1 This file records noteworthy changes.
4 * Added const_preorder_iterator and const_postorder_iterator classes (and
5 associated methods ex::pre/postorder_begin/end()) providing tree traversal
7 * Fixed the LaTeX output of the varidx class.
8 * Symbolic functions without any eval(), evalf() etc. functions now work
12 * Fixed infinite recursion in atan2_evalf() and improved atan2_eval().
13 * Added automatic evaluations for trigonometric functions with negative
14 arguments (e.g. sin(-2) -> -sin(2)).
15 * Fixed a static initialization order goof-up.
16 * Fixed various bugs in series expansion.
19 * Added a structure<T> template class for the easy creation of user-defined
21 * Added support for (acyclic) visitors, to allow cleaner implementations of
23 * Added a const_iterator class that can be used instead of op()/nops().
24 * Completely revamped the implementation of expression output. It is now
25 possible to add new output formats, to change the behavior of predefined
26 formats at run-time, and to have different output styles for algebraic
28 * Symbols can be made non-commutative.
29 * Added a method ex::conjugate() and a function conjugate() for complex
30 conjugation. Symbols can be declared as real or complex-valued.
31 * Improved the speed of subs(), normal(), to_rational() and to_polynomial()
32 by the use of maps instead of lists. The old forms
33 subs(const lst & ls, const lst & lr, unsigned options)
34 to_rational/to_polynomial(lst & repl)
35 are still available for compatibility, but using the new forms
36 subs(const exmap & m, unsigned options)
37 to_rational/to_polynomial(exmap & repl)
38 is more efficient, especially when the number of replacements is large.
39 * quo(), rem(), prem(), sprem(), decomp_rational(), unit(), content(),
40 primpart() and matrix::charpoly() now take a "const ex &" instead of a
42 * Redundant expressions (two ex pointing to different objects are found to be
43 equal in compare()) are now actively deleted/fused to conserve memory and
44 speed up subsequent comparisons. This behavior can be suppressed on a
45 per-object level with status_flags::not_shareable. Lists and matrices are
46 not shareable by default.
47 * Lots of internal streamlining and optimizations.
48 * Caveats for class implementors:
49 - basic::copy() and basic::destroy() are gone; classes derived from
50 basic can use the defaults for the assignment operator and copy
52 - basic::subs(), basic::normal(), basic::to_rational() and
53 basic::to_polynomial() take 'exmap' objects instead of lists.
54 - basic::subs() now descends into subexpressions (if accessible via
55 nops()/op()/let_op()). If you have a custom implementation of subs()
56 that calls basic::subs() after substituting subexpressions, this needs
57 to be changed to a call to subs_one_level().
58 - lst::thislst() and exprseq::thisexprseq() renamed to thiscontainer().
59 - thiscontainer() and associated constructors now take a std::auto_ptr.
60 - Overloading basic::print() is now deprecated. You should use
61 print_func<>() class options instead.
64 * Fixed a bug in canonicalize_clifford().
65 * Series expansion now works predictably. All terms with the exponent of the
66 expansion variable smaller than the given order are calculated exactly. If
67 the series is not terminating, the Order function is (at least) of the given
70 1.1.6 (22 January 2004)
71 * Added a function option "dummy()" which means "no options". This simplifies
72 the implementation of symbolic functions which are not to be further
74 * Removed a bug in the numerical evaluation of Li() that caused the system
75 to hang for certain parameter combinations.
76 * Fixed a bug in the calculation of hash values for indices that could lead
77 to wrong results or bogus error messages from simplify_indexed().
78 * Fixed a bug in the evaluation of harmonic polylogarithms for complex
79 arguments with positive imaginary part.
81 1.1.5 (5 November 2003)
82 * Harmonic polylogarithms now numerically evaluate for arbitrary arguments
83 (parameter must still be positive integers).
84 * The zeta function now can also be given a lst as a parameter in which case
85 it becomes a multiple zeta value. The use of mZeta is deprecated.
86 * The order of parameters for the multiple polylogarithm has been corrected.
