1 This file records noteworthy changes.
4 * Added a structure<T> template class for the easy creation of user-defined
6 * Added support for (acyclic) visitors, to allow cleaner implementations of
8 * Added a const_iterator class that can be used instead of op()/nops().
9 * Completely revamped the implementation of expression output. It is now
10 possible to add new output formats, to change the behavior of predefined
11 formats at run-time, and to have different output styles for algebraic
13 * Symbols can be made non-commutative.
14 * Added a method ex::conjugate() and a function conjugate() for complex
15 conjugation. Symbols can be declared as real or complex-valued.
16 * Improved the speed of subs(), normal(), to_rational() and to_polynomial()
17 by the use of maps instead of lists. The old forms
18 subs(const lst & ls, const lst & lr, unsigned options)
19 to_rational/to_polynomial(lst & repl)
20 are still available for compatibility, but using the new forms
21 subs(const exmap & m, unsigned options)
22 to_rational/to_polynomial(exmap & repl)
23 is more efficient, especially when the number of replacements is large.
24 * quo(), rem(), prem(), sprem(), decomp_rational(), unit(), content(),
25 primpart() and matrix::charpoly() now take a "const ex &" instead of a
27 * Redundant expressions (two ex pointing to different objects are found to be
28 equal in compare()) are now actively deleted/fused to conserve memory and
29 speed up subsequent comparisons. This behavior can be suppressed on a
30 per-object level with status_flags::not_shareable. Lists and matrices are
31 not shareable by default.
32 * Lots of internal streamlining and optimizations.
33 * Caveats for class implementors:
34 - basic::copy() and basic::destroy() are gone; classes derived from
35 basic can use the defaults for the assignment operator and copy
37 - basic::subs(), basic::normal(), basic::to_rational() and
38 basic::to_polynomial() take 'exmap' objects instead of lists.
39 - basic::subs() now descends into subexpressions (if accessible via
40 nops()/op()/let_op()). If you have a custom implementation of subs()
41 that calls basic::subs() after substituting subexpressions, this needs
42 to be changed to a call to subs_one_level().
43 - lst::thislst() and exprseq::thisexprseq() renamed to thiscontainer().
44 - thiscontainer() and associated constructors now take a std::auto_ptr.
45 - Overloading basic::print() is now deprecated. You should use
46 print_func<>() class options instead.
49 * Fixed a bug in canonicalize_clifford().
50 * Series expansion now works predictably. All terms with the exponent of the
51 expansion variable smaller than the given order are calculated exactly. If
52 the series is not terminating, the Order function is (at least) of the given
55 1.1.6 (22 January 2004)
56 * Added a function option "dummy()" which means "no options". This simplifies
57 the implementation of symbolic functions which are not to be further
59 * Removed a bug in the numerical evaluation of Li() that caused the system
60 to hang for certain parameter combinations.
61 * Fixed a bug in the calculation of hash values for indices that could lead
62 to wrong results or bogus error messages from simplify_indexed().
63 * Fixed a bug in the evaluation of harmonic polylogarithms for complex
64 arguments with positive imaginary part.
66 1.1.5 (5 November 2003)
67 * Harmonic polylogarithms now numerically evaluate for arbitrary arguments
68 (parameter must still be positive integers).
69 * The zeta function now can also be given a lst as a parameter in which case
70 it becomes a multiple zeta value. The use of mZeta is deprecated.
71 * The order of parameters for the multiple polylogarithm has been corrected.
72 * Documentation for the nested sums functions zeta, harmonic polylog, multiple
73 polylog, etc. has been added.
75 1.1.4 (17 October 2003)
76 * Lists and matrices can now be initialized from comma-separated lists of
77 expressions, like this:
87 This is both faster and produces much smaller code than the old constructors
88 lst(ex, ex, ...) and matrix(unsigned, unsigned, lst), especially in the case
89 of matrices, and is now the recommended way to create these objects.
