2011-03-10 bboyer
* Implement DublinBox Mother model for dense matrices (BlasM* own, BlasS* share)
* Change BlasSubmatrix private members (and add one to make Iterators faster)
* Change <> includes to "" so I have no surprise when the compiler looks them in a non expected place
* add a const in Field Categories for a weird bug...
* RawIterators are renamed to Iterators (impossible to pronounce, easy to misspell)
* add functions to blas-domain and blas-matrix
2011-24-09 bboyer
* replace DenseMatrix(Base) by Blas(Matrix/Blackbox)
* put these classes (and DenseSubmatrix) in a Protected namespace (users will be warned...)
* update Blas(Mat/Bb) interface
* clean DenseMatrix(Base) from LinBox
* Delete BlasTag (use LinBoxTag class instead, avoids branching)
* Update to Givaro gcd new clean intefarce
* Update to one/zero in (some) Fflas-Ffpack fields
* Rename Flaflas to more serious FFLAS-FFPACK
* Minor updates in some classes.
* martin albretch patch to update linbox-sage.h
* M4RI checks
* Ekopath (pathscale) checks
2011-05-11 bboyer
linbox-1.2.0 released:
* Many many bug fixes, many memleaks, much code beautify.
* Licence/copyright added to (hopefully) every file.
* Vi/emacs magic lines for auto-indentation.
* Some documentation added + improved Doxygen framework (dev/user
documentation, more *.doxy items, bug/todo list, bibliography list)
* Upgrade to givaro 3.4.0 series ( new namespace, new API for integer
operations, a few new functions (naturallog...) )
* Upgrade to fflas-ffpack 1.4.0 (new namespace, new includes)
* Some initial work introduced here and there (matrix permutation
representations, symbolic/numeric solver, sparse matrix
representation...)
* CRA algorithms reworked and tested, some added.
* Timer moved to Givaro
* Some debug/util moved to FflasFfpack
* Better configure.ac/Makefiles (more options, icc/gcc
customisation,...)
* More m4 rules
* Possibility to use IML functions as a shared library (and some
examples given)
* 'make fullcheck' checks all possible tests given and reports
* Auto-installer from sources/svn with plenty of options
* Some new field functions like characteristic()
* Work on synchronising field member functions signatures for
better compatibility between all fields (in progress)
2010-11-03 cpernet
linbox-1.1.7 released:
* bug fixes and minor enhancements
* New interface to sage (supporting native floating point
representation)
* new rebind interface
* incorporate the code for computations over the rationals (by A
Ubranska), still waiting to be cleaned-up (it is not used yet by default)
2008-09-22 cpernet
linbox-1.1.6 released:
* compilation pb, and bug fixes, for cygwin support and debianization
2008-06-10 cpernet
* Upgrade fflas-ffpack to v1.3.3 (safer and tighter bounds, bug
fixes,...)
* Design changes for balanced finite field implementations
* Improve efficiency of Chinese Remainder Algorithm
* Add new interface to Sage
* bugs and documentation fixes
* remove gmp++ from LinBox, and thus make Givaro a necessary package
2008_03-02 cpernet
linbox-1.1.5 released:
* drivers are back to the source code,
* shared libraries are enabled by default.
* new code in fflas-ffpack for rank profiles
* new double-determinant algorithm
* changed the rawIndexedIterator interface, fixing the broken previous one.
* many minor fixes, mostly memleaks.
2007-10-27 cpernet
linbox-1.1.4 has been released including the following most important
changes:
* lots of memory leaks fixed
* better compliance with BLAS
* increased system solver functionalities (rectangular, rank deficient
matrices)
* support for g++/gcc v4.2
2007-09-28 cpernet
* Add support for more general dense system solving:
The system can be rectangular (over/underdetermined), singular.
2007-09-07 bds
* This changelog is being widely ignored. Today buildbot test is on.
2007-01-31 bds
* This entry is just to record that version 1.1.0 has been released by
Jean-Guillaume Dumas and the current svn version is just a couple of
steps from the release version.
2006-08-08 bds
* examples/det.C, solutions/det.h, methods.h modified in preparation for
* mpi based paralle cra loop in algorithms/cra-domain.h (by Bryan Youse).
* also a stub of util/mpicpp.h added (will become the mpi interface).
2006-07-27 Hui Wang and bds
* Add blackbox/quad-matrix.h which defines the ZOQuad matrix,
* which is a cache efficient {0,1}-matrix rep.
* Add it's tests, modify also blackbox/zo.h,
* and accumulate-special in field/modular.h to better support sums of field elements efficiently.
2006-06-01 bds
* rank.h adjusted to work around faulty PIVOT_LINEAR code in sparse elim.
* test-rank.C now tests blackbox, elimination, hybrid methods as well as...
2006-04-18 bds and Hui Wang
* examples/smith.C minor changes
* linbox/blackbox/apply.h - don't recall what we did to wreck this.
* linbox/blackbox/archetype.h linbox/blackbox/scalar-matrix.h linbox/blackbox/zero-one.h linbox/randiter/modular.h linbox/vector/stream.h tests/Makefile.am tests/Matio.h - minor touches
* cvs thinks we changed just about everything in tests, but diff shows no change. We're checking in anyway
* Where are our added files blackbox/zo.h zo.inl quad-matrix.h ? - which are a new version of zero-one and a block decomposed sparse matrix, resp.
2006-01-20 bds and Hui Wang
* linbox/algorithms/gauss.inl - we think we made it better but it still isn't working :-)
2005-12-13 Pascal Giorgi
* linbox/algorithms/Makefile.am: add new files
* linbox/algorithms/lifting-container.h: minor change
* linbox/algorithms/matpoly-mult.h: adding FFT and Karatsuba based polynomial matrix multiplication and middle product
* linbox/algorithms/rational-reconstruction.h: use macro to enable timing
* linbox/algorithms/rational-solver.h: minor change
* linbox/algorithms/rational-solver.inl: minor chane
* linbox/algorithms/short-vector.h: work in progress for ternary lattice reduction
* linbox/algorithms/sigma-basis.h: add muti sigma basis computation and special Pade solver.
* linbox/blackbox/block-hankel-inverse.h: enable use of new special Pade solver
* linbox/blackbox/block-hankel.h: use macro to enable timing
* linbox/randiter/{Makefile.am,random-fftprime.h}: add fftprime generator
2005-12-8 bds
* minor cleanups
2005-11-17 Pascal Giorgi:
* macros/blas-check.m4: adding new support for blas libraries
* macros/linbox-opt.m4: adding new support for blas libraries
* macros/linbox-misc.m4: adding new support for blas libraries
* linbox/config-blas.h: linbox headers for BLAS
* linbox/Makefile.am: add new BLAS support
* linbox/fflas/Makefile.am: remove dependencies to ATLAS
* linbox/fflas/fflas.h: add new BLAS support
* linbox/ffpack.Makefile.am: remove dependencies to ATLAS
* configure.in: remove ATLAS checking and replace with BLAS checking
* linbox-config.in: remove ATLAS libs and replace with BLAS libs
* tests/Makefile.am: remove ATLAS dependencies
2005-10-26 Pascal Giorgi:
* linbox/algorithms/Makefile.am: add sigma-basis.h file
* linbox/algorithms/sigma-basis.h: make sigma basis algorithm available
* linbox/algorithms/blackbox-block-container.h: add support for special structured block
* linbox/algorithms/blas-domain.h: fix bug
* linbox/algorithms/lifting-container.h: add support for sparse solver using a Block Hankel inverse
* linbox/algorithms/rational-solver.h(.inl): add new algorithm for solving sparse system
* linbox/blackbox/Makefile.am: add block-hankel.h. block-toeplitz.h and block-hankel-inverse.h
* linbox/blackbox/apply.h: fix optimization compatibility
* linbox/blackbox/block-hankel.h : add a blackbox for block Hankel matrices
* linbox/blackbox/block-hankel-inverse.h: add blackbox to compute the inverse of a block Hankel blackbox
* linbox/blackbox/block-toeplitz.h: add a blackbox for block Toeplitz matrices
* linbox/element/gmp-rational.h: fix this stupid bug using signed long integer to convert rational !!!
* linbox/field/hom.h: add support for PID_integer homomorphism
* linbox/ring/givaro-polynomial.h: fix bug
* linbox/solutions/det.h: add handler for rational domain
* linbox/solutions/charpoly.h: add handler for rational domain
* linbox/solutions/minpoly.h: add handler for rational domain and hybrid method for blas blackbox
* linbox/solutions/rank.h: add handler for rational domain
* linbox/solutions/solve.h: add handler for rational domain and fix indentation
* linbox/solutions/methods.h: fix indentation and add block hankel method
2005-09-01 Dan Roche
* linbox/algorithms/toeplitz-det.h: Forgot to add this before.
2005-08-31 Pascal Giorgi:
* linbox/algorithms/blas-domain.inl: add inverse specialization for multimod double domain
* linbox/algorithms/cia.h: call domain specialized minpoly instead of calling API and make code generic according to integer ring
* linbox/algorithms/lifting-container.h: support for multimod double domain
* linbox/algorithms/matrix-hom.h: support for multimod double domain
* linbox/algorithms/matrix-inverse.h: support for multimod double domain
* linbox/algorithms/rational-solver.inl: support for multimod double domain
* linbox/blackbox/apply.h: support for multimod double domain and new optimization
* linbox/blackbox/blas-blackbox.h: support for multimod double domain, add construction from MatrixReader and finalize apply function
* linbox/blackbox/submatrix.h: fixed indent style problem
* linbox/ffpack/ffpack.h: fix memory leak problem
* linbox/field/modular-double.h: add support for multimod double domain
* linbox/field/ntl-ZZ.h: fix indent style problem
* linbox/field/ntl-ZZ_p.h: set FieldTraits categoryTag
* linbox/matrix/blas-matrix.h: add constructor from MatrixStream
* linbox/matrix/dense-submatrix.h: add constructor from non const MatrixDenseBase
* linbox/matrix/dense-submatrix.inl: add constructor from non const MatrixDenseBase, fix bug in ConstRawIterator operator++(int)
* linbox/matrix/transpose.h: make the reference to the matrix to be non const
* linbox/ring/givaro-polynomial.h: add support for PID-integer and NTL-ZZ domain
* linbox/solutions/charpoly.h: throw an exception if matrix are not square
* linbox/solutions/det.h: fix indentation and throw an exception if matrix are not square
* linbox/solutions/minpoly.h: fix indentation, throw an exception if matrix are not square and make minpoly over the integer generic
* linbox/solutions/rank.h: fix indentation
* linbox/solutions/smith-form.h" fix indentation
* linbox/util/debug.h: allow the PreconditionFailed to be a real LinBox exception
2005-08-19 Z. Wan
* linbox/element/gmp-rational.h Fixed the error which occurs duing conversion
between gmp integer and gmp rational
2005-08-18 Dan Roche
* Added functionality for computing the determinant of a toeplitz matrix
* Added field/ntl-ZZ_pX.h and field/ntl-lzz_pX.h, rings over the
respective NTL polynomial types
* Added functions pow and powin to field/ntl-ZZ_p.h and
field/ntl-lzz_p.h (could be added to all fields)
* Added blackbox/toeplitz.h and blackbox/toeplitz.inl, very similar to
the existing Toeplitz blackbox class, but generalized to use any
polynomial ring. Didn't delete the old blackbox/ntl-toeplitz.h, inl
so old code doesn't break
* Added tests/test-toeplitz-det.C to test all this
* Modified tests/test-ntl-toeplitz.C to use the new Toeplitz class
2005-08-1 bds
* numerous doc tweaks
2005-07-15 bds and zw
* Now modular-balance-int.h and modular-int32.h define different classes.
* Modular-int had been replaced by modular-int32.
* Modular-int is now deprecated.
* Added randiter/modular-balance.h
2005-07-13 bds and zw
* solutions/solve.h call block-lanczos in bbsolve when cardinality small
* algorithms/bbsolve.h call mg-block-lanczos
* algorithms/mg-block-lanczos commentating number of trials.
2005-07-13 Bradford Hovinen <hovinen@localhost.localdomain>
* linbox/solutions/solve.h: Cast A.coldim () to double before
calling log
(solve): Initialize status
* linbox/algorithms/rational-solver.inl: Use std::max
* linbox/blackbox/apply.h: Move #include ntl-ZZ.h into #ifdef
__LINBOX_HAVE_NTL
* examples/solve.C: Add using namespace std
2005-07-13 bds and zw
* tests/Makefile.am move test-la-block-lanczos to failing tests list
* test-la-block-lanczos use case n = 10 to show problem
* test-mg-block-lanczos restore maxtries to default, comment out nullspace
test which has memory probs.
* test-solve change default blocking factor to 16
2005-07-13 Bradford Hovinen <hovinen@math.utoronto.ca>
* tests/Makefile.am (NTL_TESTS): Moved test-rational-solver,
test-rational-solver-adaptive, test-smith-form-binary,
test-last-invariant-factor here because they depend on NTL
(GIVARO_TESTS): Moved test-ffpack, test-blas-domain,
test-givaro-zpz here because they depend on Givaro
* tests/Makefile.am (NTL_TESTS): Fix definition so that automake
does not complain
2005-07-13 Pascal Giorgi
fix and remove NTL dependencies
- remove NTL dependecies: use of PID_integer
* tests/test-last-invariant-factor.C
* tests/test-rational-solver.C
* tests/test-rational-solver-adaptive.C
* tests/test-smith-form-binary.C
- fix NTL dependencies: by use of __LINBOX_HAVE_NTL preprocessing
* tests/test-smith-form.C
* tests/test-smith-form-iliopoulos.C
* tests/test-smith-form-local.C
* tests/test-smith-form-adaptive.C
* linbox/field/hom.h: specialization for NTL
* linbox/algorithms/smith-form-adaptative.h
- linbox/ring/givaro-polynomial.h: to be done by Jean-Guillaume and Clement
2005-07-12 Pascal Giorgi
* linbox/matrix/factorized.inl: add support for singular system solving (inconsistent system is defined by the zero solution)
fix also bug in nonsingular solving
2005-07-11 Pascal Giorgi (authorized by dictator Dumas))
* linbox/solutions/solve.h: change the interface for integer systems
Make the solver interface working with Wiedemann method, Lanczos method
* linbox/algorithms/solve.h: add support for Lanczos method
* linbox/algorithms/bbsolve.h: fix compilation error
* linbox/algorithms/wiedemann.h: fix compilation error
* linbox/algorithmd/wiedemann.inl: fix compilation error
* linbox/algorithms/lanczos.h: fix compilation error
* linbox/algorithmd/lanczos.inl: fix compilation error
2005-07-07 Dan Roche
* linbox/blackbox/sparse.h
* linbox/blackbox/dense.h
* linbox/matrix/sparse.h: Changed the RawIndexedIterator code so that
dereferencing returns a reference to the value
(same as RawIterator)
* linbox/blackbox/lambda-sparse.h
* tests/test-matrix-stream.C
* linbox/algorithms/matrix-hom.h: Updated the code here to reflect
change above
2005-07-07 Dan Roche
* linbox/matrix/sparse.h: Went through and fixed more iterator problems that could be causing segfaults.
2005-07-06 bds and zw
* linbox/solutions/is-positive-definite.h, is-positive-semidefinite.h, added
functions isPositiveDefinite and isPositiveSemiDefinite.
* linbox/tests/test-isposdef.C, test-ispossemidef.C, Makefile.am, added
tests for them.
* linbox/field/hom.h, linbox/algorithms/signature.h, minor tweaks.
* linbox/algorithms/algorithms.doxy, docs related
2005-07-06 Dan Roche
* examples/rank.C
* examples/det.C: Modified these to use MatrixStream and thereby support multiple file formats.
* linbox/matrix/sparse.h
* linbox/matrix/sparse.inl: Fixed some problems with the iterators: seg faults were caused by poor handling of empty rows.
* linbox/util/matrix-stream.inl: minor clean-ups
2005-07-4 Pascal Giorgi
* linbox/blackbox/Makefile.am: add polynomial.h
* linbox/blackbox/apply.h: add optimization
* linbox/blackbox/blas-blackbox.h: add the use of BLAS asap
* linbox/solutions/methods.h: extend dixon method
* linbox/solutions/solve.h: add rational-solver and diophantine-solver support for BlasElimination method (need to fix denominator output).
also add support for BlasElimination over finite field (not fully implemented)
* linbox/algorithms/rational-reconstruction.h: define early termination according to threshold
(0 means no early termination, x > 0 means early termination with checking each x steps)
2005-07-01 Pascal Giorgi
* configure.in: add linbox/util/formats/Makefile.am support
* add linbox/util/formats/Makefile.am file
* linbox/util/Makefile.am: fixe comment problem
2005-07-01 Dan Roche
* linbox/blackbox/dense.h
* linbox/matrix/dense.inl
* linbox/matrix/dense.h: Added Constructors from a matrix-stream
* linbox/blackbox/sparse.h
* linbox/blackbox/sparse.inl
* linbox/matrix/sparse.h: Added Constructors from a matrix-stream
Changed to use size_t for indexing rather than ints:
* linbox/util/matrix-stream.h
* linbox/util/matrix-stream.inl
* linbox/util/formats/generic-dense.h
* linbox/util/formats/maple-dense1.h
* linbox/util/formats/maple-sparse1.h
* linbox/util/formats/matrix-market.h
* linbox/util/formats/sms.h
* linbox/util/formats/sparse-row.h
* tests/test-matrix-stream.C: Included sparse and dense blackbox matrix constructors in testing
* tests/data/sms.matrix: Fixed a formatting error
2005-07-01 W. J. Turner
* linbox/until/commentator.C: removed using namespace std directive
2005-06-30 Pascal Giorgi
resolve symbol error in libutil.a (commentator.o needed timer.o)
* linbox/util/commetator.C: include timer.C
* linbox/util/Makefile.am: remove timer.c from libutil.a source (included through commentator)
2005-06-29 W. J. Turner
Removed using declarations within several classes and replaced with
this->... to preserve backwards compatability and allow compilation
with gcc 3.4 and newer:
* linbox/algorithms/blackbox-block-container.h
* linbox/algorithms/blackbox-container-symmetric.h
* linbox/algorithms/blackbox-container-symmetrize.h
* linbox/algorithms/blackbox-container.h
* linbox/algorithms/lifting-container.h
* linbox/blackbox/frobenius.h
* linbox/blackbox/lambda-sparse.h
* linbox/blackbox/ntl-hankel.{h,inl}
* linbox/blackbox/submatrix.h
* linbox/field/givaro-zpz.h
* linbox/matrix/blas-matrix.h
* linbox/util/formats/generic-dense.h
* linbox/util/formats/maple-dense1.h
* linbox/util/formats/maple-sparse1.h
* linbox/util/formats/matrix-market.h
* linbox/util/formats/sms.h
* linbox/util/formats/sparse-row.h
2005-06-27 W. J. Turner
* linbox/ffpack/ffpack_charpoly_kgfast.inl: removed using namespace
std directive
* linbox/*/lidia-gfq.h: removed using namespace LiDIA directives
* linbox/vector/sparse.h: removed using namespace LinBox directive
* linbox/util/formats/*.h: removed using declarations
* linbox/util/*.h: removed using declarations
* linbox/util/xml/*.h: removed using declarations
2005-06-25 Bradford Hovinen <hovinen@math.utoronto.ca>
* tests/test-mg-block-lanczos.C (testRandomSolve, testSampleNullspace):
Use preconditioner SYMMETRIZE
(main): Use more sensible defaults
* examples/dot-product.C (main): Declare v1 and v2 to be of length n
(fixes segfault)
2005-06-24 W. J. Turner
* linbox/blackbox/*: removed using declarations and directives
* linbox/field/*.h: removed using declarations and directives
* linbox/field/ntl-lzz_pE.h: added explicit call to namespace in NTL::BuildIrred_zz_pX
* linbox/field/ntl-GF2E.h: added explicit call to namespace in NTL::BuildSparseIrred_GF2X
* linbox/randiter/*.h: removed using declarations
2005-06-17 Pascal Giorgi
* /linbox/field/PID-integer.h: add sqrt function and homomorphism support
* /linbox/field/hom.h: identation fixed
* /example/solver/t-rdisove.C: comment givaro.h to fix compilation errors
2005-06-16 Pascal Giorgi
* /linbox/blackbox/apply.h: add blas-blackbox compatibility
* /linbox/blackbox/blas-blackbox.h: modify constructor and add category support
* /linbox/matrix/blas-matrix.h: new creator and add category support
* /linbox/matrix/matrix-category.h: move MatrixCategory from blas-matrix.h
* /linbox/algorithms/blas-domain.h: fancy indentation
* /linbox/algorithms/blas-domain.inl: add support for BlasBlackbox
* /linbox/algorithms/matrix-hom.h: add support for BlasBlackbox
* /linbox/algorithms/rational-solver.inl: use of MatrixHom and BlasBlackbox instead of MatrixMod and BlasMatrix
2005-06-15 Pascal Giorgi
* /linbox/blackbox/apply.h : compatiblity with 64bits architecture
2005-06-14 Pascal Giorgi
* /doc directory: modification to documentation (linbox.html become central)
* macros/linbox-doc.m4: new documentation checking (use of Doxygen)
* configure.in: take into consideration new doc checking
2005-06-10 bds
methods.h working for det.h in good form - other solutions may be broken:
Can consider det.h a good model for solutions funs now.
2005-06-07 Z. Wan
* linbox/blackbox/matrix-blackbox.h, linbox/matrix/dense.h, sparse.h,
added reind to DenseMatrixBase and SparseMatrixBase.
2005-05-02 Pascal Giorgi
* configure.in , macros/linbox-opt.m4 : add optimization process during linbox configuration
* linbox/fflas/fflas.h : allow the use of optimized Winograd threshold
2005-04-26 -bds
* Name changes *ntl-zz* --> *ntl-lzz* to avoid clashes with ZZ in case
insensitive situations.
* Also change test*.cpp to test*.C for consistency
2005-03-32 Pascal Giorgi
* linbox/fflas/fflas_fgemm.inl: fix compilation error with WinoMain function call
* linbox/algorithms/Makefile.am, rational-solver.h,.inl: set default prime size to 20 bits and make number of prime to try as a user parameter
in function solve
* tests/Makefile.am, test-rational-solver.C : retrieve rational solver test (make commentator progress worrking)
2005-03-30 bds
* removed xml read/write code in numerous places
* misc small stuff (I forgot what I may have changed, to tell the truth!)
2005-03-17 PG and ZW
* linbox/algorithms/cra.h, cra-det-integer.h, minpoly-integer.h, Makefile.am, added
functionalities for integer cases using modular arithmetic and CRA
* linbox/solutions/det.h, minpoly.h, rank.h, unified the interface, added blas elimination
for det and rank
2005-03-15 bds
*tests/test-rank.C linbox/solutions/rank.h methods.h Introduce rank
by blas-elimination, together with check for too small prime in wiedemann rank.
2005-02-17 Zhendong Wan (wan@cis.udel.edu)
*linbox/algorithms/rational-reconstruction.h, implemented the fast early termination techniques, which was described
in a preprint paper submitted to ISSAC'05.
2005-02-16 Zhendong Wan (wan@cis.udel.edu)
*linbox/algorithms/rational-solver.h, rational-solver.inl, rational-solver-adaptive.h, Merged rational solver using numerical methods with others
2005-02-14 Clement Pernet (Clement.Pernet@imag.fr)
* Improvements to FFLAS::fgemm routine: Winograd's fast matrix
multiplication is now working for any operation of the type
C<-alpha.A*B + beta.C
* Dynamic pealing is now generic wrt the field
* The recursive block algorithm is now working over the field if the
BLAS condition is not met
2005-02-07 Clement Pernet (Clement.Pernet@imag.fr)
* Change the name fflapack -> ffpack
* bug fixes in blackbox-block-container.h
* compilation failure of test-rational-solver due to "assert" in
rational-solver.h should be fixed
* compilation failuer in test-matrix-stream due to 888888888888888 being to big
should be fixed
2005-02-04 Zhendong Wan (wan@cis.udel.edu)
*Add linbox/solutions/cra.h, minpoly-integer.h, add codes to compute the minpoly of an integer matrix.
*Modify linbox/test-minpoly.C, add test for minpoly of an integer matrix.
*Modify linbox/solutions/minpoly.h, it can hanle integer case.
** linbox/solutions/cra.h and linbox/algorithms/cra.h should merge in future
** minpoly-integer.h and minpoly.h should merge, and support blas in future
2005-02-03 Zhendong Wan (wan@cis.udel.edu)
*linbox/field/PID-double.h, PID-integer.h, PIR-modular-int.h,
PIR-modular-int32.h, PIR-ntl-ZZ_p.h, double-fmod.h,
field-traits.h gf2.h, givaro-gfq.h, givaro-montg.h,
givaro-rational.h, givaro-zpz.h, gmp-integers.h, gmp-rational.h,
lidia-gfq.h local2_32.h, modular-balance-int.h,
modular-balance-int32.h, modular-byte.h, modular-double.h,
modular-int.h, modular-int32.h, modular-short.h, modular.h,
ntl-GF2E.h ntl-RR.h, ntl-ZZ.h, ntl-ZZ_p.h, ntl-ZZ_pE.h,
ntl-pid-zz_p.h, ntl-zz_p.h, ntl-zz_pE.h, param-fuzzy.h,
unparametric.h,
added information in field traits in order to classify all rings.
2005-03 Zhendong Wan (wan@cis.udel.edu)
*linbox/algorithms/matrix-mod.h, change class MatrixMod to namespace,
so that some special mod functions can be addaed
*linbox/algorithms/rational-reconstruction.h, rational-solver.inl, fixed
compiling complaints
2005-01-28 Dan Roche
* examples/linbox-program.C - various fixes
* field/unparametric.h - was casting to long and then to integer - fixed
* field/hom.h - added mapping from UnparametricField to Modular
* util/matrix-stream.h Provides a new, generic way of reading matrices in multiple formats.