87 * Documentation for the nested sums functions zeta, harmonic polylog, multiple
88 polylog, etc. has been added.
90 1.1.4 (17 October 2003)
91 * Lists and matrices can now be initialized from comma-separated lists of
92 expressions, like this:
102 This is both faster and produces much smaller code than the old constructors
103 lst(ex, ex, ...) and matrix(unsigned, unsigned, lst), especially in the case
104 of matrices, and is now the recommended way to create these objects.
105 * The function mZeta now evaluates much faster for arbitrary parameters. The
106 harmonic and multiple polylogarithms evaluate considerably faster and check
107 for convergence. The order of parameters for the harmonic polylogarithm
110 1.1.3 (22 August 2003)
111 * Added new symbolic functions for better integration with nestedsums:
112 (multiple) polylogarithm Li(), Nielsen's generalized polylogarithm S(),
113 harmonic polylogarithm H(), and multiple zeta value mZeta().
114 * New exhashmap<T> template intended as a drop-in replacement for
115 std::map<ex, T, ex_is_less> using GiNaC's hashing algorithms.
117 1.1.2 (11 August 2003)
118 * Fixed a bug in the unarchiving of sums and products: terms were not
119 reordered in a canonical way.
120 * Fixed a bug in normal()/numer_denom(): denominator was not made unit
121 normal if it was a simple number.
122 * Improved the speed of subs() in some cases.
125 * lst (and exprseq) provide iterators for read-only element access. For
126 sequential access this is one order faster than using op().
127 * Implemented relational::subs() (this was done in 1.0.9 but inadvertently
128 omitted from the 1.1 branch).
129 * pole_error and do_taylor are available to library users.
130 * Added on-line help and Tab-completion for print(), iprint(), print_latex()
131 and print_csrc() in ginsh.
134 * Removed deprecated macros is_ex_a, is_ex_exactly_a and friends for good.
135 * The scalar_products mechanism allows the specification of an index dimension.
136 * Removed dirac_gamma6/7().
137 * Added ex::to_polynomial().
138 * subs() accepts an optional "options" argument. The option
139 subs_option::subs_algebraic enables "smart" substitutions in products and
141 * Added stream manipulators "dflt", "latex", "python", "python_repr", "tree",
142 "csrc", "csrc_float", "csrc_double", "csrc_cl_N", "index_dimensions" and
143 "no_index_dimensions" to control the output format. Calling basic::print()
144 directly is now deprecated.
145 * Made the hashing more simple and efficient.
146 * Caveats for class implementors:
147 - basic::subs(): third argument changed from "bool" to "unsigned"
148 - unarchiving constructor and basic::unarchive(): "const" removed from
150 - basic::let_op() should only be implemented if write access to
151 subexpressions is desired
152 - simplify_ncmul() renamed to eval_ncmul()
153 - simplified_ncmul() renamed to hold_ncmul()
154 - nonsimplified_ncmul() renamed to reeval_ncmul()
156 1.0.14 (1 March 2003)
157 * Improved the C-source output: complex numbers are printed correctly (using
158 the STL complex<> template or cln::complex()), rational numbers use cl_RA()
159 in the CLN output, and small integers are printed in a more compact format
160 (e.g. "2.0" instead of "2.0000000e+00").
161 * function_options::set_return_type() and function_options::do_not_evalf_params()
164 1.0.13 (27 January 2003)
165 * Contracting epsilon tensors with Euclidean indices now works.
166 * Improved dummy index symmetrization in sums.
167 * Added dirac_gammaL/R(), which can be used instead of dirac_gamma6/7()
168 but are single objects, to allow for a more compact notation of Dirac
170 * Powers with negative numeric exponents are printed as fractions in the
172 * Added symbolic_matrix() for the convenient creation of matrices filled
174 * Added collect_common_factors() which collects common factors from the
176 * simplify_indexed() converts "gamma~mu*p.mu" to "p\".
178 1.0.12 (30 October 2002)
179 * Fixed a bug in power::expand() that could produce invalid expressions.
180 * The input parser no longer ignores extra data following accepted input.