90 * The function mZeta now evaluates much faster for arbitrary parameters. The
91 harmonic and multiple polylogarithms evaluate considerably faster and check
92 for convergence. The order of parameters for the harmonic polylogarithm
95 1.1.3 (22 August 2003)
96 * Added new symbolic functions for better integration with nestedsums:
97 (multiple) polylogarithm Li(), Nielsen's generalized polylogarithm S(),
98 harmonic polylogarithm H(), and multiple zeta value mZeta().
99 * New exhashmap<T> template intended as a drop-in replacement for
100 std::map<ex, T, ex_is_less> using GiNaC's hashing algorithms.
102 1.1.2 (11 August 2003)
103 * Fixed a bug in the unarchiving of sums and products: terms were not
104 reordered in a canonical way.
105 * Fixed a bug in normal()/numer_denom(): denominator was not made unit
106 normal if it was a simple number.
107 * Improved the speed of subs() in some cases.
110 * lst (and exprseq) provide iterators for read-only element access. For
111 sequential access this is one order faster than using op().
112 * Implemented relational::subs() (this was done in 1.0.9 but inadvertently
113 omitted from the 1.1 branch).
114 * pole_error and do_taylor are available to library users.
115 * Added on-line help and Tab-completion for print(), iprint(), print_latex()
116 and print_csrc() in ginsh.
119 * Removed deprecated macros is_ex_a, is_ex_exactly_a and friends for good.
120 * The scalar_products mechanism allows the specification of an index dimension.
121 * Removed dirac_gamma6/7().
122 * Added ex::to_polynomial().
123 * subs() accepts an optional "options" argument. The option
124 subs_option::subs_algebraic enables "smart" substitutions in products and
126 * Added stream manipulators "dflt", "latex", "python", "python_repr", "tree",
127 "csrc", "csrc_float", "csrc_double", "csrc_cl_N", "index_dimensions" and
128 "no_index_dimensions" to control the output format. Calling basic::print()
129 directly is now deprecated.
130 * Made the hashing more simple and efficient.
131 * Caveats for class implementors:
132 - basic::subs(): third argument changed from "bool" to "unsigned"
133 - unarchiving constructor and basic::unarchive(): "const" removed from
135 - basic::let_op() should only be implemented if write access to
136 subexpressions is desired
137 - simplify_ncmul() renamed to eval_ncmul()
138 - simplified_ncmul() renamed to hold_ncmul()
139 - nonsimplified_ncmul() renamed to reeval_ncmul()
141 1.0.14 (1 March 2003)
142 * Improved the C-source output: complex numbers are printed correctly (using
143 the STL complex<> template or cln::complex()), rational numbers use cl_RA()
144 in the CLN output, and small integers are printed in a more compact format
145 (e.g. "2.0" instead of "2.0000000e+00").
146 * function_options::set_return_type() and function_options::do_not_evalf_params()
149 1.0.13 (27 January 2003)
150 * Contracting epsilon tensors with Euclidean indices now works.
151 * Improved dummy index symmetrization in sums.
152 * Added dirac_gammaL/R(), which can be used instead of dirac_gamma6/7()
153 but are single objects, to allow for a more compact notation of Dirac
155 * Powers with negative numeric exponents are printed as fractions in the
157 * Added symbolic_matrix() for the convenient creation of matrices filled
159 * Added collect_common_factors() which collects common factors from the
161 * simplify_indexed() converts "gamma~mu*p.mu" to "p\".
163 1.0.12 (30 October 2002)
164 * Fixed a bug in power::expand() that could produce invalid expressions.
165 * The input parser no longer ignores extra data following accepted input.
166 * Improved the CLN C-source output (integers are printed as integers, and
167 floating point numbers include the precision).
168 * Fixed a problem in the LaTeX-output of negative fractions.
169 * Added print_latex() and print_csrc() to ginsh.
170 * The sprem() function is now public.
172 1.0.11 (18 September 2002)
173 * Fixed a possible memory corruption in contractions of indexed objects with
174 delta or metric tensors.
