]> www.ginac.de Git - ginac.git/log
ginac.git
23 years agotrancendental function branch-point sanity
Richard Kreckel [Wed, 7 Feb 2001 19:16:21 +0000 (19:16 +0000)]
trancendental function branch-point sanity

23 years agoChangeLog: mention cvs log (must be non-empty for Debian)
Richard Kreckel [Wed, 7 Feb 2001 19:15:51 +0000 (19:15 +0000)]
ChangeLog: mention cvs log (must be non-empty for Debian)
INSTALL: update cint-status
configure.in, NEWS: Prepare for release

23 years agoExplain the branch-cut conventions in some more detail.
Richard Kreckel [Sun, 4 Feb 2001 18:08:26 +0000 (18:08 +0000)]
Explain the branch-cut conventions in some more detail.

23 years agoAdded correct series expansion for atan and atanh.
Richard Kreckel [Fri, 2 Feb 2001 21:39:19 +0000 (21:39 +0000)]
Added correct series expansion for atan and atanh.

23 years agoFixed most bugs with building Debian packages.
Richard Kreckel [Wed, 31 Jan 2001 21:07:43 +0000 (21:07 +0000)]
Fixed most bugs with building Debian packages.

23 years ago- throw at atan()'s poles.
Richard Kreckel [Wed, 31 Jan 2001 00:55:27 +0000 (00:55 +0000)]
- throw at atan()'s poles.

23 years agodocumentation update
Christian Bauer [Wed, 31 Jan 2001 00:37:44 +0000 (00:37 +0000)]
documentation update

23 years agolist of completion work break characters set correctly
Christian Bauer [Wed, 31 Jan 2001 00:37:13 +0000 (00:37 +0000)]
list of completion work break characters set correctly

23 years agonormal() applied to a power expression also normalizes the exponent
Christian Bauer [Tue, 30 Jan 2001 19:50:16 +0000 (19:50 +0000)]
normal() applied to a power expression also normalizes the exponent

23 years agoexpand() always expands the exponent and transforms x^(a+b) -> x^a*x^b
Christian Bauer [Mon, 29 Jan 2001 21:46:01 +0000 (21:46 +0000)]
expand() always expands the exponent and transforms x^(a+b) -> x^a*x^b

23 years agoadded a check that expand(x^((a+b)^2-a^2-b^2-2*a*b)) -> 1
Christian Bauer [Mon, 29 Jan 2001 21:43:45 +0000 (21:43 +0000)]
added a check that expand(x^((a+b)^2-a^2-b^2-2*a*b)) -> 1

23 years agoupdated changes
Christian Bauer [Mon, 29 Jan 2001 21:42:36 +0000 (21:42 +0000)]
updated changes

23 years ago- Lewis-Wester tests D and E now benchmark the time to expand the
Richard Kreckel [Mon, 29 Jan 2001 18:52:35 +0000 (18:52 +0000)]
- Lewis-Wester tests D and E now benchmark the time to expand the
  rational function.  Just assembling it is boring...

23 years ago- gcd() and lcm() always checked the second argument, even when check_args
Christian Bauer [Thu, 25 Jan 2001 20:53:54 +0000 (20:53 +0000)]
- gcd() and lcm() always checked the second argument, even when check_args
  was false
- add::normal() handles large expressions (that collapse to much smaller ones)
  better

23 years agoreplaced fraction addition algorithm with a slightly faster one (no polynomial
Christian Bauer [Tue, 23 Jan 2001 19:14:36 +0000 (19:14 +0000)]
replaced fraction addition algorithm with a slightly faster one (no polynomial
divisions in loop)

23 years ago- See if __GNUC__ < 2.97 before using std::vector<..,malloc_alloc>. Sorry,
Richard Kreckel [Mon, 22 Jan 2001 18:11:54 +0000 (18:11 +0000)]
- See if __GNUC__ < 2.97 before using std::vector<..,malloc_alloc>.  Sorry,
  there is no way having this depend on whether we are setting up Cint or
  not because the malloc_alloc goes into the mangled signature and the
  result won't link if one tries to be more clever.  It really sucks.
  Let's all just pray Masaharu throws away that old STL implementation
  really soon now.  (Jeez, he still calls this old junk 'reference'!)
- Bumped up a few dates.  :-)

23 years ago- Updated Cint mileage report.
Richard Kreckel [Tue, 16 Jan 2001 21:06:47 +0000 (21:06 +0000)]
- Updated Cint mileage report.