181 * Improved the CLN C-source output (integers are printed as integers, and
182 floating point numbers include the precision).
183 * Fixed a problem in the LaTeX-output of negative fractions.
184 * Added print_latex() and print_csrc() to ginsh.
185 * The sprem() function is now public.
187 1.0.11 (18 September 2002)
188 * Fixed a possible memory corruption in contractions of indexed objects with
189 delta or metric tensors.
190 * Computing the derivative of a power series object with respect to a symbol
191 that is not the expansion variable now works correctly.
192 * Several bugfixes in code generation.
194 1.0.10 (24 July 2002)
195 * Powers of indexed objects are now parenthesized correctly in LaTeX output.
196 * Input parser handles indices (they have to be specified in the same list
198 * Added some limited support for subspaces in the idx and tensor classes.
199 * Fixed a bug in canonicalize() (antisymmetric canonicalization of an already
200 sorted list containing two or more equal objects failed to return 0).
203 * simplify_indexed() now raises/lowers dummy indices to canonicalize the index
204 variance. This allows some simplifications that weren't possible before,
205 like eps~a.b~c~d*X.a*X~b -> 0 and X.a~a-X~a.a -> 0.
206 * Implemented relational::subs().
207 * Fixed bug in simplify_ncmul() for clifford objects.
209 1.0.8 (31 March 2002)
210 * Improvements in memory usage of the expand() methods.
212 1.0.7 (18 March 2002)
213 * Fixed LaTeX output of indexed and matrix objects.
214 * Fixed matrix::pow(n) for n==0 and added helper functions to create unit
215 matrices "ex unit_matrix(unsigned, unsigned)".
218 * "(x+1).subs(x==x-1)" now returns the correct result "x" instead of "x-1".
220 1.0.5 (27 January 2002)
221 * (l)degree(s), coeff(s, n) and collect(s) were extended to accept expressions
222 of any class (except add/mul/ncmul/numeric) for "s". They should even work
223 if "s" is a "power" object, as long as the exponent is non-integer, but with
224 some limitations. For example, you can "collect(a*2^x+b*2^x, 2^x)" to get
225 "(a+b)*2^x", but "degree(2^(3*x), 2^x)" yields 0 instead of 3).
226 * Fixed a small output bug.
228 1.0.4 (24 January 2002)
229 * Speedup in expand().
230 * Faster Bernoulli numbers (Markus Nullmeier).
231 * Some minor bugfixes and documentation updates.
233 1.0.3 (21 December 2001)
234 * Fixed a bug where quo() would call vector::reserve() with a negative
236 * Fix several bugs in code generation.
238 1.0.2 (19 December 2001)
239 * Input parser recognizes "sqrt()", which is also used in the output.
240 * divide(a,b,q) only modifies q if the division succeeds; also, divide(a,b,a)
242 * Fixed small bug in dummy index renaming which could cause it to not
243 recognize renamable indices in some cases.
244 * power::degree() and power::ldegree() throw an exception when encountering
245 a non-integer exponent.
246 * Add output-support for Python bindings.
248 1.0.1 (22 November 2001)
249 * Function sqrfree() handles a few more cases now.
250 * Class relational has real canonical ordering now.
251 * Handle obscene libreadline version numbers when building ginsh.
253 1.0.0 (6 November 2001)
254 * Some internal reorganization resulting in a general speed-up.
255 * The last 3 evaluated expressions in ginsh are now referred to with the
256 tokens '%', '%%' and '%%%'. The old '"', '""' and '"""' remain for
257 compatibility but may be removed in a future version of GiNaC.
259 0.9.4 (20 September 2001)
260 * Functions have better support for external scripting languages.
261 * Interface cleanups and bugfixes.
262 * Fix silly bug in evalf() that prevented things like 2^Pi being computed.
264 0.9.3 (16 August 2001)
265 * series expansion now much more consistent for small order expansion.
266 * lsolve() accepts algorithmic hint as parameter.
269 * Epsilon tensor is more functional.
270 * simplify_indexed() is better at detecting expressions that vanish for
272 * Several little bugfixes and consistency enhancements.