* util/formats New directory for matrix-stream format specifications
* util/formats/matrix-stream-readers.h Place to include all the different format specification files:
* util/formats/generic-dense.h
* util/formats/maple-dense1.h
* util/formats/maple-sparse1.h
* util/formats/matrix-market.h
* util/formats/matrix-stream-readers.h
* util/formats/sms.h
* randiter/gmp-random-prime.h Generates random numbers up to a given size
* tests/test-matrix-stream.C Tests to see if matrix-stream is working
* Sample matrices used by test-matrix-stream:
* tests/data/generic-dense.matrix
* tests/data/maple-dense1.matrix
* tests/data/maple-sparse1.matrix
* tests/data/matrix-market-array.matrix
* tests/data/matrix-market-coordinate.matrix
* tests/data/sms.matrix
* tests/data/sparse-row.matrix
2005-01-07 ZW
* linbox/field/gmp-integers.h, fixed init
* linbox/field/modular-double.h, implemented a fast init
2004-12-03 bds
* matrix/sparse.h .inl - add reading of Magma Compact form
* vector/vector-domain.h .inl - can't read into const vectors, so rm 'const
2004-10-08 bds
* examples/blackbox/Makefile.am smith.C - minor
* field/field-traits.h givaro-gfq.h ntl-GF2E.h - put in maxmodulus
* doc/Doxyfile mainpage - minor doc changes
2004-09-30 Pascal Giorgi
* linbox/algorithms/lifting-container.h: take modification of MatrixApplyDomain setup function
* linbox/algorithms/rational-solver.inl: use of BlasMatrix instead of BlasBlackbox
* linbox/algorithms/rational-reconstruction.h: add lifting commentator progress in getRational3 function
* linbox/blackbox/apply.h: redefine MatrixApplyDomain s.t. specialization for dense storage is easier
* linbox/matrix/dense.inl: make reading with sparse format work and fix bug in writing over domain with non positive cardinality
2004-09-21 Pascal Giorgi
* linbox/algorithms/Makefile.am: Add headers for (rationnal-solver.inl, diophantine-solver.inl, vector-fraction.h)
* linbox/algorithms/blackbox-block-container-base.h: Use BlasDomain and BlasMatrix
* linbox/algorithms/blackbox-block-container.h: Make BlackboxBlockContainerRecord class (handling precomputed sequence and row-colum updating)
* linbox/algorithms/block-massey-domain.h: Compute generating matrix polynomial using Sigma Basis through algorithm of GJV-ISSAC'03
* linbox/algorithms/block-wiedemann.h: Add block Wiedemann solver
* linbox/algorithms/lifting-container.h: Add specialisation for Block Wiedemann method and clean optimizations
* linbox/algorithms/rational-solver.h: Add specialisation for Block Wiedemann method
* linbox/algorithms/rational-solver.inl: Add specialisation for Block Wiedemann method
* linbox/blackbox/Makefile.am: Add headers for lambda-sparse.h and blas-blackbox.h
* linbox/blackbox/apply.h: Put optimizations on matrix-vector product - moved from lifting-container
* linbox/fflas/fflas_fgemm.inl: Fix bug on Winomain function
* linbox/fflas/fflas_ftrsm.inl: Fix bug on normalisation before BLAS call
* linbox/field/Makefile.am: Add ntl-GF2E.h header
* linbox/field/ntl-GF2E.h: Add wrapper for GF(2^n) of NTL library
* linbox/field/modular-double.h: Improve DotProductDomain specialization
* linbox/field/ntl.h: Add ntl-GF2E.h header
* linbox/matrix/blas-matrix.h: Take into account modifications of dense-submatrix
* linbox/matrix/dense-submatrix.h: Use pointer instead of reference
* linbox/matrix/dense-submatrix.inl: Use pointer instead of reference
* linbox/matrix/factorized-matrix.inl: Fix the extractions of L, U and S
* linbox/solutions/methods.h: Add BlockWiedemann method
* linbox/util/gmp++/Makefile.am: Remove gmp++_int.c ( redudant for compilation)
2004-09-02 Bradford Hovinen <hovinen@vega.pc.cis.udel.edu>
* linbox/algorithms/la-block-lanczos.h (class
LABlockLanczosSolver): Add _iter, _total_dim, _rank
* linbox/algorithms/la-block-lanczos.inl (cleanup): Update _rank
(rank): Implement
* linbox/blackbox/diagonal.h (class Diagonal): Compilation fix
* linbox/blackbox/dense.h (class DenseMatrix): Make _AT use
DenseMatrix, not DenseMatrixBase
* tests/test-la-block-lanczos.C (main):
* tests/test-mg-block-lanczos.C (main): Use better defaults
* linbox/algorithms/la-block-lanczos.inl (solve, sampleNullspace):
Clear history when done
* tests/Makefile.am (BASIC_TESTS): Add test-mg-block-lanczos
* linbox/algorithms/la-block-lanczos.h (class
LABlockLanczosSolver::InnerProductArray): Add reset
* linbox/algorithms/la-block-lanczos.inl (iterate): Reset uAv
* linbox/algorithms/Makefile.am (pkgincludesub_HEADERS): Add
eliminator.{h|inl}
* tests/Makefile.am (BASIC_TESTS): Add test-la-block-lanczos
* linbox/algorithms/Makefile.am (pkgincludesub_HEADERS): Add
la-block-lanczos.{h|inl}, mg-block-lanczos.{h|inl}; remove
block-lanczos.{h|inl}
2004-09-01 bds
* solutions/minpoly.h tests/test-minpoly.C: add minpolySymmetrize
2004-09-01 bds wan zw
* linbox/algorithms/rational-solver-adaptive.h, adaptive solve
2004-08-30 bds and zw
* linbox/algorithms/one-invariant-factor.h, smith-form-adaptive.inl, make bonus works
2004-08-30 bds and zw
* linbox/algorithms/smith-form-adaptive.h, inl. Implementation of an adaptive algorithm for
Smith form
2004-08-30 bds
* linbox/randiter/mersenne-twister.h linbox/util/commentator.h
linbox/util/gmp++/gmp++_int_misc.C tests/Makefile.am tests/test-common.C:
LinBoxSrcOnly flag also used in tests/ now (to avoid incompatibility
with use in examples/)
2004-08-30 bds
* examples/ex-integers.C examples/Makefile.am linbox/util/gmp++/Makefile.am: Turn on
LinBoxSrcOnly flag for examples compilations so that user does not confront
load-library problems.
* doc/Makefile.am examples/Matio.h linbox/integer.h linbox/blackbox/*.h etc.:
switch documentation system to doxygen. (it still needs plenty of work)
* add .doxy files (replacing .dxx files) to doc system
* examples/blackbox/smith.C and smithbig.C: smith form examples.
2004-08-28 Pascal Giorgi
* examples/solver/t-rdisolve.C: add sparse matrix format for file and make seed to work
* linbox/algorithms/diophantine-solver{.h,.inl}: adding parameters information of solving
* linbox/algorithms/rational-reconstruction.h: clear bugs in timings
2004-08-26 Z. Wan
* linbox/algoriths/last-invariant-factor.h, implement the bonus idea \ref{....}
2004-08-26 Pascal Giorgi
* examples/fields/ex-fields-archetype.C : put std namespace
* examples/run/run.C : put std namespace
* examples/solver/Makefile.am : change order of libs for compilation
* examples/solver/t-rdisolve.C : change to use the new design of rational solver
* gmp++/gmp++_int.h : Added new function to get the size of integer in base B and base 2
* gmp++/gmp++_int.inl : Added implementation of bitsize() and size_in_base( int base) functions
* linbox/algorithms
* diophantine-solver.h : use new design of rationnal solver
* diophantine-solver.inl : use new design of rationnal solver ( disabling conversion for VectorFraction)
* lifting-container.h : Added fast apply over ring when entries of the matrix are small enough ( say less than 2^16)
* rational-reconstruction.h : Added direct method for reconstruction, added switch between early termination and direct method
* rational-solver.inl : Added blas optimization for preconditionning matrices in random solution solver
* linbox/fflapack/fflapack.h : Maked triangular inverse available
* linbox/field/PID-double.h : Added abs and compare functions, + indentation
* linbox/field/PID-integer.h : Added abs and compare functions, + indentation
* linbox/matrix/dense.inl : problem with resizing in read function, leave code as it was but add the older version as comment
2004-08-24 Z. Wan (wan@cis.udel.edu)
* change ith-invariant-factor.h to one-invariant-factor.h
* Change the implementation of the alogrithm for oneinvariantfactor, use the enhanced algorithm in ISSSAC'04 papre by
bds and ZW
2004-08-24 Z. Wan (wan@cis.udel.edu)
*linbox/algorithms/rational-solver.h, inl, rational-reconstruction.h, linbox/tests/test-rational-solver.C, change
the format of the rational solution a vector of rationals ====> (a vector of the numerator, the common denominator).
2004-08-19 Z. Wan (wan@cis.udel.edu)
* linbox/algorithms/rational-reconstruction.h, implement the V. Pan's idea for final reconstruction step.
* linbox/field/ntl-ZZ.h, add abs and compare member function.
2004-08-19 David Pritchard (dpritcha@ens-lyon.fr, daveagp@mit.edu)
* gmp++/gmp++_int.inl - Added wrapper for mpz_import (convert word array to LinBox::Integer)
* linbox/algorithms - Added timing to rational solving routines (#define RSTIMING in rational-solver.h)
* diophantine-solver.h - made QSolver field into a reference type
* diophantine-solver.inl - changed comments
* lifting-container.inl - Added blas optimization for muliplication of large integer matrices in Dixon
* rational-reconstruction.h - Changed getRational2 again so that early termination num/denbounds work
* rational-solver.h - Added timing, reportTimes()
* rational-solver.inl - Added timing
* vector-fraction.h - Added comment, about how RandIter may have unrandom low-order bits
* linbox/blackbox/apply.h - Removed bad overload of applyV
* linbox/fflapack/fflapack.h - changed name of function to LQUPtoInverseOfFullRankMinor
* linbox/field/modular-double.h - added constr. from long int, to avoid ambiguity warning when making t-rdisolve
* linbox/matrix/dense.inl - attempted to fix problems with Const{Row/Col}Iterator operator++ (int)
* linbox/matrix/factorized-matrix.h - added comment about how GetQ returns Qt
* linbox/util/timer.C - added a count field to the timer (how many intervals does this timer measure)
* linbox/util/timer.h - added a count, stopped small negative rounding errors from printing
* gmp++/gmp++_int_cstor.C - added mpz_import function wrapper
* tests/test-common.C - fixed reading of large integers
* examples/solver/t-rdisolve.C - added flag to make some columns zero, made better random entries
2004-08-11 bds
* examples/blackbox/smith.C illustrates iliopolous' and local methods
* matrix/dense.inl read() now adjusts to new input size rather
than complaining. Also read() accepts sparse file format.
2004-08-11 David Pritchard (dpritcha@ens-lyon.fr, daveagp@mit.edu)
* gmp++/gmp++_int.inl - Added nth root function declaration
* linbox/algorithms
* diophantine-solver.inl - minor changes
* lifting-container.h - fixed reconstruction lifting bound
* rational-reconstruction.h - added getRational2; will eventually be las-vegas style reconstructor
- changed early termination so that 2*numbound*denbound is always less than modulus
* rational-solver.inl - more updates, fixes. now only one LQUP is used to check consistency
- added optimization wherein LQUP is used to compute a modular inverse more quickly
* linbox/fflapack/fflapack.h - added InverseFromLQUP function; fixed invL bug
* linbox/matrix/blas-matrix.h - moved indexDomain here (allows application of BlasPermutation to indices, using BlasMatrixDomain<indexDomain>)
* linbox/matrix/dense-submatrix.inl - minor fix
* linbox/util/gmp++/gmp++_int_misc.C - added nth root function
2004-08-09 Dan Roche
* gmp++/gmp++_int.inl: Add arithmetic function support for long long and unsigned long long
* linbox/blackbox/dense.h,inl: Add hadamardBound method to DenseMatrixFactory
* linbox/blackbox/sparse.h: Make hadamardBound return a reference to res, and avoid overflow in computing hadamardBound
* linbox/solutions/det.h: little fixes
2004-08-09 Pascal Giorgi
* /linbox/blackbox/sparse.h : Fix bugs in SparseMatrixFactory. it compiles right now
* /linbox/util/gmp++/gmp++_int_io.C : avoid memory leaks in Integer2string function
* /linbox/field/ntl-zz_pE.h : adding new wrapper for NTL algebraic extension of single precision prime field
* /linbox/field/ntl-ZZ_pE.h : write a class to handle parametric field
* /linbox/field/ntl.h, Makefile.am : adding new field ntl-zz_pE.h
2004-08-08 Dan Roche (roche@cis.udel.edu)
* examples/linbox-program.C - now works for determinant also - except that someone recently broke solutions/det.h and blackbox/sparse.h
* gmp++/gmp++_int.h, linbox/util/gmp++/gmp++_int_misc.C - added prevprime so that prime-stream.h works
* linbox/integer.h - added an abs function here so it works for all types from int8 up to integer.
* linbox/algorithms/cra.h - changed invmod to inv so that this compiles
* linbox/blackbox/dense.h,.inl - added DenseBlackboxFactory class so I can use det.h to find determinant of a dense matrix over the integers.
* linbox/solutions/det.h - NUM_BITS is determined from FieldTraits (linbox/field/field-traits.h)
2004-08-06 David Pritchard (dpritcha@ens-lyon.fr, daveagp@mit.edu)
* examples/solver/t-rdisolve.C - added use of 3 'solver levels' and check of certificate of inconsistency
* linbox/algorithms
* diophantine-solver.h - added use of 3 'solver levels', improved documentation
* diophantine-solver.inl - added use of 3 solver levels
* lifting-container.h - changed normblackbox to properly compute "norm" of composed boxes
* rational-reconstruction.h - added getRational2 for solver use, hopefully will be las-vegas
* rational-solver.h - added use of 3 solver levels, certifying of inconsistency, documentation
* rational-solver.inl - added use of 3 solver levels, certifying inconsistency
- combined 'solve' and 'getRandomSolution' to be both variants of 'solveMonolithic'
* vector-fraction.h - simplified constructor
* linbox/fflapack/fflapack_ludivine.inl - bugfix for LQUP of rectangular matrices (specifically M x 1 case)
* linbox/field/unparametric.h - removed dumb comment
* linbox/matrix/blas-matrix.h - added method extendTrivially to embed permutation in larger row/column size
* linbox/tests/test-common.C - bugfix
2004-07-30 David Pritchard (dpritcha@ens-lyon.fr, daveagp@mit.edu)
* examples/solver/t-rdisolve.C - added diophantine checking+certification, more options
- moved VectorFraction class to own file in algorithms
- moved new command-line stuff into test-common.C (should be backwards-compatible)
* linbox/algorithms
* vector-fraction.h - added new file. for storing fractions like [n1 n2 n3 n4 ... ]/d
- has various combination functions used by diophantine-solver
* blas-domain.inl - fixed a broken linbox_check for vector-matrix product
* diophantine-solver.h - diophantine checking, min-denom certification works now
* diophantine-solver.inl - diophantine checking, min-denom certification works now
* rational-solver.h - made SolverReturnStatus a global enum, preceded by SS_
* rational-solver.inl - added generation of a partial certificate to randomSolve
- solveDeterministic is now fully Las Vegas
* apply.h - added applyVTrans(vector, const matrix, vector)
* linbox/fflas/fflas_fgemv.inl - fixed errors in transpose matrix product
* linbox/matrix
* blas-matrix.h - fixed flipped row/coldim in constructor
* dense-submatrix.h - fixed typo, removed unimplemented operator[]
* dense.h - removed debug output
* transpose.h - replaced Column with Col
* tests/test-common.h - example string can now be passed as null and intelligently generated by default
* tests/test-common.C - boolean options can now be turned off like "-b false", "-b -", "-b N"
- added a string to write all current switch options to an ostream
* tests/test-rational-solver.C - updated to reflect new return type of rational solver
2004-07-26 David Pritchard (dpritcha@ens-lyon.fr, daveagp@mit.edu)
* linbox/algorithms
* added diophantine-solver.{h,inl} (no diophantine functionality yet)
* Makefile.am - added diophantine-solver.h
* blas-domain.h - added forms of inv and invin that return the nullity of the matrix
* blas-domain.inl - added forms of inv and invin that return the nullity of the matrix, fixed typo
* last-invariant-factor.h - changed type of b vector from int to Integer
* lifting-container.h - changed blackbox norm to square of 'induced norm', fixed lifting length calculation
- added graceful failure to digit (eg if <double> Ring is used improperly)
* rational-reconstruction.h - added debug switch DEBUG_RR, graceful failure
* matrix-rank.h - changed tabbing
* rational-solver.h - many changes to Dixon; Wiedemann may not work
- added DEFAULT_PRIMESIZE = 14 (since 30 bit default size would screw up some fields)
- added maxPrimes argument to solving routines, will retry that many primes
- added const modifier to solving routines, made _prime mutable
* rational-solver.inl - many changes to Dixon; Wiedemann may not work
- added debug switches, graceful failure for solving routines
- changed BlasMatrixDomain routines to BlasApplyDomain to avoid erroneous calculations
- in singular/random solve, check solution before returning, and multiply by conditioner P correctly
- added consistency check to random solver
* linbox/blackbox/blas-blackbox - added to CVS
* linbox/blackbox/dense.h - matrix reading bugfix
* linbox/fflapack/fflapack.h - pass nullity back in Invert2, and return NULL when matrix is singular
* linbox/fflas/fflas_fgemv.inl - remove assumption that alpha=-1, beta=1
* linbox/fflas/fflas_ftrmm.inl - fixed computation of matrix size bound for embedding into FFLAS
* linbox/field
* PID-double.h - changed 'c=abs(c)' to 'if (c<0) c = -c', no abs found for double
* PID-integer.h - changed documentation of isDivisor(a, b) (returns yes if b|a)
* givaro-montg.h - fixed init from integer
* givaro-zpz.h - fixed init from integer, double and Log16 specializations
* givaro-gfq.h - fixed init from integer
* modular.h - added init from double
- added several specializations of Modular<NTL::ZZ>
* modular-double.h - fixed operator<<
* modular-int.h - fixed init/convert with double
* modular-int32.h - fixed init/convert with double
* ntl-ZZ.h - changed const to static for convert method
- changed documentation of isDivisor(a, b) (returns yes if b|a)
* ntl-ZZ_p.h - added init/convert from/to integer, double
- put NTL_ZZ_p at top and specialization at bottom for compiler
- changed random generator to give fuller range
* ntl-zz_p.h - added init from double, integer
* unparametric.h - added convert to double
* linbox/matrix/dense-submatrix.h - added mapleFormat option to DenseSubmatrix writing
* linbox/matrix/dense-submatrix.inl - added mapleFormat option to DenseSubmatrix writing
* linbox/randiter/random-prime.h - the class now takes a 'bits' argument on construction and makes primes of that size
* linbox/vector/vector-domain.inl - fixed reading bug
* tests/Matio.h - added mapleFormat option to write_field
* tests/test-fflapack.C - matched Invert2 to new 'nullity' argument format
* tests/test-iliopoulos-elimination.C - #define int32 LinBox::int32 to avoid ambiguity with Givaro
* tests/test-last-invariant-factor.C - int32 -> LinBox::int32 to avoid ambiguity with Givaro
* tests/test-last-rational-solver.C - int32 -> LinBox::int32 to avoid ambiguity with Givaro
- now handles solver failures without crashing
- added check to ensure that system is solvable (ie no zero elements on diagonal)
* tests/test-smith-form.C - int32 -> LinBox::int32 to avoid ambiguity with Givaro
* examples/solver/t-rdisolve.C - test driver for rational solver, compile and run t-rdisolve -?
* examples/solver/Makefile.am - make file
2004-07-16 Pascal Giorgi
* linbox/field/lidia-gfq.h : fixing bug and improvement
* tests/Makefile.am : retrieving test-lidia-gfq, fixing modular and gssv redundancy
* tests/test-modular.C : using namespace std
2004-07-16 Pascal Giorgi
* linbox-config.in : switching order of LiDIA in libs
* linbox/randiter/envelope.h : modify random function ( assignement is not needed)
* linbox/matrix/dense-submatrix.h : setting input and output functions public
* linbox/field/givaro-zpz.h : adding linbox/utils/debug.h header
* linbox/field/lidia-gfq.h : setting default degree extension to 1
* linbox/field/ntl-ZZ_p.h : overloading the constructor of UnparamField
2004-07-08 Dan Roche
* Adding examples/linbox-program.C which will eventually do a wide variety of operations on integer matrices
* gmp++/gmp++_int.h,inl : Little fixes
* linbox/algorithms/wiedemann.inl : Little fixes
* linbox/matrix/blas-matrix.h : Little fixes
* linbox/util/gmp++/gmp++_int_compare.C, gmp++_int_mod.C : Little fixes
* linbox/vector/sparse.h : Little fixes
* tests/test-fields.C : Little fixes
* tests/Makefile.am : Moved everything that uses rational-solver to ATLAS_TESTS rational-solver won't work without BLAS.
* fields/field-traits.h : Took implementation out of this file and put it in each field header file.
* fields/givaro-gfq.h : Added init and convert from/to a padic number
* fields/modular.inl : Fixed some problems with MVProductDomain functions over unsigned types
2004-07-06 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/blackbox/dense.h (class DenseMatrix): Add _AT, a
TransposeMatrix of this
* linbox/blackbox/dense.inl (DenseMatrix::applyTranspose): Use _AT
* linbox/matrix/matrix-domain.h (class MatrixDomain): Get rid of
template argument MatrixTrait on specialized permute
* linbox/vector/vector-domain.h (class VectorDomain): Get rid of
template argument VectorTrait on specialized permute, swap
* linbox/matrix/sparse.h (struct MatrixTraits):
* linbox/vector/vector-domain.inl (class VectorDomain):
* linbox/vector/vector-domain.h (class VectorDomain):
* linbox/matrix/matrix-domain.h (class MatrixDomain):
* linbox/matrix/transpose.h (struct MatrixTraits):
* linbox/matrix/dense.h (struct MatrixTraits): Remove templates
from matrix tags
2004-07-02 pascal giorgi
* blackbox/apply.h : adding BlasApply domain
* blackbox/compose.h : adding accessor to left an right blackbox
* blackbox/sparse.h : make write function const
* field/ all givaro file : fixe bugs
* field/lidia-gfq.h : make it works
* field/modular-double.h : add reset function in axpy domain
* field/field-traits.h : add specialization for some fields
* field/double-fmod.h : include <math.h>
* field/ntl-ZZ.h : remove all static and replace by const function
* field/PID-double field/PID-integer.h : adding 2 new PID
* matrix/dense.h.inl matrix/sparse.h.inl : make interators work
* randiter/ntl-ZZ.h randiter/lidia-gfq.h : fix bugs and make clearer
* solutions/method.h : add new traits for rational solvers
* algorithms/
* blackbox-container-base.h : adding new constructor allowing size of sequence as parameter
* blackbox-container.h : adding new constructor allowing size of sequence as parameter
* lifting-containe.h : make a new interface for lifting (give specialization for Wiedemann and Dixon methods)
* matrix-inverse.h : put Field as template and require a field as parameter (not restrictive to blackboxes)
* matrix-mod.h : giving a specializatino
* rational-reconstruction.h : make it clear with lifting interface
* rational-solver.h : make a new interface for rational solver
* rational-solver.inl : give specialization for Wiedemann and Dixon methods
2004-06-21 bds
* removed NAGSparse blackbox which has been replaced by Triplesbb.
Mods to linbox/util/xml/reader-blackbox-factory.h tests/Makefile.am
Removed Files: blackbox/nag-sparse.h tests/test-nag-sparse.C
2004-06-16 Zhendong Wan
* linbox/solutions/lu.h, tests/test-lu.C, fix the block lu and make test work.
2004-06-10 Dan Roche
* Added linbox/field/modular-balance-int32.h to use int32 type
* Added more fields to tests/test-fields.C
* Added more fields and functionality to linbox/field/field-traits.h
* Changed most fields so that every field constructor takes two
arguments, a modulus and an exponent. Default values are
put in so everything should still work the same.
2004-06-10 bds
* New version 0.1.4
* final doc mods before release 0.1.4
2004-06-09 bds
* Little tweaks.
2004-06-08 Dan Roche
* Updated Modular<int> to use the int32 type in integer.h
* Added linbox/field/modular-int32.h
* Added linbox/field/PIR-modular-int32.h
* Updated tests/test-modular-int.h
* Updated tests/test-iliopoulos-elimination.C
* Updated tests/test-last-invariant-factor.C
* Updated tests/test-rational-solver.C
* Updated tests/test-smith-form.C
* Updated tests/test-fields.C to use types from integer.h and to
determine number of iterations based on machine speed.
Also uses goodModulus method from FieldTraits struct.
* Added linbox/field/field-traits.h, containing struct FieldTraits
<class Field> which can do other things but for now just
provides information about valid moduli for a given field.
2004-06-08 bds
* assorted documentation mods.
* introduced BlackboxInterface, similar to FieldInterface, an empty base
class solely for the purpose of organizing the heirarchy in the documentation.
2004-06-07 bds
* assorted documentation mods.
2004-06-05 bds and Zhendong Wan
* linbox/field/double-fmod.h is now redundant
* linbox/field/modular-double.h uses the fmod approach
* tests/test-generic.h now has a basic concept check for randiter
and a chi-sq randomiterator check that doesn't fail on reasonably
uniform random iterators over int elements.
* tests/test-double-fmod.C and tests/test-modular-double.C now work.
2004-06-04 Dan Roche
* Added tests/test-double-fmod.C
Added tests/test-fields.C which benchmarks ops for fields
Added tests/test-modular-double.C
Added linbox/randiter/generic.h which is a generic random iterator
used in double-fmod
Added linbox/field/double-fmod.h which is like modular-double, but
uses the fmod command from math.h
2004-06-04 bds
* comments in rational.h suggesting that it is nowhere used.
Shall we delete?
2004-04-20 Zhendong Wan <wan@udel.edu>
* Add linbox/blackbox/blackbox_parallel.h, blackbox_thread.h, subrowmatrix.h, parallel blackbox apply
and applyTranspose by using posix threads.
* Modify linbox/blackbox/sparse.h, dense.h, dense.inl, linbox/vector/subvector.h, add parallel apply
and applyTranspose for SparseMatrix and DenseMatrix.
2004-03-25 bds
* documentation mods
2004-03-17 bds
* gmp++/gmp++_int.h linbox/util/gmp++/gmp++_int.C, gmp++_int_add.C, gmp++_int_compare.C,
gmp++_int_cstor.C, gmp++_int_div.C, gmp++_int_gcd.C, gmp++_int_io.C, gmp++_int_misc.C,
gmp++_int_mod.C, gmp++_int_mul.C, gmp++_int_pow.C, gmp++_int_sub.C, tests/test-common.h:
mods to make flag -DLinBoxSrcOnly work.
2004-03-04 Zhendong Wan
* linbox/blackbox/sparse.h, set default Row to be SparseMap, which is more efficient to call setEntry, refEntry member function.
* tests/test-minpoly.h, make it works with change in sparse.h.
2004-03-04 Zhendong Wan
* linbox/blackbox/diagonal.h linbox/blackbox/hilbert.h,
linbox/blackbox/scalar-matrix.h linbox/blackbox/submatrix.h,
linbox/field/PIR-modular-int.h linbox/field/PIR-ntl-ZZ_p.h,
linbox/field/modular-byte.h linbox/field/modular-int.h,
linbox/field/modular-short.h linbox/field/modular.h,
linbox/field/modular.inl linbox/matrix/matrix-domain.h,
linbox/matrix/matrix-domain.inl linbox/matrix/sparse.h,
linbox/matrix/sparse.inl linbox/vector/stream.h,
linbox/vector/vector-domain.h linbox/vector/vector-domain.inl,
linbox/vector/vector-traits.h tests/test-common.h, make these Tag classes non-template, since they are only used to distinguish vector types and for this purpose, ordinary classes suffice.
2004-03-04 Zhendong Wan
* linbox/blackbox/dense.h linbox/blackbox/sparse.h, linbox/matrix/archetype.h linbox/matrix/dense-submatrix.h, linbox/matrix/dense.h linbox/matrix/matrix-domain.h, linbox/matrix/sparse.h linbox/matrix/transpose.h, tests/test-matrix-domain.C, make struct RowMatrixTag, struct ColMatrixTag, struct RowColMatrixTag non-template classes, since they are only used to distinguish the matrix types and in this case, non-template classes suffice and codes become simpler.
2004-03-02 Zhendong Wan
* remove linbox/vector/const-iterator-type.h, since in new gcc, std::vector<Element>::iteretor is not same type as Element*, it doesnot work anymore.
* modify linbox/vector/subvector.h, add one template parameter const_iterator.
* linbox/blackbox/dense.h, dense.inl, use MatrixDomain to conform SparseMatrix implementation.
2004-03-02 Zhendong Wan
* linbox/blackbox/sparse.h, remove field VectorDomain<Field> _VD, since it is not used in any member function.
