3.0200 June 18, 2010
- Added skewness and kurtosis
- https://rt.cpan.org/Ticket/Display.html?id=58187
- Thanks to Shawn Laffan.
3.0102 June 15, 2010
- Add the $VERSION variable to Statistics::Descriptive::Sparse and
Statistics::Descriptive::Full. This was done to silence the CPAN indexer.
3.0101 June 15, 2010
- Moved the trimmed_mean caching test (that used the Benchmark.pm module)
to rejects/descr.t , because it kept failing.
math/p5-Math-Random-MT.
The Mersenne Twister is a pseudorandom number generator developed by
Makoto Matsumoto and Takuji Nishimura. It is described in their paper at
<URL:http://www.math.keio.ac.jp/~nisimura/random/doc/mt.ps>.
CHANGES MADE TO MATHOMATIC 15.2.0 TO BRING IT UP TO THE NEXT VERSION:
All makefiles were improved. Library test/example program is renamed to "testmain".
Package maintainers please take note: support for the DESTDIR environment variable was
added to the makefiles; for proper operation when packaging version 15.2.1 or higher,
please remove any patches for missing DESTDIR support.
m4 Mathomatic should work now when included in the Mathomatic package (make m4install).
Thank you for packaging Mathomatic! If I did anything wrong, please let me know.
8/26/10 - Added the -e option, which processes mathematical expressions and Mathomatic commands
instead of input files on the shell command line. For example, entering
"mathomatic -eq 2+3" gives "answer = 5". This functionality has been requested
many times by Mathomatic command line users. A complete example:
CHANGES MADE TO MATHOMATIC 15.1.6 TO BRING IT UP TO THE NEXT VERSION:
Minor improvements were made to the user documentation.
8/22/10 - Removed "Complex number roots approximated" warning message, since this happens often.
Capitalized E, I, PI, and Pi are now accepted as the universal constants e, i, and pi,
without needing to enter "set no case". This allows Mathomatic to easily
accept Mathematica style expression input.
m4 Mathomatic now additionally accepts Mathematica style capitalized function input.
matho and rmath now display elapsed, CPU, and system times in seconds upon exit.
8/23/10 - Fixed #equation-number entry at the main prompt to always work and allow an expression
or equation following on the same line to be entered at that equation space.
For example: "#10 y=1/x" will work now;
previously only worked if equation space number 10 was previously allocated and used.
The way it works is: all equation spaces up to and including number 10 are allocated,
if not already allocated, upon entry of "#10".
Equations spaces are allocated with the memory allocator malloc(3).
CHANGES MADE TO MATHOMATIC 15.1.5 TO BRING IT UP TO THE NEXT VERSION:
7/26/10 - Disabled ncurses call for auto-color detection when CYGWIN is defined while
compiling the source code, due to a reported problem of readline failing with
ncurses in Cygwin.
7/28/10 - Disabled readline history save file for the Cygwin port,
because it is a filename that starts with a period.
7/31/10 - Added "set fractions_display" option, to allow disabling the automatic conversion of
fractions like .5 to 1/2 for display.
Developer requested and useful in the symbolic math library,
when numerical fraction output isn't wanted.
8/01/10 - Fixed a memory leak when ignoring the output string in the symbolic math library.
8/04/10 - Preserve overflowed powers like 2^2222 rather than aborting with an error message.
Allow simplification of math like 2*2^2222 and 2/2^2222.
8/08/10 - matho-primes runs twice as fast with the -ffast-math gcc compilation option,
which is now enabled by default. Don't try -ffast-math with the main Mathomatic
program though, because then Mathomatic won't work properly.
CHANGES MADE TO MATHOMATIC 15.1.4 TO BRING IT UP TO THE NEXT VERSION:
Many minor tweaks and improvements.