275 * Ctors of class numeric are not explicit any more. All built-in callers for
276 pseudofunctions are now templated and default to ex arguments which relaxes
277 the need for explicit ctors.
278 * New functions/methods:
280 - remove_first(), remove_last(), sort() and unique() for lists
281 - symmetrize_cyclic()
283 * Instead of just totally symmetric or antisymmetric, complex symmetries
284 can now be defined for indexed objects. Symmetries are described by a
285 tree of "symmetry" objects that is constructed with the sy_none(),
286 sy_symm(), sy_anti() and sy_cycl() functions. The symmetry of a function
287 with respect to its arguments can also be defined (this is currently
288 only used for the Beta function).
289 * Generalized map() to take a function object instead of a function pointer.
290 This allows passing an arbitrary number of additional state to the
291 function being called.
292 * color_trace(), dirac_trace(), diff(), expand(), evalf() and normal() work
293 better with container classes, e.g. using color_trace() on a relation will
294 take the trace on both sides, using diff() on a matrix differentiates every
296 * diff() works properly with non-commutative products and indexed objects.
297 * New option flag "expand_function_args" for expand().
298 * Supplement some (now deprecated) macros by inlined template functions:
299 - is_of_type(foo, type) -> is_a<type>(foo)
300 - is_ex_of_type(foo, type) -> is_a<type>(foo)
301 - is_exactly_of_type(foo, type) -> is_exactly_a<type>(foo)
302 - is_ex_exactly_of_type(foo, type) -> is_exactly_a<type>(foo)
303 - ex_to_foobar(baz) -> ex_to<foobar>(baz)
304 * rem(c, p[x], x) (c: numeric, p[x]: polynomial) erroneously returned p[x]
306 * Small bugfixes in pattern matching.
307 * Updated libtool to version 1.4.
310 * In the output and in ginsh, lists are now delimited by { } braces, and
311 matrices are delimited by single [ ] brackets.
312 * simplify_indexed() renames dummy indices so, e.g., "a.i*a.i+a.j*a.j" gets
313 simplified to "2*a.i*a.i" (or "2*a.j*a.j").
314 * New functions/methods:
315 - canonicalize_clifford() (helpful when comparing expressions containing
317 - symmetrize() and antisymmetrize()
318 - numer_denom() (return numerator and denominator in one call)
319 - map() (apply function to subexpressions)
320 - evalm() (evaluate sums, products and integer powers of matrices)
321 * Added a new function match() for performing pattern matching. subs() and
322 has() also accept patterns as arguments. A pattern can be any expression,
323 optionally containing wildcard objects. These are constructed with the
324 call "wild(<unsigned>)" and are denoted as "$0", "$1" etc. in the output
326 * Positive integer powers of non-commutative expressions (except matrices)
327 are automatically expanded.
328 * Removed cint subdirectory, ginaccint is a separate package now due to
329 packaging considerations.
330 * Several little bugfixes.
333 * color and clifford classes are functional and documented.
334 * New "spinidx" class for dotted/undotted indices.
335 * Predefined spinor metric tensor (created by spinor_metric()).
336 * Symbols can have a LaTeX name, e.g. symbol s("s", "\\sigma");
337 * LaTeX output of indexed objects is much nicer.
338 * Fixed some build problems (with recent libreadline).
339 * Semantics of arithmetic operators now follows the C++ rules more strictly.
341 0.8.2 (24 April 2001)
342 * degree(), ldegree(), coeff(), lcoeff(), tcoeff() and collect() work with
343 non-symbols as the second argument in ginsh.
344 * the argument to collect() can be a list of objects in which case the
345 result is either a recursively collected polynomial, or a polynomial in
346 a distributed form with terms like coeff*x1^e1*...*xn^en, as specified by
347 the second argument to collect().
348 * Several bugfixes (including a nasty memory leak in .normal()).
349 * class matrix: solve() doesn't call algorithms redundantly any more and
350 inverse() falls back to solve() which works in more general cases.
