1 This file records noteworthy changes.
4 * Added variants of dirac_trace() and color_trace() that take the trace over
5 more than one representation label by specifying a set or list of labels.
6 * diracgamma::contract_with() uses Chisholm identities in 4 dimensions to
7 produce more compact results.
8 * The --enable-html-doc and --enable-ps-doc configure options are gone.
9 Documentation in HTML, DVI, PostScript, and PDF formats is now built with
10 "make html", "make dvi", "make ps", and "make pdf", respectively. The only
11 documentation that gets built by default are the tutorial in .info format
15 * Added const_preorder_iterator and const_postorder_iterator classes (and
16 associated methods ex::pre/postorder_begin/end()) providing tree traversal
18 * Fixed the LaTeX output of the varidx class.
19 * Fixed bugs in series expansion and complex conjugation.
20 * (p.i*p.i).get_free_indices() returns the correct result () instead of (.i).
21 * Symbolic functions without any eval(), evalf() etc. functions now work
23 * integer_content(), content(), and primpart() now also work for polynomials
24 with rational coefficients (calculating the LCM of coefficients'
26 * Added method matrix::rank().
27 * Added function resultant().
28 * Added integer_content() function to ginsh.
31 * Fixed infinite recursion in atan2_evalf() and improved atan2_eval().
32 * Added automatic evaluations for trigonometric functions with negative
33 arguments (e.g. sin(-2) -> -sin(2)).
34 * Fixed a static initialization order goof-up.
35 * Fixed various bugs in series expansion.
38 * Added a structure<T> template class for the easy creation of user-defined
40 * Added support for (acyclic) visitors, to allow cleaner implementations of
42 * Added a const_iterator class that can be used instead of op()/nops().
43 * Completely revamped the implementation of expression output. It is now
44 possible to add new output formats, to change the behavior of predefined
45 formats at run-time, and to have different output styles for algebraic
47 * Symbols can be made non-commutative.
48 * Added a method ex::conjugate() and a function conjugate() for complex
49 conjugation. Symbols can be declared as real or complex-valued.
50 * Improved the speed of subs(), normal(), to_rational() and to_polynomial()
51 by the use of maps instead of lists. The old forms
52 subs(const lst & ls, const lst & lr, unsigned options)
53 to_rational/to_polynomial(lst & repl)
54 are still available for compatibility, but using the new forms
55 subs(const exmap & m, unsigned options)
56 to_rational/to_polynomial(exmap & repl)
57 is more efficient, especially when the number of replacements is large.
58 * quo(), rem(), prem(), sprem(), decomp_rational(), unit(), content(),
59 primpart() and matrix::charpoly() now take a "const ex &" instead of a
61 * Redundant expressions (two ex pointing to different objects are found to be
62 equal in compare()) are now actively deleted/fused to conserve memory and
63 speed up subsequent comparisons. This behavior can be suppressed on a
64 per-object level with status_flags::not_shareable. Lists and matrices are
65 not shareable by default.
66 * Lots of internal streamlining and optimizations.
67 * Caveats for class implementors:
68 - basic::copy() and basic::destroy() are gone; classes derived from
69 basic can use the defaults for the assignment operator and copy
71 - basic::subs(), basic::normal(), basic::to_rational() and
72 basic::to_polynomial() take 'exmap' objects instead of lists.
73 - basic::subs() now descends into subexpressions (if accessible via
74 nops()/op()/let_op()). If you have a custom implementation of subs()
75 that calls basic::subs() after substituting subexpressions, this needs
76 to be changed to a call to subs_one_level().
77 - lst::thislst() and exprseq::thisexprseq() renamed to thiscontainer().
78 - thiscontainer() and associated constructors now take a std::auto_ptr.
79 - Overloading basic::print() is now deprecated. You should use
80 print_func<>() class options instead.
83 * Fixed a bug in canonicalize_clifford().
84 * Series expansion now works predictably. All terms with the exponent of the
85 expansion variable smaller than the given order are calculated exactly. If
86 the series is not terminating, the Order function is (at least) of the given
89 1.1.6 (22 January 2004)
90 * Added a function option "dummy()" which means "no options". This simplifies
91 the implementation of symbolic functions which are not to be further
93 * Removed a bug in the numerical evaluation of Li() that caused the system
94 to hang for certain parameter combinations.