7/03/10 - Makefiles and compile scripts were corrected and enhanced
per http://www.gnu.org/prep/standards/
7/06/10 - Changed all "#if true" and "#if false" conditional commenting to "#if 1" and "#if 0"
in the C source code, thanks to Min Sik Kim of NetBSD pkgsrc.
CHANGES MADE TO MATHOMATIC 15.1.3 TO BRING IT UP TO THE NEXT VERSION:
All of the Unix man pages and user manuals for Mathomatic were fixed.
The compare and "solve verify" commands now simplify more thoroughly with "repeat simplify"
for better expression equality determination.
6/17/10 - Greatly improved file operation error reporting by using the perror(3) function.
6/18/10 - Fixed categories in "icons/mathomatic.desktop";
Mathomatic now goes under valid categories, mainly Education.
CHANGES MADE TO MATHOMATIC 15.1.2 TO BRING IT UP TO THE NEXT VERSION:
6/6/10 - I made mistakes in the improvement to the simplify command of version 15.1.2,
the original working simplify logic of version 15.1.1 is now restored, sorry.
CHANGES MADE TO MATHOMATIC 15.1.1 TO BRING IT UP TO THE NEXT VERSION:
A general cleanup was done.
A small improvement was made to the final result of the simplify and fraction commands.
Showing intermediate results in the calculate, sum, and product commands is now done with "set debug 1".
5/28/10 - Added "tests/collatz.in", the Collatz conjecture as an automatically computable equation.
CHANGES MADE TO MATHOMATIC 15.1.0 TO BRING IT UP TO THE NEXT VERSION:
Code and documentation cleanup.
5/21/10 - Added "primes/matho-sum", a utility that sums its command line arguments or standard input.
Use "matho-primes 0 2000000 | matho-sum" to find the sum of all primes less than 2,000,000.
Solves Project Euler problem #10: http://projecteuler.net/index.php?section=problems&id=10
5/23/10 - Primes (') are allowed in variable names now, if not using the symbolic math library, so that the
derivative, integrate, and nintegrate commands can change the dependent variable to y', y'', etc.
This can be turned on in the symbolic math library by the command "set special_variable_characters='".
Non-alphanumeric characters in variable names are now converted to underline characters (_)
when exporting to a programming language or to a different program.
CHANGES MADE TO MATHOMATIC 15.0.8 TO BRING IT UP TO THE NEXT VERSION:
5/10/10 - Added "help constants" command.
5/11/10 - Integer variables are now specified by using a variable name that starts with "integer",
like "integer1", "integer_x", etc. Currently only the modulus operator "%" checks
for integer variables, to help with simplification.
5/12/10 - Corrected the output string type of the symbolic math library API. It was erroneously declared
as type "const", possibly causing a memory leak.
5/13/10 - Moved and adapted "makefile.lib" to "lib/makefile", so the symbolic math library build is isolated.
Previously "make clean" was required between different builds. All makefiles require GNU make now.
CHANGES MADE TO MATHOMATIC 15.0.7 TO BRING IT UP TO THE NEXT VERSION:
4/26/10 - Allow "make pdf" to generate PDF documentation from the HTML documentation with htmldoc.
Please read the comments in the makefile for all available options.
4/29/10 - For every makefile, CFLAGS has been modified to include OPTFLAGS as required by the Fedora Linux
build system, and OPTFLAGS defaults to the optional gcc specific flags like optimization.
In the symbolic math library, made available the equation number of the result of calling the API,
if also stored in an equation space. The result equation number is stored in the global "result_en".
Useful if you want to know where the result was stored, to act on it with further commands.
CHANGES MADE TO MATHOMATIC 15.0.6 TO BRING IT UP TO THE NEXT VERSION:
Corrections and improvements to the documentation were made.
4/3/10 - Vastly improved the "factor number" user interface, now factors integer expressions like 2^32-1.
"factor number" is disabled in library mode.
4/13/10 - The real and imaginary commands no longer fail when the expression is not complex,
just a warning is given.