2004-02-07 Zhendong Wan
* linbox/algorithms/rational-reconstruction.h, speed it up bu only reconstructing the rationals at every threshold position.
2004-02-06 bds
* linbox/gen_superlu/gssv_rank.h linbox/gen_superlu/trsv.h
tests/Makefile.am tests/test-gssv_rank.C
tests/data/gssv_rank_data: fix gssv_rank test (works except for A'+A case).
2004-02-04 Zhendong Wan <wan@udel.edu>
* Add linbox/field/PIR-modular-int.h, which is an extension of Modular<int> with PIR functions.
2004-02-03 Zhendong Wan <wan@udel.edu>
* Modify linbox/field/PIR-ntl-ZZ_p.h, add MVProductDomain, DotProductDomain for PIR_ntl_ZZ_p
2004-01-28 Zhendong Wan <wan@udel.edu>
* Modify linbox/algorithms/iliopoulos-elimination.h, making it work in a PIR suggest by bds.
* Modify linbox/tests/test-iliopoulos-elimination.C, making it work upon change of iliopoulos-elimination
* Add linbox/field/PIR-ntl-ZZ_p.h, adpating ntl-ZZ_p, and adding PIR interface (gcd, xgcd, dxgcd).
2004-01-27 Zhendong Wan <wan@udel.edu>
* Modify linbox/algorithms/bitonic-sort.h, minor name change
* Add linbox/tests/test-bitonic.C, test for bitonic-sort
2004-01-27 BDS <saunders@udel.edu>
* Modify linbox.dxx doc/.cvsignore doc/doc.dxx, min doc change
2004-01-27 BDS <saunders@udel.edu>
*add linbox/field/ntl-pid-zz_p.h, linbox/field/local2_32.h, linbox/algroithms/local-smith.h, code for local smith form computation.
*add linbox/tests/test-local-smith.C, modify linbox/tests/Makefile.am, add test for local smith algorithm
2004-01-26 Zhendong Wan <wan@udel.edu>
* add linbox/field/ntl-ZZ.h, linbox/randiter/ntl-ZZ.h, adapter NTL::ZZ to linbox.
* add linbox/algorithms/lifting-container.h, linbox/algorithms/rational-reconstruction.h, linbox/algorithms/rational-solver.h, solve the linear systems over the rationals by lifting.
* add linbox/algorithms/bitonic-sort.h, implement bitonic-sort, a sort network sequential implemenation.
* add linbox/algorithms/iliopoulos-elimination.h, implement the Iliopoulos algorithm for Smith form.
* add linbox/blackbox/apply.h, linbox/blackbox/scompose.h, linbox/blackbox/compose-traits.h, linbox/blackbox/random-matrix-traits.h, linbox/blackbox/random-matrix.h, linbox/blackbox/submatrix-traits.h, linbox/algorithms/ith-invariant-factor.h, linbox/algorithms/last-invariant-factor.h, linbox/algorithms/matrix-inverse.h, linbox/algorithms/matrix-mod.h, linbox/algorithms/smith-form.h, implement EGV/EGV+/EGV++ algorithm for Smith Form computation.
* add linbox/tests/test-last-invariant-factor.C, linbox/tests/test-smith-form.C, linbox/tests/test-iliopoulos-elimination.C, add test for smith form computation and iliopoulos elimination.
2004-01-26 Zhendong Wan <wan@udel.edu>
* linbox/blackbox/submatrix.h, add specialization for submatrix of denese matrices.
* linbox/matrix/dense-submatrix.h, linbox/matrix/dense-submatrix.inl, fixed output / input
2004-1-15 bds
* field/archetype.h field/field-interface.h: doc tweaks
* matrix/dense-submatrix.h matrix/dense-submatrix.inl: fixed write trouble.
* vector/subvector.h: fixed swap
2004-01-07 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
Fix some printing messages in :
* macro/atlas-check.m4
* macro/expat-check.m4
* macro/givaro-check.m4
* macro/lidia-check.m4
* macro/ntl-check.m4
* macro/saclib-check.m4
2003-12-29 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/matrix/transpose.h (class TransposeMatrix): Remove const
requirement on input matrix; add in regular iterators
* tests/test-matrix-domain.C (transpose): Remove const
(testInvMulOver, testInvMulUnder, testInvLeftMulinOver)
(testInvRightMulinOver): Remove const from parameter M
2003-12-28 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/vector/vector-domain.h (class VectorDomain): Add permute,
swap functions
* linbox/vector/vector-domain.inl (swapSpecialized,
permuteSpecialized): Implement
* linbox/matrix/matrix-domain.inl (permuteRowsByRow)
(permuteRowsByCol, permuteColsByRow, permuteColsByCol): Implement
* linbox/matrix/matrix-domain.h (class MatrixDomain): Add
row/column permutation functions
2003-12-19 bds
* tests/Makefile.am: just clean up some comments
2003-12-19 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
* added files :
linbox/gen_superlu/Makefile.am: missing file
macros/config-header.m4: create linbox-config.h from config.h by prefixing all variable by __LINBOX_
macros/linbox-misc.m4: allow --with-all and --with-default options
Modify Makefile such they use pkg variable
* gmp++/Makefile.am
* interfaces/Makefile.am
* interfaces/maple/Makefile.am
* linbox/Makefile.am
* linbox/algorithms/Makefile.am
* linbox/blackbox/Makefile.am
* linbox/element/Makefile.am
* linbox/fflapack/Makefile.am
* linbox/fflas/Makefile.am
* linbox/field/Makefile.am
* linbox/iterators/Makefile.am
* linbox/matrix/Makefile.am
* linbox/randiter/Makefile.am
* linbox/solutions/Makefile.am
* linbox/switch/Makefile.am
* linbox/util/Makefile.am
* linbox/util/xml/Makefile.am
* linbox/vector/Makefile.am
Modify all macro such that they use --with-package terminology + some new features
* macros/Makefile.am
* macros/atlas-check.m4
* macros/expat-check.m4
* macros/givaro-check.m4
* macros/gmp-check.m4
* macros/lidia-check.m4
* macros/maple-check.m4
* macros/ntl-check.m4
* macros/saclib-check.m4
modify sources and headers files such that the use the new terminology variable from linbox-config.h
* linbox/blackbox/archetype.h
* linbox/blackbox/compose.h
* linbox/blackbox/dense.h
* linbox/blackbox/diagonal.h
* linbox/blackbox/nag-sparse.h
* linbox/blackbox/ntl-toeplitz.h
* linbox/blackbox/ntl-toeplitz.inl
* linbox/blackbox/permutation.h
* linbox/blackbox/scalar-matrix.h
* linbox/blackbox/sparse.h
* linbox/blackbox/triplesbb.h
* linbox/blackbox/zero-one.h
* linbox/blackbox/zero-one.inl
* linbox/fflas/fflas.h
* linbox/fflas/fflas_fgemm.inl
* linbox/field/abstract.h
* linbox/field/archetype.h
* linbox/field/envelope.h
* linbox/field/gf2.h
* linbox/field/givaro-gfq.h
* linbox/field/givaro-zpz.h
* linbox/field/gmp-rational.h
* linbox/field/lidia-gfq.h
* linbox/field/modular.h
* linbox/field/ntl-RR.h
* linbox/field/ntl-ZZ_p.h
* linbox/field/ntl-ZZ_pE.h
* linbox/field/unparametric.h
* linbox/matrix/dense.h
* linbox/matrix/dense.inl
* linbox/matrix/sparse.h
* linbox/matrix/sparse.inl
* linbox/randiter/gf2.h
* linbox/randiter/gmp-rational.h
* linbox/randiter/lidia-gfq.h
* linbox/randiter/modular.h
* linbox/randiter/nonzero.h
* linbox/randiter/unparametric.h
* linbox/util/xml/field-reader-analyzer.h
* linbox/util/xml/reader-blackbox-factory.h
* configure.in: Modification for new features
* tests/Makefile.am: disable some tests and put tests in right place according to their dependencies
* tests/test-lsp.C: changing the verification of result to be faster
* Hacking: Adding new way of installation
* maple stuff installation have been improved
2003-12-15 Saunders/Wan/Lobo
* lbmaple.C: getField -> field, template for triplebb.
2003-11-18 Saunders/Wan
* merge newBB branch into main branch (note: there is an indented section of
branch changelog notes down below.)
2003-10-21 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
Modify few Makefile.am in order to synchronise make dist and make check
* linbox/algorithm/Makefile.am adding lsp.h
* linbox/field/Makefile.am add modular-balance-int.h hom.h image-field.h
* linbox/blackbox/Makefile.am add missing files
* tests/Makefile.am fixe some bugs. move test-gssv.C to EXTRA PROGRAM
2003-10-08 wan@udel.edu
* linbox/matrix/dense.inl, remove the width in write member function, since it causes problem for some ring and it is not necessary, for example, linbox integer doesnot have fixed length
2003-09-29 wan@udel.edu
* linbox/util/field-zxpy.h, linbox/vector/vector-domain.h,.inl, linbox/field/givaro-zpz.h,modular-balance-int.h,modular-byte.h,modular-int.h,modular-short.h,modular.h,unparametric.h, add a new member reset to FieldAXPY, set initial accumulator's value to zero, thus we can put an instantiation of FieldAXPY into field of VectorDomain, instead of temporily instantiating FieldAXPY whenever we need it.
2003-09-27 wan@udel.edu
* linbox/field/ntl* linbox/randiter/mersenne-twister.* linbox/randiter/unparametric.h, make them work with the new random memebr function.
2003-09-27 wan@udel.edu
* linbox/blackbox/dense.h, change the parameter RandIter &iter --> const RandIter &iter in its constructor.
2003-09-27 wan@udel.edu
* linbox/randiter/*.h, make each random member function to be a const member function in each random class, since the member will not change the instantiation of the random class at all, just generate a random element.
* linbox/matrix/dense.inl, make RowIterator and ColIterator support operator --() and operator-- (int).
2003-09-26 wan@udel.edu
* linbox/matrix/dense.h, fit to g++3 change, add std to endl.
2003-08-15 saunders@udel.edu
* sundry minor doc++ tweaks.
2003-08-14 saunders@udel.edu
* sundry doc++ comments in doc/ examples/ and tests/
* added files (in newbB branch)
linbox.dxx doc/fix-index examples/ex-integers.C
examples/examples.dxx examples/blackbox/blackbox.dxx
examples/fields/fields.dxx tests/tests.dxx
* linbox/blackbox/sparse.h added default sizes
* removed FROZEN, it's contents added to CHANGED-INTERFACES
2003-08-13 saunders@udel.edu
* test-generic.h blackbox tests simplified, testBB() added
* ntl-zz_p.h field fixed
* tests/test-ntl-toeplitz/hankel to newBB.
begin newBB branch notes
2003-11-05 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
* add file /linbox/field/modular-double.h
* /linbox/field/modular.h include modular-double.h
2003-10-23 wan@udel.edu
* linbox/blackbox/dense.h, in the operator= , I forced the assignment for field _F.
2003-10-22 wan@udel.edu
* tests/Makefile.am, comment test-lsp, since g++ complains its syntax error.
* linbox/blackbox/compose.h, correct a few typos.
2003-10-22 wan@udel.edu
* linbox/field/modular-byte.h, minor change for write member function, output the modulus as a int instead of a char.
2003-10-22 wan@udel.edu
* linbox/blackbox/compose.h, add specialization for Compose<Blackbox, Blackbox>
2003-10-08 wan@udel.edu
* linbox/matrix/dense.inl, remove the width in write member function, since it causes problem for some ring and it is not necessary, for example, linbox integer doesnot have fixed length
2003-09-29 wan@udel.edu
* linbox/util/field-zxpy.h, linbox/vector/vector-domain.h,.inl, linbox/field/givaro-zpz.h,modular-balance-int.h,modular-byte.h,modular-int.h,modular-short.h,modular.h,unparametric.h, add a new member reset to FieldAXPY, set initial accumulator's value to zero, thus we can put an instantiation of FieldAXPY into field of VectorDomain, instead of temporily instantiating FieldAXPY whenever we need it.
2003-09-27 wan@udel.edu
* linbox/field/ntl* linbox/randiter/mersenne-twister.* linbox/randiter/unparametric.h, make them work with the new random memebr function.
2003-09-27 wan@udel.edu
* linbox/blackbox/dense.h, change the parameter RandIter &iter --> const RandIter &iter in its constructor.
2003-09-27 wan@udel.edu
* linbox/randiter/*.h, make each random member function to be a const member function in each random class, since the member will not change the instantiation of the random class at all, just generate a random element.
* linbox/matrix/dense.inl, make RowIterator and ColIterator support operator --() and operator-- (int).
2003-09-26 wan@udel.edu
* linbox/matrix/dense.h, fit to g++3 change, add std to endl.
2003-08-15 saunders@udel.edu
* sundry minor doc++ tweaks.
2003-08-14 saunders@udel.edu
* sundry doc++ comments in doc/ examples/ and tests/
* added files (in newbB branch)
linbox.dxx doc/fix-index examples/ex-integers.C
examples/examples.dxx examples/blackbox/blackbox.dxx
examples/fields/fields.dxx tests/tests.dxx
* linbox/blackbox/sparse.h added default sizes
* removed FROZEN, it's contents added to CHANGED-INTERFACES
2003-08-13 saunders@udel.edu
* test-generic.h blackbox tests simplified, testBB() added
* ntl-zz_p.h field fixed
* tests/test-ntl-toeplitz/hankel to newBB.
end newBB branch notes
2003-08-09 saunders@udel.edu
* sundry documentation mods
* blackbox/companion.h fix
2003-08-08 Rich Seagraves <seagrave@cis.udel.edu>
* linbox/util/xml/field-reader-analyzer.C:
* linbox/util/xml/reader-blackbox-factory.C: Removed files (depricated)
* gmp++/gmp++_int.h:
* linbox/util/gmp++/gmp++_int_misc.C: Added operator bool to LinBox Integer class (casts integer to remainder upon division by 2)
* linbox/blackbox/archetype.h: Add XML support to BlackboxArchetype
* linbox/blackbox/companion.h: Add XML support
* linbox/blackbox/compose.h: Added partial XML support (incomplete)
* linbox/blackbox/scalar-matrix.h:
* linbox/blackbox/triplesbb.h:
* linbox/blackbox/zero-one.h:
* linbox/blackbox/zero-one.inl:
* linbox/blackbox/dense.h:
* linbox/blackbox/nag-sparse.h:
* linbox/blackbox/ntl-toeplitz.h:
* linbox/blackbox/ntl-toeplitz.inl:
* linbox/matrix/sparse.inl: Modified XML Reading, Writing methods for these Blackboxes
* linbox/field/abstract.h:
* linbox/field/envelope.h: Added reading & writing methods to envelope & abstract
* linbox/field/givaro-zpz.h:
* linbox/field/modular.h:
* linbox/field/unparametric.h: Modified XML Reading, Writing methods for these Blackboxes
* linbox/field/archetype.h: Added XML support to Field archetype (incomplete)
* linbox/util/xml/field-reader-analyzer.h:
* linbox/util/xml/reader-blackbox-factory.h: Completed experimental version of FieldReaderAnalyzer & ReaderBlackBoxFactory
* linbox/util/xml/Makefile.am: Added entries for reader-blackbox-factory.h & field-reader-analyzer.h
2003-08-07 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/element/Makefile.am (EXTRA_DIST): Remove abstract.dxx
* linbox/Makefile.am (EXTRA_DIST): Remove interfaces.dxx
* RELEASE-INSTRUCTIONS: Add this file
* tests/Makefile.am (BASIC_TESTS): Remove test-givaro-zpz
(EXTRA_PROGRAMS): Add $(GIVARO_TESTS)
* linbox/algorithms/wiedemann.inl (precondition): Allocate all
Compose objects dynamically
* tests/test-solve.C (main): Restore call to
testPreconditionedSolve
* linbox/field/gmp-rational.h (GMPRationalField::read):
Compilation fix
2003-07-25 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/algorithms/block-lanczos.inl: Add calls for detailed
timing information
2003-07-20 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/field/modular.inl (mulColDenseSpecialized): Compilation
fixes
* linbox/algorithms/block-lanczos.h (class BlockLanczosSolver):
Initialize _MD in all constructors
2003-07-23 wan@udel.edu
* linbox/field/modular-int.h, linbox/field/modular-short.h, linbox/field/modular-byte.h add MVProductionDomain specialization.
2003-07-21 Rich Seagraves <seagrave@cis.udel.edu>
* linbox/util/xml/field-reader-analyzer.C:
* linbox/util/xml/field-reader-analyzer.h: added files for Reader->Field conversion
* linbox/util/xml/reader-blackbox-factory.h:
* linbox/util/xml/reader-blackbox-factory.C: added files for Reader->BlackBox conversion
* linbox/blackbox/dense.h:
* linbox/matrix/dense.h: Added XML support to DenseMatrix class
* linbox/blackbox/diagonal.h: "" Diagonal class
* linbox/blackbox/nag-sparse.h: "" NAGSparse class
* linbox/blackbox/ntl-toeplitz.h:
* linbox/blackbox/ntl-toeplitz.inl: "" NTLToeplitz class
* linbox/blackbox/permutation.h: "" Permutation class
* linbox/blackbox/scalar-matrix.h: "" ScalarMatrix class
* linbox/blackbox/sparse.h:
* linbox/matrix/sparse.h: "" SparseMatrix class
* linbox/blackbox/triplesbb.h: "" TriplesBB class
* linbox/blackbox/zero-one.h:
* linbox/blackbox/zero-one.inl: "" ZeroOne class
* linbox/field/gf2.h: "" GF2 class
* linbox/field/gf2.inl:
* linbox/field/givaro-gfq.h: "" GivaroGfq class
* linbox/field/givaro-zpz.h:
* linbox/field/givaro-zpz.inl: "" GivaroZpz class
* linbox/field/gmp-rational.h: "" GMPRationalField class
* linbox/field/lidia-gfq.h: "" LidiaGfq class
* linbox/field/modular.h: "" Modular class
* linbox/field/ntl-RR.h: "" NTL::RR specilization of Unparametric
* linbox/field/ntl-ZZ_p.h: "" NTL::ZZ_p ""
* linbox/field/ntl-zz_p.h: "" NTL::zz_p ""
* linbox/field/ntl-ZZ_pE.h: "" NTL::ZZ_pE ""
* linbox/field/unparametric.h: Added XML support to UnparametricField
* linbox/randiter/gmp-rational.h: "" GMPRationalRandIter class
* linbox/randiter/lidia-gfq.h: "" LidiaGfq randiter
* linbox/randiter/modular.h: "" Modular randiter
* linbox/randiter/nonzero.h: "" NonZero randiter
* linbox/randiter/unparametric.h: "" UnparametricRanditer
* linbox/util/xml/linbox-reader.h:
* linbox/util/xml/linbox-writer.h:
* linbox/util/xml/xml-tree.h: Updated XML reading & writing API
2003-07-21 saunders@udel.edu
* Documentation modifications in doc/doc.dxx linbox/archetypes.dxx linbox/linbox.dxx
linbox/blackbox/blackbox.dxx linbox/blackbox/direct-sum.h
linbox/element/abstract.h linbox/element/archetype.h
linbox/element/envelope.h
Removed Files:
linbox/interfaces.dxx linbox/wrappers.dxx
linbox/element/abstract.dxx linbox/element/archetype.dxx
2003-07-21 wan@udel.edu
* linbox/field/modular-balance-int.h, add balance representation for modular field.
* tests/test-modular-balance-int.C, add test for balance representation for modular field.
2003-07-14 saunders@udel.edu
* tests/Makefile.am givaro test now works
* various.dxx some doc tweaks
2003-07-11 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/algorithms/block-lanczos.inl (iterate): Add TIMER_* calls
to report detailed timings when DETAILED_PROFILE is set
* linbox/randiter/mersenne-twister.h:
* linbox/randiter/mersenne-twister.C: Eliminate #pragma directives
2003-07-10 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/blackbox/matrix-blackbox.h (class MatrixBlackbox): Add
this file
2003-07-06 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/matrix/transpose.h (class TransposeMatrix): Add missing
Row, Col, Column declarations
* linbox/field/modular.h (class MVProductDomain): Add
specializations for uint8, uint16, and uint32
2003-07-05 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/matrix/matrix-domain.h (class MVProductDomain): Create
this class and move mulColSpecialized for dense vectors into it
* linbox/algorithms/block-lanczos.inl (solve): Try checking system
solution even if iterate returned false, just in case enough of
the Krylov basis was obtained to get the true solution.
2003-07-03 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/blackbox/diagonal.h (Diagonal::apply): Don't resize the
output
* linbox/blackbox/transpose.h (class Transpose): Parameterize by
black box; add versions of apply and applyTranspose separately
parameterized by vector
* linbox/blackbox/compose.h (class Compose): Add versions of apply
and applyTranspose separately parameterized by vector
(class Compose): Parameterize by blackboxes
* tests/test-sparse.C (runSparseMatrixTests): Disable testing
sparse matrix against non-dense vectors, since I don't support
them any more
* linbox/blackbox/sparse.h (class SparseMatrix): Add versions of
apply and applyTranspose parameterized by two vector types; add
MatrixDomain _MD; Use _MD for arithmetic and eliminate
specializations as they are no longer needed; eliminate
applyTransposeSpecialized
* linbox/matrix/matrix-domain.h (MatrixDomain::blackboxMul):
Rename blackboxMulLeft, blackboxMulRight; parameterize by black
box and not by vector
2003-07-02 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/solutions/solve.h (solve): Eliminate Vector from
BlockLanczosSolver declaration; parameterize by black box
* linbox/algorithms/block-lanczos.inl (iterate): Declare tmp and
tmp1 locally
(mulin): Use Vector<Field>::Dense::iterator rather than Vector
* linbox/algorithms/block-lanczos.h (class BlockLanczosSolver):
Eliminate v and Av
(class BlockLanczosSolver): Eliminate _t1; replace _t with _tmp
* linbox/algorithms/block-lanczos.inl (solve): Parameterize by
black box and vector
(iterate): Ditto
(solve): Eliminate v and Av
* linbox/algorithms/block-lanczos.h:
(class BlockLanczosSolver): Make generic wrt internal matrix type
used; no longer parameterize by vector
* linbox/algorithms/block-lanczos.inl (compute_Winv_S): Compute
and return rank of V^T AV
(iterate): Keep track of Ni and add trace statements to report it;
abort if total of Ni's is too large
(iterate): Use Vi = 0 as termination condition
(iterate): Add variable progress_interval giving number of
iterations between each call to commentator.progress
(compute_Winv_S): Don't call commentator.progress unless
DETAILED_TRACE is set
2003-07-09 saunders@cis.udel.edu
* field/ntl-zz_p.h unparametric.h fix yesterday's version - syntax of new constructor from p,e.
2003-07-08 saunders@cis.udel.edu
* Documentation tweaks, especially vis a vis @memo entries.
* See doc/doc.dxx for discussion of documentation conventions.
* For now, let's try to make the html documentation work and leave tex for later.
2003-07-07 Rich Seagraves <seagrave@cis.udel.edu>
* macros/expat-check.m4: Disabled automatic detection of expat in system standard locations, expat must now be manually specified in configure step
2003-06-30 Rich Seagraves <seagrave@cis.udel.edu>
* macros/expat-check.m4:
* linbox/blackbox/triplesbb.h:
* linbox/field/modular.h: Files altered to use "XMLENABLED" pre-processor flag instead of "RWENABLED"
* linbox/matrix/sparse.h:
* linbox/matrix/sparse.inl:
* linbox/blackbox/sparse.h:
* linbox/blackbox/sparse.inl: Added XML support to all SparseMatrix variants
* linbox/blackbox/nag-sparse.h: Added XML support to NAGSparse blackbox
* linbox/blackbox/zero-one.h: Added XML support to ZeroOne blackbox
* linbox/util/xml/linbox-reader.h: Added constructor from initalized Writer
* linbox/util/xml/linbox-writer.h: begun adding ability to traverse written Tag structure, add tags BEFORE as well as after the current Tag
2003-06-26 Rich Seagraves <seagrave@cis.udel.edu>
* add macros/expat-check.m4,
* add linbox/util/xml,
* add linbox/util/xml/linbox-reader.h,
* add linbox/util/xml/linbox-writer.h,
* add linbox/util/xml/xml-tree.h,
* add linbox/util/xml/README,
* add linbox/util/xml/Makefile.am: Defines LinBox/XML framework
* configure.in: Alters build system to make expat package optional dependency
* linbox/util/Makefile.am: make xml a subdir of util
* linbox/blackbox/triplesbb.h: Add XML reading & Writing to TriplesBB class
* linbox/field/modular.h: Add XML reading & Writing to ModularBase class
* linbox/solutions/det.h: commented out debug output
2003-06-24 wan@udel.edu
* linbox/field/modular-int.h, linbox/field/modular-short.h, linbox/field/modular-byte.h, use a fast way to compute 2^64 % modulus.
* linbox/tests/test-modular-int.h, change the modulus to a 30 big integer.
2003-06-20 wan@udel.edu
* linbox/blackbox/butterfly.h, linbox/blackbox/dense.h
* linbox/blackbox/dif.h,linbox/blackbox/inverse.h
* linbox/blackbox/sparse.h, linbox/blackbox/submatrix.h : for high performance reason, and a field only occupies a small memory, change const Field& to const Field in private field.
2003-06-16 wan@udel.edu
* add linbox/field/modular-int.h, modular-short.h , modular-byte, combining the trick of NTL for arithmetic operation and dot product of modluar<uint32>
* add linbox/tests/test-modular-int.C, test-modular-short.C, test-modular-byte.C.
* linbox/randiter/modular.h, remove field _card, since it only uses to initialize _size.
2003-06-12 wan@udel.edu
* linbox/field/givaro-gfq.h, enforce that the cardinality must be <2^16, for givaro-gfq
* linbox/randiter/modular.h, replace static-cast with init in the random function.
2003-06-05 saunders@udel.edu
* work on documentation in linbox/blackbox (particularly blackbox.dxx and archetype.h).
* Note: renamed blackboxes.dxx to blackbox.dxx and removed some cruft files.
2003-06-03 zhendong wan
* linbox/blackbox/ntl-sylvester.inl remove the warning message.
2003-06-03 zhendong wan
* linbox/field/ntl-zz-p.h, set default prime to be 65521
* tests/test-sum.C, correct a type name error.
2003-06-03 zhendong wan
* linbox/util/commentator.C, tests/test-commentator.C, temporarily fixed the bug in test-commentator, left a little memory leaking.
2003-06-02 saunders@udel.edu
* blackbox/zero-one.h, .inl: mods to avoid g++ 3.2.1 typename warnings.
2003-05-30 saunders@udel.edu
* field/unparametric.h field/ntl-zz_p.h tests/test-ntl-zz_p.cpp: put in prototype of a constructor
* I propose to have all field classes have a constructor from a prime and optionally an exponent --
* FieldClass(integer q) if a prime field class, (allow q = 0)
* FieldClass(integer q, size_t e) if an extension of degree e.
2003-05-30 saunders@udel.edu
* tests/test-commentator.C - adjust so passes on solaris. Note: there is apparently still a bug to fix.
2003-05-30 zhendong wan <wan@mail.eecis.udel.edu>
* linbox/blackbox/null-matrix.h: remove getInstance member function. Make default constructor public, remove static member nullMatrix
* linbox/blackbox/direct-sum.h: update the include file for NullMatrix.
