1 2002-01-04, version 1.1.4
2 =========================
7 * Fix compilation issues with GMP-4.0 and on IBM s390.
9 * Updated infrastructure to autoconf-2.52.
12 2001-11-05, version 1.1.3
13 =========================
15 Implementation changes
16 ----------------------
18 * Fix a computation error of sin and sinh for small arguments.
19 (Reported by Christopher Kennedy.)
21 * Module ordering works with gcc-3.0 -fuse-cxa-atexit now.
23 * -DNO_ASM works properly on Sparc.
26 2001-07-25, version 1.1.2
28 Implementation changes
29 ----------------------
31 * Minor cleanups and portability issues.
32 (No need to upgrade if you have been happy so far.)
35 2001-05-31, version 1.1.1
36 =========================
38 Implementation changes
39 ----------------------
41 * Minor cleanups for GCC 3.0 and libstdc++-v3.
43 * Fixes several little bugs and portability issues.
46 2000-12-14, version 1.1.0
47 =========================
52 * ISO/IEC 14882 fashion adjustments:
53 Put everything into namespace cln. All fundamental data types still
54 carry their old names. Other non-macro identifiers are now written as
55 cln::foo instead of cl_foo, except where the cl_ comes from a data type
56 (as in cl_I_to_int()). Headers are installed into a separate
57 directory, #include <cln/foo.h> instead of <cl_foo.h>. Applications
58 must be manually ported to the new style. We apologize for the
61 Implementation changes
62 ----------------------
64 * Removed internal copy of mpn subdirectory tree from libgmp-2.0.2.
65 Configuring with --with-gmp now relies on an installed libgmp library
66 version 3 or above. We recommend libgmp-3.1 or newer.
68 * Adjusted several break-even points to make better use of a faster libgmp and
69 better match present-day CPUs.
71 * Fix several errors for architectures with 64-bit wide words.
72 (Others than Alpha, which worked already.)
74 * Fix compilation errors with current GCC-3.0 snapshots.
79 * Added package tools: script cln-config and autoconf macro AC_PATH_CLN
83 2000-01-13, version 1.0.3
84 =========================
90 `cl_I doublefactorial (uintL n)'.
92 Implementation changes
93 ----------------------
95 * Fix several little configuration errors.
97 * Fix some compilation errors with gcc-2.96 prereleases.
100 1999-09-07, version 1.0.2
101 =========================
107 `cl_I numerator (const cl_RA&)'
108 `cl_I denominator (const cl_RA&)'.
109 Suggested by Richard Kreckel and Sylvain Pion.
111 * New function `cl_equal_hashcode' for the classes cl_N, cl_R, cl_RA, cl_I,
112 cl_F, cl_SF, cl_FF, cl_DF, cl_LF.
113 Suggested by Richard Kreckel.
115 Implementation changes
116 ----------------------
118 * Fix an endless loop when either of the functions
119 `cl_boolean rootp (const cl_RA&, uintL, cl_RA*)',
120 `cl_boolean rootp (const cl_RA&, const cl_I&, cl_RA*)'
121 was called with an integer argument. Thanks to Richard Kreckel.
123 * Fix a bug in the addition and subtraction of rational numbers which could
124 cause unnormalized rational numbers like 3/1 to be created.
127 1999-06-09, version 1.0.1
128 =========================
133 * Speed up the functions `logand', `lognand', `logandc2'.
135 Implementation changes
136 ----------------------
138 * Fix compilation errors with gcc-2.7.2, egcs-1.1.2 and gcc-2.95.
140 * Fix compilation errors on HPPA, MIPS, some versions of DEC Alpha, OpenBSD,
141 and SPARC with LiDIA.
143 * Fix compilation errors on Win32. Thanks to Mumit Khan.