4/14/10 - Changed normal display of "-1*" to "-", for prettier 2D expression output,
so things like "-a" display properly, not as "-1*a".
CHANGES MADE TO MATHOMATIC 15.0.5:
1/28/10 - Added a Python utility called "primorial" to the Prime Number Tools install
that multiplies together the results of matho-primes, displaying the primorials
of the integers given on the command line.
3/18/10 - Changed author email address to "gesslein@linux.com".
3/23/10 - Catch SIGHUP and SIGTERM signals for proper termination of the Mathomatic program;
readline was messing up when Mathomatic was terminated by closing the shell window.
The plot command now always plots expressions with grid marks displayed for reference.
CHANGES MADE TO MATHOMATIC 15.0.4:
1/21/10 - In the makefile, changed the HTML man page generator back to rman because
groff HTML output looks really bad and rman allows linking to other man pages.
groff is no longer used.
1/24/10 - Fixed "make m4install", the installed rmath and matho programs weren't working.
1/27/10 - Added GNU LGPL license notices to every C source file with a copyright notice,
for proper protections.
CHANGES MADE TO MATHOMATIC 15.0.3:
1/9/10 - Fixed a problem only in the version 15.0.3 makefile, where it didn't respect
the CC environment variable set by the user, instead it always used "gcc"
as the C compiler.
CHANGES MADE TO MATHOMATIC 15.0.2:
12/27/09 - Moved get_screen_size() from main.c to am.c because it is used in the library when
compile-time options UNIX or CYGWIN are defined.
Thanks to Cygwin port maintainer Reini Urban for noticing and fixing this problem.
Defining UNIX or CYGWIN in library mode is not recommended.
12/31/09 - Added code to allow any command to be preceded by "repeat", which sets the
repeat flag for the following command. Most commands ignore the repeat flag.
1/1/10 - Ported divide and roots commands to be repeatable. Also repeatable are the
calculate and eliminate commands.
1/2/10 - Allow Taylor series computation even if the specified differentiation variable
is not found in the expression, giving a warning.
Ported simplify command to be a repeatable full simplify; that is, typing
"repeat simplify" repeatedly runs the simplify command until the result stabilizes
to the smallest size expression.
1/7/10 - Thanks to pretty C code submitted by Simon Geard,
the code and variables commands have been made much more readable.
1/8/10 - Made Mathomatic easier to compile under Solaris, thanks to Michael Pogue of Sun.
Fixed failure to compile under BSD Unix when compiling with readline support.
CHANGES MADE TO MATHOMATIC 15.0.1:
12/19/09 - The last few versions fix the ugliness caused by the GCD factoring change made on 6/22/09.
Today's change factors out the numerical GCD of rational coefficients as needed to simplify.
Most simplification results should be beautiful and the simplest possible again now,
without the misleading observed magnitude caused by always factoring out the GCD,
which was why the change of 6/22/09 was made.
CHANGES MADE TO MATHOMATIC 15.0.0:
12/12/09 - Fixed a problem with the -q (quiet mode) option being ignored if the session options
were ever saved with the "set save" command.
12/13/09 - Added code to allow Mathomatic output to be redirected by default.
Fixed the derivative command to be successful even when the result is 0,
when compiled as a library.
12/14/09 - Added ability to log symbolic math library results, and made command behavior
more consistent in the library by always returning the final result string.
12/16/09 - The factor command now factors more by factoring out the GCD of rational coefficients.
CHANGES MADE TO MATHOMATIC 14.6.3 TO BRING IT UP TO 15.0.0:
Cleanup and more bug fixes.
11/26/09 - Added detection of the terminal's ANSI color availability, when readline is enabled.
11/28/09 - Added detection of divide by zero and NaN when using the "solve verify" command,
for more correct results.
Solving now factors out the GCD of rational coefficients, for improved results.
The fixes today are from errors solving equations like (2*x/(x - 3)) + 3 = 6/(x - 3)
11/29/09 - Disallow the variable named "nan". NaN cannot be directly entered into Mathomatic.