* tests/Makefile.am: comment some tests lacking of source codes
* linbox/blackbox/Makefile.am: put in null-matrix.h direct-sum.h
2003-05-29 saunders@udel.edu
* change name to null-matrix.h and depreciate applyIn and a doc++ comment.
2003-05-29 zhendong wan <wan@mail.eecis.udel.edu>
* linbox/blackbox/nullmatrix.h : added a new singelton class. a matrix of size 0
* linbox/black/direct-sum.h : solved some ugly problem
2003-05-28 zhendong wan <wan@mail.eecis.udel.edu>
* tests/test-direct-sum.c: update the tests for DirectSum
2003-05-28 austin lobo <alobo@udel.edu> david saunders <saunders@udel.edu>
* Changed blackbox/direct-sum.h to allow for 0 x 0 blocks
* Changed blackbox/scalar-matrix.h to properly handle 0 x 0 matrix
* Added test for 0 x 0 in tests/test-direct-sum.C
* Added test for 0 x 0 in tests/test-scalar-matrix.C
* New blackbox/frobenius.h and tests/test-frobenius.C
2003-05-27 zhendong wan <wan@mail.eecis.udl.edu>
* linbox/blackbox/companion.h, tests/test-companion.h: revised the Compa
nion class, remove template parameter Polynomial, modify the constructor with ra
ndom genertor.
2003-05-27 saunders@udel.edu
* blackbox/companion.h, tests/test-companion.h: added class inherits from triplesbb
* tests/test-triplesbb.C: add test from other constructor
2003-05-26 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
* linbox/FFLAS/fflas.h : modify field_trsm to be closed to blas standard
* linbox/FFLAS.fflas.inl : put macro LEFT,RIGHT,UNIT,NOUNIT,LOWER,UPPER, to specialize field_trsm function
* linbox/algorithms/lsp.h : adding function to retrieve index of zero and non-zero rows in S.
* linbox/algorithms/lsp-tools.h: adding function to handle row permutation , synchronization with field_trsm
2003-05-23 saunders@udel.edu
* tests/test-generic.h: make testSmallBlackbox work, put it in with linearity and transpose in testBlackbox.
2003-05-20 saunders@udel.edu
* linbox/blackbox/compose.h: add constructor from references, &A &B.
* linbox/blackbox/sum.h: add constructor from refs, &A &B; remove cloning; copy field for speed.
* linbox/field/ntl-zz_p.h: typdef to name NTL_zz_p
* macros/atlas-check.m4: remove f77blas refs which break things for me. Hope not needed by others.
* tests/test-butterfly.C: also check Compose constructor with refs (vs ptrs)
* tests/test-generic.h: fix vector size bug in testTranspose().
* tests/test-rank.C: add testZeroAndIdentRank().
* tests/test-sum.C: also check Sum constructor with refs (vs ptrs), calls testBlackbox() too.
* Added Files: linbox/blackbox/direct-sum.h tests/test-direct-sum.C:
DirectSum BB is direct sum of 2 BBs, similar in structure to Compose, Sum.
2003-05-14 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* tests/Makefile.am (BASIC_TESTS): Reintroduce test-solve
* tests/test-solve.C (testSingularInconsistentSolve): Check status
== ...INCONSISTENT, not status == SINGULAR
2003-04-23 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
* linbox/algorithms/lsp.h: delete typo errors
* linbox/algorithms/lsp-tools.h : clarify Compute-G function
* tests/Makefile.am: retrieve test-lsp
2003-04-19 -bds - comment out non working tests in tests/Makefile.am
2003-04-11 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
* linbox/FFLAS/FFFMMBLAS.h: use of init and convert with double in MatGfq2MatDouble,
MatGfq2MatDouble_Triangular and MatDouble2MatGfq
* linbox/field/givaro-zpz.h: define convert to double
2003-04-11 Austin Lobo <alobo@cis.udel.edu>
* Added rules for ntl-sylvester and test-ntl-sylvester
* Added ntl-sylvester.h ntl-sylvester.inl to the repository
2003-04-10 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
* linbox/algorithms/lsp.h : optimization
* linbox/FFLAS.fflas.inl : new version of Field_trsm (more effficient)
2003-04-04 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
* linbox/field/givaro-zpz.h: fix bug with init from double
* linbox/field/givaro-zpz.inl: fix some warnings
* linbox/FFLAS/FFFMMBLAS.h: fix some bugs
* linbox/FFLAS/fflas.inl: fix some bugs
* linbox/algorithms/lsp.h: adding <iostream.h> header
* linbox/tests/test-lsp.C: adding test for LSP factorization with atlas library
* linbox/tests/Makefile.am: adding tests and variables for atlas library
2003-04-02 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
* linbox/matrix/dense.h: adding parameter _ptr as a pointer on the first Element of the matrix
adding FullIterator() function (returning _ptr)
2003-04-02 Jean-Guillaume Dumas <Jean-Guillaume.Dumas@imag.fr>
* gmp++/gmp++_int.h: Backed out my commit
* linbox/util/gmp++/gmp++_int_misc.C: Backed out my commit
* macros/gmp-check.m4 : added GMP_VERSION Makefile variable to enable
gmp-3.xxx compatibility
* All this is made to have fully SIMILAR gmp++ files within givaro
* and linbox (e.g. no "#include <linbox-config.h>" nor
* "#include <givaro-config.h>"
2003-04-02 Giorgi Pascal <pascal.giorgi@ens-lyon.fr>
* linbox/macros/atlas-check.m4: adding checking for atlas tuning library
* linbox/configure.in: adding check for atlas
* linbox/field/givaro-zpz.h: adding init function from double
* linbox/algorithms/lsp.h :
* linbox/algorithms/lsp-tools.h : algorithm to compute the LSP decomposition of a dense matrix stored contiguously by row.
* linbox/FFLAS/ : new directory containing FFLAS routine of jean-guillaume and clement and also LinBox wrappers
* linbox/FFLAS/FFFMMBLAS.h: FFLAS routine (j-g and clement code)
* linbox/FFLAS/ FFFMMBLAS.inl: FFLAS routine (j-g and clement code)
* linbox/FFLAS/lin_wrap_c++.h: FFLAS routine (j-g and clement code)
* linbox/FFLAS/lin_wrap_c++.inl: FFLAS routine (j-g and clement code)
* linbox/FFLAS/fflas.h: wrappers including field_dgemm and field_trsm
* linbox/FFLAS/fflas.inl: wrappers including field_dgemm and field_trsm
2003-04-01 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/matrix/matrix-domain.inl:
* linbox/matrix/matrix-domain.h: Back out previous changes
2003-03-31 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* tests/test-vector-domain.C (main): Add test for uint8
* linbox/field/modular.h (class FieldAXPY<uint8>): typedef uint8
Element
2003-03-29 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/vector/bit-vector.h: Be generic wrt word length, so we
can use 64-bit arithmetic when available
* linbox/util/debug.h: Remove GCCisms
2003-03-27 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* gmp++/gmp++_int.h: Add operator unsigned char
* gmp++/gmp++_int.h: Backed out previous commit, since it broke
the build for the entire library
* linbox/util/gmp++/gmp++_int_misc.C: Backed out previous commit,
since it broke the build for this file
* tests/test-modular.C (main): Add test for Modular<uint8>
* linbox/field/modular.h (class FieldAXPY): Add specialization for
uint8
(class Modular): Add specialization for uint8
(class DotProductDomain): Add specialization for uint8
2003-03-26 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/matrix/matrix-domain.h (class MatrixVectorDomain): Create
this class and move mulRowSpecialized and mulColSpecialized
functions into it
(class MatrixDomain): Inherit MatrixVectorDomain
(class MatrixDomain): Move _VD to MatrixVectorDomain
* linbox/field/gf2.inl (class Diagonal): Add specialization for GF2
* linbox/blackbox/diagonal.h (class Diagonal): Use RandIter, not
RandomIterator
* linbox/field/gf2.inl (addinSpecialized): Remember to increment
other iterators
* tests/test-gf2.C (testDotProductGF2): Create new version for
dense/sparse test
(testDotProductGF2): Use generic vector names
(main): Use unsigned int rather than size_t for n
* linbox/field/gf2.inl (dotSpecializedDD): Use mask at the end of
the vector
* linbox/vector/bit-vector.inl (class BitVector::const_iterator):
Add operator = (iterator)
(class BitVector::iterator): Make const_iterator a friend
(class BitVector::reference): Ditto
(class BitVector::iterator, class BitVector::const_iterator): Add
equality checks with each other
* tests/test-gf2.C (testDotProduct): Enable this
* linbox/field/gf2.inl (class RandomSparseStreamGF2, class
RandomDenseStreamGF2): Require seed as a parameter
(addSpecialized): Swap x and y in if statement
(VectorDomain<GF2>::addinSpecialized): Fix call to copy
* linbox/vector/bit-vector.inl (operator ==): Make this operation
a bit smarter
2003-03-23 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/field/gf2.inl (class VectorDomain): Sprinkle this->
before copy liberally to make SGI CC happy
* linbox/vector/bit-vector.inl (operator =): Return *this
* tests/test-dif.C (testZeroApply): Remove neg_one
* linbox/util/field-axpy.h (FieldAXPY::operator =): Don't try to
assign the field
* tests/test-sparse.C (main): Don't use NonzeroRandIter for
RandomSparseStream -- that's already implied
* linbox/field/gf2.inl (class RandomDenseStreamGF2, class
RandomSparseStreamGF2): Rename from specializations of
RandomDenseStream and RandomSparseStream
* tests/Makefile.am: Enable test-gf2
* linbox/vector/bit-vector.inl (BitVector::resize): Set _size
* linbox/randiter/mersenne-twister.h
(MersenneTwister::randomDoubleRange): Move this to the header file
* linbox/randiter/mersenne-twister.C (randomInt): Make this not
inline
* linbox/field/gf2.inl (class RandomDenseStream): Use the Mersenne
Twister
(class RandomSparseStream): Use the Mersenne Twister
* linbox/randiter/gf2.h (class GF2RandIter): Use the Mersenne
Twister
(GF2RandIter::random): Overload with two reference types rather
than using template <reference>
* linbox/field/gf2.h (class GF2): Eliminate template <reference>
declarations; instead use manual overloading. This means we only
support bool & and BitVector::reference. Use a straight bool on
const parameters
* linbox/vector/bit-vector.inl: Add inserter and extractor
overloading for BitVector::(const_)reference
* linbox/vector/Makefile.am (include_HEADERS): Add missing
bit-vector.inl
* linbox/randiter/Makefile.am (BASIC_HDRS): Add missing gf2.h
2003-03-16 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/util/commentator.C (Commentator::setMaxDepth,
Commentator::setMaxDetailLevel): Use ++i, not i++
* linbox/vector/subvector.h (Subvector::at): Use const_reference
rather than const reference
* linbox/vector/reverse.h (ReverseVector): Use const_reference
rather than const reference
* linbox/util/commentator.h (class Commentator): Add setMaxDepth
and setMaxDetailLevel to disabled commentator
2003-03-21 Austin Lobo <alobo2@washcoll.edu>
* test-generic.h: added testApply(..) and testApplyTranspose(...)
to testBlackBox, so that timings for these two
operations can be obtained.
* test-ntl-toeplitz.C, test-ntl-hankel.C: cosmetic changes to
give slightly cleaner status message.
* ntl-toeplitz.inl: modifications in the
limits of iterations to take into account the
possibility of rectangular matrices.
2003-03-20 Giorgi Pascal <pascal.giorgi@ens-lyon.fr>
* linbox/field/givaro-zpz.h fix bugs for init function
2003-03-15 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/randiter/modular.h (RandIter::RandIter): Copy _size and
_seed in copy constructor
* linbox/vector/stream.h (class RandomSparseStream): Use
MersenneTwister
(class RandomSparseStream): Declare the base random iterator
locally
* linbox/randiter/modular.h (Modular<uint32>::RandIter,
Modular<uint16>::RandIter): New specializations; use the new
Mersenne Twister random number generator
* linbox/randiter/Makefile.am (noinst_LTLIBRARIES): Add libranditer.la
(libranditer_la_SOURCES): Add mersenne-twister.C
(BASIC_HDRS): Add mersenne-twister.h
* linbox/Makefile.am (liblinbox_la_LIBADD): Add libranditer.la
2003-02-07 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/matrix/Makefile.am (BASIC_HDRS): Add matrix-domain.{h|inl}
* linbox/vector/Makefile.am (include_HEADERS): Add
vector-domain.{h|inl}
2002-03-04 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
* linbox/field/givaro-zpz.h added specialization for DotProductDomain and FieldAxpy
* linbox/field/givaro-zpz.inl added specialization for DotProductDomain and FieldAxpy
2003-03-03 Zhendong Wan <wan@cis.udel.edu>
*linbox/matrix/matrix-domain.h, include header file linbox/vector/vector-domain.h
2003-02-26 Austin Lobo <alobo@cis.udel.edu>
* linbox/blackbox/Makefile.am (include_HEADERS) : added ntl-hankel.h
* linbox/blackbox/ntl-hankel.h: added to library after testing
* linbox/blackbox/ntl-hankel.inl: added to library after testing
* linbox/tests/test-ntl-hankel.C: added to library after testing
* linbox/tests/Makefile.am: (test_ntl_hankel_SOURCES) added
* linbox/tests/test-generic.h: minor changes to make output clearer.
2003-02-25 Zhendong Wan <wan@mail.eecis.udel.edu>
*linbox/blackbox/zero-one.h, zero-one.inl. Correct one typo.
2003-02-25 Zhendong Wan <wan@mail.eecis.udel.edu>
*linbox/blackbox/zero-one.h, zero-one.inl. Seperate the Intereface and i
implementation, put them in seperate files.And make apply specialization for Modular<uint32> work.
2003-02-24 David Saunders and Zhendong Wan
* linbox/util/timer.h: #define __LINBOX_TIMER_H to avoid conflict with givaro timer
* tests/Makefile.am: reinstate givaro test
2003-02-22 David Saunders
* tests/test-solve.C: comment out "Sparse preconditioner", MethodTrait::Wiedemann::SPARSE
It's segfaulting, likely due to Compose changes.
* linbox/blackbox/transpose.h compose.h: remove cloning
* linbox/blackbox/triplesbb.h: coupla "typename"s to remove warnings.
* linbox/field/hom.h image-field.h: Introduce a scheme for homomorphic images
with injections in extension fields in mind, isoms in mind, ints to modular in mind.
Specializations are now needed.
* tests/test-hom.C test-image-field.C: Add prototype tests for 'em.
* tests/test-generic.h: make statistical randiter test more likely to pass.
2003-02-05 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* tests/Makefile.am (NTL_TESTS): Move test-ntl-toeplitz here
(NTL_TESTS): Move test-gssv_rank here
* tests/test-solve.C (testSingularPreconditionedSolve)
(testSingularInconsistentSolve): Use version of solve without
exceptions
* linbox/solutions/solve.h (solve): Return WiedemannSolver return
status rather than reference to solution vector. Require parameter
for certificate of inconsistency. New version that throws exceptions
* linbox/algorithms/wiedemann.h (class WiedemannSolver): Remove
exception code; have all functions return a new enum ReturnStatus
2003-02-04 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/vector/stream.h (class RandomSparseStream): Remove
specialization for dense vectors
* tests/test-generic.h: Remove testFieldAXPY
(runFieldTests): Remove call to testFieldAXPY
* linbox/vector/vector-traits.h (struct VectorCategories): Add
ZeroOneVectorTag, SparseVectorTag for GF(2)
2003-02-03 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* tests/test-vector-domain.C: Include test-generic.h
* tests/test-generic.h (testRandomIterator): Move from
test-modular.C here
(testDotProduct, testAddMul, testSubMul, testAXPY, testCopyEqual):
Move from test-vector-domain.C here
* tests/test-rank.C (testEliminationRank): New test
* linbox/matrix/sparse.h (class SparseMatrixBase): Add operator []
for row access
(class SparseMatrixBase): Add typedef ConstRow
* examples/blackbox/Makefile.am (EXTRA_PROGRAMS): Add random-rank
* linbox/matrix/Makefile.am (BASIC_HDRS): Add dense-rows-matrix.h
* linbox/vector/Makefile.am (include_HEADERS): Add bit-vector.h
* linbox/field/Makefile.am (BASIC_HDRS): Add gf2.h, gf2.inl
* linbox/algorithms/wiedemann.inl (makeLambdaSparseMatrix): Update
RandomSparseStream
* linbox/vector/stream.h (RandomSparseStream::RandomSparseStream):
Put probability parameter before vector dimension
* linbox/algorithms/gauss.h: Rename gauss_rankin as rankin and
gauss_rank as rank
* linbox/algorithms/block-lanczos.h (class BlockLanczosSolver):
* linbox/algorithms/block-lanczos.inl (solve):
* linbox/algorithms/lanczos.h (class LanczosSolver):
* linbox/algorithms/lanczos.inl (solve):
* linbox/solutions/solve.h (solve):
* linbox/algorithms/wiedemann.h (class WiedemannSolver):
* linbox/algorithms/wiedemann.inl: Port to new SolverTraits interface
* linbox/solutions/methods.h (struct WiedemannTraits): Move
Wiedemann-dependent traits here
(struct LanczosTraits): Move Lanczos-dependent traits here
(struct BlockLanczosTraits): New structure
(struct SolverTraits): Parameterize by MethodTraits and include an
instance thereof; remove all method-specific traits
* linbox/matrix/dense.h (DenseMatrixBase::columnDensity): New
function
* linbox/matrix/archetype.h (MatrixArchetype::columnDensity,
MatrixArchetype::transpose): New functions
* linbox/matrix/sparse.h (SparseMatrixBase::columnDensity): New
function
* linbox/solutions/methods.h (struct RankTraits): New structure
(struct MethodTrait): Add BlockLanczos
2003-03-4 Rich Seagraves <seagrave@cis.udel.edu>
* interfaces/maple/demonstration.mws: Finished
* interfaces/maple/lbmaple.C: Fixed the final bugs with LinBoxV
handling code.
* interfaces/maple/lbmaple.mpl.tail: ditto. Also, renamed LBApply and
LBApplyTranpose as LBaply and LBapplyTranspose to be like the rest of
the names
* interfaces/maple/lbmaple.mpl.tail: ditto.
* configure.in: Removed a double reference to matrix/Makefile in
AC_OUTPUT[...]
2003-03-4 Rich Seagraves <seagrave@cis.udel.edu>
* interfaces/maple/demonstration.mws: Updated file, should be working
* interfaces/maple/README: Added file
* interfaces/maple/lbmaple.mpl.tail: Edited final line, lbmaple.mpl
script now puts LinBox module into global maple namespace when run
2003-02-3 Rich Seagraves <seagrave@cis.udel.edu>
* interfaces/maple/lbmaple.mpl.tail: Fixed some interface bugs
* interfaces/maple/demosntration.mws: Updated file, it's almost finished
2003-01-30 Rich Seagrves <seagrave@cis.udel.edu>
* interfaces/maple/Makefile.am:
* interfaces/maple/Makefile.in.2:
* macros/maple-check.m4: Fixed bug in maple interface build setup
* linbox/blackbox/triplesbb.h: Added missing _max() function
2003-01-30 Rich Seagraves <seagrave@cis.udel.edu>
* tests/Makefile.am: Added entry for test-zero-one.C
* tests/test-zero-one.C: Fixed file, now complies with standard for tests
* macros/maple-check.m4: Now install maple interface in "prefix" directory
* interfaces/maple/Makefile.am: Ditto
* interfaces/maple/Makefile.in.2: Ditto as well :-)
* linbox/blackbox/triplesbb.h: Added file (copy of maplebb.h)
* interfaces/maple/maplebb.h: Removed file
* interfaces/maple/lbmaple.C: Now use TriplesBB blackbox instead of MapleBB
* linbox/blackbox/zero-one.h: Removed some warnings
2003-01-30 Saunders
* tests/test-scalar-matrix.C: simplify test to generic testBlackbox.
* tests/test-*.C: modify " ... test suite" header line of output to be preceeded
by blank line rather than followed.
* linbox/blackbox/hilbert.h, tests/test-hilbert.C: fix hilbert example
(for dense vector only) and insert a generic test
2003-01-29 Rich Seagraves <seagrave@cis.udel.edu>
* interfaces/maple/lbmaple.C: Killed all bugs
* interfaces/maple/lbmaple.mpl.tail: Ditto
* macros/maple-check.m4: Fixed problem with check whether specified maple dir
IS actually valid maple distribution
2003-01-29 Ahmet Duran and David Saunders
* examples/gen_superlu/*: examples added and also removed as a subdir within
linbox, i.e. linbox/gen_superlu/example/ subdir.
2003-01-29 Ahmet Duran and David Saunders
* linbox/gen_superlu/: directory and contents are a generic version of
superlu created by Ahmet. Basics work but header comments and the like
will be brought up to form in the near future,
* tests/test-gssv_rank: a proto test for gen_superlu, improvements will come.
* test/data/gssv_rank_data: input matrix for test-gssv_rank.
2003-01-29 Rich Seagraves <seagrave@cis.udel.edu>
* interfaces/maple/Makefile.am: Remove "if HAVE_MAPLE" check
* interfaces/maple/lbmaple.C: Fixed some bugs, warnings
* interfaces/maple/maplebb.h: Fixed some bugs, warnings
* interfaces/maple/lbmaple.mpl.head: Changed formatting a bit
* interfaces/maple/lbmaple.mpl.tail: Changed formatting a bit
* macros/maple-check.m4: Requires the flag "--enable-shared" to build maple interface,
made some fixes to ensure proper building of lbmaple.mpl, interface script
* linbox/blackbox/zero-one.h: Fixed in-file test functions
* tests/test-zero-one.C: Added test for ZeroOne blackbox
2003-01-20 Austin Lobo <alobo@cis.udel.edu>
* linbox/field/ntl-RR.h: Corrected clashes with LinBox namespace
2003-01-20 Austin Lobo <alobo@cis.udel.edu>
* tests/test-ntl-toeplitz.C: Corrected clashes with LinBox namespace
* linbox/blackbox/ntl-toeplitz.inl: Corrected clashes with LinBox namespace
* linbox/blackbox/ntl-toeplitz.h: Corrected clashes with LinBox namespace
2003-01-20 Austin Lobo <alobo@cis.udel.edu>
* tests/Makefile.am: reinstated test-ntl-toeplitz
* tests/test-generic.h: added LinBox namespace to commentator calls
2003-01-17 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* configure.in: Give a sane default for CFLAGS and CXXFLAGS
2003-01-17 David Saunders
* CHANGED-INTERFACES: coupla edits to bring this summary up to date a bit
2003-01-16 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* configure.in: Disable shared again
2003-01-14 David Saunders
*linbox/linbox.dxx, interfaces.dxx: Text edits. interfaces.dxx should be renamed archetypes.dxx.
Much text added to interfaces.dxx, some of it speculative. Further editing definitely desirable.
2003-01-14 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/matrix/matrices.dxx: Add missing
* interfaces/maple/Makefile.am (mapledir): Put all build rules
inside a conditional
* configure.in: Add linbox/matrix/Makefile
* linbox/matrix/matrix-domain.h: Move from linbox/field
* linbox/matrix/matrix-domain.inl: Move from linbox/field
* linbox/vector/vector-domain.h: Move from linbox/field
* linbox/vector/vector-domain.inl: Move from linbox/field/vector-domain.C
2003-01-11 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/blackbox/Makefile.am (BASIC_HDRS): Remove files listed
below
* */*.{h|C|inl}: Update to reflect changes below;
s/SparseMatrix0/SparseMatrix/
* linbox/matrix/dense.h: Move from blackbox/dense-base.h
* linbox/matrix/dense.inl: Move from blackbox/dense-base.inl
* linbox/matrix/sparse.h: Move from blackbox/sparse-base.h
* linbox/matrix/sparse.inl: Move from blackbox/sparse-base.inl
* linbox/matrix/transpose.h: Move from blackbox/transpose-matrix.h
* linbox/matrix/dense-submatrix.h: Move from blackbox/dense-submatrix.h
* linbox/matrix/dense-submatrix.inl: Move from blackbox/dense-submatrix.inl
* linbox/Makefile.am (SUBDIRS): Add matrix
2003-01-09 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/solutions/det.h (det): Give correct solution when deg
minpoly is odd
2003-01-08 Brad Botting <bbotting@math.uwaterloo.ca>
* linbox/blackbox/dense.h (DenseMatrix::read, DenseMatrix::write):
Use fully-qualified name for calling overloaded read and write
2003-01-08 Rich Seagraves <seagrave@cis.udel.edu>
* linbox/blackbox/dif.h (very top): changed __TRANSPOSE_H to __DIF_H
in the preprocessor test directive at the top of the file
* linbox/blackbox/sum.h (very top): changed __TRANSPOSE_H to __SUM_H
in the preprocessor test directive at the top of the file
* linbox/blackbox/zero-one.h: begin implementing modular template
specialization
* linbox/util/timer.h (class Timer): change members ut, st, and rt
into private members instead of public members (changed second
appearance of "public" to "private").
2003-01-08 Saunders
* tests/test-dif.C, blackbox/dif.h: difference of two matrices. Minor tweak on sum.h.
