönhage-Strassen). The improvement is noticeable when both
480 factors have at least 1250 words = 40000 bits = 12000 decimal digits.
482 * Speed up the functions `cl_I gcd (const cl_I&, const cl_I&)' and
483 `cl_I xgcd (const cl_I&, const cl_I&, cl_I*, cl_I*)' by a factor
484 of 1.5. Simple trick.
486 * Speed up the function `cl_boolean sqrtp (const cl_I&, cl_I*)' using a
487 trick from Henri Cohen.
489 * Fixed an endless loop in the function `uint32 gcd (uint32, uint32)'
490 which occurred when one of the arguments is zero.
492 * Fixed an endless loop in the function `cl_I factorial (uintL)'.
494 Implementation changes
495 ----------------------
497 * <cl_io.h> now uses iostreams by default instead of stdio FILE pointers.
498 (Reported by Tito Flagella.)
500 * Fixed linking error when compiling without -O.
501 (Reported by Tito Flagella.)
503 * Fixed syntax error in <cl_input.h>.
504 (Reported by Tito Flagella.)
506 * Fixed syntax errors in src/cl_DS.h and src/cl_FF_plus.h.
507 (Reported by Marcus Daniels.)
509 * More portability fixes.
511 * Configuration now supports the --srcdir option.
512 (Reported by Marcus Daniels.)
518 * Compiles well on Sparc, using gcc.
520 * Mail to beta testers.
526 * Compiles well on Linux, using gcc.