11/30/09 - Added shell scripts "t" and "tests/t" to easily test Mathomatic by typing "./t".
12/2/09 - The fraction command now factors out the GCD of rational coefficients like the
solve command does, so that coefficients in algebraic fractions become integers.
The GCD verifying routine was perfected by making it very strict, like it should be.
From PR pkg/43790 by Kamel Derouiche
pkgsrc changes:
- re-set LICENSE (modified-bsd).
upstream changes:
Networkx-1.2
Release date: 28 July 2010
See: https://networkx.lanl.gov/trac/timeline
New features
* Ford-Fulkerson max flow and min cut
* Closness vitality
* Eulerian circuits
* Functions for isolates
* Simpler s_max generator
* Compatible with IronPython-2.6
* Improved testing functionality: import networkx; networkx.test() tests
entire package and skips tests with missing optional packages
* All tests work with Python-2.4
* and more, see
https://networkx.lanl.gov/trac/query?status=closed&group=milestone&milestone=networkx-1.2
Networkx-1.1
Release date: 21 April 2010
See: https://networkx.lanl.gov/trac/timeline
New features
* Algorithm for finding a basis for graph cycles
* Blockmodeling
* Assortativity and mixing matrices
* in-degree and out-degree centrality
* Attracting components and condensation.
* Weakly connected components
* Simpler interface to shortest path algorithms
* Edgelist format to read and write data with attributes
* Attribute matrices
* GML reader for nested attributes
* Current-flow (random walk) betweenness and closeness.
* Directed configuration model, and directed random graph model.
* Improved documentation of drawing, shortest paths, and other
algorithms
* Many more tests, can be run with ?import networkx; networkx.test()?
* and much more, see
https://networkx.lanl.gov/trac/query?status=closed&group=milestone&milestone=networkx-1.1
API Changes
Returning dictionaries
Several of the algorithms and the degree() method now return dictionaries keyed
by node instead of lists. In some cases there was a with_labels keyword which is
no longer necessary. For example,
>>> G=nx.Graph()
>>> G.add_edge('a','b')
>>> G.degree() # returns dictionary of degree keyed by node
{'a': 1, 'b': 1}
Asking for the degree of a single node still returns a single number
>>> G.degree('a')
1
The following now return dictionaries by default (instead of lists) and the
with_labels keyword has been removed:
* Graph.degree(), MultiGraph.degree(), DiGraph.degree(),
DiGraph.in_degree(), DiGraph.out_degree(), MultiDiGraph.degree(),
MultiDiGraph.in_degree(), MultiDiGraph.out_degree().
* clustering(), triangles()
* node_clique_number(), number_of_cliques(), cliques_containing_node()
* eccentricity()
The following now return dictionaries by default (instead of lists)
* pagerank()
* hits()
Adding nodes
add_nodes_from now accepts (node,attrdict) two-tuples
>>> G=nx.Graph()
>>> G.add_nodes_from([(1,{'color':'red'})])
Examples
* Mayvi2 drawing
* Blockmodel
* Sampson?s monastery
* Ego graph
Bug fixes
* Support graph attributes with union, intersection, and other graph
operations
* Improve subgraph speed (and related algorithms such as
connected_components_subgraphs())
* Handle multigraphs in more operators (e.g. union)
* Handle double-quoted labels with pydot
* Normalize betweeness_centrality for undirected graphs correctly
* Normalize eigenvector_centrality by l2 norm
* read_gml() now returns multigraphs
Changes in Maxima 5.22.1
New items in core:
* function gensym: new
* function makelist: extensions
Other revisions in core:
* improvements in specint (Laplace transform)
* improvements in Xmaxima user interface
* inverse properties for erf and friends