2003-01-08 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/blackbox/dense.h (DenseMatrix::write): Use os, not file
* tests/test-matrix-domain.C (testMatrixDomain): Don't refer to
stream.dim when constructing stream (remaining two
specializations)
2003-01-07 Brad Botting <bbotting@math.uwaterloo.ca>
* linbox/switch/Makefile.am (includedir): Install to switch, not
vector
2003-01-07 Bradford Hovinen <hovinen@algol.cis.udel.edu>
* linbox/algorithms/block-lanczos.h: Undefine _S, _M, _N for the
benefit of Solaris
2003-01-07 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/algorithms/wiedemann.h: #include sparse.h
* tests/test-matrix-domain.C (testMatrixDomain): Don't refer to
stream.dim when constructing stream
2002-1-7 Saunders
* tests/test-subvector.cpp: removed. Superceded by test-subvector.C
2002-1-6 Rich Seagraves <seagrave@cis.udel.edu>
* linbox/algorithms/massey-domain.h: changed swap() to std::swap()
* linbox/randiter/modular.h: changed endl to std::endl in commentator line
* linbox/blackbox/zero-one.h: added RawIterator and RawIndexIterator support to ZeroOne blackbox
2003-01-06 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* tests/test-common.C: #include <cmath>
2002-12-23 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* tests/test-common.h: Add incompleteGamma, chiSquaredCDF
* tests/test-modular.C (testRandomIterator): New test
(main): Call testRandomIterator
2002-12-18 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/* (LinBox): Doc++ fixes
* configure.in: Add check for latex
2002-12-16 Rich Seagraves <seagrave@cis.udel.edu>
* linbox/blackbox/zero-one.h: Added file
* linbox/blackbox/nag-sparse.h: Fixed some errors
2002-12-11 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/util/commentator.h: Include streambuf
* linbox/blackbox/butterfly.h (buildIndices): Remove check that _n
is nonnegative, since _n is unsigned
* linbox/blackbox/dense-base.inl (ColIterator::operator +): Make
this non-const
(ConstColIterator::operator +): Ditto
* linbox/blackbox/diagonal.h (class Diagonal):
* linbox/blackbox/moore-penrose.h (class MoorePenrose):
* linbox/blackbox/butterfly.h (class Butterfly):
* linbox/blackbox/dense.h (class DenseMatrix):
* linbox/blackbox/compose.h (class Compose):
* linbox/blackbox/scalar-matrix.h (class ScalarMatrix):
Parameterize on _Vector and typedef Vector
* linbox/field/modular.h: Cast to signed quantities when comparing
with 0
(LinBox): Add template <> when needed
* linbox/blackbox/butterfly.h (setButterfly): Eliminate unused
j_1, j_2
(setButterfly): Eliminate check for j < 0, as we are dealing with
unsigned quantities
* linbox/blackbox/archetype.h (class BlackboxArchetype): Remove
std:: from size_t
* linbox/algorithms/block-lanczos.inl (eliminate_col)
(find_pivot_row, permute): Take an std::vector for indices rather
than a pointer
(compute_Winv_S): Use _indices; don't declare indices
(compute_Winv_S, find_pivot_row, eliminate_col): Make these
non-const
* linbox/algorithms/block-lanczos.h (class BlockLanczosSolver):
Add vector _indices
* linbox/blackbox/sparse-base.inl (refEntry):
* linbox/field/vector-domain.C: Make sure there's an std:: before
pair
* linbox/util/debug.C:
* linbox/field/lidia-gfq.h:
* linbox/util/debug.h:
* linbox/blackbox/archetype.h: Get rid of 'using namespace std'
* linbox/blackbox/transpose.h (class Transpose): Use linbox_check
* linbox/blackbox/sparse.h:
* linbox/vector/stream.h: Parameterize all classes by _Vector and
typedef to Vector
* linbox/field/modular.h:
* tests/test-common.C:
* linbox/util/debug.h:
* linbox/field/vector-domain.C:
* linbox/field/vector-domain.h:
* linbox/blackbox/sparse-base.inl: Make sure there's an std::
before iostream references
* gmp++/gmp++_int.inl: Move constructor and destructor definitions
to gmp++_int.h to avoid compiler complaints
2002-12-09 Bradford Hovinen <bghovine@math.uwaterloo.ca>
* linbox/blackbox/dense-base.inl (RowIterator): Add operator +=
(ColIterator): Ditto
2002-12-10 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* tests/test-solve.C (main): Add CLI option for blocking factor
* linbox/algorithms/lanczos.inl (iterate): Use DETAILED_TRACE ala
block-lanczos.inl
* linbox/solutions/methods.h (SolverTraits::Preconditioner): Add
DEFAULT
* linbox/algorithms/wiedemann.inl (precondition): Add default case
label for unsupported preconditioners
* linbox/solutions/solve.h (solve): Add support for Block Lanczos
* tests/test-solve.C (testRandomSolve): New test
(main): Add calls to testRandomSolve for Lanczos and Block Lanczos
* linbox/solutions/methods.h (SolverTraits::Method): Add
BLOCK_LANCZOS
(struct SolverTraits): Add _blockingFactor and related methods
* linbox/algorithms/Makefile.am (include_HEADERS): Add
block-lanczos.{h|inl}
* linbox/algorithms/lanczos.inl (solve): Import code from
block-lanczos.inl to support more preconditioner types
2002-12-09 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* configure.in: Bump version number to 0.1.4
2002-12-07 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/util/commentator.C (Commentator::finishActivityReport):
Show the timings in the finalized report, if requested
* linbox/field/modular.h (Modular::negin): Fix in case where x ==
0
* linbox/util/commentator.C (Commentator::printActivityReport):
Reset _last_line_len if no length is given
* linbox/field/vector-domain.C (axpyinSpecialized): Add explicit
specializations for Dense/*
* linbox/field/matrix-domain.inl (mulColSpecialized): Fix
non-dense code
* linbox/field/matrix-domain.h (MatrixDomain::mulColSpecialized):
Create specializations for dense/* and have everything route to
that combination
* linbox/solutions/methods.h (SolverTraits::Preconditioner): Add
SYMMETRIZE, PARTIAL_DIAGONAL, and FULL_DIAGONAL
2002-12-06 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/blackbox/sparse-base.inl (write): Tweak output slightly
* linbox/blackbox/sparse.h (class MatrixTraits): Define
specialization for SparseMatrix0
* linbox/blackbox/dense-base.inl (ConstRowIterator::operator +):
Don't return a reference
(write): Tweak output slightly
2002-12-05 Rich Seagaves
* interfaces/maple/Makefile.in.1: fixed a small break in the build
stage
2002-12-01 Rich Seagraves
* interfaces/maple/Makefile.blank: removed, not needed
* interfaces/maple/lbmaple.mpl: also removed
* interfaces/maple/Makefile.in.1:
* interfaces/maple/Makefile.in.2: added files
* macros/maple-check.m4: modified for new maple build system
* interfaces/maple/lbmaple.C: fixed a bunch of boo-boo's
* interfaces/maple/maplebb.h: fixed a boo-boo
2002-12-01 Rich Seagraves
* interfaces/maple/lbmaple.C:
* interfaces/maple/lbmaple.mpl.head:
* interfaces/maple/lbmaple.mpl.tail:
* interfaces/maple/maplebb.h:
* interfaces/maple/Makefile.am:
* interfaces/maple/Makefile.blank:
* interfaces/maple/demonstration.mws:
* macros/maple-check.m4: Added files for interface
* configure.in: Added checks for interface build, installation
2002-12-01 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* configure.in (AC_OUTPUT): Add missing interfaces/Makefile
2002-11-30 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/field/matrix-domain.inl (axpyinRowRowRow)
(axpyinColColCol): Fix allocation sizes for t
* tests/test-matrix-domain.C (testInvMul, testInvLeftMulin,
testInvRightMulin): Split these into three cases: square,
over-determined, and under-determined. Use the product of the
matrix and its transpose (in the appropriate order) in the latter
cases
* linbox/algorithms/block-massey-domain.h (class
MasseyBlockDomain): Make _MD a MatrixDomain<Field>
* linbox/field/matrix-domain.inl: Comment out pow_apply and pow_horn
* linbox/blackbox/sparse-base.h (class SparseMatrix0WriteHelper)
(class SparseMatrix0ReadWriteHelper): Compilation fixes
(class SparseMatrix0Base): Update iterator interfaces to match the
dense matrix archetype
(class SparseMatrix0Base): Make Row a typedef of _Row
* linbox/blackbox/dense.h (DenseMatrix::read, DenseMatrix::write):
Swap order of arguments for consistency
* linbox/blackbox/dense-submatrix.h
(DenseSubmatrix::DenseSubmatrix): Take a reference rather than a
pointer
* linbox/blackbox/dense-base.inl (ConstRowIterator::operator ++)
(RowIterator::operator ++): Use ++*this rather than ++_row
* tests/Makefile.am (BASIC_TESTS): Add test-matrix-domain
* linbox/randiter/modular.h (ModularRandIter::ModularRandIter):
clock was always returning 0, so go back to time (NULL)
* linbox/blackbox/sparse.h (SparseMatrix0::write): Make
FORMAT_PRETTY the default
* linbox/blackbox/sparse-base.h (SparseMatrix0Base::write): Make
FORMAT_PRETTY the default
* linbox/field/Makefile.am (BASIC_HDRS): Add matrix-domain.h,
matrix-domain.inl
* linbox/util/commentator.C (Commentator::printActivityReport):
(Commentator::updateActivityReport):
(Commentator::finishActivityReport): Don't indent
* tests/test-trace.C:
* tests/test-sum.C:
* tests/test-submatrix.C:
* tests/test-sparse.C:
* tests/test-solve.C:
* tests/test-rank.C:
* tests/test-moore-penrose.C:
* tests/test-minpoly.C:
* linbox/blackbox/dense-base.inl:
* linbox/blackbox/sparse-base.inl:
* tests/test-matrix-domain.C:
* tests/test-inverse.C:
* tests/test-diagonal.C:
* tests/test-det.C:
* tests/test-dense.C:
* tests/test-butterfly.C:
* tests/test-vector-domain.C:
* tests/test-generic.h: Eliminate calls to commentator.indent
* linbox/util/commentator.C (Commentator::report): Don't initially
indent
* linbox/util/commentator.h (class MessageClass): Define
smartStreambuf, build _stream from that; add Commentator to all
constructors; add copy constructor
* linbox/util/commentator.C (Commentator::indent): Make this const
2002-11-29 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/field/vector-domain.h (class VectorDomain): Add support
for add/sub/mul/axpy with vectors of different representations,
using GenericVectorTag to control the number of specializations
required
* linbox/vector/vector-traits.h (struct VectorCategories): Add
GenericVectorTag; make everything else inherit that
* linbox/blackbox/dense-base.inl (DenseMatrixBase::read): Return
the input stream
* linbox/blackbox/dense-base.h (DenseMatrixBase::read,
DenseMatrixBase::write): Swap order of arguments, to be consistent
with SparseMatrix0Base
* linbox/blackbox/dense-base.h:
* linbox/blackbox/dense-submatrix.h:
* linbox/blackbox/sparse-base.h: Add struct MatrixTraits
* linbox/blackbox/dense-base.h:
* linbox/blackbox/sparse-base.h:
* linbox/blackbox/dense-submatrix.h: Make Element a typedef from
_Element
2002-11-28 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/blackbox/dense-submatrix.h (class DenseSubmatrix): Use a
reference rather than a pointer
* linbox/blackbox/sparse-base.h (class
SparseMatrix0ReadWriteHelper): Make all methods const
2002-11-26 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/blackbox/Makefile.am (BASIC_HDRS): Add transpose-matrix.h
* linbox/blackbox/dense-submatrix.h (struct MatrixTraits): Define
* linbox/blackbox/dense-base.h (struct MatrixTraits): Define
* linbox/field/matrix-domain.h (struct MatrixCategories)
(struct MatrixTraits): Define
* linbox/field/matrix-domain.inl: Moved all the arithmetic from
matrix-domain.C here, in accordance with the convention
* linbox/field/matrix-domain.h: Added detailed documentation,
cleaned up the interface slightly, and added support for matrix
traits. Added read, write, neg, negin, axpy, matrix-vector and
matrix-black box operations
* linbox/field/vector-domain.h (class VectorDomain): Extend axpyin
so that it can work with vectors of more than one
representation. Create specializations for input and output
vectors of different representations that just use copy. For the
benefit of MatrixDomain
2002-11-27 Rich Seagraves and David Saunders
* interfaces, interfaces/maple: directories added to hold the Maple
linbox interface (which Rich will put there soon ).
* Makefile.am interfaces/Makefile.am: subdir entries.
2002-10-29 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/blackbox/dense-base.inl (ColIterator::operator +): Make this
work like a real + operator
(RowIterator::operator +): Ditto
(RowIterator, ConstRowIterator): Fixed some ID10T errors
2002-10-28 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/blackbox/dense-submatrix.inl: Ditto below
* linbox/blackbox/dense-base.h: Rename ColOfRowsIterator as
RowIterator; similarly with RowOfColsIterator
2002-10-28 Zhendong Wan and David Saunders
* linbox/blackbox/dense.h: tweak doc++, delete RowIterator and
ColIterator types in matrix class. Change name of RawIndexIterator
to RawIndexedIterator.
* linbox/blackbox/dense.inl: update impl to reflect dense.h changes.
2002-10-28 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* tests/test-trace.C (testDiagonalTrace): Make sure vectors get
declared to the right length
* tests/test-sum.C (testZeroApply): Make sure vectors get declared
to the right length
* linbox/vector/stream.h (RandomSparseStream::get,
RandomDenseStream::get): Don't try to resize the vector
* linbox/blackbox/dense-base.inl (write): Make output format a
little prettier
* linbox/algorithms/lanczos.inl (solve): Length of b1 should be
row dimension of A
* linbox/field/vector-domain.h (neg, negin): Add these interfaces
and implement
(class VectorDomain): Parameterize arithmetic methods by separate
vector types for each vector argument. Though they do not support
arithmetic on different vector *representations* at the same time,
one can use this to, say, add a column vector into a standalone
vector
2002-10-27 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/blackbox/dense.inl:
* linbox/blackbox/dense.h: Split out container/iterator
functionality into new DenseMatrixBase
* linbox/blackbox/Makefile.am (BASIC_HDRS): Add dense-submatrix.h,
dense-submatrix.inl
* linbox/blackbox/dense.inl (read): Compilation fix
(write): Prefer ISO standard outputs
* linbox/blackbox/dense-submatrix.inl: Rename from densesubmatrix.C
* linbox/blackbox/dense-submatrix.h (class DenseSubmatrix): Fix
constructor interface; add interfaces missing from DenseMatrix;
add Doc++ documentation; reindent
* linbox/blackbox/dense-submatrix.h: Rename from densesubmatrix.h
* linbox/blackbox/dense.h (DenseMatrix::refEntry): Define
(class DenseMatrix): Reorder the declarations to match
SparseMatrix0Base; reindent; add Doc++ documentation to the
interfaces
(DenseMatrix::getEntry): Add a form that returns a read-only
reference to the entry
(class DenseMatrix): Make generic wrt vector type
(class DenseMatrix): Move some of the one-liners from dense.inl
into the class
(DenseMatrix::resize): Implement
* macros/saclib-check.m4:
* macros/givaro-check.m4:
* macros/lidia-check.m4:
* macros/ntl-check.m4:
* macros/gmp-check.m4: Preserve user-specified CXXFLAGS and LIBS
when running test
2002-10-26 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* tests/Makefile.am (CLEANFILES): Add $(TESTS)
* .cvsignore: Update
* acconfig.h: Remove this file
* macros/gmp-check.m4:
* macros/ntl-check.m4:
* macros/lidia-check.m4:
* macros/givaro-check.m4:
* macros/saclib-check.m4: Use modernized AC_DEFINE
* configure.in: Bump version to 0.1.3
* .cvsignore:
* tests/.cvsignore: Update
* tests/test-solve.C (testIdentitySolve, testNonsingularSolve):
Don't explicitly give whether the matrix is symmetric
* linbox/algorithms/Makefile.am (include_HEADERS): Add
lanczos.{h|inl}
* linbox/solutions/solve.h (solve): Add support for Lanczos
iteration
* tests/test-solve.C (testIdentitySolve, testNonsingularSolve)
(testSingularConsistentSolve, testSingularInconsistentSolve): Add
support for specifying the solver to be used
(testSingularPreconditionedSolve): Add support for specifying the
preconditioner to be used
2002-10-14 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/algorithms/wiedemann.inl (solveNonsingular): Add support
for symmetric containers, using _traits.symmetric
* linbox/algorithms/blackbox-container-symmetric.h (class
BlackboxContainerSymmetric): Add parameter for the field
* linbox/solutions/methods.h (struct SolverTraits): Add parameter
symmetric, telling whether the matrix is symmetric
* tests/test-solve.C: Update with changes to solve
2002-10-13 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/solutions/methods.h (struct SolverTraits): Add
documentation; remove unimplemented string constructor
* linbox/solutions/solve.h (solve): Add a variant that does not
throw any exceptions
* linbox/solutions/methods.h (struct SolverTraits): Make default
maxTries 100
* linbox/solutions/methods.h (class InconsistentSystem, class
SolveFailed): Move these from WiedemannSolver here, since they
apply to more than just Wiedemann
2002-10-09 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/blackbox/ntl-toeplitz.inl: Compilation fixes
* tests/test-ntl-toeplitz.C: Compilation fixes
2002-10-03 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/switch/cekstv.h (applyTranspose, apply): Fix formulae
* linbox/algorithms/wiedemann.inl (findNullspaceElement)
(findRandomSolution): Pass _randiter to RandomDenseStream
constructor
* linbox/algorithms/wiedemann.h (class WiedemannSolver): Add
constructor that accepts random iterator
* linbox/algorithms/wiedemann.inl (solveNonsingular)
(precondition, makeLambdaSparseMatrix): Use _randiter
(solveNonsingular): Add parameter useRandIter to determine whether
to use the random iterator, or to use b for the Krylov sequence
computation
(findNullspaceElement, certifyInconsistency): Remove
preconditioner support
(solveSingular): Use non-preconditioned A for certificate of
inconsistency
(makeLambdaSparseMatrix): Use the existing randiter, since
RandomSparseStream already constructs a NonzeroRandIter
(certifyInconsistency): Pass _randiter to WiedemannSolver
constructor
* linbox/algorithms/wiedemann.h (class WiedemannSolver): Add _randiter
* linbox/algorithms/wiedemann.inl (solveSingular): Throw a
BadPreconditioner exception if the reason for failure was a bad
preconditioner, since this means the rank computation was probably
correct and can be reused
(solveSingular): Move rank computation to solve
(solve): Compute rank and try to reuse when possible; set rank to
(unsigned long) -1 when SolveFailed thrown
* tests/test-solve.C (testSingularPreconditionedSolve): Return
success value rather than always returning true
* linbox/randiter/modular.h (LinBox): Use clock () rather than
time (NULL) to get a higher-resolution timer
* tests/test-solve.C (testSingularPreconditionedSolve): Don't fill
matrix with 1's; just build a diagonal black box
2002-10-02 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/algorithms/wiedemann.h: Refactoring: Put everything
inside a WiedemannSolver class, with the following interface
solveNonsingular - Solve a nonsingular system
solveSingular - Solve a general singular system
findRandomSolution - Find a random solution to a singular preconditioned problem
findNullSpaceElement - Find an element of the right nullspace
certifyInconsistency - Find a certificate of inconsistency for a linear system
precondition - Form a preconditioner and return it
* tests/test-solve.C (testSingularPreconditionedSolve)
(testSingularInconsistentSolve, testSingularConsistentSolve):
Update according to interface changes in wiedemann.h
* linbox/field/vector-domain.C (VectorDomain::writeSpecialized):
Remove debugging code
* tests/test-butterfly.C (main): Use Modular<uint32> and make
GF(2147483647) the default
(testCekstvSwitch): Use preconditioner PDQ rather than just PD
* tests/test-solve.C (testSingularPreconditionedSolve): Always
return true
* linbox/blackbox/moore-penrose.h (MoorePenrose::MoorePenrose):
Update use of Submatrix
* tests/test-moore-penrose.C (testRandomApply1): Update use of
Submatrix
* tests/test-solve.C (main): Reenable tests other than
preconditioned solve
* linbox/algorithms/wiedemann.h (solveWiedemannSingular):
Implement sparse preconditioner
* linbox/blackbox/sparse-base.h (class SparseMatrix0Base): Add
method transpose
(class SparseMatrix0Base): Rename VectorTrait as RowTrait in
specializations
* linbox/blackbox/sparse-base.inl (transpose): Implement
2002-09-30 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/solutions/methods.h (struct SolverTraits): Add enum
Preconditioner; rename precondition to preconditioner; add methods to
set parameters
(SolverTraits::SolverTraits): Initialize _singular with singular, not
UNKNOWN
* linbox/vector/stream.h (RandomSparseStream::setP): Implement
* tests/test-butterfly.C (testSetButterfly): Reset the stream
* tests/test-solve.C (testSingularPreconditionedSolve): Use a
SparseMatrix0 and fill in the entries below the zeros with 1's
* tests/test-butterfly.C (testCekstvSwitch): Ditto
2002-09-26 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/algorithms/wiedemann.h (solveWiedemannSingular):
Construct a CekstvSwitchFactory and use it as input for the
CekstvSwitch constructor
(solveWiedemannSingular): Commentator fixes
* linbox/blackbox/butterfly.h (class Butterfly): Parameterize on
the field, give a reasonable default for Vector
* linbox/switch/boolean.h (class BooleanSwitch):
* linbox/switch/cekstv.h (class CekstvSwitch): Don't maintain
vector of 2x2 block information; just hold data for one 2x2 block;
eliminate field and temporary
* linbox/blackbox/butterfly.h (Butterfly): Make the vector of
switches
(class Butterfly): Constructor reorganization; see header file
2002-09-25 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/switch/cekstv.h (class CekstvSwitch): Add copy
constructor
(CekstvSwitch::CekstvSwitch): Implement copy constructor
* tests/test-butterfly.C (testRandomTranspose)
(testRandomLinearity): Implement these tests
(testCekstvSwitch): Construct a random vector and pass it to
CekstvSwitch
* linbox/algorithms/wiedemann.h (solveWiedemannSingular): Prepare
rhs before obtaining minpoly
* linbox/blackbox/submatrix.h (Submatrix::apply,
Submatrix::applyTranspose): Fill unused portions of temporary
input vector with zeros
(class Submatrix): Parameterize by field
(class Submatrix): Give a reasonable default for the vector type
* tests/test-submatrix.C (testRandomLinearity)
(testRandomTranspose): Implement these tests
* linbox/blackbox/submatrix.h (Submatrix::Submatrix): Initialize
_z with column dimension
* tests/test-submatrix.C (testRandomApply): Make vector output
LEVEL_UNIMPORTANT
(testRandomTranspose): Implement this
* linbox/blackbox/submatrix.h (class Submatrix): Make _z, _y
mutable; delete _zt, _yt
(Submatrix::apply, Submatrix::applyTranspose): Don't declare local
vectors
* tests/test-submatrix.C (testRandomApply): Reset stream when done
(testRandomLinearity): Implement this test
2002-09-23 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/algorithms/wiedemann.h (solveWiedemannSingular): Don't
use DenseVector
(solveWiedemann): Compute m_A rather than m_A,b
(solveWiedemannSingular): Fix generator for random solution
(solveWiedemannSingular): Add butterfly preconditioner support
(solveWiedemannSingular): Cancel this loop iteration if the
leading principal minor turned out to be singular
* tests/test-solve.C (testSingularConsistentSolve): Implement
(testNonsingularSolve): Rename from testDiagonalSolve
(testSingularInconsistentSolve): Implement
* linbox/algorithms/cra.h (cra): Use unsigned long in casts
(cra): Fix order of arguments to invmod
(cra): Reduce the result modulo the product of the primes
* linbox/blackbox/sparse-base.h (_RawIndexIterator::operator ++):
Update _curr.second properly
* linbox/util/prime-stream.h (PrimeStream::next): Move _curr
by 2
* linbox/algorithms/cra.h (cra): Invoke commentator
* linbox/solutions/det.h (det): Invoke commentator
* tests/test-det.C (testIntegerDet): Implement this test
(main): Call testIntegerDet
* linbox/blackbox/sparse.inl (makeBlackbox): Implement (smarter
version now, taking over for the constructor that was removed)
* linbox/blackbox/sparse.h (class SparseMatrix0): Remove
constructor from SparseMatrix0Base -- this will be handled by
SparseMatrixFactory from now on
* linbox/blackbox/sparse-base.h (class
SparseMatrix0Base::RawIterator): Remove _A, add _A_end; add
default constructor; update other constructors
(class SparseMatrix0Base::RawIndexIterator): Ditto
(class SparseMatrix0Base): Remove implementations from default
class
(class SparseMatrix0Base::RawIterator): Make this a parameterized
class to allow const version to be defined easily (N.B. I hate
C++)
(class SparseMatrix0Base::RawIndexIterator): Same as RawIterator
(N.B. Same note applies, doubly so)
* linbox/blackbox/sparse.h (SparseMatrixFactory): Pull declaration
of tmp out of the loop
* tests/test-det.C (testDiagonalDet2): Don't specify template
parameters explicitly
(testSingularDiagonalDet): Ditto
* linbox/blackbox/factory.h: Include vector-traits.h
* gmp++/gmp++_int.h (Integer::invmod): Make this static
* linbox/algorithms/cra.h (cra): Compilation fixes
* linbox/solutions/det.h (det): Compilation fixes
* linbox/blackbox/sparse.h (class SparseMatrixFactory): Make sure
SparseMatrix0Base is used with its own element and row types
* linbox/blackbox/factory.h (BlackboxFactory::makeBlackbox): Make
F const
* linbox/algorithms/cra.h (cra): Use integer:: on integer
operations
* linbox/blackbox/sparse-base.h (class SparseMatrix0Base): Provide
a reasonable default for Row; clean up template declaration
* linbox/blackbox/sparse.h (class SparseMatrix0): Clean up
template declaration
2002-09-22 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/blackbox/sparse-base.inl (write): Add support for
FORMAT_MATLAB in sparse parallel specialization
(write): Write closing bracket; put opening bracket at the
beginning of the matrix, not the beginning of each row
* linbox/blackbox/sparse.h (SparseMatrixFactory::maxNorm,
SparseMatrixFactory::rowdim, SparseMatrixFactory::coldim):
Implement
* linbox/util/gmp++/gmp++_int_gcd.C (invmod): Implement
* gmp++/gmp++_int.h (class Integer): Add invmod
* linbox/field/modular.h (class ModularBase): Remove assignment
operator
(Modular::operator =): Copy documentation from ModularBase; use F
rather than rhs; copy _two_64 rather than calling init again
* linbox/util/gmp++/gmp++_int_misc.C (nextprime, prevprime):
Implement
Include cmath
* gmp++/gmp++_int.h (class Integer): Add nextprime, prevprime
* linbox/blackbox/factory.h (class BlackboxFactory): Make a
reasonable default for Vector
* linbox/Makefile.am (BASIC_HDRS): Add rational.h
* linbox/rational.h: Add this file
* tests/test-butterfly.C (testBooleanSwitch): Remove this test; it
is redundant
* linbox/switch/cekstv.h (class CekstvSwitch): Make _iter,
_riter, and _R mutable
* linbox/blackbox/diagonal.h (class Diagonal): Give a reasonable
default for Vector
* tests/test-butterfly.C (testSetButterfly): Use stock typedef for
dense vectors
(testCekstvSwitch): Implement
2002-09-21 Bradford Hovinen <bghovinen@math.uwaterloo.ca>
* linbox/solutions/det.h (det): Add versions for integer and
rational determinant
Comilation fixes
* examples/dot-product.C (main): Update according to new
interfaces; remove unnecessary typedefs
* linbox/vector/vector-traits.h (class RawVector): Rename from
Vector; parameterize on Element rather than Field
(class Vector): Declare to inherit RawVector<typename Field::Element>
* linbox/blackbox/sparse.h (class SparseMatrixFactory): Don't
require that the field and the SparseMatrixBase share the same
element type
* examples/blackbox/.cvsignore: Update
* linbox/field/modular.h: Reinsert two deleted lines
* linbox/util/commentator.C (Commentator::printActivityReport):
Don't print percentage unless _len is positive
2002-09-21 Rich Seagraves <seagrave@cis.udel.edu>
* linbox/field/modular.h: Added = assignment operator.
2002-09-21 Bradford Hovinen <bghovine@math.uwaterloo.ca>
* examples/blackbox/Makefile.am (EXTRA_PROGRAMS): Add load-minpoly
* linbox/blackbox/sparse-base.inl (readMatlab): Use is.putback,
not is.unget
* linbox/blackbox/sparse-base.h
(SparseMatrix0WriteHelper::Format): Add FORMAT_MATLAB
2002-09-20 Bradford Hovinen <bghovine@math.uwaterloo.ca>
* linbox/randiter/param-fuzzy.h (ParamFuzzyRandIter::random):
Uncomment ElementAbstract version
Include element/abstract.h, element/envelope.h
* linbox/field/param-fuzzy.h: Reintroduce forward declaration of
RandIter; move include randiter/... to the end
* linbox/randiter/param-fuzzy.h (class ParamFuzzyRandIter): Add a
version of the constructor that *really* takes a field as an
argument
* tests/test-param-fuzzy.C (main): Add arguments n and i; call
runFieldTests
* tests/test-lidia-gfq.C (main): Add arguments n and i; fix
description text for argument q; call runFieldTests
* tests/test-ntl-RR.cpp (main):
* tests/test-unparametric-field.C (main): Return 0 always
* linbox/randiter/unparametric.h (UnparametricRandIter::random):
Remove debugging code
* tests/test-unparametric-field.C (main):
* tests/test-ntl-RR.cpp (main):
* tests/test-ntl-ZZ_p.cpp (main):
* tests/test-ntl-zz_p.cpp (main): Add arguments n and i; call
runFieldTests
* tests/test-gmp-rational.C (main):
* tests/test-givaro-zpz.C (main):
* tests/test-modular.C (main): Reverse order of iterations and n
when calling runFieldTests
* tests/test-generic.h: Do b / a rather than a / b
* tests/test-givaro-zpz.C (main): Add arguments n and i; fix
description text for argument q; call runFieldTests
(main): Don't print out separate headers for each test; just use
the runFieldTests wrapper.