146 1999-01-12, version 1.0
147 =======================
152 * New include files, containing I/O functions:
156 <cl_sfloat_io.h> <cl_ffloat_io.h> <cl_dfloat_io.h> <cl_lfloat_io.h>
159 The file <cl_io.h> now contains only I/O functions for characters and
162 * To access the base ring of a univariate polynomial ring, now use
163 `R->basering()' instead of `R->basering'.
165 * Implement `plusp', `max', `min' for the classes cl_F, cl_SF, cl_FF, cl_DF,
168 * Implement `abs' for the class cl_RA.
170 * Replaced `read_number' with specialized functions `read_complex',
171 `read_real', `read_float', `read_rational', `read_integer'.
173 * Replaced the functions
174 `void fprint (cl_ostream stream, int x)'
175 `void fprint (cl_ostream stream, unsigned int x)'
176 `void fprint (cl_ostream stream, long x)'
177 `void fprint (cl_ostream stream, unsigned long x)'
179 `void fprintdecimal (cl_ostream stream, int x)'
180 `void fprintdecimal (cl_ostream stream, unsigned int x)'
181 `void fprintdecimal (cl_ostream stream, long x)'
182 `void fprintdecimal (cl_ostream stream, unsigned long x)'
187 * The function `cl_I xgcd (const cl_I& a, const cl_I& b, cl_I* u, cl_I* v)'
188 now normalizes the resulting u and v to be of minimum absolute value.
189 (Suggested by the LiDIA group.)
191 * The conversion from string to cl_F, cl_R, cl_N now chooses the format of
192 the resulting float depending on the number of significant digits present
195 * Speed up the functions
196 `cl_R operator* (const cl_RA&, const cl_LF&)'
197 `cl_R operator* (const cl_LF&, const cl_RA&)'
198 `cl_R operator/ (const cl_RA&, const cl_LF&)'
199 `cl_LF operator/ (const cl_LF&, const cl_RA&)'
201 Implementation changes
202 ----------------------
204 * Fix for `cl_I ash (const cl_I& x, const cl_I& n)' when |n| is between
206 (Reported by Thomas Papanikolaou.)
208 * Fixed a memory leak in the long-float truncate function.
209 (Reported by Keith Briggs.)
211 * Fixed a memory leak in the modular integer exponentiation.
213 * Reduced the stack space consumption. The default stack size (typically
214 8 MB on Unix) should now be sufficient in order to run all CLN programs.
216 * Portability fixes for m68k platforms.
217 (Reported and fixed by Roman Hodek.)
222 * Restructured the sources. The subpackages
225 - base + integer + rational
226 - base + integer + rational + float
227 - base + integer + rational + float + real
228 - base + integer + rational + float + real + complex
229 - base + integer + modinteger
230 are now self-contained.
232 * The configuration script can be called with option `--enable-shared',
233 to build CLN as a shared library.
235 * Updated the LiDIA interface.
238 1998-05-07, version 0.98
239 ========================
241 Implementation changes
242 ----------------------
244 * Portability fixes for 64-bit platforms like DEC Alpha.
245 (Reported by John Cremona and Thomas Papanikolaou.)
254 * Portability fixes for new ISO/ANSI C++ compilers.
256 * Updated the LiDIA interface for LiDIA version 1.3 and fixed a couple of
263 Implementation changes
264 ----------------------
266 * Portability fix for i686 platforms.
267 (Reported by Michael Pfeifer.)
277 <cl_ring.h>, declaring general rings,
278 <cl_GV.h>, <cl_SV.h>, defining vectors,
279 <cl_univpoly.h>, defining univariate polynomial rings.
284 * Speed up the internal computation of ln(2) and ln(10) by about 20%.
286 * Speed up the computation of exp (for 1000 digits or more) by about 10%.
288 Implementation changes
289 ----------------------
291 * Portability fix for 64-bit platforms like DEC Alpha:
292 Fixed a bug: endless loop in `sqrt' for large precision.
293 (Reported by Blake Jones.)
295 * Portability fix for RS/6000 platforms.