* function transpose: faster algorithm
Other revisions in share:
* package graphs: new graph definitions, tutte and flow polynomials, bug fixes, other updates
* package draw: new object geomap, rename mesh to elevation_grid, other updates
* package descriptive: new functions smin, smax; functions mini, maxi now deprecated
* package simplify_sum: try to detect telescoping sums
* package ezunits: additional conversions, faster conversion algorithm
* package lapack: new function dgesv
* package colnew: improved interface functions
* package dynamics: bug fix
* package mnewton: bug fix
* package finance: update
Bug fixes:
3035313 - some array references translated incorrectly
3029610 - integrate and %e_to_numlog
3025038 - gruntz needs logexpand:true
3024797 - @POSIX_SHELL@ always substitued by /bin/sh in configure.lisp
3014545 - submatrix does not work as expected
3012427 - tex2ooo.lisp invalid output
3010829 - numerical evaluation of elliptic_ec fails for argument > 1
3010525 - abs documentation doesn't explain mapping behavior
3009011 - plot3d syntax
3007061 - log(x^2) simplifies to 2 log(x) for all x
3005820 - full_listify
3002971 - limit fails where rat+subst works
2999635 - trigrat(sin(1)) makes mess
2998621 - conjugate(atan(x+%i*y)) wrong
2997276 - zeta(3),numer; gives Lisp error
2996542 - log(x) integration is incorrect.
2996106 - at(diff(f(x,y),x,1,y,1),[x=a,y=b]) is wrong
2996065 - Unevaluated numerical hypergeometric() call
2992398 - sort doesn't give error for invalid comparison
2991924 - Incorrect integration of rational functions
2990307 - error for solve_rat_ineq
2988544 - integrate(signum(abs(x)),x,-2,2);
2988190 - atan2(1b20,-1b0); badly wrong
2985866 - derivatives of functions of taylor polys
2933097 - polyfactor not documented
2905526 - lcm(0,6,0) --> divide by zero
2806446 - ev_diff in vect.mac
2805600 - depends() partially prevents diff() to work
2789110 - solve, tan and atan depend on order of variables
2784283 - psi[0](negative float)
2781127 - bfpsi0 of complex
2556133 - "at" should do parallel substitutions
2036462 - Very long calculation time, normal ?
2014941 - compositions of 'at'
2011228 - vect redefines "." as commutative, was:Matrix multiplication
1994295 - errormsg
1961494 - translated functions & values list
1959214 - integrate() and array having lisp style name
1928142 - keepfloat breaks ratsubst() in some cases
1758005 - compiler warning for rat3c
1677217 - composistions of 'at'
1663385 - declare multiplicative - wrong simplification
1281740 - declare/featurep - strange behaviour
1212598 - bug in the VECT.MAK - VECTORSIMP cross product
1117533 - letsimp complains about assignment to %pi
1114128 - 2nd argument in permanent
1078046 - Alias atoms displayed after unorder()
1046653 - input prompt appearing when it should not
1003494 - scalefactors broken Maxima 5.9.0 (win)
947808 - logcontract and ratfac
838301 - vect negate cross product simplification
754220 - Featurep and Declare inconsistent /FIX
742909 - trigrat(sin(x/2)) makes a mess
unnumbered - is(f("x")>0) causes an error.
to trigger/signal a rebuild for the transition 5.10.1 -> 5.12.1.
The list of packages is computed by finding all packages which end
up having either of PERL5_USE_PACKLIST, BUILDLINK_API_DEPENDS.perl,
or PERL5_PACKLIST defined in their make setup (tested via
"make show-vars VARNAMES=..."), minus the packages updated after
the perl package update.
sno@ was right after all, obache@ kindly asked and he@ led the
way. Thanks!
COMMENT.
Changes:
Andreas:
* Clarify autofilter dialog. [#625688]
* Make the tooltips theme-independent. [#625966]
* Adjust and fix WORKDAY and NETWORKDAY.