175 * Computing the derivative of a power series object with respect to a symbol
176 that is not the expansion variable now works correctly.
177 * Several bugfixes in code generation.
179 1.0.10 (24 July 2002)
180 * Powers of indexed objects are now parenthesized correctly in LaTeX output.
181 * Input parser handles indices (they have to be specified in the same list
183 * Added some limited support for subspaces in the idx and tensor classes.
184 * Fixed a bug in canonicalize() (antisymmetric canonicalization of an already
185 sorted list containing two or more equal objects failed to return 0).
188 * simplify_indexed() now raises/lowers dummy indices to canonicalize the index
189 variance. This allows some simplifications that weren't possible before,
190 like eps~a.b~c~d*X.a*X~b -> 0 and X.a~a-X~a.a -> 0.
191 * Implemented relational::subs().
192 * Fixed bug in simplify_ncmul() for clifford objects.
194 1.0.8 (31 March 2002)
195 * Improvements in memory usage of the expand() methods.
197 1.0.7 (18 March 2002)
198 * Fixed LaTeX output of indexed and matrix objects.
199 * Fixed matrix::pow(n) for n==0 and added helper functions to create unit
200 matrices "ex unit_matrix(unsigned, unsigned)".
203 * "(x+1).subs(x==x-1)" now returns the correct result "x" instead of "x-1".
205 1.0.5 (27 January 2002)
206 * (l)degree(s), coeff(s, n) and collect(s) were extended to accept expressions
207 of any class (except add/mul/ncmul/numeric) for "s". They should even work
208 if "s" is a "power" object, as long as the exponent is non-integer, but with
209 some limitations. For example, you can "collect(a*2^x+b*2^x, 2^x)" to get
210 "(a+b)*2^x", but "degree(2^(3*x), 2^x)" yields 0 instead of 3).
211 * Fixed a small output bug.
213 1.0.4 (24 January 2002)
214 * Speedup in expand().
215 * Faster Bernoulli numbers (Markus Nullmeier).
216 * Some minor bugfixes and documentation updates.
218 1.0.3 (21 December 2001)
219 * Fixed a bug where quo() would call vector::reserve() with a negative
221 * Fix several bugs in code generation.
223 1.0.2 (19 December 2001)
224 * Input parser recognizes "sqrt()", which is also used in the output.
225 * divide(a,b,q) only modifies q if the division succeeds; also, divide(a,b,a)
227 * Fixed small bug in dummy index renaming which could cause it to not
228 recognize renamable indices in some cases.
229 * power::degree() and power::ldegree() throw an exception when encountering
230 a non-integer exponent.
231 * Add output-support for Python bindings.
233 1.0.1 (22 November 2001)
234 * Function sqrfree() handles a few more cases now.
235 * Class relational has real canonical ordering now.
236 * Handle obscene libreadline version numbers when building ginsh.
238 1.0.0 (6 November 2001)
239 * Some internal reorganization resulting in a general speed-up.
240 * The last 3 evaluated expressions in ginsh are now referred to with the
241 tokens '%', '%%' and '%%%'. The old '"', '""' and '"""' remain for
242 compatibility but may be removed in a future version of GiNaC.
244 0.9.4 (20 September 2001)
245 * Functions have better support for external scripting languages.
246 * Interface cleanups and bugfixes.
247 * Fix silly bug in evalf() that prevented things like 2^Pi being computed.
249 0.9.3 (16 August 2001)
250 * series expansion now much more consistent for small order expansion.
251 * lsolve() accepts algorithmic hint as parameter.
254 * Epsilon tensor is more functional.
255 * simplify_indexed() is better at detecting expressions that vanish for
257 * Several little bugfixes and consistency enhancements.
260 * Ctors of class numeric are not explicit any more. All built-in callers for
261 pseudofunctions are now templated and default to ex arguments which relaxes
262 the need for explicit ctors.