(main): Add a parameter e and a field F5 which is GF(q^e); run a
test on that
* linbox/randiter/gmp-rational.h: Reindent
(GMPRationalRandIter::random): Initialize value properly
(GMPRationalRandIter::random): Canonicalize before returning
* linbox/field/gmp-rational.h:
* linbox/element/gmp-rational.h: Reindent
* linbox/randiter/gmp-rational.h (class GMPRationalRandIter): Fix
compile errors
* tests/test-generic.h (runFieldTests): Move from test-modular.C
here
* linbox/field/gmp-rational.h
(GMPRationalField::GMPRationalField): Initialize all the const
integers
(class GMPRationalField): Make integers not static const; fix
compile errors
* linbox/blackbox/dense.inl (apply): Remove unused variables
(applyTranspose): Remove unused variables
* linbox/blackbox/dense.h: Include vector-domain.h
* linbox/blackbox/sparse1.h: Fix compile warning
* examples/mmZZ_pE.C: Fix header and indentation; include dense.h
rather than dense-matrix1.h
2002-09-18 zhendong wan <wan@aphrodite.cis.udel.edu>
* linbox/blackbox/dense.h
Add RawIndexIterator, which looks like RawIterator
expect that it has two more member function.
rowIndex returns the row index,
colIndex returns the column index.
2002-09-17 Bradford Hovinen <bghovine@math.uwaterloo.ca>
* linbox/solutions/det.h (det): Use commentator.start /
commentator.stop
* linbox/field/vector-domain.h (VectorDomain::copySpecialized):
Don't call resize
* linbox/blackbox/dense.h (class DenseMatrix): Remove template
parameter from VectorStream constructor; support a different
vector type in the representation
* tests/.cvsignore: Update
* linbox/blackbox/dense.h (class DenseMatrix): Add constructor
that takes a VectorStream
2002-09-16 Bradford Hovinen <bghovine@math.uwaterloo.ca>
* linbox/blackbox/dense.h (class DenseMatrix): Define a vector
domain _VD
* linbox/blackbox/dense.inl (applyTranspose, apply): Use vector
domain
2002-09-15 Bradford Hovinen <bghovine@math.uwaterloo.ca>
* linbox/blackbox/sparse-base.h (class
SparseMatrix0ReadWriteHelper): Add readMatlab
(SparseMatrix0WriteHelper::Format): Add FORMAT_MATLAB
(SparseMatrix0Base::Format): Add FORMAT_MATLAB
* linbox/blackbox/sparse-base.inl (write, read): Add support for
FORMAT_MATLAB
(readMatlab): Implement
2002-09-12 Bradford Hovinen <bghovine@math.uwaterloo.ca>
* configure.in (EIGHT_BYTES): Fix typo
2002-09-11 Rich Seagraves <seagrave@cis.udel.edu>
* linbox/blackbox/nag-sparse.h: Add RawIterator and
RawIndexIterator support, altered test suite appropriately
* linbox/blackbox/archetype.h: include cstddef to ensure support
for size_t type.
2002-09-10 Bradford Hovinen <bghovine@math.uwaterloo.ca>
* tests/test-generic.h (testFieldInversion): Make sure a is
nonzero
2002-09-09 Bradford Hovinen <bghovine@math.uwaterloo.ca>
* configure.in: Support older versions of autoconf that give
AC_SIZEOF results in bits
2002-09-09 Saunders & Wan
* solutions/lu.h: add comments
2002-09-08 Bradford Hovinen <bghovine@math.uwaterloo.ca>
* examples/blackbox/load-det.C: Cut out unnecessary typedefs
* linbox/blackbox/sparse.h (class SparseMatrixFactory): Define
(class SparseMatrix0): Introduce default of Dense for vector type
2002-09-08 Bradford Hovinen <bghovine@math.uwaterloo.ca>
* .cvsignore: Update
* linbox/vector/stream.h (class RandomDenseStream, class
RandomSparseStream): Set default vector types to dense and sparse,
respectively
* linbox/vector/vector-traits.h: Add class Vector with canonical
vector types
* linbox/vector/Makefile.am (include_HEADERS): Add stream.h
* linbox/util/Makefile.am (include_HEADERS): Remove vector-factory.h
* tests/test-trace.C:
* tests/test-sum.C:
* tests/test-submatrix.C:
* tests/test-minpoly.C:
* tests/test-inverse.C:
* tests/test-diagonal.C:
* linbox/solutions/trace.h:
* linbox/algorithms/wiedemann.h:
* tests/test-butterfly.C:
* linbox/algorithms/wiedemann.h:
* tests/test-generic.h:
* tests/test-vector-domain.C:
* tests/test-solve.C:
* tests/test-moore-penrose.C:
* tests/test-sparse.C: Update with changes to linbox/vector/stream.h
* tests/test-moore-penrose.C (buildRandomSparseMatrix): Work with
sparse parallel vectors
2002-09-05 Bradford Hovinen <bghovine@math.uwaterloo.ca>
* linbox/vector/stream.h:
- Renamed to stream.h and moved to linbox/vector
- VectorFactory is now called VectorStream, which fits its purpose
somewhat better
- The interface is now closer to the interface for istream
- RandomDenseVectorFactory, et al. are refactored into classes
parameterized on the vector type and specialized appropriately. This
allows, e.g. construction of a random dense vector in sparse
representation and so on.
- New constructor interface for RandomSparseStream accepts proportion of
nonzero entries
- Reindented, since the other changes are so numerous that diffs aren't a
big deal anyway
* linbox/blackbox/archetype.h: s/stddef.h/cstddef/
2002-09-04 <hovinen@cis.udel.edu>
* tests/test-vector-domain.C (main): Remove srand call
* examples/blackbox/load-symmetrize.C (main): Remove srand call
(main): Fix help text
* examples/blackbox/load-det.C (main): Remove srand call
2002-08-28 Bradford Hovinen <hovinen@cis.udel.edu>
* NEWS: Updated text
* tests/Makefile.am (BASIC_TESTS): Remove test-lu due to
compilation problems
* configure.in: Bump version to 0.1.2
* linbox/field/lidia-gfq.h (LidiaGfq::read, LidiaGfq::write): Use
std::ostream, std::istream
* macros/gmp-check.m4: Add directives to allow use of uninstalled
GMP
* configure.in: Use AM_CONDITIONAL statements for Doc++; Add
arguments --enable-doc and --enable-doc-tex for building
documentation
* doc/Makefile.am (SUBDIRS): Mandary inclusion of html and tex
* doc/html/Makefile.am:
* doc/tex/Makefile.am: Add mechanism for optionally compiling in
targets if USE_DOCPLUSPLUS is set
2002-08-27 Bradford Hovinen <hovinen@cis.udel.edu>
* NEWS: Update text
* linbox/*/Makefile.am (EXTRA_DIST): Merge in contents of
EXTRA_HEADERS
* NEWS: Update text
* configure.in: Bump version number to 0.1.1
* tests/test-commentator.C (testBriefReport): Output brief report
data to the same file used for the other data
* linbox/util/commentator.C (MessageClass::checkConfig): Make
depth and level unsigned
(Commentator::Commentator): Set default max depth of BRIEF_REPORT
to 1
(Commentator::start): Push new activity after printing activity
report
(Commentator::stop): Pop new activity before finishing activity
report
(Commentator::progress): Test on _activities.size () - 1 for
updateActivityReport
(Commentator::progress): Tweak output
(Commentator::setPrintParameters, Commentator::setMaxDepth,
Commentator::setMaxDetailLevel): Omit BRIEF_REPORT
(Commentator::printActivityReport,
Commentator::updateActivityReport,
Commentator::finishActivityReport): Tweak output parameters
(Commentator::updateActivityReport): Don't show progress if deeper
messages are printed
* linbox/util/commentator.h (Commentator::isPrinted,
MessageClass::isPrinted): Make depth and level unsigned
* tests/.cvsignore: Update
* tests/Makefile.am (BASIC_TESTS): Add test-commentator.C
* linbox/algorithms/blackbox-container-base.h
(BlackboxContainerBase::~BlackboxContainerBase): Delete _BB, as it
was a clone
* linbox/util/commentator.h (Commentator::MessageLevel): Renumber
so that LEVEL_ALWAYS is 0
* linbox/util/commentator.C (MessageClass::fixDefaultConfig): Add
a directive to make sure LEVEL_ALWAYS messages always get printed
* linbox/util/commentator.h (Commentator::setPrintParameters,
MessageClass::setPrintParameters): Interface adjustment: only one
depth is specified, the maximum for which the directive is valid;
make depth and level unsigned
2002-08-26 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/randiter/modular.h (ModularRandIter::random): Fix call to
linbox_check
* tests/test-sparse.C (runSparseMatrixTests): Tweak text
* linbox/util/vector-factory.h
(RandomSparseSeqVectorFactory::next,
RandomSparseParVectorFactory::next): Initialize i to -1 so that
the first index might get an entry
* tests/test-generic.h (testTranspose, testLinearity): Declare new
reports inside the loop
* tests/test-sparse.C (testRandomApply2): Reset A_factory when
done
* linbox/util/field-axpy.h (class FieldAXPY): Make _F mutable
* tests/test-sparse.C (testRandomApply2): Remove this test
(testRandomApply3): Rename as testRandomApply2
(testRandomApply1, testRandomApply2): Accept a factory as a
parameter and use it to build the matrix
(testRandomTranspose, testRandomLinearity): Accept a factory as a
parameter and use it to build the matrix
* tests/test-generic.h (testGeometricSummation): Make sure a != 1
* tests/test-modular.C (runFieldTests): Implement; copy code from
relevant section of main
(main): Use runFieldTests
2002-08-25 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/.cvsignore: Update
* tests/.cvsignore: Update
* linbox/blackbox/butterfly.h (applyTranspose, apply): Call
std::copy (x.begin (), x.end (), y.begin ())
* configure.in (AC_OUTPUT): Add linbox/switch/Makefile
* macros/ntl-check.m4: Remove EX_FIELDS_NTL, as it is not needed
* linbox/Makefile.am (BASIC_HEADERS): Rename from include_HEADERS;
have include_HEADERS use the mechanism in tests/Makefile.am for
optional inclusion
* macros/ntl-check.m4 (NTL_HEADERS_*): Remove
* macros/givaro-check.m4 (GIVARO_HEADERS_*): Remove
* macros/lidia-check.m4 (LIDIA_HEADERS_*): Remove
* linbox/field/Makefile.am (BASIC_HEADERS): Rename from
include_HEADERS; have include_HEADERS use the mechanism in
tests/Makefile.am for optional inclusion
* tests/test-ntl-toeplitz.C: Rename from test-toeplitz.C
* tests/Makefile.am: Rename test-toeplitz as test-ntl-toeplitz
* macros/lidia-check.m4 (HAVE_LIDIA): Define
* macros/givaro-check.m4 (HAVE_GIVARO): Define
* macros/ntl-check.m4 (HAVE_NTL): Define
* tests/Makefile.am: Put {NTL|GIVARO|LIDIA}_TESTS in here; use
conditionals to make inclusion in TESTS optional
2002-08-20 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/blackbox/butterfly.h (setButterfly): Remove Commentator
parameter; update uses of commentator to current interface; remove
TRACE code
* linbox/Makefile.am (SUBDIRS): Add switch
* examples/.cvsignore: Update
2002-08-19 Bradford Hovinen <hovinen@algol.cis.udel.edu>
* linbox/randiter/modular.h (ModularRandIter::random): Protect
from getting a result equal to the size of the sample
* tests/test-sparse.C (testRandomApply1): Remove debugging code
* linbox/randiter/modular.h (class ModularRandIter): Make _size a
double and _seed an Element, as they can never be larger than
Element
(ModularRandIter::randIter): Initialize _card
(ModularRandIter::random): Don't use temp_integer; don't mod out
by cardinality as it should not be necessary; don't convert _size
to a double
2002-08-18 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/modular.inl (dotSpecializedDSP): Add optimized
variant for very short vectors, to avoid cost of additional mod
operation
(dotSpecializedDSP, dotSpecializedDD): Defer declarations until
needed
* configure.in (AC_DEFINE_UNQUOTED): Don't quote description
* examples/blackbox/ex-mat0.C (main): Use Modular<uint32>
* examples/blackbox/example.C: Use Modular<uint32>
* examples/fields/ex-fields0.C (main): Use Modular<uint32>
* examples/fields/ex-fields.C (main): Use Modular<uint32>
* examples/blackbox/load-symmetrize.C: Use Modular<uint32>
* examples/blackbox/load-det.C: Use Modular<uint16>
* examples/dot-product.C: Use Modular<uint32>
* linbox/field/modular.inl (dotSpecializedDD, dotSpecializedDSP):
Update signatures of specialization
* tests/test-trace.C (main): Use Modular<uint32>
* tests/test-sum.C (main):
* tests/test-submatrix.C (main):
* tests/test-sparse.C (main):
* tests/test-solve.C (main):
* tests/test-scalar-matrix.C (main):
* tests/test-minpoly.C (main):
* tests/test-inverse.C (main):
* tests/test-diagonal.C (main):
* tests/test-det.C (main):
* tests/test-rank.C (main):
* tests/test-dense.C (main):
* tests/test-randiter-nonzero.C (main):
* tests/test-moore-penrose.C (main):
* tests/test-vector-domain.C: Use Modular<uint*>
* tests/test-modular.C:
* linbox/field/modular.h: Change Modular, FieldAXPY, and
DotProductDomain specializations to use new Linbox uint* types
* linbox/integer.h: Add typedefs for int*
* acconfig.h: Add #undef __LINBOX_INT*
* configure.in: Define __LINBOX_INT* based on results of
AC_SIZEOF
* linbox/field/modular.inl (dotSpecializedDD, dotSpecializedDSP):
Reintroduce faster modding out
* linbox/field/modular.h (Modular<long>::init_two_64): Start
counter at 0
* tests/test-vector-domain.C (main): Test vector domains for all
basic modular types
(testVectorDomain): Implement
* examples/blackbox/.cvsignore: Update
* examples/Makefile.am (EXTRA_PROGRAMS): Add dot-product
2002-08-13 Bradford Hovinen <hovinen@cis.udel.edu>
* configure.in: Bump version to 0.1.0
* linbox/field/modular.inl (dotSpecializedDSP, dotSpecializedDD):
Use traditional modding out pending getting the more advanced
version working correctly
* linbox/algorithms/wiedemann.h (solveWiedemannSingular): Fix
start/stop pairing
* linbox/field/modular.h (Modular<long>::inv): Keep the inputs
from sign-extending to long long
* linbox/util/commentator.C (Commentator::startIteration): Store
string in _iteration_str so that it is not deallocated when the
call returns
* linbox/util/commentator.h (class Commentator): Add _iteration_str
* tests/test-diagonal.C (testRandomMinpoly): Make vector reports
at LEVEL_UNIMPORTANT
(testRandomTranspose): Use factory to generate diagonal vector;
make it a dense vector, use VectorDomain::write
(testRandomLinearity): Implement
* tests/test-vector-domain.C (main): Used unsigned short field
* tests/test-modular.C (main): Make default n 10000; reorder test
invocations to group fields together
2002-08-12 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-vector-domain.C (testDotProduct): Make vector output
LEVEL_UNIMPORTANT
(testCopyEqual, testAXPY, testSubMul, testAddMul, testDotProduct):
Terminate vector displays with endl's
* linbox/field/modular.inl (dotSpecializedDD, dotSpecializedDSP):
Make sure arithmetic takes place without sign-extension
* tests/test-generic.h (testFieldAssociativity): Implement
2002-08-11 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-modular.C (main): Call tests indicated below for each
modular type
* tests/test-generic.h (testFieldNegation, testFieldInversion)
(testFieldAxioms, testGeometricSummation)
(testFieldCharacteristic, testFreshmansDream)
(testArithmeticConsistency, testAxpyConsistency): Implement
* tests/test-minpoly.C (main): Make A_factory sparse parallel
(testIdentityMinpoly, testNilpotentMinpoly): Define Row to be
sparse parallel
* linbox/field/modular.h (read): Do arithmetic in the domain
integer
* linbox/randiter/modular.h (ModularRandIter::random): Remove
check for negativity
* linbox/field/modular.h (Modular<short>::init): Fix in case
modulus > 2^15
(Modular<short>::init): Perform modding out operation in class
integer
(class Modular<short>): Convert to Modular<unsigned short> to
allow larger moduli
* tests/test-modular.C (main): Test Modular<integer> and
Modular<short> as well; add CLI parameters for that
* linbox/field/modular.h (class Modular<short>): Make _k a long
long
(class FieldAXPY<Modular<short> >): Make sure _y uses long long
arithmetic
(class Modular<short>): Do add and sub in a long to remove
restriction on modulus
* tests/test-vector-domain.C (main): Use a typedef for the field
* linbox/field/modular.h (class Modular<short>): Rename 'k' as '_k'
* linbox/field/modular.inl (dotSpecializedDSP): Fix compiler errors
(dotSpecializedDD): Ditto
* linbox/field/modular.h: Add specializations of DotProductDomain;
include modular.inl
(class Modular): Make DotProductDomain<...> a friend
(Modular<long>::init_two_64): Implement
(class Modular<long>): Add field _two_64
(Modular<long>::Modular): Call init_two_64
* linbox/blackbox/sparse-base.inl (setEntry): Use v.first.size ()
* examples/blackbox/load-symmetrize.C: Ditto
* examples/blackbox/load-det.C: Don't explicitly specify row type
* linbox/field/vector-domain.h (class VectorDomain): Pull out the
dense/dense and dense/sparse parallel dot products into a separate
class DotProductDomain, which VectorDomain inherits. Put the field
in a virtual base class VectorDomainBase
(VectorDomain::field): Implement this
* linbox/blackbox/sparse.h (class SparseMatrix0): Make default row
type a sparse parallel
* examples/blackbox/Makefile.am (EXTRA_PROGRAMS): Add
load-symmetrize
* linbox/algorithms/blackbox-container-symmetrize.h (class
BlackboxContainerSymmetrize): Fix compiler warnings
* linbox/blackbox/sparse-base.inl (read): Fix arguments to
readPretty
(read): Remove unused variables
(read): Add dummy return statement
* examples/blackbox/load-det.C: Fix order of template arguments;
add using namespace std
* linbox/blackbox/sparse-base.h (class SparseMatrix0Base): Fix
member initializer order
2002-08-10 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/util/vector-factory.h (class StandardBasisFactory): Add
specialization for sparse parallel vectors
* tests/test-generic.h (testLinearity): Use VD.write
(testTranspose): Use VD.write
(testFieldAXPY): Use VD.write, commentator.startIteration
* tests/test-sparse.C (test*): Use VD.write rather than
printVector
* linbox/blackbox/sparse-base.inl (write): Implement
specialization for sparse parallel vectors
* linbox/blackbox/sparse-base.h (class
SparseMatrix0ReadWriteHelper): Implement class NoField to avoid
having twice as many implementations as otherwise
(class SparseMatrix0ReadWriteHelper): Split out writing portions
into SparseMatrix0WriteHelper, which is partially specialized for
sparse parallel vectors
* linbox/blackbox/sparse-base.inl (getEntry, refEntry, setEntry):
Implement specializations for sparse parallel vectors
* linbox/blackbox/sparse.h: Implement specialization for sparse
parallel vectors
* linbox/blackbox/sparse.h (class SparseMatrix0): Specialize
applyTranspose so that it can handle sparse parallel vectors,
without having to specialize the entire class
* linbox/algorithms/Makefile.am (include_HEADERS): Add
wiedmemann.h
* linbox/solutions/Makefile.am (include_HEADERS): Add solve.h
2002-08-09 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/Makefile.am (BASIC_TESTS): Remove test-dense-matrix
* tests/.cvsignore: Update
* linbox/blackbox/Makefile.am (include_HEADERS): Remove
dense-matrix.h
(include_HEADERS): Change dense-matrix1.[Ch] to dense.inl, dense.h
* tests/test-submatrix.C: #include dense.h
* tests/test-inverse.C:
* linbox/blackbox/dense.h (class DenseMatrix): Make F const
2002-08-09 Dave Saunders
* linbox/vector/subvector.h: homage to the indentation police.
2002-08-09 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/blackbox/dense.h (class DenseMatrix): Make _F a const
reference
(DenseMatrix::DenseMatrix): Make F const
* linbox/blackbox/dense.h:
* linbox/blackbox/dense.inl: Reindent and remove compiler warnings
* tests/test-dense.C: Rename from test-dense-matrix1.C; update
header
* linbox/algorithms/wiedemann.h (solveWiedemannSingular): Give up
if tries exceeds maxTries
(solveWiedemannSingular, solveWiedemann): Fix order of arguments
to VD.axpy
* linbox/solutions/methods.h (struct SolverTraits): Add maxTries
* linbox/algorithms/wiedemann.h (solveWiedemann): Output minimal
polynomial at LEVEL_UNIMPORTANT
* tests/test-solve.C (main): Correct output
(testDiagonalSolve): Correct output
(testDiagonalSolve): Print output vector too
* linbox/util/commentator.C (Commentator::start): Don't output
starting activity method if messages at that activity's depth
aren't printed
(Commentator::stop): Give a different message if this is the
greatest depth at which messages are printed
* linbox/algorithms/wiedemann.h (solveWiedemann): Use traditional
indexing approach; eliminate count
(solveWiedemannSingular): Ditto
* linbox/solutions/methods.h (struct SolverTraits): Add singular,
checkResult, certificate
* tests/Makefile.am (BASIC_TESTS): Remove test-solver, add
test-solve
* linbox/solutions/solve.h: Move solveWiedemann[Singular] to
linbox/algorithms/wiedemann.h; solve is now just a wrapper
* linbox/solutions/solve.h: Rename from solver.h, tweak
indentation, add header
(solve): Add argument SolverTraits
(solve): Remove randiter
(solve): Use P.front () rather than P[0]
(solve): Use iterator to go through polynomial
(solve): Don't use applyIn and axpyin; allocate another vector
(solve): Split into nonsingular case
(solveWiedemann): Move Wiedemann-based method here
(solveWiedemann): Rename y as b
* linbox/solutions/methods.h (struct SolverTraits): Add a
constructor that takes a string
* linbox/blackbox/inverse.h (apply, applyTranspose): Remove
template spec on minpoly call
* linbox/solutions/rank.h: Update header
* linbox/solutions/trace.h (trace): Reorder template parameters so
that only Vector need be specified explicitly
* tests/test-moore-penrose.C: Remove compiler warnings
* tests/test-det.C: Remove compiler warnings
* tests/test-rank.C: Remove compiler warnings
* linbox/solutions/rank.h (rank): Make i a size_t
* linbox/solutions/det.h (det): Make i a size_t
* tests/test-trace.C: Remove compiler warnings
* linbox/solutions/trace.h (trace): Remove unused int i
* linbox/blackbox/dense-matrix.h (DenseMatrix::DenseMatrix): Make
F a const
(class DenseMatrix): Make _F a const
* tests/test-common.h (multiEvalPoly): Make j a size_t
* tests/test-inverse.C: Fix compiler warnings
(testHilbertInverse, testIdentityInverse, testVandermondeInverse)
(testDiagonalInverse): Make generic wrt vector type; use
VectorFactory
(main): Add CLI argument -N for number of vectors to apply
Vandermonde inverse to
* tests/test-diagonal.C: Update header; remove compiler warnings
(testIdentityApply): Use commentator.startIteration; make generic
wrt vector type; use VectorFactory
(testRandomMinpoly): Ditto above
* linbox/blackbox/scalar-matrix.h: Remove unused y_end
* linbox/field/vector-domain.C (VectorDomain::writeSpecialized):
Make idx a size_t
* linbox/blackbox/inverse.h (Inverse::apply): Fix order of
arguments to axpy
(Inverse::applyTranspose): Ditto
* linbox/field/vector-domain.C (VectorDomain::axpySpecialized):
Fix order of arguments to _F.axpy
* linbox/algorithms/blackbox-container-base.h (class
BlackboxContainerBase): Add virtual destructor
* linbox/algorithms/massey-domain.h (MasseyDomain::massey): Cast
to avoid compiler warning
* tests/test-common.h (printPolynomial): Make val a size_t
* tests/test-minpoly.C (testRandomMinpoly1): Use VD in more places
(testRandomMinpoly1): Use startIteration
(testRandomMinpoly1): Use a vector factory to build the matrix
(testRandomMinpoly): Rename from testRandomMinpoly1
(testRandomMinpoly2): Eliminate this test, as it is now redundant
(main): Remove CLI argument -N; construct vector factories
(testNilpotentMinpoly): Make i a size_t
* tests/test-common.h (applyPoly): Fix order of arguments to
VD.axpy
(applyPoly): Make generic wrt vector type
2002-08-08 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/.cvsignore: Remove test-dense-vector-domain, add
test-subiterator, test-param-fuzzy, test-dense-matrix1
* linbox/vector/subvector.h (Subvector::Subvector): Use Iterator
() rather than 0
* linbox/field/param-fuzzy.h (class ParamFuzzy): Use std:: for
istream and ostream
* tests/test-scalar-matrix.C (testScalarApply): Make j a size_t
* tests/test-submatrix.C (testRandomApply): Make i, j, k, l
size_t's
* tests/test-subvector.C (testSubvector): Remove unused variable
i, make z a size_t
* linbox/vector/const-iterator-type.h: Qualify vector with std::
* tests/test-subiterator.C (test): Use
Subiterator<std::vector<int>::iterator>
* linbox/vector/subiterator.h: Reindent
* linbox/vector/const-iterator-type.h: Rename from
constiteratortype.h; reindent
* tests/test-randiter-nonzero.C: Remove #include <strstream>
2002-08-06 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/.cvsignore: Update
* examples/blackbox/example.C: Compilation fixes
* linbox/blackbox/sparse.h: Rename from sparse0.h
* linbox/blackbox/sparse.inl: Rename from sparse0.inl
* linbox/blackbox/sparse-base.h: Rename from sparse0-base.h
* linbox/blackbox/sparse-base.inl: Rename from sparse0-base.inl
* tests/test-sparse.C: Rename from test-sparse0.C
2002-08-05 Bradford Hovinen <hovinen@cis.udel.edu>
* Change local variable header so that it actually works
2002-08-05 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/blackbox/Makefile.am (include_HEADERS): Update
* tests/Makefile.am (BASIC_TESTS): Remove test-toeplitz and
test-lu, since they depends on an optional library (i.e. NTL)
(EXTRA_PROGRAMS): Add test-toeplitz, test-lu
(BASIC_TESTS): Remove test-stub (test-stub.C is missing)
2002-08-03 Dave Saunders <saunders@cis.udel.edu>
* linbox/field/param-fuzzy.h, linbox/randiter/param-fuzzy.h: update.
* tests/test-param-fuzzy.C: Added this test. Passes, but fails thru archetype
We need to sharpen our testing. Fields should work thru the arch and
there should be exact field tests that fuzzy fields fail.