* Fix crash on certain handcrafted ODF files. [#626632]
* Fix import of some ODF 1.0 charts. [#626632]
* Fix crash on paste into an empty expression entry. [#626648]
* Import and export boxplots to and from ODF files. [#626653]
* Improve handling of corrupted ODF files. [#626654][#626655]
* Fix autocorrection when entering in ranges. [#626781]
* Fix column wdth and row height adjustments when entering in ranges.
[#626789]
* Fix autofitting on data entry. [#626861]
* Avoid some of the multiple undos when entering existing text in
ranges. [#626797]
* Adjust formats when autofitting new entries. [#626903]
* Autofit rows and columns when entering an array expression. [#626904]
* Fix column and row fitting on paste. [#359404][#611181]
* Export images to ODF.
* Fix import of graphs from rich ODF files. [#626961]
* Some ring and pie plot ODF export improvements.
Jean:
* Import Guppi graphs from gnumeric-1.0.x files. [#567953]
* Survive malformed guppi graphs. [#625726]
* Use theme foreground color for row and column headers. [#625727]
* Force column span recalculation when switching between number and
error. [#625994]
* Fix scroll when selecting a merged cell. [#626065]
* Fix objects dragging when zoomed. [#626484]
* Fix control points behavior when zooming.
* Fix in-cell editing when zoom < 100%. [#626279]
Morten:
* Survive malformed xml files.
* Fix win32 random problem. [#533779]
* Fix IMPOWER problem. [#627014]
* Avoid criticals in ODS import.
* Fix problem with collect cache. [#627079]
Based on PR#43736 by Kamel Derouiche.
V 4.9.5 - Dec 10, 2007 Updated software license, allowing distribution
of modified source code.
The actual library source code is identical to
version 4.9.2 (except the obvious library version
identifier, and Digital Mars #ifdef).
Added support for the Digital Mars Compiler.
Added build script for Intel Linux Compiler.
V 4.9.5a - Feb 21, 2010 Updated authors email address.
Note that there are no changes to the library
source code - so the library will still identify
itself as "4.9.5".
Added Makefile for Microsoft Visual Studio 2005.
Added another Makefile for MAC OSX (for 10.6).
Changes in muParser 1.32
------------------------
Changes:
* "example3" renamed to "example2"
* Project/Makefiles files are now provided for:
- msvc2003
- msvc2005
- msvc2008
- watcom (makefile)
- mingw (makefile)
- bcc (makefile)
* Project files for borland cpp builder were removed
New features:
* Added function returning muparsers version number
* Added function for resetting the locale
Bugfixes:
* Changes example1 in order to fix issues with irritating memory leak reports.
Added conditional code for memory leak detection with MSVC in example1.
(see: http://www.codeproject.com/KB/recipes/FastMathParser.aspx?msg=3286367#xx3286367xx)
* Fixed some warnings for gcc
Changes in muParser 1.31cp
--------------------------
Changes:
* Archive structure changed
* C# wrapper added
* Fixed issued that prevented compiling with VS2010 Beta2
Contributed by Kamel Derouiche via PR pkg/43741
Changes in this version:
It works with Borland Builder version 6, 8;
Microsoft Visual C++ 6,7,8; Open Watcom 1.7a;
Gnu version 3, 4 and Intel for Windows & Linux compilers 9, 10.
Options and work-arounds for older compilers are being removed.
You can enter values from an integer array with statements like
Matrix A(3,2); int a[] = { 11,12,21,22,31,33 }; A << a;
There is a new matrix type SquareMatrix. You can use this where
you know a matrix is square. The only place where the information
that a matrix is used is in the Kronecker product KP(A,B).
If A is band and B is any square matrix type (i.e. type SquareMatrix
or band, triangular, symmetric or diagonal) the result is band.
If B is type Matrix then the result is of type Matrix.
Apart from this it may be worth using the SquareMatrix type to
improve readability of a program and to make sure that a matrix
that is intended to be square really is square.