263 * New functions/methods:
265 - remove_first(), remove_last(), sort() and unique() for lists
266 - symmetrize_cyclic()
268 * Instead of just totally symmetric or antisymmetric, complex symmetries
269 can now be defined for indexed objects. Symmetries are described by a
270 tree of "symmetry" objects that is constructed with the sy_none(),
271 sy_symm(), sy_anti() and sy_cycl() functions. The symmetry of a function
272 with respect to its arguments can also be defined (this is currently
273 only used for the Beta function).
274 * Generalized map() to take a function object instead of a function pointer.
275 This allows passing an arbitrary number of additional state to the
276 function being called.
277 * color_trace(), dirac_trace(), diff(), expand(), evalf() and normal() work
278 better with container classes, e.g. using color_trace() on a relation will
279 take the trace on both sides, using diff() on a matrix differentiates every
281 * diff() works properly with non-commutative products and indexed objects.
282 * New option flag "expand_function_args" for expand().
283 * Supplement some (now deprecated) macros by inlined template functions:
284 - is_of_type(foo, type) -> is_a<type>(foo)
285 - is_ex_of_type(foo, type) -> is_a<type>(foo)
286 - is_exaclty_of_type(foo, type) -> is_exaclty_a<type>(foo)
287 - is_ex_exaclty_of_type(foo, type) -> is_exaclty_a<type>(foo)
288 - ex_to_foobar(baz) -> ex_to<foobar>(baz)
289 * rem(c, p[x], x) (c: numeric, p[x]: polynomial) erroneously returned p[x]
291 * Small bugfixes in pattern matching.
292 * Updated libtool to version 1.4.
295 * In the output and in ginsh, lists are now delimited by { } braces, and
296 matrices are delimited by single [ ] brackets.
297 * simplify_indexed() renames dummy indices so, e.g., "a.i*a.i+a.j*a.j" gets
298 simplified to "2*a.i*a.i" (or "2*a.j*a.j").
299 * New functions/methods:
300 - canonicalize_clifford() (helpful when comparing expressions containing
302 - symmetrize() and antisymmetrize()
303 - numer_denom() (return numerator and denominator in one call)
304 - map() (apply function to subexpressions)
305 - evalm() (evaluate sums, products and integer powers of matrices)
306 * Added a new function match() for performing pattern matching. subs() and
307 has() also accept patterns as arguments. A pattern can be any expression,
308 optionally containing wildcard objects. These are constructed with the
309 call "wild(<unsigned>)" and are denoted as "$0", "$1" etc. in the output
311 * Positive integer powers of non-commutative expressions (except matrices)
312 are automatically expanded.
313 * Removed cint subdirectory, ginaccint is a separate package now due to
314 packaging considerations.
315 * Several little bugfixes.
318 * color and clifford classes are functional and documented.
319 * New "spinidx" class for dotted/undotted indices.
320 * Predefined spinor metric tensor (created by spinor_metric()).
321 * Symbols can have a LaTeX name, e.g. symbol s("s", "\\sigma");
322 * LaTeX output of indexed objects is much nicer.
323 * Fixed some build problems (with recent libreadline).
324 * Semantics of arithmetic operators now follows the C++ rules more strictly.
326 0.8.2 (24 April 2001)
327 * degree(), ldegree(), coeff(), lcoeff(), tcoeff() and collect() work with
328 non-symbols as the second argument in ginsh.
329 * the argument to collect() can be a list of objects in which case the
330 result is either a recursively collected polynomial, or a polynomial in
331 a distributed form with terms like coeff*x1^e1*...*xn^en, as specified by
332 the second argument to collect().
333 * Several bugfixes (including a nasty memory leak in .normal()).
334 * class matrix: solve() doesn't call algorithms redundantly any more and
335 inverse() falls back to solve() which works in more general cases.