95 * Fixed a bug in the calculation of hash values for indices that could lead
96 to wrong results or bogus error messages from simplify_indexed().
97 * Fixed a bug in the evaluation of harmonic polylogarithms for complex
98 arguments with positive imaginary part.
100 1.1.5 (5 November 2003)
101 * Harmonic polylogarithms now numerically evaluate for arbitrary arguments
102 (parameter must still be positive integers).
103 * The zeta function now can also be given a lst as a parameter in which case
104 it becomes a multiple zeta value. The use of mZeta is deprecated.
105 * The order of parameters for the multiple polylogarithm has been corrected.
106 * Documentation for the nested sums functions zeta, harmonic polylog, multiple
107 polylog, etc. has been added.
109 1.1.4 (17 October 2003)
110 * Lists and matrices can now be initialized from comma-separated lists of
111 expressions, like this:
121 This is both faster and produces much smaller code than the old constructors
122 lst(ex, ex, ...) and matrix(unsigned, unsigned, lst), especially in the case
123 of matrices, and is now the recommended way to create these objects.
124 * The function mZeta now evaluates much faster for arbitrary parameters. The
125 harmonic and multiple polylogarithms evaluate considerably faster and check
126 for convergence. The order of parameters for the harmonic polylogarithm
129 1.1.3 (22 August 2003)
130 * Added new symbolic functions for better integration with nestedsums:
131 (multiple) polylogarithm Li(), Nielsen's generalized polylogarithm S(),
132 harmonic polylogarithm H(), and multiple zeta value mZeta().
133 * New exhashmap<T> template intended as a drop-in replacement for
134 std::map<ex, T, ex_is_less> using GiNaC's hashing algorithms.
136 1.1.2 (11 August 2003)
137 * Fixed a bug in the unarchiving of sums and products: terms were not
138 reordered in a canonical way.
139 * Fixed a bug in normal()/numer_denom(): denominator was not made unit
140 normal if it was a simple number.
141 * Improved the speed of subs() in some cases.
144 * lst (and exprseq) provide iterators for read-only element access. For
145 sequential access this is one order faster than using op().
146 * Implemented relational::subs() (this was done in 1.0.9 but inadvertently
147 omitted from the 1.1 branch).
148 * pole_error and do_taylor are available to library users.
149 * Added on-line help and Tab-completion for print(), iprint(), print_latex()
150 and print_csrc() in ginsh.
153 * Removed deprecated macros is_ex_a, is_ex_exactly_a and friends for good.
154 * The scalar_products mechanism allows the specification of an index dimension.
155 * Removed dirac_gamma6/7().
156 * Added ex::to_polynomial().
157 * subs() accepts an optional "options" argument. The option
158 subs_option::subs_algebraic enables "smart" substitutions in products and
160 * Added stream manipulators "dflt", "latex", "python", "python_repr", "tree",
161 "csrc", "csrc_float", "csrc_double", "csrc_cl_N", "index_dimensions" and
162 "no_index_dimensions" to control the output format. Calling basic::print()
163 directly is now deprecated.
164 * Made the hashing more simple and efficient.
165 * Caveats for class implementors:
166 - basic::subs(): third argument changed from "bool" to "unsigned"
167 - unarchiving constructor and basic::unarchive(): "const" removed from
169 - basic::let_op() should only be implemented if write access to
170 subexpressions is desired
171 - simplify_ncmul() renamed to eval_ncmul()
172 - simplified_ncmul() renamed to hold_ncmul()
173 - nonsimplified_ncmul() renamed to reeval_ncmul()
175 1.0.14 (1 March 2003)
176 * Improved the C-source output: complex numbers are printed correctly (using
177 the STL complex<> template or cln::complex()), rational numbers use cl_RA()
178 in the CLN output, and small integers are printed in a more compact format
179 (e.g. "2.0" instead of "2.0000000e+00").
180 * function_options::set_return_type() and function_options::do_not_evalf_params()
183 1.0.13 (27 January 2003)
184 * Contracting epsilon tensors with Euclidean indices now works.
185 * Improved dummy index symmetrization in sums.