304 * Speed up the computation of the elementary transcendental functions:
305 exp, sinh, cosh, asinh, atanh (for 100 digits) by about 30%,
306 log (for 100 digits) by about 25%,
307 sin, cos (for 1000 or 10000 digits) by about 20%,
308 asin, acos, atan (for 1000 digits) by about 25%.
310 Implementation changes
311 ----------------------
313 * Portability fix for SunOS 4 and other BSD platforms.
322 * Fixed a bug: endless recursion in `cl_boolean sqrtp (const cl_RA&, cl_RA*)'.
324 * Fixed a bug: Only the first 85% of the digits returned by `cl_eulerconst'
325 for moderate precision (< 11370 digits) were correct.
326 (Reported by Thomas Papanikolaou.)
328 Implementation changes
329 ----------------------
331 * Fix static initialization order bug. Requires g++ 2.7.0 or newer
332 (g++ 2.8.0 or newer on HPPA).
334 * New method `debug_print()', for pretty printing of CLN objects, intended
335 to be called from the debugger.
344 * New functions `float cl_float_approx (const cl_R&)',
345 `double cl_double_approx (const cl_R&)'
346 for converting a real number to a C `float' or `double'.
351 * Speed up `cl_zeta' using Cohen-Villegas-Zagier convergence acceleration.
353 Implementation changes
354 ----------------------
356 * New function `cl_print', for pretty printing of CLN objects, intended to be
357 called from the debugger.
359 * Portability fixes for 64-bit platforms like DEC Alpha.
369 <cl_string.h>, defining strings,
370 <cl_symbol.h>, defining symbols,
371 <cl_proplist.h>, defining property lists,
372 <cl_condition.h>, defining conditions/exceptions.
374 * New functions `cl_F cl_catalanconst ()',
375 `cl_F cl_catalanconst (const cl_F&)',
376 `cl_F cl_catalanconst (cl_float_format_t)'
377 which return Catalan's constant.
378 (Suggested by Thomas Papanikolaou.)
380 * New functions `cl_F cl_zeta (int s)',
381 `cl_F cl_zeta (int s, const cl_F&)',
382 `cl_F cl_zeta (int s, cl_float_format_t)'
383 which return the Riemann zeta function at an integral point s>1.
385 * New functions `cl_F cl_exp1 ()',
386 `cl_F cl_exp1 (const cl_F&)',
387 `cl_F cl_exp1 (cl_float_format_t)'
388 which return e = exp(1).
390 * New function `cl_I binomial (uintL n, uintL k)' which returns the
391 binomial coefficient (n choose k).
393 * New functions `int cl_I_to_int (const cl_I&)',
394 `unsigned int cl_I_to_uint (const cl_I&)',
395 `long cl_I_to_long (const cl_I&)',
396 `unsigned long cl_I_to_ulong (const cl_I&)'
397 for converting an integer to a C `int' or `long'.
399 * New functions `float cl_float_approx (const cl_I&)',
400 `float cl_float_approx (const cl_RA&)',
401 `double cl_double_approx (const cl_I&)',
402 `double cl_double_approx (const cl_RA&)'
403 for converting a rational number to a C `float' or `double'.
405 Implementation changes
406 ----------------------
408 * Moved the sources to subdirectories.
409 (Suggested by Jörg Arndt.)
414 * New benchmark for better comparing LiDIA, Pari and CLN.
416 * Added an interface to LiDIA, allows using CLN as a replacement of libI.
417 (Suggested by Thomas Papanikolaou.)
419 * Added an ILOG Talk interface, for interactive use of the library.
428 * New include file <cl_modinteger.h>, defining modular integer rings.
430 * New functions `cl_F cl_eulerconst ()', `cl_F cl_eulerconst (const cl_F&)',
431 `cl_F cl_eulerconst (cl_float_format_t)' which return Euler's constant.
436 * Speed up square root of large numbers by use of Newton's algorithm.
438 * Speed up multiplication and division of large long-floats by small integers.