337 0.8.1 (16 April 2001)
338 * degree(), ldegree(), coeff(), lcoeff(), tcoeff() and collect() can now
339 be used with constants, functions and indexed expressions as well, so you
340 can use it to collect by powers of Pi or sin(x), or to find the coefficient
343 - it only works with symbols, constants, functions and indexed expressions,
344 trying to find the coefficient of, e.g., "x^2" or "x+y" won't work;
345 - it does not know about dummy index summations; the coefficient of
346 gamma~0 in p.mu*gamma~mu should be p.0 but is returned as 0;
347 - using coeff(), tcoeff(), lcoeff() or collect() on elements of
348 noncommutative products might return wrong or surprising results.
349 * subs() no longer only substitutes symbols and indices but performs a more
350 general "syntactic substitution", i.e. it substitutes whole objects in sub-
351 expressions. You can subs((a+b)^2,a+b==3) and get 9, but subs(a+b+c,a+b==3)
354 - substituting numerics (subs(expr, 2==4)) will not replace then in all
355 occurences; in general, you shouldn't substitute numerics, though.
356 * Added preliminary (re)implementations of color and clifford classes.
357 * simplify_indexed(): contraction of symmetric and antisymmetric tensors
359 * Replaced the various print*() member functions by a single print() that
360 takes a print_context object that determines the output formatting. This
361 should make it easier to add more output types, such as LaTeX output,
362 which is based on work by Stefan Weinzierl.
363 * Added functions to retrieve the properties stored in archive objects
364 outside of unarchive() (for printing or debugging purposes).
365 * Some bugfixes (indexed objects, archive writing).
366 * .collect() on non-polynomials is now algebraically correct.
368 0.8.0 (24 March 2001)
369 * Complete revamp of indexed objects. Instead of multiple classes for
370 indexed things and their indices there is now only one "indexed" class
371 and two types of indices: "idx" for simple indices and "varidx" for
372 indices with variance. There are predefined delta, epsilon and metric
373 tensors, and a function simplify_indexed() that performs canonicalization
374 and dummy index summations. Matrix objects can be indexed for doing simple
376 * Added an option "expand_indexed" to expand() to perform expansion of
377 indexed objects like (a+b).i -> a.i + b.i
378 * Renamed get_indices() to get_free_indices(), which no longer returns
379 dummy indices and checks the consistency of indices in sums.
380 * sqrfree() factorization fixed and improved syntactically.
381 * subs() works on matrices.
382 * Matrices can be constructed from flat list of elements; diagonal matrices
383 can be constructed from list of diagonal elements with diag_matrix().
384 * Fixed memory leak in expand().
385 * Operator% for objects of class ncmul has gone. Use operator* now for that
386 case too, which is much more natural.
388 0.7.3 (28 February 2001)
389 * Several bugfixes and minor performance tunings.
390 * Added a section to the tutorial about adding new algebraic classes to GiNaC.
391 * Closed many in-source documentation gaps.
393 0.7.2 (17 February 2001)
394 * Several bugfixes in power series expansion, one of them critical.
396 0.7.1 (7 February 2001)
397 * Fix problems with Cint that were caused by CLN's overloaded operator new.
398 * Fix compilation errors with GCC3.
399 * normal() handles large sums of fractions better and normalizes the exponent
400 of power expressions.
401 * expand() always expands the exponent and transforms x^(a+b) -> x^a*x^b.
402 * Some bugfixes of series expansion around branch cuts of special functions.
404 0.7.0 (15 December 2000)
405 * Requires CLN 1.1 now. Class numeric doesn't use an indirect pointer to the
406 actual representation any more. This is a speedup.
407 * mul::expand() was reengineered to not allocate excess temporary memory.
408 * Non-integer powers of a symbol are treated as constants by (l)degree() and
409 coeff(). Using these functions on an expression containing such powers used
410 to fail with an internal error message. The side-effect is that collect()
411 can be used on expressions which are not polynomials.
412 * Added a man page for the ginac-config script.
413 * Ctor of numeric from char* honors Digits.
415 0.6.4 (10 August 2000)
416 * Complete revamp of methods in class matrix. Some redundant (and poor)
417 implementations of elimination schemes were thrown out. The code is now
418 highly orthogonal, more flexible and much more efficient.