There is a new routine for carrying out cross products of
ColumnVectors or RowVectors of length 3 and new routines for
updating a triangular matrix derived from a Cholesky decomposition.
Interfacing with old C functions involving one and two
dimensional arrays is simplified.
Additional QR functions, extend_orthogonal function.
Lower-case versions of functions.
You can use the manipulators scientific and fixed in the matrix
output expressions.
changes:
-Add new Excel and Openformula units ("pica" and "picapt")
-New functions DATE2HDATE, DATE2HDATE_HEB and DATE2JULIAN
-UI improvements
-bugfixes
pkgsrc changes:
- adjust dependencies
- remove executable file bits from perl modules
- adjust installer type
Upstream changes:
0.32 Mon Jul 5 18:58:05 CEST 2010
Official release for development release 0.31_1
0.31_1 Mon Jun 28 16:51:44 CEST 2010
- Add autofilter to worksheet
- Swichting to Module::Install
pkgsrc changes:
- add license definition
- adjust regex for removing instant time point in version number
Upstream changes:
- 1.6.A6FGHKE Tue Jun 15 16:17:20:14 2010
* bumped minor version number so they'll keep ascending (without PT
comprehension)
as math/mpcomplex.
Packaged by Marko Schuetz for pkgsrc-wip.
MPC is a C library for the arithmetic of complex numbers with
arbitrarily high precision and correct rounding of the result.
It is built upon and follows the same principles as MPFR.
changes:
-new compressor called Blosc (speed-up)
-A new `tables.Expr` module (based on Numexpr) that allows to do
persistent, on-disk computations on many algebraic operations.
-Support for HDF5 hard links, soft links and automatic external links
-Suport for 'fancy' indexing
-bugfixes
Contributed by Stathis Kamperis on IRC.
Changes from versions 2.4.* to version 3.0.0:
- The "boudin aux pommes" release.
- MPFR 3.0.0 is binary incompatible with previous versions but (almost)
API compatible. More precisely the obsolete functions mpfr_random
and mpfr_random2 have been removed, the meaning of the return type
of the function mpfr_get_f has changed, and the return type of the
function mpfr_get_z is now int instead of void. In practice, this
should not break any existing code.
- MPFR is now distributed under the GNU Lesser General Public License
version 3 or later (LGPL v3+).
- Rounding modes GMP_RNDx are now MPFR_RNDx (GMP_RNDx kept for
compatibility).
- A new rounding mode (MPFR_RNDA) is available to round away from zero.
- The rounding mode type is now mpfr_rnd_t (as in previous versions,
both mpfr_rnd_t and mp_rnd_t are accepted, but mp_rnd_t may be
removed in the future).
- The precision type is now mpfr_prec_t (as in previous versions, both
mpfr_prec_t and mp_prec_t are accepted, but mp_prec_t may be removed
in the future) and it is now signed (it was unsigned in MPFR 2.*, but
this was not documented). In practice, this change should not affect
existing code that assumed nothing on the precision type.
- MPFR now has its own exponent type mpfr_exp_t, which is currently
the same as GMP's mp_exp_t.
- Functions mpfr_random and mpfr_random2 have been removed.
- mpfr_get_f and mpfr_get_z now return a ternary value.
- mpfr_strtofr now accepts bases from 37 to 62.
- mpfr_custom_get_mantissa was renamed to mpfr_custom_get_significand
(mpfr_custom_get_mantissa is still available via a #define).
- Functions mpfr_get_si, mpfr_get_ui, mpfr_get_sj, mpfr_get_uj,
mpfr_get_z and mpfr_get_z_2exp no longer have cases with undefined
behavior; in these cases, the behavior is now specified, and in
particular, the erange flag is set.
- New functions mpfr_buildopt_tls_p and mpfr_buildopt_decimal_p giving
information about options used at MPFR build time.
- New function mpfr_regular_p.
- New function mpfr_set_zero.
- New function mpfr_digamma.