186 * Added dirac_gammaL/R(), which can be used instead of dirac_gamma6/7()
187 but are single objects, to allow for a more compact notation of Dirac
189 * Powers with negative numeric exponents are printed as fractions in the
191 * Added symbolic_matrix() for the convenient creation of matrices filled
193 * Added collect_common_factors() which collects common factors from the
195 * simplify_indexed() converts "gamma~mu*p.mu" to "p\".
197 1.0.12 (30 October 2002)
198 * Fixed a bug in power::expand() that could produce invalid expressions.
199 * The input parser no longer ignores extra data following accepted input.
200 * Improved the CLN C-source output (integers are printed as integers, and
201 floating point numbers include the precision).
202 * Fixed a problem in the LaTeX-output of negative fractions.
203 * Added print_latex() and print_csrc() to ginsh.
204 * The sprem() function is now public.
206 1.0.11 (18 September 2002)
207 * Fixed a possible memory corruption in contractions of indexed objects with
208 delta or metric tensors.
209 * Computing the derivative of a power series object with respect to a symbol
210 that is not the expansion variable now works correctly.
211 * Several bugfixes in code generation.
213 1.0.10 (24 July 2002)
214 * Powers of indexed objects are now parenthesized correctly in LaTeX output.
215 * Input parser handles indices (they have to be specified in the same list
217 * Added some limited support for subspaces in the idx and tensor classes.
218 * Fixed a bug in canonicalize() (antisymmetric canonicalization of an already
219 sorted list containing two or more equal objects failed to return 0).
222 * simplify_indexed() now raises/lowers dummy indices to canonicalize the index
223 variance. This allows some simplifications that weren't possible before,
224 like eps~a.b~c~d*X.a*X~b -> 0 and X.a~a-X~a.a -> 0.
225 * Implemented relational::subs().
226 * Fixed bug in simplify_ncmul() for clifford objects.
228 1.0.8 (31 March 2002)
229 * Improvements in memory usage of the expand() methods.
231 1.0.7 (18 March 2002)
232 * Fixed LaTeX output of indexed and matrix objects.
233 * Fixed matrix::pow(n) for n==0 and added helper functions to create unit
234 matrices "ex unit_matrix(unsigned, unsigned)".
237 * "(x+1).subs(x==x-1)" now returns the correct result "x" instead of "x-1".
239 1.0.5 (27 January 2002)
240 * (l)degree(s), coeff(s, n) and collect(s) were extended to accept expressions
241 of any class (except add/mul/ncmul/numeric) for "s". They should even work
242 if "s" is a "power" object, as long as the exponent is non-integer, but with
243 some limitations. For example, you can "collect(a*2^x+b*2^x, 2^x)" to get
244 "(a+b)*2^x", but "degree(2^(3*x), 2^x)" yields 0 instead of 3).
245 * Fixed a small output bug.
247 1.0.4 (24 January 2002)
248 * Speedup in expand().
249 * Faster Bernoulli numbers (Markus Nullmeier).
250 * Some minor bugfixes and documentation updates.
252 1.0.3 (21 December 2001)
253 * Fixed a bug where quo() would call vector::reserve() with a negative
255 * Fix several bugs in code generation.
257 1.0.2 (19 December 2001)
258 * Input parser recognizes "sqrt()", which is also used in the output.
259 * divide(a,b,q) only modifies q if the division succeeds; also, divide(a,b,a)
261 * Fixed small bug in dummy index renaming which could cause it to not
262 recognize renamable indices in some cases.
263 * power::degree() and power::ldegree() throw an exception when encountering
264 a non-integer exponent.
265 * Add output-support for Python bindings.
267 1.0.1 (22 November 2001)
268 * Function sqrfree() handles a few more cases now.
269 * Class relational has real canonical ordering now.
270 * Handle obscene libreadline version numbers when building ginsh.
272 1.0.0 (6 November 2001)
273 * Some internal reorganization resulting in a general speed-up.
274 * The last 3 evaluated expressions in ginsh are now referred to with the
275 tokens '%', '%%' and '%%%'. The old '"', '""' and '"""' remain for
276 compatibility but may be removed in a future version of GiNaC.
278 0.9.4 (20 September 2001)
279 * Functions have better support for external scripting languages.