352 0.8.1 (16 April 2001)
353 * degree(), ldegree(), coeff(), lcoeff(), tcoeff() and collect() can now
354 be used with constants, functions and indexed expressions as well, so you
355 can use it to collect by powers of Pi or sin(x), or to find the coefficient
358 - it only works with symbols, constants, functions and indexed expressions,
359 trying to find the coefficient of, e.g., "x^2" or "x+y" won't work;
360 - it does not know about dummy index summations; the coefficient of
361 gamma~0 in p.mu*gamma~mu should be p.0 but is returned as 0;
362 - using coeff(), tcoeff(), lcoeff() or collect() on elements of
363 noncommutative products might return wrong or surprising results.
364 * subs() no longer only substitutes symbols and indices but performs a more
365 general "syntactic substitution", i.e. it substitutes whole objects in sub-
366 expressions. You can subs((a+b)^2,a+b==3) and get 9, but subs(a+b+c,a+b==3)
369 - substituting numerics (subs(expr, 2==4)) will not replace then in all
370 occurences; in general, you shouldn't substitute numerics, though.
371 * Added preliminary (re)implementations of color and clifford classes.
372 * simplify_indexed(): contraction of symmetric and antisymmetric tensors
374 * Replaced the various print*() member functions by a single print() that
375 takes a print_context object that determines the output formatting. This
376 should make it easier to add more output types, such as LaTeX output,
377 which is based on work by Stefan Weinzierl.
378 * Added functions to retrieve the properties stored in archive objects
379 outside of unarchive() (for printing or debugging purposes).
380 * Some bugfixes (indexed objects, archive writing).
381 * .collect() on non-polynomials is now algebraically correct.
383 0.8.0 (24 March 2001)
384 * Complete revamp of indexed objects. Instead of multiple classes for
385 indexed things and their indices there is now only one "indexed" class
386 and two types of indices: "idx" for simple indices and "varidx" for
387 indices with variance. There are predefined delta, epsilon and metric
388 tensors, and a function simplify_indexed() that performs canonicalization
389 and dummy index summations. Matrix objects can be indexed for doing simple
391 * Added an option "expand_indexed" to expand() to perform expansion of
392 indexed objects like (a+b).i -> a.i + b.i
393 * Renamed get_indices() to get_free_indices(), which no longer returns
394 dummy indices and checks the consistency of indices in sums.
395 * sqrfree() factorization fixed and improved syntactically.
396 * subs() works on matrices.
397 * Matrices can be constructed from flat list of elements; diagonal matrices
398 can be constructed from list of diagonal elements with diag_matrix().
399 * Fixed memory leak in expand().
400 * Operator% for objects of class ncmul has gone. Use operator* now for that
401 case too, which is much more natural.
403 0.7.3 (28 February 2001)
404 * Several bugfixes and minor performance tunings.
405 * Added a section to the tutorial about adding new algebraic classes to GiNaC.
406 * Closed many in-source documentation gaps.
408 0.7.2 (17 February 2001)
409 * Several bugfixes in power series expansion, one of them critical.
411 0.7.1 (7 February 2001)
412 * Fix problems with Cint that were caused by CLN's overloaded operator new.
413 * Fix compilation errors with GCC3.
414 * normal() handles large sums of fractions better and normalizes the exponent
415 of power expressions.
416 * expand() always expands the exponent and transforms x^(a+b) -> x^a*x^b.
417 * Some bugfixes of series expansion around branch cuts of special functions.
419 0.7.0 (15 December 2000)
420 * Requires CLN 1.1 now. Class numeric doesn't use an indirect pointer to the
421 actual representation any more. This is a speedup.
422 * mul::expand() was reengineered to not allocate excess temporary memory.
423 * Non-integer powers of a symbol are treated as constants by (l)degree() and
424 coeff(). Using these functions on an expression containing such powers used
425 to fail with an internal error message. The side-effect is that collect()
426 can be used on expressions which are not polynomials.
427 * Added a man page for the ginac-config script.
428 * Ctor of numeric from char* honors Digits.
430 0.6.4 (10 August 2000)
431 * Complete revamp of methods in class matrix. Some redundant (and poor)
432 implementations of elimination schemes were thrown out. The code is now
433 highly orthogonal, more flexible and much more efficient.