23 years ago- Import namespace std globally.
Richard Kreckel [Sun, 7 Jan 2001 19:47:41 +0000 (19:47 +0000)]
- Import namespace std globally.

23 years agowhen normal() doesn't recur any further because the specified recursion level
Christian Bauer [Sat, 6 Jan 2001 19:18:29 +0000 (19:18 +0000)]
when normal() doesn't recur any further because the specified recursion level
was reached, the remaining subexpression is replaced by a symbol to avoid
continuing with non-polynomial expressions

23 years agojust a test
Christian Bauer [Wed, 3 Jan 2001 00:55:35 +0000 (00:55 +0000)]
just a test

23 years ago- cleanup, removed some obsolte comments and warnings.
Richard Kreckel [Tue, 2 Jan 2001 21:40:29 +0000 (21:40 +0000)]
- cleanup, removed some obsolte comments and warnings.

23 years ago- Add Denny-Fliegner test.
Richard Kreckel [Tue, 2 Jan 2001 17:48:50 +0000 (17:48 +0000)]
- Add Denny-Fliegner test.

23 years ago- GiNaC-cint: Change to new (still experimental) scheme for handling 2-arg
Richard Kreckel [Tue, 2 Jan 2001 17:48:31 +0000 (17:48 +0000)]
- GiNaC-cint: Change to new (still experimental) scheme for handling 2-arg
  operator new since <cln/object.h> defines such beasts.

23 years ago- power::derivative(): slight optimization.
Richard Kreckel [Tue, 2 Jan 2001 17:46:07 +0000 (17:46 +0000)]
- power::derivative(): slight optimization.

23 years ago-DIN_GINAC moved from CPPFLAGS to INCLUDES (as suggested in automake docs)
Christian Bauer [Sat, 30 Dec 2000 19:20:07 +0000 (19:20 +0000)]
-DIN_GINAC moved from CPPFLAGS to INCLUDES (as suggested in automake docs)

23 years agoman pages are included even if they are compressed
Christian Bauer [Sat, 30 Dec 2000 19:00:50 +0000 (19:00 +0000)]
man pages are included even if they are compressed

23 years ago- fixed typo.
Richard Kreckel [Fri, 22 Dec 2000 20:46:56 +0000 (20:46 +0000)]
- fixed typo.

23 years ago- utils.cpp, version.h.in: add version information to the library itself. release_0-7-0
Richard Kreckel [Fri, 15 Dec 2000 20:47:51 +0000 (20:47 +0000)]
- utils.cpp, version.h.in: add version information to the library itself.
- numeric.*: It was not a good idea to have a cast from numeric to cl_N,
  make a method to_cl_N, instead, in order not to confuse compilers.

23 years ago- Some minor fixes.
Richard Kreckel [Fri, 15 Dec 2000 20:46:48 +0000 (20:46 +0000)]
- Some minor fixes.

23 years ago- Makefile.am: target clean should remove G__*.
Richard Kreckel [Fri, 15 Dec 2000 20:46:29 +0000 (20:46 +0000)]
- Makefile.am: target clean should remove G__*.
- ginaccint.bin.cpp: check for recent enough cint.

23 years ago- Updated dependencies in INSTALL.
Richard Kreckel [Fri, 15 Dec 2000 20:45:59 +0000 (20:45 +0000)]
- Updated dependencies in INSTALL.
- Last fixes to NEWS.
- New config.guess and config.sub from FSF.
- Rewritten ginac.m4 to make it idiot-proof.

23 years ago- Eduardo Kalinowskis documentation add-on.
Richard Kreckel [Fri, 8 Dec 2000 03:41:56 +0000 (03:41 +0000)]
- Eduardo Kalinowskis documentation add-on.

23 years ago- ctor numeric from char* now honors Digits variable.
Richard Kreckel [Thu, 7 Dec 2000 23:13:51 +0000 (23:13 +0000)]
- ctor numeric from char* now honors Digits variable.