440 * Speed up the computation of pi, e, exp, log, sin, cos, tan, sinh, cosh, tanh,
441 asin, acos, atan, asinh, acosh, atanh.
442 All of them are now computed with asymptotically fast algorithms, of
443 bit complexity O(log(N)^2*M(N)) = O(N^(1+epsilon)) for N digits.
445 * Fixed several bugs in the transcendental functions routines. In particular,
446 the `log' function went into an endless loop.
448 * Fixed a bug: The cl_I -> cl_DF and cl_I -> cl_LF conversions didn't work
449 correctly for negative integers.
458 * New include file <cl_numtheory.h>, containing number theoretic functions,
459 for now only the Jacobi symbol.
464 * Speed up squaring of large numbers by 30%.
466 * Speed up division of large numbers by use of Newton's algorithm. The
467 improvement is noticeable for numbers with at least about 1000 words =
468 32000 bits = 10000 decimal digits.
470 * Speed up the binary-to-decimal conversion of large numbers by use of
471 a divide-and-conquer method. The improvement is noticeable for numbers
472 of at least 1250 words = 40000 bits = 12000 decimal digits.
474 * Fixed a bug: The functions `cl_boolean minusp (const type&)' and
475 `cl_boolean zerop (const type&)' were missing for type=cl_I and type=cl_RA.
477 * Fixed a bug: The function `cl_boolean logtest (const cl_I&, const cl_I&)'
478 didn't work correctly if both arguments were bignums.
480 * Fixed a bug: The function `cl_I mod (const cl_I&, const cl_I&)', when called
481 as `mod(-n*m,m)' (n>0), returned m instead of 0.
483 * Fixed a bug: The function `uintL power2p (const cl_I&)' returned wrong
484 values for arguments x = 2^n, n >= 29.
486 Implementation changes
487 ----------------------
489 * Speed up CLN by using the low-level addition/subtraction, shift and
490 multiplication loops from GNU gmp-2.0.2.
492 * Fixed a severe reference counting bug in the constructors
493 `type::type (const long)' and `type::type (const unsigned long)'
494 for type=cl_number, type=cl_N, type=cl_RA, type=cl_I that most often
497 * CLN's word sequences can be chosen to be stored big-endian in memory
498 (like it was up to now) or little-endian (needed for interfacing to GMP).
507 * New include file <cl_timing.h>, contains timing facilities.
512 * Speed up the multiplication of large numbers by use of an FFT based
513 algorithm (Schönhage-Strassen). The improvement is noticeable when both
514 factors have at least 1250 words = 40000 bits = 12000 decimal digits.
516 * Speed up the functions `cl_I gcd (const cl_I&, const cl_I&)' and
517 `cl_I xgcd (const cl_I&, const cl_I&, cl_I*, cl_I*)' by a factor
518 of 1.5. Simple trick.
520 * Speed up the function `cl_boolean sqrtp (const cl_I&, cl_I*)' using a
521 trick from Henri Cohen.
523 * Fixed an endless loop in the function `uint32 gcd (uint32, uint32)'
524 which occurred when one of the arguments is zero.
526 * Fixed an endless loop in the function `cl_I factorial (uintL)'.
528 Implementation changes
529 ----------------------
531 * <cl_io.h> now uses iostreams by default instead of stdio FILE pointers.
532 (Reported by Tito Flagella.)
534 * Fixed linking error when compiling without -O.
535 (Reported by Tito Flagella.)
537 * Fixed syntax error in <cl_input.h>.
538 (Reported by Tito Flagella.)
540 * Fixed syntax errors in src/cl_DS.h and src/cl_FF_plus.h.
541 (Reported by Marcus Daniels.)
543 * More portability fixes.
545 * Configuration now supports the --srcdir option.
546 (Reported by Marcus Daniels.)
552 * Compiles well on Sparc, using gcc.
554 * Mail to beta testers.
560 * Compiles well on Linux, using gcc.