434 * Some long standing and quite nasty bugs were discovered and fixed in the
435 following functions: add::normal(), heur_gcd(), sr_gcd() and Order_eval().
438 * Derivatives are now assembled in a slightly different manner (i.e. they
439 might 'look' different on first sight). Under certain circumstances this
440 can result in a dramatic speedup because it gives hashing a better chance,
441 especially when computing higher derivatives.
442 * Some series expansions of built-in functions have been reengineered.
443 * The algorithm for computing determinants can be chosen by the user. See
444 ginac/flags.h and ginac/matrix.h.
445 * The Dilogarithm (Li2) now has floating point evaluation, derivative and a
446 proper series expansion.
447 * Namespace 'std' cleanly disentangled, as demanded by ISO/EIC 14882-1998(E).
448 * Some minor bugfixes, one major lsolve()-bugfix and documentation updates.
451 * ginaccint.bin is now launched by a binary program instead of by a scripts.
452 This allows us to write #!-scripts. A small test suite for GiNaC-cint was
454 * Several minor bugfixes.
457 * Cleanup in the interface to Cint. The required version is now Cint 5.14.38.
458 * Several bugfixes in target install.
461 * IMPORTANT: Several interface changes make programs written with GiNaC
462 much clearer but break compatibility with older versions:
463 - f(x).series(x,p[,o]) -> f(x).series(x==p,o)
464 - series(f(x),x,p[,o]) -> series(f(x),x==p,o)
465 - gamma() -> tgamma() (The true Gamma function, there is now also
466 log(tgamma()), called lgamma(), in accord with ISO/IEC 9899:1999.)
467 - EulerGamma -> Euler
468 * #include'ing ginac.h defines the preprocessor symbols GINACLIB_MAJOR_VERSION,
469 GINACLIB_MINOR_VERSION, and GINACLIB_MICRO_VERSION with the respective GiNaC
470 library version numbers.
471 * Expressions can be constructed from strings like this:
472 ex e("2*x+y", lst(x, y));
473 * ex::to_rational() provides a way to extend the domain of functions like
474 gcd() and divide() that only work on polynomials or rational functions (the
475 good old ex::subs() method reverses this process)
476 * Calling diff() on a function that has no derivative defined returns the
477 inert derivative function "Derivative".
478 * Several new timings in the check target. Some of them may be rather rude
479 at your machine, feel free to interrupt them.
481 0.5.4 (15 March 2000)
482 * Some algorithms in class matrix (notably determinant) were replaced by
483 less brain-dead ones and should now have much better performance.
484 * Checks were completely reorganized and split up into three parts:
485 a) exams (small regression tests with predefined input)
486 b) checks (lenghty coherence checks with random input)
487 c) timings (for coherence and crude benchmarking)
488 * Behaviour of .evalf() was changed: it doesn't .evalf() any exponents.
489 * Expanded expressions now remember they are expanded to prevent
490 superfluous expansions.
491 * Small bugfixes and improvements in the series expansion.
493 0.5.3 (23 February 2000)
494 * A more flexible scheme for registering functions was implemented,
495 allowing for remembering, too.
498 0.5.2 (16 February 2000)
499 * Mainly fixes a bug in the packaging of release 0.5.1.
501 0.5.1 (14 February 2000)
502 * Fixes a small number of bugs.
504 0.5.0 (7 February 2000)
505 * Expressions can be written ("archived") to files and read therefrom.
506 * Addition of GiNaC-cint, which lets you write complete programs in
507 an interactive shell-like manner in your favoured programming
510 0.4.1 (13 December 1999)
511 * Series Expansion of Gamma function and some other trigonometric
512 functions at their poles works now.
513 * Many more evaluations of special functions at points where
515 * info_flags::rational doesn't return true for complex extensions
516 any more---use info_flags::crational for the old behaviour.
517 info_flags::integer and -::cinteger work similarly, the same
518 holds for types like info_flags::rational_polynomial.
520 0.4.0 (26 November 1999)
521 * First public release.