- New function mpfr_ai (incomplete, experimental).
- New functions mpfr_set_flt and mpfr_get_flt to convert from/to the
float type.
- New function mpfr_urandom.
- New function mpfr_set_z_2exp (companion to mpfr_get_z_2exp, which
was renamed from mpfr_get_z_exp in previous versions).
- Speed improvement for large operands in the trigonometric functions
(mpfr_sin, mpfr_cos, mpfr_tan, mpfr_sin_cos): speedup of about 2.5
for 10^5 digits, of about 5 for 10^6 digits.
- Speed improvement for large operands of the inverse trigonometric
functions (arcsin, arccos, arctan): about 2 for 10^3 digits, up to
2.7 for 10^6 digits.
- Some documentation files are installed in $docdir.
- The detection of a GMP build directory (more precisely, the internal
header files of GMP) was previously done separately from the use of
the --with-gmp-build configure option. This was not consistent with
the documentation and with other parts of the configure script. So,
as of MPFR 3.0.0, the internal header files of GMP are now used if
and only if the --with-gmp-build configure option is given.
- The configure script recognizes some extra "long double" formats
(double big endian, double little endian, double-double big endian).
- MPFR manual: added "API Compatibility" section.
- Test coverage: 97.1% lines of code.
- Bug fixes.
The following suffixes for variables and constraints were
implemented in the MathProg language:
.lb (lower bound),
.ub (upper bound),
.status (status in the solution),
.val (primal value), and
.dual (dual value).
Now the MathProg language allows comment records (marked by
'#' in the very first position) in CSV data files read with the
table statements. Note that the comment records may appear only
in the beginning of a CSV data file.
The API routine glp_cpp to solve the Critical Path Problem was
added and documented.
Notable changes (compared to version 1.0.9) include:
- New domains and packages: VectorSpaceBasis domain,
DirichletRing domain, 3D graphic output in Wavefront .obj format,
specialized machine precision numeric vectors and matrices (faster
then general vectors and matrices), Html output.
- Support Clifford algebras corresponding to non-diagonal matrix,
added new operations.
- 'normalize' now tries to simplify logarithms of algebraic constants.
- New functions: Fresnel integrals, carmichaelLambda.
- Speed improvements: several polynomial operations are faster,
faster multiplication in Ore algebras, faster computation of
strong generating set for permutation groups, faster coercions.
- Several improvements to the guessing package (in particular new
option Somos for restricting attention to Somos-like
sequences
Bug fixes, in particular:
- FriCAS can now compute multiplicative inverse of a power series
with constant term not equal to 1.
- Fixed a problem with passing interpreter functions to algebra.
- Two bugs causing crashes in HyperDoc interface are fixed.
- FriCAS now ignores sign when deciding if number is prime.
- A failing coercion that used to crash FriCAS is now detected.
- 'has' test sometimes gave wrong result.
- Plotting fixes.
* R CMD INSTALL checks if dependent packages are available early on
* in the installation of source packages, thereby giving clearer
error messages.
* R CMD INSTALL --build now names the file in the format used
for Mac OS X binary files on that platform.
* BIC() in package stats4 now also works with multiple fitted models,
analogously to AIC().
* Use of file extension .C for C++ code in packages is now
deprecated: it has caused problems for some 'make's on
case-insensitive file systems (although it currently works
with the recommended toolkits).
* The C function mkCharLenCE now no longer reads past 'len' bytes
(unlikely to be a problem except in user code).
* On systems without any default LD_LIBRARY_PATH (not even
/usr/local/lib), [DY]LIB_LIBRARY_PATH is now set without a trailing colon.
* More efficient utf8ToInt() on long multi-byte strings with
many multi-byte characters.
* aggregate.ts() gave platform-depedent results due to rounding
error for ndeltat != 1.
* package.skeleton() sometimes failed to fix filenames for .R or
.Rd files to start with an alphanumeric.