2002-08-02 Dave Saunders <saunders@cis.udel.edu>
* linbox/field/field-interface.h: empty parent class for fields added for the sake of doc++
* linbox/interfaces.dxx: summary of the various interface concepts. Many have archtypes, but not all
* consequential changes to these files: doc/Makefile.am doc/html/Makefile.am linbox/Makefile.am linbox/linbox.dxx
linbox/field/Makefile.am linbox/field/archetype.h linbox/field/fields.dxx linbox/field/givaro-gfq.h linbox/field/givaro-zpz.h
linbox/field/gmp-rational.h linbox/field/lidia-gfq.h linbox/field/modular.h linbox/field/param-fuzzy.h linbox/field/unparametric.h
2002-08-01 Dave Saunders <saunders@cis.udel.edu>
* linbox/field/Makefile.am: delete ref to the previously removed dense-vector-domain
* tests/Makefile.am: add comments
* HACKING: mod to indent policy, more "cvs up" info.
2002-07-31 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/blackbox/nag-sparse.h: Tweak header
* linbox/blackbox/sparse0-base.h (class SparseMatrix0Base): Remove
copyToRow and swapRows, as they are unnecessary
2002-07-31 Dave Saunders <saunders@cis.udel.edu>
* linbox/blackbox/ntl-toeplitz.h: fix indentation
* linbox/blackbox/ntl-toeplitz.inl: replace .C and fix indentation
2002-07-30 Saunders and Lobo
* linbox/blackbox/ntl-toeplitz.* and it's test. Adding a toeplitz black box.
* First version, more general capability is coming.
2002-07-26 Bradford Hovinen <hovinen@algol.cis.udel.edu>
* linbox/field/modular.h (class Modular, class FieldAXPY): Add
specializations for int element type
* linbox/util/commentator.h
(Commentator::nullstreambuf::showmanyc): Return std::streamsize
2002-07-29 Rich Seagraves <seagrave@cis.udel.edu>
* linbox/blackbox/Makefile.am (include_HEADERS): Added nag-sparse.h
2002-07-26 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/blackbox/sparse0-base.C (readTurner, readGuillaume): Use
A.refEntry
* linbox/blackbox/sparse0.[Ch]: Move apply, applyTranspose
definitions to sparse0.C
* tests/test-minpoly.C (testRandomMinpoly1, testRandomMinpoly2):
Update as per change below
* linbox/blackbox/sparse0-base.C (getEntry): Make the Row
reference const
(getEntry): Have specialization for sparse associative vectors use
find to avoid const issues
* tests/test-sparse0.C (testRandomApply1, testRandomApply2)
(testRandomApply3, testRandomTranspose, testRandomLinearity):
Update as per change below
* tests/test-moore-penrose.C (buildRandomSparseMatrix): Update as
per change below
* linbox/blackbox/sparse0-base.h (class SparseMatrix0Base): Change
non-const version of getEntry to refEntry
2002-07-26 Bradford Hovinen <hovinen@stimpy.eecis.udel.edu>
* linbox/blackbox/Makefile.am (include_HEADERS): Remove
densesubmatrix.H
2002-07-25 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/util/commentator.C (Commentator::stop): Make sure all
timings get the same precision
* linbox/field/modular.h (clsas Modular<short>): Define k for
delayed modding out
(class FieldAXPY<short>): Use static modding out procedure; make
_y a long long
2002-07-24 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-vector-domain.C (main): Add calls to test sparse
parallel vectors
Use VD.write rather than printVector
* linbox/util/vector-factory.h (class
RandomSparseParVectorFactory): Added this class for random sparse
parallel vectors
* tests/test-vector-domain.C (testDotProduct, testAddMul)
(testSubMul, testAXPY, testCopyEqual): Reset factories when done
* linbox/field/vector-domain.C (VectorDomain::addinSpecialized,
VectorDomain::subinSpecialized, VectorDomain::axpyinSpecialized):
Use a temporary vector and copy when done for the sparse sequence
specialization (O(n) vs. O(n^2))
Don't use iterator operator < unless necessary
* linbox/field/vector-domain.h (class VectorDomain): Add support
for SparseParallelVectorTag
(VectorDomain::copySpecialized): Use std::copy for inline
implementations, just in case there are two different vector types
with the same tag
* linbox/vector/vector-traits.h (struct VectorCategories): Add
SparseParallelVectorTag
(struct VectorTraits): Add specialization for sparse parallel
vectors
(refSpecialized, constRefSpecialized, ensureDimSpecialized): Add
specialization for sparse parallel vectors
* tests/test-moore-penrose.C (testRandomApply1): Add commentator
activity for constructing the matrix
(testRandomApply1): Delete A when done
* linbox/field/vector-domain.C (VectorDomain::copySpecialized):
Add version optimized for copying to vectors of the same type
2002-07-23 Bradford Hovinen <hovinen@cis.udel.edu>
* examples/blackbox/load-det.C: Fix include directive
* linbox/field/vector-domain.C (VectorDomain::write): Write out
i->second, not *i
* tests/test-moore-penrose.C (testRandomApply1): Make this generic wrt
vector and row type
(buildRandomSparseMatrix): Use VectorDomain and facilities of new
SparseMatrix0; make generic wrt row type
(testRandomApply1): Indent correct and computed output messages
* linbox/field/vector-domain.C (VectorDomain::copySpecialized):
Implement version described below
* linbox/field/vector-domain.h (class VectorDomain): Add a form of copy
that allows copying a subvector into a subvector
* linbox/util/commentator.C (MessageClass::fixDefaultConfig):
Clear entire list
2002-07-23 Saunders
* tests/Makefile.am includes test-subiterator. Advise me if it doesn't
work for you.
2002-07-22 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-minpoly.C: Updated to reflect changes to
SparseMatrix0
(testNilpotentMinpoly, testIdentityMinpoly): Use a
StandardBasisFactory object
* tests/test-sparse0.C: Updated to reflect changes to SparseMatrix0
* linbox/blackbox/sparse0.h:
* linbox/blackbox/sparse0-base.h: Refactoring:
- Eliminated SparseMatrix0Aux and moved that functionality into Sparse0
- Made SparseMatrix0Base parameterized only on the element type
- New read/write implementations for SparseMatrix0Base, supporting multiple
formats
- Eliminated Gaussian elimination code
- Added iterators, including ColOfRowsIterator, RawIterator, and
RawIndexIterator
- Eliminated operator []; added getEntry; changed put_value to setEntry
* linbox/util/commentator.C (Commentator::progress): Enable
act->_timer.stop ()
* linbox/util/timer.C (stop, start): Use _start_t
* linbox/util/timer.h (class BaseTimer): Added _start_t
* linbox/vector/subvector.h (struct VectorTraits): Use
Subvector<Iterator> to get value_type
2002-07-18 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/vector/vector-traits.h: Add #include <algorithm>
* linbox/util/commentator.h (class Commentator): Add form of
isPrinted that does not require activity depth; add isNullStream
* macros/*-check.m4: Don't add a -I and -L arguments if they refer
only to /usr or to /usr/local
* macros/*-check.m4: Add rules for cross-compiling
* macros/*-check.m4: Tweak argument documentation
* macros/gmp-check.m4: Remove note that GMP installation is
optional
2002-07-12 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/ntl-RR.h (LinBox): Ditto
* linbox/field/ntl-ZZ_p.h (LinBox): Ditto
* linbox/field/ntl-zz_p.h (LinBox): Ditto
* linbox/blackbox/submatrix.h (LinBox): Ditto
* linbox/blackbox/hilbert.h (LinBox): Ditto
* linbox/blackbox/diagonal.h (LinBox): Ditto
* linbox/blackbox/scalar-matrix.h (LinBox): Ditto
* tests/test-scalar-matrix.C: Update as per C++ standard
* linbox/solutions/methods.h (struct SolverTraits): First stab at
this structure
* linbox/algorithms/blackbox-container.h (LinBox): Ditto
* linbox/vector/reverse.h (LinBox): Ditto
* linbox/vector/subvector.h: Ditto
* linbox/vector/subiterator.h (LinBox): Ditto
* linbox/field/unparametric.h (LinBox): Update as per C++ standard
* linbox/util/commentator.C (Commentator::updateActivityReport):
Make i signed
* linbox/util/error.C: Ditto
* linbox/util/timer.C: Updates to C++ standard; place everything
inside LinBox namespace
* linbox/util/gmp++/gmp++_int_misc.C: Ditto
* linbox/util/gmp++/gmp++_int_io.C: Update header file names
Add using namespace std
Updates to conform to C++ standard
* linbox/util/gmp++/gmp++_int_cstor.C: s/iostream.h/iostream/
* linbox/vector/vector-traits.h (struct
SparseSequenceVectorPairLessThan): Use std:: namespace qualifier
(class CompareSparseEntries): Use std:: namespace qualifier
(refSpecialized): Use std:: namespace qualifier
* linbox/util/error.h: s/iostream.h/iostream/
Use std:: namespace qualifier for iostream
* linbox/field/envelope.h (class FieldEnvelope): Ditto
* linbox/field/archetype.h (class FieldArchetype): Ditto
* linbox/field/abstract.h (class FieldAbstract): Use std::
namespace qualifier for ostream/istream
* gmp++/gmp++_int.inl (Integer::random): Eliminate default from
definition
* gmp++/gmp++_int.h: Use namespace qualifier std:: for parts of
the STL; update header names in accordance with C++ standard
* linbox/util/timer.h: s/iostream.h/iostream/ as per updated C++
standard; put everything inside the LinBox namespace; use
std::ostream
2002-07-08 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/solutions/methods.h: Put everything inside the LinBox
namespace
* linbox/solutions/rank.h (rank): Update
(rank): Use VectorWrapper::ensureDim; add commentator calls
* linbox/solutions/det.h (det): Update
* linbox/solutions/minpoly.h (minpoly): Update
* linbox/solutions/methods.h: Changed the name
_DEFAULT_EarlyTerm_THRESHOLD_ to the more standard-consistent
DEFAULT_EARLY_TERM_THRESHOLD; changed the name
Early_Term_Threshold to earlyTermThreshold, also in keeping with
the standard.
2002-07-05 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/blackbox/sparse0-aux.h
(SparseMatrix0Aux::applyTranspose): Avoid use of operator [];
tweak performance with iterators
* linbox/util/field-axpy.h (class FieldAXPY): Add assignment
operator; make _F a real copy and not a reference
* linbox/field/modular.h (FieldAXPY::FieldAXPY): Add copy
constructor; use direct assignment for 0
(class FieldAXPY): Make _F a real copy; add assignment operator
* linbox/util/field-axpy.h (class FieldAXPY): Add copy constructor
* linbox/blackbox/sparse0-base.h (SparseMatrix0Base::prettyPrint):
Make i, j size_t's
* linbox/blackbox/sparse0-aux.h (class SparseMatrix0Aux): Add
_faxpy
(SparseMatrix0Aux::applyTranspose): Use _faxpy vector to enable
delayed modding out
2002-07-02 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/modular.h (class FieldAXPY): Add template <>
* examples/blackbox/.cvsignore: Update
* tests/.cvsignore: Update
* tests/test-moore-penrose.C (testIdentityApply): Make display of
vectors LEVEL_UNIMPORTANT
(testRandomApply1): Ditto
(testRandomApply1): Put the construction of Adagger inside a
commentator activity; change "Computing" to "Applying" for the
activity surrounding the application
* HACKING: Update instructions for adding a test
* tests/Makefile.am (BASIC_TESTS): Move all mandatory tests here
(TESTS): Include BASIC_TESTS
(EXTRA_PROGRAMS): Include BASIC_TESTS
(noinst_PROGRAMS): Remove
* linbox/util/vector-factory.h
(RandomSparseSeqVectorFactory::next): Make i a size_t
* linbox/field/vector-domain.C (VectorDomain::mulSpecialized):
Eliminate unused j
(VectorDomain::addinSpecialized, VectorDomain::subinSpecialized,
VectorDomain::axpyinSpecialized): Make i a size_t
(VectorDomain::copySpecialized,
VectorDomain::areEqualSpecialized): Make idx a size_t
* linbox/util/vector-factory.h
(RandomSparseSeqVectorFactory::RandomSparseSeqVectorFactory):
Reorder initializers
(class VectorFactory): Add virtual destructor
* tests/test-vector-domain.C (testSubMul): Remove unused i, j
(testAXPY): Ditto
* linbox/blackbox/inverse.h (Inverse::apply): Eliminate unused j
* tests/test-vector-domain.C (testCopyEqual, testAddMul): Remove
unused i, j
(testDotProduct): Make j a size_t
* linbox/util/vector-factory.h (StandardBasisFactory::next): Make
idx a size_t
* linbox/blackbox/sparse0-aux.h
(SparseMatrix0Aux::applyTranspose): Initialize y to all zeros
first
* tests/test-sparse0.C (testRandomApply3, testRandomApply2,
testRandomApply1): Remove unused VD
* linbox/util/vector-factory.h
(RandomSparseSeqVectorFactory::next): Don't die if skip is 0; just
increment i
* tests/test-generic.h (testTranspose, testLinearity): Remove
unused i, j
2002-07-01 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/vector-domain.C (dotSpecialized): Don't initialize
res; it's not necessary here
* linbox/algorithms/blackbox-container-base.h (class
BlackboxContainerBase): Tweak indentation
* linbox/blackbox/sparse0-aux.h
(SparseMatrix0Aux::applyTranspose): Use linbox_check and _F.axpyin
* linbox/algorithms/Makefile.am (include_HEADERS): Add
blackbox-container-symmetrize.h
* linbox/field/modular.h (FieldAXPY::get, FieldAXPY::assign):
inline these
* linbox/util/field-axpy.h (FieldAXPY::get, FieldAXPY::assign):
inline these
* linbox/field/vector-domain.C (VectorDomain::dotSpecialized):
inline this
* linbox/blackbox/sparse0-aux.h (SparseMatrix0Aux::apply):
Eliminate Element temp
* linbox/util/commentator.C (Commentator::progress): Show elapsed time
(Commentator::progress): Get a pointer to the top activity first
* linbox/algorithms/massey-domain.h (MasseyDomain::massey): Use
COMMOD for progress reports
* linbox/util/commentator.C (Commentator::updateActivityReport):
Use a strstream and update _last_line_len
(Commentator::finishActivityReport): Use _last_line_len
(Commentator::setReportStream): Disable brief report if we are
outputing to the same stream it is
* linbox/util/commentator.h (class Commentator): Add _last_line_len
* linbox/util/commentator.C (Commentator::stop): Increase output
precision; make sure output is nonnegative
* examples/blackbox/Makefile.am (EXTRA_PROGRAMS): Add load-det
* linbox/field/givaro-zpz.h: Reindent
* examples/run/Makefile.am: Kill _DEPENDENCIES lines
* examples/integers/Makefile.am: Kill _DEPENDENCIES lines
* examples/fields/Makefile.am: Kill _DEPENDENCIES lines
* examples/blackbox/Makefile.am: Kill _DEPENDENCIES lines
* examples/Makefile.am: Kill _DEPENDENCIES lines
* HACKING: Remove note about _DEPENDENCIES lines
* tests/Makefile.am: Get rid of _DEPENDENCIES lines; turns out
they aren't necessary
* linbox/util/commentator.C (Commentator::stop): Modified report
format; display all three timings
2002-06-28 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/util/commentator.C (Commentator::updateActivityReport):
Use floor (percent + 0.5) rather than round (percent)
* linbox/util/commentator.h (Commentator): Redefine Commentator to
CommentatorDisabled when disabled; define the existing commentator
regardless of whether it is disabled -- the new #defines will
deflect any access
* tests/test-minpoly.C (main): Re-enable commentator
* linbox/util/commentator.C (Commentator::Commentator): Set
default max depth for BRIEF_REPORT to 2; use macros for detail
levels
(MessageClass::fixDefaultConfig): Make sure j is not erased from
the list before insertion
* gmp++/gmp++_int.h: Add short, unsigned short implicit
conversions
2002-06-27 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/util/commentator.C (Commentator::updateActivityReport):
Use round (percent) rather than setting precision
(MessageClass::checkConfig): Use depth < (*i).first
(MessageClass::fixDefaultConfig): Erase everything after
_max_depth and everything before _max_depth that is not more
restrictive
* tests/test-vector-domain.C (testAXPY): Update use of VD.axpy
* tests/test-common.h (applyPoly): Update use of VD.axpy
* tests/test-generic.h (testLinearity): Update use of VD.axpy
* linbox/blackbox/inverse.h (Inverse::apply,
Inverse::applyTranspose): Update
* linbox/field/vector-domain.h (class VectorDomain): Change order
of parameters to match FieldArchetype::axpy
2002-06-26 Bradford Hovinen <hovinen@eos.cis.udel.edu>
* tests/test-minpoly.C: Report minpoly degree; tweak report
parameters
* linbox/algorithms/blackbox-container.h
(BlackboxContainer::BlackboxContainer): Initialize _applyTime and
_dotTime if INCLUDE_TIMING is defined
2002-06-26 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/vector-domain.C (VectorDomain::addinSpecialized,
VectorDomain::subinSpecialized): Use i->first < j->first
* linbox/util/vector-factory.h
(RandomSparseSeqVectorFactory::next): Add check that skip > 0
* linbox/field/vector-domain.C (VectorDomain::axpySpecialized):
Fix order of arguments to F.axpy
* tests/test-generic.h (testLinearity): Fix variable names to be
more descriptive
* linbox/util/vector-factory.h (LinBox): Use v[] rather than
v.insert
* tests/test-generic.h (testDenseConsistency): Eliminate this test
(testLinearity): Fix formulae
* linbox/solutions/trace.h (trace): Make sure v has correct
dimension
* tests/test-inverse.C (testRandomTranspose): Update according to
changes in testTranspose
* tests/test-diagonal.C (testRandomTranspose): Update according to
changes in testTranspose
* linbox/blackbox/sparse0-aux.h
(SparseMatrix0Aux::SparseMatrix0Aux): Initialize _VD for sparse
associative specialization
* tests/test-generic.h (testTranspose): Use VectorFactory, make
this generic wrt vector type
* tests/test-sparse0.C (testRandomTranspose, testRandomLinearity):
New tests
(main): Call tests on all combinations of vector types
* linbox/util/vector-factory.h (StandardBasisFactory::next):
Initialize the whole vector for the dense case
* linbox/blackbox/sparse0-base.h (SparseMatrix0Base::put_value,
SparseMatrix0Base::operator[]): Use linbox_check
* tests/test-sparse0.C (testRandomApply1, testRandomApply2,
testRandomApply3): Make these generic wrt vector and row type
* */*.[Ch]: Change top Emacs identifier to mode: C++ so that it
does the correct thing with templates and so on
* linbox/util/vector-factory.h (class RandomDenseVectorFactory):
Use RandIter, not Field::RandIter for _r
2002-06-25 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-sparse0.C (testIdentityApply): Use VectorFactory
(testNilpotentApply): Use VectorFactory
* linbox/blackbox/sparse0-aux.h (class SparseMatrix0Aux): Use
VectorDomain for the sparse associative specialization; replace
Field1 with VectorDomain<Field>
(SparseMatrix0Aux::apply): Use VectorDomain::dot for all
specializations
(class SparseMatrix0Aux): Restore compilation of sparse
associative specialization
(SparseMatrix0Aux::apply): Use iterator rather than integer
* tests/Makefile.am (TESTS): Add test-subvector
* linbox/algorithms/blackbox-container.h (class
BlackboxContainer): Add optional timing facilities, enabled with
INCLUDE_TIMING
* linbox/algorithms/massey-domain.h (class MasseyDomain): Add
optional timing facilities, enabled with INCLUDE_TIMING
* linbox/vector/Makefile.am (include_HEADERS): Add subvector.h,
subiterator.h, reverse.h
* linbox/vector/subvector.h: Reindent; fix headers
* linbox/algorithms/massey-domain.h (class MasseyDomain): Replace
_field with _F; declare VectorDomain _VD; use const long rather
than #define when possible
* linbox/algorithms/blackbox-container-base.h (class
BlackboxContainerBase): Add member VectorDomain _VD; rename _field
to _F (for consistency); eliminate operators; align members
(BlackboxContainerBase::size): Make this const
* linbox/algorithms/blackbox-container.h
(BlackboxContainer::_launch): Use _VD.dot rather than home-cooked
dot product
(class BlackboxContainer): Make RandIter a template parameter,
defaulted to Field::RandIter
* linbox/algorithms/Makefile.am (include_HEADERS): Add
blackbox-container-symmetric.h
* linbox/solutions/trace.h (trace): Use factory.j () - 1
2002-06-24 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-vector-domain.C (main): Reset factory 5 before sparse
sequence/sparse associative dot product; factory 6 before sparse
associative add/mul test
(testDotProduct): Add timing
* linbox/util/vector-factory.h (VectorFactory): Fix output of j ()
* HACKING: Update with consolidated ChangeLogs
* linbox/vector/vector-traits.h (refSpecialized,
cosntRefSpecialized): Check that size is nonzero before conducting
binary search; check that return iterator is not past-the-end
2002-06-21 Bradford Hovinen <hovinen@cis.udel.edu>
* examples/integers/.cvsignore: Update
* examples/.cvsignore: Update
2002-06-21 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/vector-domain.C (VectorDomain::axpyinSpecialized): Don't use an
iterator here, since we invalidate iterators by calling insert
* linbox/field/modular.h (Modular::neg): Make sure -0 = 0
* linbox/field/vector-domain.C (VectorDomain::axpySpecialized): Use _F.axpy
when possible; avoid use of temporaries
(VectorDomain::addSpecialized, VectorDomain::addinSpecialized,
VectorDomain::subSpecialized, VectorDomain::subinSpecialized,
VectorDomain::isZeroSpecialized,
VectorDomain::areEqualSpecialized,
VectorDomain::convertSpecialized, VectorDomain::readSpecialized,
VectorDomain::writeSpecialized, VectorDomain::copySpecialized):
Implement
(VectorDomain::mulSpecialized, VectorDomain::mulinSpecialized):
Implement specialization for sparse associative vectors
* linbox/randiter/nonzero.h (class NonzeroRandIter): Set default RandIter to
Field::RandIter
* linbox/util/vector-factory.h (class RandomSparseSeqVectorFactory, class
RandomSparseMapVectorFactory): Use NonzeroRandIter
* linbox/vector/vector-traits.h (class CompareSparseEntries): Implement
(refSpecialized, constRefSpecialized): Use lower_bound
* tests/test-generic.h (testLinearity): New test
* tests/test-vector-domain.C (testAddMul, testSubMul, testCopyEqual):
New tests; call these from main
* tests/test-vector-domain.C (testAXPY): Use VD.isZero
* tests/.cvsignore: Update
2002-06-20 Bradford Hovinen <hovinen@cis.udel.edu>
* examples/Makefile.am (EXTRA_PROGRAMS): Add mmZZ_pE
* examples/integers/Makefile.am (EXTRA_PROGRAMS): Don't compile programs by
default
2002-06-20 Bradford Hovinen <hovinen@cis.udel.edu>
* examples/blackbox/Makefile.am (EXTRA_PROGRAMS): Don't compile examples by default
* examples/fields/Makefile.am (EXTRA_PROGRAMS): Don't compile programs by default
* linbox/blackbox/dense-matrix1.h (DenseMatrix::Row::Row): Fix typos in default
initialization
* linbox/blackbox/dense-matrix1.C (DenseMatrix::write): Remove default for
parameter
(DenseMatrix::ColOfRowsIterator::ColOfRowsIterator): Remove
default for first parameter
(DenseMatrix::RowOfColsIterator::RowOfColsIterator): Remove
default for first parameter
(class DenseMatrix::RowOfColsIterator): Add constructor with no
parameters
(class DenseMatrix::ColOfRowsIterator): Add constructor with no
parameters
* linbox/field/Makefile.am (EXTRA_DIST): Add EXTRA_HEADERS
2002-06-20 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
* linbox/field/givaro-gfq.h: Replace cast according to gmp++ .
2002-06-20 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/randiter/Makefile.am (include_HEADERS): Add nonzero.h
* linbox/util/vector-factory.h (class RandomDenseVectorFactory, class
RandomSparseSeqVectorFactory, class RandomSparseMapVectorFactory):
Add constructor specifying a particular random iterator
2002-06-20 Dave Saunders <saunders@cis.udel.edu>
* insert LinBoxSrcOnly flag for all-source compilation (fixed)
2002-06-20 Bradford Hovinen <hovinen@cis.udel.edu>
* macros/ntl-check.m4 (AC_TRY_LINK): Don't include iostream -- it was
causing compiler trouble wrt autoconf
(NTL_HEADERS_FIELD): Add ntl-ZZ_pE.h
* tests/Makefile.am (TESTS): Add test-randiter-nonzero
(test_diagonal_DEPENDENCIES): Remove duplicate definition
2002-06-19 Bradford Hovinen <hovinen@cis.udel.edu>
* configure.in (AC_OUTPUT): Remove linbox/iterators/Makefile,
remove examples/run/Makefile
2002-06-19 Bradford Hovinen <hovinen@cis.udel.edu>
* examples/Makefile.am (SUBDIRS): Add integers
* examples/fields/Makefile.am (ex_fields_wrapper_SOURCES): Don't comment out this
line
* linbox/util/gmp++/Makefile.am (INCLUDES): Add -I$(top_srcdir)
* linbox/util/vector-factory.h (RandomSparseSeqVectorFactory::next): Use
simpler formulas ala Knuth
(class RandomSparseSeqVectorFactory): Remove extraneous variables
* tests/test-sum.C (main): Use GF(101) by default
(testZeroApply): Use VD.mul rather than VD.axpy
2002-06-18 Bradford Hovinen <hovinen@cis.udel.edu>
* Makefile.am (SUBDIRS): Add gmp++
* configure.in (AC_OUTPUT): s/integer/gmp++/
2002-06-18 Bradford Hovinen <hovinen@cis.udel.edu>
* examples/blackbox/ex-mat0.C: Add using namespace std
* linbox/integer.h: Use GMP++
2002-06-18 Zhendong Wan <wan@cis.udel.edu>
* dense-matrix1.h dense-matrix1.C
* FieldBASL.h FieldBLAS.C
* linbox/field/ntl-ZZ_pE.h
2002-06-18 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/util/vector-factory.h (RandomSparseMapVectorFactory::next): Use a
simpler mechanism for constructing vectors
* linbox/util/gmp++/gmp++_int.h: Define GMPXX_PREFIX if not already done
* linbox/util/gmp++/*.[Ch]: Prefix #include lines with GMPXX_PREFIX
* linbox/util/Makefile.am (SUBDIRS): s/integer/gmp++/
(libutil_la_LIBADD): s/integer/gmpxx/
2002-06-14 Bradford Hovinen <hovinen@cis.udel.edu>
* examples/blackbox/.cvsignore: Update
2002-06-14 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/vector-domain.C (VectorDomain::dotSpecialized): Make sure not to
accumulate if j is at v2.end.
