gh-2973 Fix `1` is printed during numpy.test()
gh-2983 BUG: gh-2969: Backport memory leak fix 80b3a34.
gh-3007 Backport gh-3006
gh-2984 Backport fix complex polynomial fit
gh-2982 BUG: Make nansum work with booleans.
gh-2985 Backport large sort fixes
gh-3039 Backport object take
gh-3105 Backport nditer fix op axes initialization
gh-3108 BUG: npy-pkg-config ini files were missing after Bento build.
gh-3124 BUG: PyArray_LexSort allocates too much temporary memory.
gh-3131 BUG: Exported f2py_size symbol prevents linking multiple f2py
modules.
gh-3117 Backport gh-2992
gh-3135 DOC: Add mention of PyArray_SetBaseObject stealing a reference
gh-3134 DOC: Fix typo in fft docs (the indexing variable is 'm', not 'n').
gh-3136 Backport 3128
Notable changes (compared to version 1.1.8):
- New MatrixManipulation package.
- New ParallelIntegrationTools package.
- Gruntz algorithm in now used also for finite one-sided limits.
- FriCAS has now true 2-dimensional arrays (previously they were
emulated using vectors of vectors).
- Speedups in some matrix operations and in arithmetic with
algebraic expressions.
- FreeModule is now more general, it allows Comparable as second
argument.
- Changed Spad parser, it now uses common scanner with interpreter.
Spad language is now closer to interpreter language and Aldor.
'leave' is removed, 'free', 'generate' and 'goto' are now keywords.
Pile rules changed slightly, they should be more intuitve now.
Error messages from Spad parser should be slightly better.
Bug fixes, in particular:
- Fixed a few build problems.
- Eliminated division by 0 during 'normalize'.
- 'nthRootIfCan' removes leading zeros from generalized series
(this avoids problems with power series expanders).
- Fixed corruption of formal derivatives.
- Fixed two problems with fortan output.
- Fixed ')untrace' and ')undo'. Fixed ')trace' with ECL.
- Fixed problem with calling efricas if user default shell is (t)csh.
Differences from previous version (5.5.2):
2013.02.15: Changes between NTL 5.5.2 and 6.0
Replaced the old template-like macros for vectors, matrices, and pairs
with true template classes: Vec<T>, Mat<T>, and Pair<S,T>. For
backwards compatibilty, all the names that were used in previous
versions (e.g., vec_ZZ_p, mat_ZZ_p) have been replaced with
appropriate typedefs.
For many years, I resisted the temptation of using templates, because
compiler support was very inconsistent. But that no longer seems to
be the case.
This change, while rather sweeping, should create very few, if any,
incompatibilities with existing software. The biggest issue would be
for software that uses the old template-like macros: such macro
invocations can simply be replaced with appropriate typedefs.
Made the conversion interface more complete and uniform. Also, using
template notation, one can and should now write conv<ZZ>(a) instead of
to_ZZ(a) (for backward compatibility, all the old names to_XXX are
still there, but many new conversions are not available under these
old names). There are many new conversions provided. Moreover,
whenever there is a conversion from a ring R to a ring S, there is a
corresponding, coefficiet-wise conversion from the polynomial ring
R[X] to the polynomial ring R[X].
In addition, using the template mechanism, there are generic
conversions for vectors and matrices. For example, if there is a
conversion from S to T, then there is automatically a corresponding
component-wise conversion from Vec<S> to Vec<T>.
Introduced a more general mechanism for accessing GF2's in packed
structures via indexing (see the class ref_GF2 in the GF2 module).
Employed ideas from David Harvey to make the single-precision FFT
faster (about twice as fast in many cases). This speeds up many
higher-level operations.
Fixed all known bugs.
Summary of changes since 0.10.1:
This is a major release from 0.10.1 and includes many new features and
enhancements along with a large number of bug fixes. The methods of
Selecting Data have had quite a number of additions, and Dtype support
is now full-fledged. There are also a number of important API changes
that long-time pandas users should pay close attention to.
* New precision indexing fields loc, iloc, at, and iat, to reduce
occasional ambiguity in the catch-all hitherto ix method.
* Expanded support for NumPy data types in DataFrame.
* NumExpr integration to accelerate various operator evaluation.
* Improved DataFrame to CSV exporting performance.
For a full list refer to the "what's new" page.
Also fixes PLIST errors introduced in last update.
PR pkg/47783 by Litvinov Sergey.
METIS 4.0.3, 3/19/11
------------------------------------------------------------------------------
r9587 | karypis | 2011-03-19 12:22:36 -0500 (Sat, 19 Mar 2011) | 1 line
- Renamed log2() to ilog2() to remove conflicts with C99 log2() function
- Fixed I/O routines to eliminate compilation warnings
- Fixed error reporting routines to eliminate compilation warnings
METIS 4.0.2, 3/10/04
------------------------------------------------------------------------------
- Fixed a problem with weighted graphs and ometis.c
METIS 4.0.1, 11/29/98
------------------------------------------------------------------------------
This is mostly a bug-fix release
- Fixed some bugs in the multi-constraint partitioning routines
- Fixed some bugs in the volume-minimization routines
Major changes in 2.4:
- Improved HDF5 error logging management.
- Added support for the float16 data type.
- Leaf nodes now have attributes for retrieving the size of data in memory
and on disk.
- Configurable maximum number of threads for Blosc and Numexpr.
- ndim (read-only) attribute added to Leaf, Atom and Col objects.
- Added read support for variable length string attributes.
Full changelog for 2.4.0, see:
http://pytables.github.io/release-notes/RELEASE_NOTES_v2.4.x.html
Major changes in 2.3:
- Integrated functionality from PyTablesPro (re-licensed under a BSD license).
- OPSI is a powerful and innovative indexing engine allowing PyTables
to perform fast queries on arbitrarily large tables.
- A fine-tuned LRU cache for both metadata (nodes) and regular data.
Full changelog for 2.3.x, see:
http://pytables.github.io/release-notes/RELEASE_NOTES_v2.3.x.html
The new API routine glp_mincost_relax4, which is a driver to
relaxation method of Bertsekas and Tseng (RELAX-IV), was added
to the package. RELAX-IV is a code for solving minimum cost
flow problems. On large instances it is 100-1000 times faster
than the standard primal simplex method. Prof. Bertsekas, the
author of the original RELAX-IV Fortran code, kindly permitted
to include a C translation of his code in GLPK under GPLv3.
A bug (wrong dual feasibility test) was fixed in API routine
glp_warm_up. Thanks to David T. Price <dtprice@speakeasy.net>
for bug report.
Obsolete API routine lpx_check_kkt was replaced by new routine
glp_check_kkt.
IMPORTANT: All old API routines whose names begin with 'lpx_'
were removed from API level and NO MORE AVAILABLE.
File too long (should be no more than 24 lines).
Line too long (should be no more than 80 characters).
Trailing empty lines.
Trailing white-space.
Trucated the long files as best as possible while preserving the most info
contained in them.
COMMENT should not be longer than 70 characters.
COMMENT should not begin with 'A'.
COMMENT should not begin with 'An'.
COMMENT should not begin with 'a'.
COMMENT should not end with a period.
COMMENT should start with a capital letter.
pkglint warnings. Some files also got minor formatting, spelling, and style
corrections.
Math_BigInteger provices support of base-2, base-10, base-16, and base-256
numbers. Uses the GMP or BCMath extensions, if available, and an internal
implementation, otherwise.