419 * Some long standing and quite nasty bugs were discovered and fixed in the
420 following functions: add::normal(), heur_gcd(), sr_gcd() and Order_eval().
423 * Derivatives are now assembled in a slightly different manner (i.e. they
424 might 'look' different on first sight). Under certain circumstances this
425 can result in a dramatic speedup because it gives hashing a better chance,
426 especially when computing higher derivatives.
427 * Some series expansions of built-in functions have been reengineered.
428 * The algorithm for computing determinants can be chosen by the user. See
429 ginac/flags.h and ginac/matrix.h.
430 * The Dilogarithm (Li2) now has floating point evaluation, derivative and a
431 proper series expansion.
432 * Namespace 'std' cleanly disentangled, as demanded by ISO/EIC 14882-1998(E).
433 * Some minor bugfixes, one major lsolve()-bugfix and documentation updates.
436 * ginaccint.bin is now launched by a binary program instead of by a scripts.
437 This allows us to write #!-scripts. A small test suite for GiNaC-cint was
439 * Several minor bugfixes.
442 * Cleanup in the interface to Cint. The required version is now Cint 5.14.38.
443 * Several bugfixes in target install.
446 * IMPORTANT: Several interface changes make programs written with GiNaC
447 much clearer but break compatibility with older versions:
448 - f(x).series(x,p[,o]) -> f(x).series(x==p,o)
449 - series(f(x),x,p[,o]) -> series(f(x),x==p,o)
450 - gamma() -> tgamma() (The true Gamma function, there is now also
451 log(tgamma()), called lgamma(), in accord with ISO/IEC 9899:1999.)
452 - EulerGamma -> Euler
453 * #include'ing ginac.h defines the preprocessor symbols GINACLIB_MAJOR_VERSION,
454 GINACLIB_MINOR_VERSION, and GINACLIB_MICRO_VERSION with the respective GiNaC
455 library version numbers.
456 * Expressions can be constructed from strings like this:
457 ex e("2*x+y", lst(x, y));
458 * ex::to_rational() provides a way to extend the domain of functions like
459 gcd() and divide() that only work on polynomials or rational functions (the
460 good old ex::subs() method reverses this process)
461 * Calling diff() on a function that has no derivative defined returns the
462 inert derivative function "Derivative".
463 * Several new timings in the check target. Some of them may be rather rude
464 at your machine, feel free to interrupt them.
466 0.5.4 (15 March 2000)
467 * Some algorithms in class matrix (notably determinant) were replaced by
468 less brain-dead ones and should now have much better performance.
469 * Checks were completely reorganized and split up into three parts:
470 a) exams (small regression tests with predefined input)
471 b) checks (lenghty coherence checks with random input)
472 c) timings (for coherence and crude benchmarking)
473 * Behaviour of .evalf() was changed: it doesn't .evalf() any exponents.
474 * Expanded expressions now remember they are expanded to prevent
475 superfluous expansions.
476 * Small bugfixes and improvements in the series expansion.
478 0.5.3 (23 February 2000)
479 * A more flexible scheme for registering functions was implemented,
480 allowing for remembering, too.
483 0.5.2 (16 February 2000)
484 * Mainly fixes a bug in the packaging of release 0.5.1.
486 0.5.1 (14 February 2000)
487 * Fixes a small number of bugs.
489 0.5.0 (7 February 2000)
490 * Expressions can be written ("archived") to files and read therefrom.
491 * Addition of GiNaC-cint, which lets you write complete programs in
492 an interactive shell-like manner in your favoured programming
495 0.4.1 (13 December 1999)
496 * Series Expansion of Gamma function and some other trigonometric
497 functions at their poles works now.
498 * Many more evaluations of special functions at points where
500 * info_flags::rational doesn't return true for complex extensions
501 any more---use info_flags::crational for the old behaviour.
502 info_flags::integer and -::cinteger work similarly, the same
503 holds for types like info_flags::rational_polynomial.
505 0.4.0 (26 November 1999)
506 * First public release.