23 years agofixed typos
Christian Bauer [Tue, 5 Dec 2000 23:02:14 +0000 (23:02 +0000)]
fixed typos

23 years agoInclude files were not properly included in package
Christian Bauer [Tue, 5 Dec 2000 23:02:05 +0000 (23:02 +0000)]
Include files were not properly included in package

23 years ago- It's CLN, not LIBCLN. sigh^2.
Richard Kreckel [Tue, 5 Dec 2000 20:29:40 +0000 (20:29 +0000)]
- It's CLN, not LIBCLN.  sigh^2.

23 years ago- It's CLN, not LIBCLN. sigh.
Richard Kreckel [Tue, 5 Dec 2000 20:23:15 +0000 (20:23 +0000)]
- It's CLN, not LIBCLN. sigh.

23 years agoTransition to the (yet to be released) CLN 1.1.
Richard Kreckel [Sat, 25 Nov 2000 01:19:44 +0000 (01:19 +0000)]
Transition to the (yet to be released) CLN 1.1.

23 years agoTransition to the (yet to be released) CLN 1.1.
Richard Kreckel [Sat, 25 Nov 2000 01:19:44 +0000 (01:19 +0000)]
Transition to the (yet to be released) CLN 1.1.

23 years agoremoved non-sources from CVS
Christian Bauer [Thu, 16 Nov 2000 22:38:21 +0000 (22:38 +0000)]
removed non-sources from CVS

23 years agodisabled automake maintainer mode
Christian Bauer [Thu, 16 Nov 2000 22:38:13 +0000 (22:38 +0000)]
disabled automake maintainer mode

23 years agoremoved non-sources from CVS
Christian Bauer [Thu, 16 Nov 2000 22:12:24 +0000 (22:12 +0000)]
removed non-sources from CVS

23 years agoall man pages now display the current GiNaC version number
Christian Bauer [Thu, 16 Nov 2000 22:08:52 +0000 (22:08 +0000)]
all man pages now display the current GiNaC version number

23 years ago- added man page for ginac-config
Christian Bauer [Thu, 16 Nov 2000 22:08:37 +0000 (22:08 +0000)]
- added man page for ginac-config
- all man pages now display the current GiNaC version number

23 years agonew iterative mul::expand()
Alexander Frink [Tue, 14 Nov 2000 23:28:27 +0000 (23:28 +0000)]
new iterative mul::expand()

23 years agodoobiidoo
Alexander Frink [Tue, 14 Nov 2000 23:28:12 +0000 (23:28 +0000)]
doobiidoo

23 years ago- inifcns_trans.cpp: Added rules for atan(+/-1).
Richard Kreckel [Sun, 29 Oct 2000 18:15:51 +0000 (18:15 +0000)]
- inifcns_trans.cpp: Added rules for atan(+/-1).

23 years ago- INSTALL: Added success report for Cint 5.14.55,56.
Richard Kreckel [Sun, 29 Oct 2000 18:15:15 +0000 (18:15 +0000)]
- INSTALL: Added success report for Cint 5.14.55,56.
- config.guess, config.sub: Updated to newer version from FSF.

23 years ago- non-integer powers are treated as 0 by (l)degree() and coeff()
Christian Bauer [Wed, 25 Oct 2000 18:33:32 +0000 (18:33 +0000)]
- non-integer powers are treated as 0 by (l)degree() and coeff()

23 years ago- Corrected use of macro AM_PATH_GINAC.
Richard Kreckel [Tue, 24 Oct 2000 20:44:27 +0000 (20:44 +0000)]
- Corrected use of macro AM_PATH_GINAC.

23 years agoupdated a little
Christian Bauer [Wed, 11 Oct 2000 00:40:48 +0000 (00:40 +0000)]
updated a little

23 years agoupdated spec file which is now also included in the tarball
Christian Bauer [Thu, 5 Oct 2000 19:48:34 +0000 (19:48 +0000)]
updated spec file which is now also included in the tarball

23 years ago- config.guess, config.sub: upgraded from FSF.
Richard Kreckel [Thu, 5 Oct 2000 19:11:59 +0000 (19:11 +0000)]
- config.guess, config.sub: upgraded from FSF.

23 years ago- target distclean removes G__*, which are sometimes left behind. :-(
Richard Kreckel [Tue, 26 Sep 2000 18:22:43 +0000 (18:22 +0000)]
- target distclean removes G__*, which are sometimes left behind.  :-(

23 years ago- Updated Cint status.
Richard Kreckel [Tue, 26 Sep 2000 18:15:17 +0000 (18:15 +0000)]
- Updated Cint status.