280 * Interface cleanups and bugfixes.
281 * Fix silly bug in evalf() that prevented things like 2^Pi being computed.
283 0.9.3 (16 August 2001)
284 * series expansion now much more consistent for small order expansion.
285 * lsolve() accepts algorithmic hint as parameter.
288 * Epsilon tensor is more functional.
289 * simplify_indexed() is better at detecting expressions that vanish for
291 * Several little bugfixes and consistency enhancements.
294 * Ctors of class numeric are not explicit any more. All built-in callers for
295 pseudofunctions are now templated and default to ex arguments which relaxes
296 the need for explicit ctors.
297 * New functions/methods:
299 - remove_first(), remove_last(), sort() and unique() for lists
300 - symmetrize_cyclic()
302 * Instead of just totally symmetric or antisymmetric, complex symmetries
303 can now be defined for indexed objects. Symmetries are described by a
304 tree of "symmetry" objects that is constructed with the sy_none(),
305 sy_symm(), sy_anti() and sy_cycl() functions. The symmetry of a function
306 with respect to its arguments can also be defined (this is currently
307 only used for the Beta function).
308 * Generalized map() to take a function object instead of a function pointer.
309 This allows passing an arbitrary number of additional state to the
310 function being called.
311 * color_trace(), dirac_trace(), diff(), expand(), evalf() and normal() work
312 better with container classes, e.g. using color_trace() on a relation will
313 take the trace on both sides, using diff() on a matrix differentiates every
315 * diff() works properly with non-commutative products and indexed objects.
316 * New option flag "expand_function_args" for expand().
317 * Supplement some (now deprecated) macros by inlined template functions:
318 - is_of_type(foo, type) -> is_a<type>(foo)
319 - is_ex_of_type(foo, type) -> is_a<type>(foo)
320 - is_exactly_of_type(foo, type) -> is_exactly_a<type>(foo)
321 - is_ex_exactly_of_type(foo, type) -> is_exactly_a<type>(foo)
322 - ex_to_foobar(baz) -> ex_to<foobar>(baz)
323 * rem(c, p[x], x) (c: numeric, p[x]: polynomial) erroneously returned p[x]
325 * Small bugfixes in pattern matching.
326 * Updated libtool to version 1.4.
329 * In the output and in ginsh, lists are now delimited by { } braces, and
330 matrices are delimited by single [ ] brackets.
331 * simplify_indexed() renames dummy indices so, e.g., "a.i*a.i+a.j*a.j" gets
332 simplified to "2*a.i*a.i" (or "2*a.j*a.j").
333 * New functions/methods:
334 - canonicalize_clifford() (helpful when comparing expressions containing
336 - symmetrize() and antisymmetrize()
337 - numer_denom() (return numerator and denominator in one call)
338 - map() (apply function to subexpressions)
339 - evalm() (evaluate sums, products and integer powers of matrices)
340 * Added a new function match() for performing pattern matching. subs() and
341 has() also accept patterns as arguments. A pattern can be any expression,
342 optionally containing wildcard objects. These are constructed with the
343 call "wild(<unsigned>)" and are denoted as "$0", "$1" etc. in the output
345 * Positive integer powers of non-commutative expressions (except matrices)
346 are automatically expanded.
347 * Removed cint subdirectory, ginaccint is a separate package now due to
348 packaging considerations.
349 * Several little bugfixes.
352 * color and clifford classes are functional and documented.
353 * New "spinidx" class for dotted/undotted indices.
354 * Predefined spinor metric tensor (created by spinor_metric()).
355 * Symbols can have a LaTeX name, e.g. symbol s("s", "\\sigma");
356 * LaTeX output of indexed objects is much nicer.
357 * Fixed some build problems (with recent libreadline).
358 * Semantics of arithmetic operators now follows the C++ rules more strictly.
360 0.8.2 (24 April 2001)
361 * degree(), ldegree(), coeff(), lcoeff(), tcoeff() and collect() work with
362 non-symbols as the second argument in ginsh.
363 * the argument to collect() can be a list of objects in which case the
364 result is either a recursively collected polynomial, or a polynomial in
365 a distributed form with terms like coeff*x1^e1*...*xn^en, as specified by
366 the second argument to collect().