23 years ago- pseries::print(): did not insert parenthesis when needed for precedence.
Richard Kreckel [Sat, 23 Sep 2000 21:36:50 +0000 (21:36 +0000)]
- pseries::print(): did not insert parenthesis when needed for precedence.
- pseries::precedence: was missing (thus being inherited from basic).
- add::print(): was not honoring upper_precedence for rest.

23 years ago- matrix::inverse(): Added a shortcut for sparse cases.
Richard Kreckel [Mon, 18 Sep 2000 22:31:26 +0000 (22:31 +0000)]
- matrix::inverse(): Added a shortcut for sparse cases.

23 years ago- matrix::gauss_elimination(): Added a shortcut for sparse cases.
Richard Kreckel [Mon, 18 Sep 2000 15:31:31 +0000 (15:31 +0000)]
- matrix::gauss_elimination(): Added a shortcut for sparse cases.

23 years ago- Fixed a memory corruption in matrix::determinant_minor().
Richard Kreckel [Tue, 29 Aug 2000 02:30:03 +0000 (02:30 +0000)]
- Fixed a memory corruption in matrix::determinant_minor().

23 years ago- removed non-sources from CVS
Christian Bauer [Mon, 28 Aug 2000 16:54:39 +0000 (16:54 +0000)]
- removed non-sources from CVS

23 years ago- Updated Cint status.
Richard Kreckel [Mon, 21 Aug 2000 20:11:19 +0000 (20:11 +0000)]
- Updated Cint status.

23 years ago- Duh, we had some style discussion again...
Richard Kreckel [Mon, 14 Aug 2000 19:44:58 +0000 (19:44 +0000)]
- Duh, we had some style discussion again...

23 years ago- removed the compatiblity stuff for broken slink linkers.
Richard Kreckel [Mon, 14 Aug 2000 18:52:39 +0000 (18:52 +0000)]
- removed the compatiblity stuff for broken slink linkers.

23 years ago- Inserted some more std:: to make it compile under GCC2.96.
Richard Kreckel [Mon, 14 Aug 2000 16:21:06 +0000 (16:21 +0000)]
- Inserted some more std:: to make it compile under GCC2.96.

23 years ago- Updated from FSF.
Richard Kreckel [Mon, 14 Aug 2000 16:20:38 +0000 (16:20 +0000)]
- Updated from FSF.

23 years ago- Corrected Cint 5.14.45 status.
Richard Kreckel [Mon, 14 Aug 2000 16:20:21 +0000 (16:20 +0000)]
- Corrected Cint 5.14.45 status.

23 years ago- Changes to make it more ANSI-conformant. Stuff detected while trying
Richard Kreckel [Sat, 12 Aug 2000 22:00:04 +0000 (22:00 +0000)]
- Changes to make it more ANSI-conformant.  Stuff detected while trying
  the compiler from Portland Group.

23 years ago- New file.
Richard Kreckel [Sat, 12 Aug 2000 21:59:17 +0000 (21:59 +0000)]
- New file.

23 years ago- Same shit as Christian did yesterday in ginac/.
Richard Kreckel [Sat, 12 Aug 2000 21:59:10 +0000 (21:59 +0000)]
- Same shit as Christian did yesterday in ginac/.

23 years ago- more indentation fixes
Christian Bauer [Fri, 11 Aug 2000 21:39:36 +0000 (21:39 +0000)]
- more indentation fixes

23 years ago- fixes to the indentation
Christian Bauer [Fri, 11 Aug 2000 19:52:41 +0000 (19:52 +0000)]
- fixes to the indentation

23 years ago- indentation is now done with tabs
Christian Bauer [Fri, 11 Aug 2000 19:07:15 +0000 (19:07 +0000)]
- indentation is now done with tabs
- printtree() prints the class names as known to the registrar, not the
  (mangled) class names from typeid().name()

23 years ago- Huh, \includegraphics doesn't like dots. Disabled. release_0-6-4
Richard Kreckel [Thu, 10 Aug 2000 18:17:38 +0000 (18:17 +0000)]
- Huh, \includegraphics doesn't like dots.  Disabled.