367 * Several bugfixes (including a nasty memory leak in .normal()).
368 * class matrix: solve() doesn't call algorithms redundantly any more and
369 inverse() falls back to solve() which works in more general cases.
371 0.8.1 (16 April 2001)
372 * degree(), ldegree(), coeff(), lcoeff(), tcoeff() and collect() can now
373 be used with constants, functions and indexed expressions as well, so you
374 can use it to collect by powers of Pi or sin(x), or to find the coefficient
377 - it only works with symbols, constants, functions and indexed expressions,
378 trying to find the coefficient of, e.g., "x^2" or "x+y" won't work;
379 - it does not know about dummy index summations; the coefficient of
380 gamma~0 in p.mu*gamma~mu should be p.0 but is returned as 0;
381 - using coeff(), tcoeff(), lcoeff() or collect() on elements of
382 noncommutative products might return wrong or surprising results.
383 * subs() no longer only substitutes symbols and indices but performs a more
384 general "syntactic substitution", i.e. it substitutes whole objects in sub-
385 expressions. You can subs((a+b)^2,a+b==3) and get 9, but subs(a+b+c,a+b==3)
388 - substituting numerics (subs(expr, 2==4)) will not replace then in all
389 occurences; in general, you shouldn't substitute numerics, though.
390 * Added preliminary (re)implementations of color and clifford classes.
391 * simplify_indexed(): contraction of symmetric and antisymmetric tensors
393 * Replaced the various print*() member functions by a single print() that
394 takes a print_context object that determines the output formatting. This
395 should make it easier to add more output types, such as LaTeX output,
396 which is based on work by Stefan Weinzierl.
397 * Added functions to retrieve the properties stored in archive objects
398 outside of unarchive() (for printing or debugging purposes).
399 * Some bugfixes (indexed objects, archive writing).
400 * .collect() on non-polynomials is now algebraically correct.
402 0.8.0 (24 March 2001)
403 * Complete revamp of indexed objects. Instead of multiple classes for
404 indexed things and their indices there is now only one "indexed" class
405 and two types of indices: "idx" for simple indices and "varidx" for
406 indices with variance. There are predefined delta, epsilon and metric
407 tensors, and a function simplify_indexed() that performs canonicalization
408 and dummy index summations. Matrix objects can be indexed for doing simple
410 * Added an option "expand_indexed" to expand() to perform expansion of
411 indexed objects like (a+b).i -> a.i + b.i
412 * Renamed get_indices() to get_free_indices(), which no longer returns
413 dummy indices and checks the consistency of indices in sums.
414 * sqrfree() factorization fixed and improved syntactically.
415 * subs() works on matrices.
416 * Matrices can be constructed from flat list of elements; diagonal matrices
417 can be constructed from list of diagonal elements with diag_matrix().
418 * Fixed memory leak in expand().
419 * Operator% for objects of class ncmul has gone. Use operator* now for that
420 case too, which is much more natural.
422 0.7.3 (28 February 2001)
423 * Several bugfixes and minor performance tunings.
424 * Added a section to the tutorial about adding new algebraic classes to GiNaC.
425 * Closed many in-source documentation gaps.
427 0.7.2 (17 February 2001)
428 * Several bugfixes in power series expansion, one of them critical.
430 0.7.1 (7 February 2001)
431 * Fix problems with Cint that were caused by CLN's overloaded operator new.
432 * Fix compilation errors with GCC3.
433 * normal() handles large sums of fractions better and normalizes the exponent
434 of power expressions.
435 * expand() always expands the exponent and transforms x^(a+b) -> x^a*x^b.
436 * Some bugfixes of series expansion around branch cuts of special functions.
438 0.7.0 (15 December 2000)
439 * Requires CLN 1.1 now. Class numeric doesn't use an indirect pointer to the
440 actual representation any more. This is a speedup.
441 * mul::expand() was reengineered to not allocate excess temporary memory.
442 * Non-integer powers of a symbol are treated as constants by (l)degree() and
443 coeff(). Using these functions on an expression containing such powers used
444 to fail with an internal error message. The side-effect is that collect()
445 can be used on expressions which are not polynomials.
446 * Added a man page for the ginac-config script.
447 * Ctor of numeric from char* honors Digits.
449 0.6.4 (10 August 2000)
450 * Complete revamp of methods in class matrix. Some redundant (and poor)
451 implementations of elimination schemes were thrown out. The code is now
452 highly orthogonal, more flexible and much more efficient.
453 * Some long standing and quite nasty bugs were discovered and fixed in the
454 following functions: add::normal(), heur_gcd(), sr_gcd() and Order_eval().
457 * Derivatives are now assembled in a slightly different manner (i.e. they
458 might 'look' different on first sight). Under certain circumstances this
459 can result in a dramatic speedup because it gives hashing a better chance,
460 especially when computing higher derivatives.
461 * Some series expansions of built-in functions have been reengineered.
462 * The algorithm for computing determinants can be chosen by the user. See
463 ginac/flags.h and ginac/matrix.h.
464 * The Dilogarithm (Li2) now has floating point evaluation, derivative and a
465 proper series expansion.
466 * Namespace 'std' cleanly disentangled, as demanded by ISO/EIC 14882-1998(E).
467 * Some minor bugfixes, one major lsolve()-bugfix and documentation updates.
470 * ginaccint.bin is now launched by a binary program instead of by a scripts.
471 This allows us to write #!-scripts. A small test suite for GiNaC-cint was
473 * Several minor bugfixes.
476 * Cleanup in the interface to Cint. The required version is now Cint 5.14.38.
477 * Several bugfixes in target install.
480 * IMPORTANT: Several interface changes make programs written with GiNaC
481 much clearer but break compatibility with older versions:
482 - f(x).series(x,p[,o]) -> f(x).series(x==p,o)
483 - series(f(x),x,p[,o]) -> series(f(x),x==p,o)
484 - gamma() -> tgamma() (The true Gamma function, there is now also
485 log(tgamma()), called lgamma(), in accord with ISO/IEC 9899:1999.)
486 - EulerGamma -> Euler
487 * #include'ing ginac.h defines the preprocessor symbols GINACLIB_MAJOR_VERSION,
488 GINACLIB_MINOR_VERSION, and GINACLIB_MICRO_VERSION with the respective GiNaC
489 library version numbers.
490 * Expressions can be constructed from strings like this:
491 ex e("2*x+y", lst(x, y));
492 * ex::to_rational() provides a way to extend the domain of functions like
493 gcd() and divide() that only work on polynomials or rational functions (the
494 good old ex::subs() method reverses this process)
495 * Calling diff() on a function that has no derivative defined returns the
496 inert derivative function "Derivative".
497 * Several new timings in the check target. Some of them may be rather rude
498 at your machine, feel free to interrupt them.
500 0.5.4 (15 March 2000)
501 * Some algorithms in class matrix (notably determinant) were replaced by
502 less brain-dead ones and should now have much better performance.
503 * Checks were completely reorganized and split up into three parts:
504 a) exams (small regression tests with predefined input)
505 b) checks (lenghty coherence checks with random input)
506 c) timings (for coherence and crude benchmarking)
507 * Behaviour of .evalf() was changed: it doesn't .evalf() any exponents.
508 * Expanded expressions now remember they are expanded to prevent
509 superfluous expansions.
510 * Small bugfixes and improvements in the series expansion.
512 0.5.3 (23 February 2000)
513 * A more flexible scheme for registering functions was implemented,
514 allowing for remembering, too.
517 0.5.2 (16 February 2000)
518 * Mainly fixes a bug in the packaging of release 0.5.1.
520 0.5.1 (14 February 2000)
521 * Fixes a small number of bugs.
523 0.5.0 (7 February 2000)
524 * Expressions can be written ("archived") to files and read therefrom.
525 * Addition of GiNaC-cint, which lets you write complete programs in
526 an interactive shell-like manner in your favoured programming
529 0.4.1 (13 December 1999)
530 * Series Expansion of Gamma function and some other trigonometric
531 functions at their poles works now.
532 * Many more evaluations of special functions at points where
534 * info_flags::rational doesn't return true for complex extensions
535 any more---use info_flags::crational for the old behaviour.
536 info_flags::integer and -::cinteger work similarly, the same
537 holds for types like info_flags::rational_polynomial.
539 0.4.0 (26 November 1999)
540 * First public release.