23 years ago- Complete revamp of methods in class matrix. Some redundant (and poor)
Richard Kreckel [Thu, 10 Aug 2000 17:14:39 +0000 (17:14 +0000)]
- Complete revamp of methods in class matrix.  Some redundant (and poor)
  implementations of elimination schemes were thrown out.  The code is now
  highly orthogonal, more flexible and much more efficient.  The function
  lsolve() now falls back to matrix::solve(), which in turn calls different
  elimination strategies.  matrix::fraction_free_elim() was thrown out for
  good.  matrix::pivot() has now two integer arguments (ro and co) to allow
  pivoting when eliminating to echelon form.
- Some new checks to guard against regressions in the new matrix code.
- Some minor bug-fixes in documentation.
- Bumped up version from 0.6.3 to 0.6.4.
- ...and the usual changes triggered by other changes.

23 years ago- fixed normal((b*a-c*a)/(4-a)) bug: heur_gcd() only works when both input
Christian Bauer [Fri, 4 Aug 2000 16:57:54 +0000 (16:57 +0000)]
- fixed normal((b*a-c*a)/(4-a)) bug: heur_gcd() only works when both input
  polynomials are non-zero

23 years ago- added check for normal((b*a-c*a)/(4-a)) bug
Christian Bauer [Fri, 4 Aug 2000 16:57:34 +0000 (16:57 +0000)]
- added check for normal((b*a-c*a)/(4-a)) bug

23 years ago- Check the normal(c-(b*a-c*a)/(4-a)) bug.
Richard Kreckel [Fri, 4 Aug 2000 16:37:25 +0000 (16:37 +0000)]
- Check the normal(c-(b*a-c*a)/(4-a)) bug.

23 years ago- Check *this for zeroness before .invert()ing it.
Richard Kreckel [Fri, 4 Aug 2000 16:36:55 +0000 (16:36 +0000)]
- Check *this for zeroness before .invert()ing it.

23 years ago- fixed "normal(2-2*(1+a)/(-1-a))" bug: add::normal() assumed that if the
Christian Bauer [Wed, 2 Aug 2000 19:38:01 +0000 (19:38 +0000)]
- fixed "normal(2-2*(1+a)/(-1-a))" bug: add::normal() assumed that if the
  common denominator is 1, all denominators would be 1 (they can be +/-1)

23 years ago- added check for "normal(2-2*(1+a)/(-1-a))" bug
Christian Bauer [Wed, 2 Aug 2000 19:37:13 +0000 (19:37 +0000)]
- added check for "normal(2-2*(1+a)/(-1-a))" bug

23 years ago- I'm fired. Fixed a really stupid bug in sr_gcd() which called
Christian Bauer [Mon, 31 Jul 2000 21:09:44 +0000 (21:09 +0000)]
- I'm fired. Fixed a really stupid bug in sr_gcd() which called
  divide_in_z(r, foo) with "var+1" as its last parameter, while the main
  variable x=*var usually appears in r. This should never have worked...

23 years ago- Fixed an assertion-thinko in matrix::fraction_free_elimination().
Richard Kreckel [Mon, 31 Jul 2000 18:18:27 +0000 (18:18 +0000)]
- Fixed an assertion-thinko in matrix::fraction_free_elimination().

23 years ago- Bernard Parisse's patch for Order_eval().
Richard Kreckel [Mon, 31 Jul 2000 18:15:29 +0000 (18:15 +0000)]
- Bernard Parisse's patch for Order_eval().

23 years ago- Fixed a thinko in atan(const numeric &).
Richard Kreckel [Thu, 27 Jul 2000 10:51:28 +0000 (10:51 +0000)]
- Fixed a thinko in atan(const numeric &).
- Many more functions throw pole_errors now, instead of domain_errors
  or other exceptions.

23 years ago- Bumped up version to 0.6.3. release_0-6-3
Richard Kreckel [Wed, 26 Jul 2000 02:01:14 +0000 (02:01 +0000)]
- Bumped up version to 0.6.3.

23 years ago- Inserted date.
Richard Kreckel [Wed, 26 Jul 2000 01:59:56 +0000 (01:59 +0000)]
- Inserted date.

23 years ago- Made determinant_algo (in flags.h) really work.
Richard Kreckel [Wed, 26 Jul 2000 00:49:22 +0000 (00:49 +0000)]
- Made determinant_algo (in flags.h) really work.
- Fixed minor bugs in matrix.cpp.
- Small Documantation updates.

23 years ago- Fixed the bug that broke xloop's po_redux.
Richard Kreckel [Sat, 22 Jul 2000 04:47:04 +0000 (04:47 +0000)]
- Fixed the bug that broke xloop's po_redux.
- Interface Change: No argument 'bool branchcut' for series methods,
  instead 'unsigned options'.  Sorry.
: ----------------------------------------------------------------------

23 years ago- Derivatives are now assembled in a slightly different manner (i.e. they
Richard Kreckel [Fri, 21 Jul 2000 01:12:25 +0000 (01:12 +0000)]
- Derivatives are now assembled in a slightly different manner (i.e. they
  'look' different on first sight).  Under certain circumstances this can
  result in a dramatic speedup because it gives hashing a better chance,
  especially when computing higher derivatives.
- Series expansion accepts an additional bool argument now, telling it
  whether branch cuts are to be honored (as Maple does it) or not (as
  Mathematica does it).  It defaults to true (i.e. the Maple way).
- Many series expansions of built-in functions have been reengineered.
- The Dilogarithm (Li2) now has floating point evaluation, derivative and a
  proper series expansion.
- Namespace 'std' cleanly disentangled, as demanded by ISO/EIC 14882-1998(E).
- Some helpers updated from upstream.

23 years ago- added a warning that things like sin(Li2(x)).series(x==0) cannot work
Richard Kreckel [Fri, 7 Jul 2000 00:40:25 +0000 (00:40 +0000)]
- added a warning that things like sin(Li2(x)).series(x==0) cannot work
  without having proper support for limits.

23 years ago- added a test for Li2(sin(x)).series(x==0,8).
Richard Kreckel [Fri, 7 Jul 2000 00:39:46 +0000 (00:39 +0000)]
- added a test for Li2(sin(x)).series(x==0,8).

23 years ago- Introduced derivative of Li2.
Richard Kreckel [Thu, 6 Jul 2000 23:34:42 +0000 (23:34 +0000)]
- Introduced derivative of Li2.
- Introduced nontrivial part of series expansion of Li2 around 0.  (Read the
  comments in the source to see why it's nontrivial!)  Missing: the branch
  point and cut.

23 years ago- Patches needed for cygwin.
Richard Kreckel [Mon, 3 Jul 2000 23:43:17 +0000 (23:43 +0000)]
- Patches needed for cygwin.

23 years ago- suppress zero-terms in expansion.
Richard Kreckel [Mon, 3 Jul 2000 23:42:50 +0000 (23:42 +0000)]
- suppress zero-terms in expansion.

23 years ago- version.h: change triggered. release_0-6-2
Richard Kreckel [Thu, 22 Jun 2000 00:09:21 +0000 (00:09 +0000)]
- version.h: change triggered.

23 years ago- stamp-vti, version.texi: bumped up version to 0.6.2.
Richard Kreckel [Thu, 22 Jun 2000 00:09:08 +0000 (00:09 +0000)]
- stamp-vti, version.texi: bumped up version to 0.6.2.

23 years ago- exam_mansamples.cpp, exam_paranoia.cpp: fixes and extensions.
Richard Kreckel [Thu, 22 Jun 2000 00:08:50 +0000 (00:08 +0000)]
- exam_mansamples.cpp, exam_paranoia.cpp: fixes and extensions.
- ginaccint.bin.cpp: extended "silent" behaviour.

23 years ago- configure, configure.in: Bumped version up to 0.6.2.
Richard Kreckel [Thu, 22 Jun 2000 00:08:09 +0000 (00:08 +0000)]
- configure, configure.in: Bumped version up to 0.6.2.
- INSTALL, NEWS: Mention new launcher for ginaccint.bin.

23 years ago- '#' as the first character of a line starts a comment
Christian Bauer [Wed, 21 Jun 2000 21:05:38 +0000 (21:05 +0000)]
- '#' as the first character of a line starts a comment

23 years ago- constructor from strings once again accepts Lisp-style numbers like
Christian Bauer [Wed, 21 Jun 2000 21:05:17 +0000 (21:05 +0000)]
- constructor from strings once again accepts Lisp-style numbers like
  "#b101010" and "#xEA31"