(VectorDomain::dotSpecialized): Add specialization for sparse
associative/sparse associative dot product
* linbox/field/modular.h (class ModularBase, class Modular): Replace
Modular[Base]::Modular[Base] (Element) with unsigned long version
2002-06-14 zhenddong wan <wan@cis.udel.edu>
* linbox/vector/subvector.h: fix the const_iterator problem
2002-06-14 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-vector-domain.C (main): Add call to test sparse
associative/sparse associative dot product
2002-06-13 Bradford Hovinen <hovinen@cis.udel.edu>
* .cvsignore: Remove config.h[.in] and add linbox-config.h[.in]
* configure.in: Change AM_CONFIG_HEADER to linbox-config.h
* */*.C: Change #include "config.h" to #include "linbox-config.h"
* Makefile.am (include_HEADERS): Add linbox-config.h
* acconfig.h: Add GMP_VERSION_4
2002-06-13 Bradford Hovinen <hovinen@cis.udel.edu>
* examples/blackbox/.cvsignore: Update
* linbox/integer.h: Only use extern "C" if GMP_VERSION_4 is not set
* linbox/vector/vector-traits.h: Add #include linbox/field/archetype.h
* macros/gmp-check.m4: Added check for GMP 4.0 or greater
* tests/test-vector-domain.C (main): Invoke testDotProduct for sparse
sequence/sparse sequence vectors
* tests/Makefile.am: Commented out subiterator and solveNonsingular as
the former has compilation problems and the latter was not added
properly
2002-06-11 Dave Saunders <saunders@cis.udel.edu>
* linbox/field/ntl-ZZ_p.h: include sys/time.h
2002-06-11 Dave Saunders <saunders@cis.udel.edu>
* tests/test-subiterator.C: new test of vector/subiterator.h
2002-06-10 Dave Saunders <saunders@cis.udel.edu>
* linbox/vector/subiterator.h, subvector.h: subvector now constructs from
subiterators, not from underlying vector. This avoids
copy and reference problems with underlying vector.
2002-06-07 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/vector-domain.C (VectorDomain::axpySpecialized): Implement
specialization for sparse associative vectors
(VectorDomain::axpyinSpecialized): Implement specialization for
sparse associative vectors
(VectorDomain::dotSpecialized): Implement specialization for
sparse sequence/sparse sequence and sparse associative/sparse
sequence vectors
2002-06-07 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-vector-domain.C (main): Add call to test sparse associative
vector axpy
2002-06-06 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/vector-domain.C (VectorDomain::dotSpecialized): Implement
specialization for sparse associative/dense dot product
* linbox/field/vector-domain.h (class VectorDomain): Add return statements
2002-06-06 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/util/vector-factory.h
(RandomSparseSeqVectorFactory::RandomSparseSeqVectorFactory): Make
sure k < n
(RandomSparseSeqVectorFactory::next): Tweak formulae
(RandomSparseMapVectorFactory::next): Tweak formulae the same way
as above
* tests/test-vector-domain.C (testAXPY): Rename from testDenseAXPY. Make
this generic with respect to the vector type being tested
(main): Update according to the above
* tests/.cvsignore: Update
* tests/test-vector-domain.C (main): Add call to test sparse associative
dot product
2002-06-05 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-vector-domain.C (testDotProduct): Rename from
testDenseDotProduct. Make this generic with respect to the vector
type being tested
2002-06-04 Bradford Hovinen <hovinen@aphrodite.cis.udel.edu>
* .cvsignore: Update
2002-06-04 Bradford Hovinen <hovinen@aphrodite.cis.udel.edu>
* examples/run/.cvsignore: Update
2002-06-04 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/blackbox/inverse.h (class Inverse): Update use of VectorDomain
* linbox/blackbox/sparse0-aux.h (class SparseMatrix0Aux): Update use of VectorDomain
* linbox/blackbox/dense-matrix.h (class DenseMatrix): Update use of VectorDomain
2002-06-04 Bradford Hovinen <hovinen@aphrodite.cis.udel.edu>
* linbox/blackbox/moore-penrose.h (class MoorePenrose): s/_B/_B1/ since _B seems
to be a macro of some type on Solaris
2002-06-04 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/vector-domain.h (class VectorDomainBase, class VectorDomain):
Merge these classes together
(class VectorDomain): Parameterize methods on the vector types
rather than parameterizing the whole class. Add methods
* Specialized for template specialization
* linbox/field/vector-domain.C: Update according to changes in vector-domain.h
2002-06-04 Dave Saunders <saunders@cis.udel.edu>
* linbox/vector/subvector.h: tweak to have reference semantics to underlying vect.
* tests/test-subvector.C: new test for check adapted from Will's
test-subvector.cpp
* tests/Makefile.am: new target test-subvector, part of check.
2002-06-04 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-moore-penrose.C (testRandomApply1):
s/VectorDomainBase/VectorDomain/ and update
* tests/test-common.h (applyPoly): Update use of VectorDomain
* tests/test-generic.h (testDenseConsisntency): Update use of
VectorDomain
(testTranspose): Update use of VectorDomain
2002-05-31 Bradford Hovinen <hovinen@cis.udel.edu>
* macros/gmp-check.m4 (min_gmp_version): s/{gmp_prefix}/${gmp_prefix}/
2002-05-28 Bradford Hovinen <hovinen@cis.udel.edu>
* configure.in: AC_OUTPUT linbox-config
* Makefile.am (bin_SCRIPTS): Add linbox-config
2002-05-28 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/blackbox/Makefile.am: Add sum.h
2002-05-25 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/blackbox/compose.h (Compose::Compose): Add call to
VectorWrapper::ensureDim
(class Compose): Make _z mutable
2002-05-25 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/Makefile.am (TESTS): Reorder tests in approximate order of
dependence
(TESTS): Add test-sum
2002-05-19 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/blackbox/sparse0-aux.h (SparseMatrix0Aux::applyTranspose): Don't
construct a new vector for y; just fill in existing values
* linbox/blackbox/inverse.h (class Inverse): Add _transposeMinpoly for the minimal
polynomial of the transpose of the matrix
(Inverse::Inverse): Don't construct minimal polynomials; just
clear them
(Inverse::apply, Inverse::applyTranspose): Construct minimal
polynomials if needed
* linbox/blackbox/moore-penrose.h (MoorePenrose::MoorePenrose): Reverse order of
arguments for call to Submatrix
(MoorePenrose::MoorePenrose): Construct _B too
* linbox/blackbox/compose.h (Compose::Compose): Don't set _z's size
(Compose::apply, Compose::applyTranspose): use
VectorWrapper::ensureDim rather than setting _z's size
* linbox/blackbox/inverse.h (Inverse::apply, Inverse::applyTranspose): Don't
resize the output vector
(Inverse::Inverse): Check that the input matrix is square
(Inverse::Inverse): Use VectorWrapper::ensureDim rather than
_z.resize
* linbox/blackbox/Makefile.am (include_HEADERS): Added moore-penrose.h
2002-05-19 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/modular.h (class FieldAXPY): Update get () form; remove obsolete
forms
* linbox/field/vector-domain.C (VectorDomain::dotprod): Fix use of FieldAXPY::get
* linbox/solutions/trace.h (trace): Use constRef
* linbox/util/field-axpy.h (class FieldAXPY): Remove obsolete get form
* linbox/util/vector-factory.h (class VectorFactory): Add interface n(), which
returns the dimension of the space from which vectors are being
generated
* linbox/vector/vector-traits.h (class VectorWrapper): Add method ensureDim
(namespace VectorWrapper): Convert this to a namespace and use
function specialization
* tests/test-moore-penrose.C (main): Switch to field GF(2147483647) by
default
* tests/test-generic.h (testTranspose): Rename from testTranpose
(testTranspose): Use commentator.startIteration
* tests/test-moore-penrose.C (testIdentityApply): Ensure w has space for
result
(buildRandomSparseMatrix): Implement
(testRandomApply1): New test
* tests/test-trace.C (testDiagonalTrace): Use constRef
2002-05-18 Bradford Hovinen <hovinen@cis.udel.edu>
* configure.in: Make disabling shared libraries the default
2002-05-18 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/util/vector-factory.h: Refactor: Create one class
StandardBasisFactory, parameterized by vector type, with
specializations for dense, sparse map, and sparse associative
vectors
* linbox/util/commentator.C (Commentator::setMaxDepth,
Commentator::setMaxDetailLevel): Implement
* linbox/vector/vector-traits.h (class VectorWrapper): Implement
2002-05-17 Bradford Hovinen <hovinen@cis.udel.edu>
* examples/.cvsignore: Update
2002-05-17 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/blackbox/submatrix.h (class Submatrix): Add Trait to template parameters
and specialize for dense vectors
(Submatrix::Submatrix): Check that the row and column dimensions
given are valid
* linbox/blackbox/sparse0-aux.h: Backed out revision 1.6
(SparseMatrix0Aux::apply): Restore use of _VD.dotprod
* linbox/field/vector-domain.h (LinBox): Added specialization for
SparseSequence, SparseSequence type
* linbox/field/vector-domain.C: Update as per changes in vector-domain.h
(VectorDomainBase::mul, VectorDomainBase::mulin): Don't resize and
don't invalidate iterators
* linbox/field/vector-domain.h (class VectorDomainBase): New class. Pull out
all functions from VectorDomain that only need one vector type.
* linbox/util/vector-factory.h (RandomDenseVectorFactory::next): Resize input
vector
* linbox/util/debug.h (PreconditionFailed::PreconditionFailed): Put out a
couple of newlines before printing the error message
* linbox/util/vector-factory.h (class VectorFactory): Add method reset ()
* linbox/util/commentator.C (Commentator::startIteration): Implement
Include strstream and not cstdio
* linbox/util/commentator.h (class Commentator): Add startIteration
* macros/ntl-check.m4 (EX_FIELDS_NTL): Define
* tests/Makefile.am (test_moore_penrose_SOURCES): New test
* tests/test-submatrix.C: Don't include fstream or cstdio
(testRandomApply): Use commentator.startIteration rather than
forming a new string and calling start
(testRandomApply): Put the declaration for report inside the loop
and remove the call to indent
(testRandomApply): Update report output
(main): Don't construct factory2
* tests/.cvsignore: Update
2002-05-14 William J. Turner <wjturner@acm.org>
* linbox/field/archetype.h, envelope.h: changed randIter to RandIter.
2002-05-14 William J. Turner <wjturner@acm.org>
* linbox/randiter/modular.h: Seeded random number generator in constructor. _seed was
never used before.
2002-05-07 William J. Turner <wjturner@acm.org>
* linbox/field/ntl-zz_p.h, ntl-ZZ_p.h: modified random generator constructor and
creating functions to use NTL functions.
2002-05-07 William J. Turner <wjturner@acm.org>
* linbox/randiter/unparametric.h: static cast added to constructor
2002-04-26 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
* configure.in, acconfig.h: updating to allow Lidia Library
2002-04-26 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
* linbox/field/lidia-gfq.h, lidia.h: adding wrapper of extensions fields for Lidia
* linbox/field/givaro-gfq.h, givaro-zpz.h: To use these wrappers you need
to get a modified version of givaro. You can find these version
at this address : http://www.ens-lyon.fr/~pgiorgi/software.html
version 3.0.1.
* linbox/field/givaro-gfq.h: adding wrapper of extensions fields for Givaro
* linbox/field/givaro.h: adding givaro-gfq.h
* linbox/field/Makefile.am: adding givaro-gfq.h
* linbox/randiter/lidia-gfq.h: adding randiter for lidia gfq field
* linbox/randiter/lidia.h: file for all randiter of lidia library
* linbox/randiter/Makefile.am: adding the header of Lidia and an EXTRA_DIST variable
* macros/givaro-check.m4: adding headers of givaro fields wrappers
and the test test-givaro-zpz.h
* macros/lidia-check.m4: creating for the library LiDIA
* macros/Makefile.am: adding lidia-check.m4 to the EXTRA_DIST
* tests/test-givaro-zpz.C: test for prime finites fields of givaro
using all type of zpz fields of givaro and an extension field of
degree 1 for gfq.
* tests/test-lidia-gfq.C: test for Lidia Gfq field.
But the Gfq field is tested as a prime field (the extension degree
is 1)
* tests/test-generic.h: Put in comment the line 56 which is inconsistant.
Also make the test for the init/convert function generic for every
fields (finite and inifinite fields).
2002-04-25 Zhendong Wan <wan@mail.eecis.udel.edu>
* tests/test-dense-vector-domain.C: test for dense vector domain.
2002-04-24 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
* linbox/givaro.h: include file for all Givaro wrappers
* linbox/lidia.h: include file for all Lidia wrappers
* linbox/Makefile.am: adding file givaro.h and lidia.h and adding
all Lidia header base
2002-04-24 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
* linbox/field/givaro-zpz.h, givaro.h: Adding wrappers of primes fields for
givaro and his file givaro.h associated.
NB: this wrappers is template class which wrapps three sorts of
givaro fields. For more information see the file itself.
* linbox/field/Makefile.am: adding givaro fields wrapper
2002-04-18 William J. Turner <wjturner@acm.org>
* linbox/vector/subvector.h, subiterator.h: New wrappers for subvector of dense
vector.
2002-04-18 William J. Turner <wjturner@acm.org>
* tests/test-subvector.cpp: New partial test program for subvector wrapper
class. This needs a lot of work, and is not in the makefile yet.
2002-04-17 Dave Saunders <saunders@cis.udel.edu>
* tweak identity case to copy
2002-04-13 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/modular.h (FieldAXPY<Modular<short|long> >): Add form of get
where the element in which to store result is also passed
2002-04-13 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/util/Makefile.am (include_HEADERS): Added vector-factory.h
* linbox/util/field-axpy.h (class FieldAXPY): Add form of get where the place
in which to store the result is passed
* tests/test-vector-domain.C (testDenseDotProduct): Convert this to use
VectorFactories
2002-04-12 Bradford Hovinen <hovinen@cis.udel.edu>
* s/element/Element/ everywhere except in comments (I hope!)
* Add a file dummy.C to fool libtool into actually building
liblinbox.la
* Move debug.[Ch] and error.[Ch] to linbox/util
2002-04-12 William J. Turner <wjturner@acm.org>
* linbox/blackbox/sparse0-aux.h: Made dense vector apply not invalidate iterators
2002-04-12 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/util/commentator.h (class Commentator): Added compatibility layer for
Jean-Guillaume's code
(class Commentator): Move cnull above protected section to fix
crashes on DISABLE_COMMENTATOR
* linbox/util/commentator.h: Added null commentator
* linbox/util/Makefile.am (include_HEADERS): Put debug.h, error.h here
(libutil_la_SOURCES): Put error.C, debug.C here
* macros/givaro-check.m4: Fix definition of GIVARO_HEADERS_*, GIVARO_TESTS
* tests/test-unparametric-field.C: Rename from test-unparametric-field.cpp
2002-04-11 Bradford Hovinen <hovinen@cis.udel.edu>
* Makefile.am: Remove EXTRA_DIST, since both of these files are
for CVS developers only
2002-04-11 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/modular.h (FieldAXPY<Modular<short|long> >::accumulate): Use -t
for overflow check
(FieldAXPY<Modular<short|long> >::accumulate): Cast to unsigned
when modding out to avoid sign flip problem
(FieldAXPY<Modular<short|long> >::get): Cast to unsigned when
modding out
* tests/test-modular.C (main): Call testFieldAXPY
* tests/test-generic.h (testField): Make title const
(testFieldAXPY): Add this test
* tests/test-diagonal.C (testRandomTranspose): Report diagonal vector
used as input
* tests/test-scalar-matrix.C: #include "modular.h"
(main): Use integers mod 2147483647; update to use Modular interface
* tests/.cvsignore: Add test-submatrix
* tests/test-submatrix.C (main): Fix banner string
* tests/test-diagonal.C (main): Enable printing of detailed messages
* tests/test-unparametric-field.cpp (main):
* tests/test-modular.C (main):
* tests/test-inverse.C (main):
* tests/test-minpoly.C (main):
* tests/test-diagonal.C (main):
* tests/test-rank.C (main): Use GF(2147483647) by default; fix relevant
documentation
* tests/test-sparse0.C (main): Fix documentation string for default modulus
2002-04-10 Bradford Hovinen <hovinen@cis.udel.edu>
* examples/example.C: Update with new Modular signature
2002-04-10 Dave Saunders <saunders@udel.edu>
* linbox/blackbox/scalar-matrix.h: New (simple) black box - sparse vector part not
checked.
2002-04-10 Bradford Hovinen <hovinen@cis.udel.edu>
* macros/ntl-check.m4: Set NTL_TESTS, NTL_HEADERS_* to empty strings by
default and to real data if NTL was found. Call AC_SUBST in all
cases
Fix check for whether NTL prefix was set
* tests/.cvsignore: Add test-unparametric-field;
s/test-sparse-matrix/test-sparse0/
2002-04-10 Dave Saunders <saunders@udel.edu>
* tests/test-scalar-matrix.C: New test for new blackbox.
2002-04-08 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/util/commentator.h (class Commentator): Use std:: namespace for
ofstream
(class Commentator): Add std:: specifiers for nullstreambuf
declarations and cnull
2002-04-08 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
* macros/gmp-check.m4: Keep the current LD_LIBRARY_PATH
* tests/test-generic.h: WARNING: The test of a field is not generic. The test
of init and convert functions use a fixed number 49193295. So the test
failed if the characteristic is less than this number.
2002-04-05 William J. Turner <wjturner@acm.org>
* linbox/Makefile.am: Install NTL and Givaro wrapper headers only if
libraries are installed.
2002-04-05 William J. Turner <wjturner@acm.org>
* linbox/blackbox/Makefile.am: Only install NTL and Givaro wrapper header files
if libraries are installed.
* linbox/blackbox/sparse0-aux.h: Fixed apply and applyTranspose for sparse vectors.
* macros/ntl-check.m4, givaro-check.m4: Only run tests of NTL and Givaro
wrappers, and install header files, if libraries are installed.
* tests/test-common.h: Fix printVector to print correct number of zeros
for sparse sequence vectors. Made functions for testing equality of
two vectors of the same type and to test if a vector is the zero
vector. These may need some work still.
* tests/Makefile.am: Only run tests of NTL and Givaro wrappers if libraries
are installed.
* tests/test-sparse0.C: Attempted to convert test functions to be generic
for the Row and Vector types. These still need more work to be
complete. In particular, there still needs to be a function to
create a vector with no zero elements.
2002-04-04 William J. Turner <wjturner@acm.org>
* configure.in: removed requirement of givaro
2002-04-04 William J. Turner <wjturner@acm.org>
* linbox/ntl.h: include file for all NTL wrappers.
* linbox/blackbox/archetype.h: Added virtual destructor to avoid warning.
Removed resize() from apply and applyTranspose.
* linbox/blackbox/sparse0-base.h: Fixed comparison between signed and unsigned
integers, accessor function to field.
* linbox/blackbox/sparse0-aux.h: Fixed constructor from SparseMatrix0Base.
2002-04-04 William J. Turner <wjturner@math.ncsu.edu>
* linbox/field/Makefile.am: Install NTL and Givaro wrapper header files only
if libraries are installed.
* linbox/field/ntl.h, ntl-RR.h, ntl-ZZ_p.h, ntl-zz_p.h: Broke NTL field wrappers
into separate files.
2002-04-04 Pascal Giorgi <pascal.giorgi@ens-lyon.fr>
* macros/ntl-check.m4, givaro-check.m4: Use GMP_CFLAGS and GMP_LIBS and
appending to LD_LIBRARY_PATH, and removed UNSET LD_LIBRARY_PATH at
end of file
* macros/gmp-check.m4: removed UNSET LD_LIBRARY_PATH at end of file
2002-04-04 William J. Turner <wjturner@acm.org>
* macros/ntl-check.m4: Use GMP_CFLAGS and GMP_LIBS
* tests/test-ntl-zz_p.cpp, test-ntl-ZZ_p.cpp, test-ntl-RR.cpp: New tests
* tests/test-common.h, test-vector-domain.C: generalized printVector to all
vector types. Created areVectorsEqual function. Both need to be
generalized to sparse associative vectors.
2002-04-03 William J. Turner <wjturner@acm.org>
* examples/example.C: changed sparse-matrix to sparse0
2002-04-03 William J. Turner <wjturner@acm.org>
* linbox/blackbox/sparse1.h: Original port; needs work!
* linbox/blackbox/sparse0*.h: Renamed from sparse-matrix*.h.
2002-04-03 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/modular.h (class Modular): Rename from LargeModular, make this a
class parameterized on the element type
2002-04-03 William J. Turner <wjturner@math.ncsu.edu>
* linbox/field/ntl.h: Fixed random iterator random(x) for zz_p
2002-04-03 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/randiter/modular.h (class Modular): Rename from LargeModular; make this a
class parameterized on the element type
2002-04-03 William J. Turner <wjturner@acm.org>
* linbox/randiter/unparametric.h: Ported UnparametricRandIter<K> from old module
* macros/givaro-check.m4: Copy from gmp-check.m4 and fix for Givaro
* macros/ntl-check.m4: Change to version >= 5.0 and C++ namespace
* tests/test-unparametric-field.cpp: New test
* tests/test-minpoly.C: Changed sparse-matrix to sparse0
* tests/test-sparse0.C: Changed sparse-matrix to sparse0
2002-04-02 Bradford Hovinen <hovinen@cis.udel.edu>
* s/*_{abstract|archetype|envelope}/Abstract|Archetype|Envelope/g
2002-04-02 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/blackbox/dense-matrix.h (DenseMatrix::apply): Signedness fix
* linbox/blackbox/submatrix.h (class Submatrix): Don't parameterize on the field;
add rowdim, coldim
(Submatrix::Submatrix): Pass rowdim, coldim; initialize vectors
directly
* linbox/field/Makefile.am (include_HEADERS): Add ntl.h
* linbox/field/vector-domain.C (VectorDomain::axpy): Remove unused iterator k
* linbox/util/commentator.C (Commentator::printActivityReport,
Commentator::indent)
(Commentator::updateActivityReport,
Commentator::finishActivityReport): Make i unsigned
(MessageClass::checkConfig): Signedness fix
* linbox/util/integer/integer-io.C: Return v
* tests/test-dense-matrix.C (test*): s/size_t/long/ for parameter n
* tests/Makefile.am (TESTS): Add test test-submatrix
* tests/test-common.h (printSparseSeqVector, printVector): Signedness
fix
* tests/test-vector-domain.C (test*): s/size_t/long/ for parameter n
* tests/.cvsignore: Added test-det, test-rank
2002-04-01 Bradford Hovinen <hovinen@cis.udel.edu>
* Makefile.am (SUBDIRS): Build documentation last
2002-04-01 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/blackbox/transpose.h (Transpose::Transpose): Make A_ptr const
(Transpose::Transpose): Use Transpose<Vector>, not Compose<Vector>
for copy constructor
* linbox/blackbox/compose.h (Compose::apply, Compose::applyTranspose): Declare
local intermediate vector, since it was not compiling (for what
reason I do not know) with the class member
* linbox/solutions/rank.h: #include blackbox-container.h
(rank): Compilation fixes
* linbox/solutions/det.h: #include blackbox-container.h
* tests/test-det.C (test*): Supply iterations to commentator.start
* tests/Makefile.am (TESTS): Add test test-rank
2002-03-30 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/blackbox/compose.h (Compose::Compose): Use const pointers for inputs
(Compose::Compose): Initialize _z directly
2002-03-30 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/solutions/det.h (det): Compilation fixes
2002-03-29 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-generic.h (testDenseConsisntency): New test
2002-03-28 Bradford Hovinen <hovinen@cis.udel.edu>
* Makefile.am (EXTRA_DIST): Typo fix
2002-03-28 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/blackbox/dense-matrix.h (DenseMatrix::clone): Implement nontrivially
(DenseMatrix::DenseMatrix): Implement copy constructor
* linbox/util/commentator.C (Commentator::Commentator): Set default max depth
for errors and warnings to 10; set max detail to 3
(Commentator::registerMessageClass): Accept parameters max_depth
and max_level for initial configuration of message class
* tests/test-inverse.C (main): Set maximum depth on internal
descriptions to 3
(testDiagonalInverse): Add this test
(main): Disable Hilbert inverse test pending test of Hilbert black
box
* tests/Makefile.am (test_large_modular_SOURCES):
s/test-field-common/test-generic/
* tests/test-large-modular.C: Update
* tests/test-field-common.h: Obsolete now; remove
* tests/test-field-common.h (testField): Move to test-generic.h
* tests/test-common.[Ch] (test_header, test_trailer): Remove
* tests/test-field-common.h (isPower): Move to test-common.c/.h
2002-03-27 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/util/commentator.h (Commentator::Activity::Activity): Initialize
_progress to 0
* linbox/util/commentator.C (Commentator::Commentator): Initialize cnull
(Commentator::Commentator): Kill customization of BRIEF_REPORT
(MessageClass::dumpConfig): Implement. Dumps the state of this
MessageClass
(Commentator::updateActivityReport): #if out estimated time
* linbox/util/commentator.h (class Commentator): Make cnull no longer static
* linbox/util/commentator.C (MessageClass::MessageClass): Accept max_depth and
max_level parameters
(Commentator::progress): Remove redundant isPrinted check
(MessageClass::checkConfig): Return false if we get all the way
through the loop
2002-03-27 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-inverse.C (test*): Use commentator, kill report parameter
* tests/test-generic.h (testTranpose): Use commentator, kill report
parameter
* tests/test-diagonal.C (test*): Use commentator, kill report parameter
* tests/test-minpoly.C (test*): Use commentator, kill report parameter
* tests/test-large-modular.C (main): Kill report
* tests/test-sparse-matrix.C (test*): Use commentator, kill report
parameter
* tests/test-dense-matrix.C (testVandermonde, testIdentity): Use
commentator, kill report parameter
2002-03-26 Bradford Hovinen <hovinen@cis.udel.edu>
* configure.in: Add default CXXFLAGS
2002-03-26 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/util/commentator.[Ch]: Complete rewrite
* tests/test-*.C (main): Fix argument list
* tests/test-large-modular.C (main): Update; add Dave Saunders to
credits
* tests/test-field-common.h (test_field): Rename to testField
(testField): Remove report parameter; rename iters as iterations
(testField): Fix indentation
(testField): Use commentator
2002-03-24 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/algorithms/massey-domain.h (class MasseyDomain): Remove commentator; use
global commentator instead. Update
2002-03-24 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-common.C (parseArguments): Remove report argument
(parseArguments): Use commentator to set default report file
* tests/test-vector-domain.C (testDenseDotProduct): Use commentator
2002-03-19 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/blackbox/dense-matrix.h (class DenseMatrix): Use VectorDomain rather than
MatrixDomain; typedef Vector and use it throughout
(DenseMatrix::setEntry): Implement
(DenseMatrix::rowdim, DenseMatrix::coldim): Fix signatures
(DenseMatrix::clone): Define this; implementation deferred
(DenseMatrix::DenseMatrix): Accept field and build VectorDomain
object
2002-03-19 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-field-common.h: Prevent multiple inclusion
* tests/Makefile.am (TESTS): Beatify listing; put test-large-modular at
top
2002-03-18 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-diagonal.C (testRandomTranspose): Add this test
* tests/test-common.h (applyPoly, multiEvalPoly): Return resulting vector
(interpolatePoly): Implement
* tests/Makefile.am (TESTS): Add test-dense-matrix
2002-03-15 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-inverse.C (testVandermondeInverse): New test; construct the
inverse of a Vandermonde matrix and check that it behaves correctly
* tests/test-common.h (applyPoly): Use VectorDomain; typedef Vector to
reduce verbosity
(multiEvalPoly): Implement
2002-03-12 Bradford Hovinen <hovinen@cis.udel.edu>
* CHANGED-INTERFACES: Updated text
* TODO: Updates
2002-03-12 Bradford Hovinen <hovinen@cis.udel.edu>
* examples/example.C: Update to new interface
* linbox/blackbox/sparse-matrix-aux.h (LinBox): Rename references to MatrixDomain
to VectorDomain
(apply): Use vector domain dotprod
* linbox/blackbox/inverse.h (LinBox): Use matrix domain vector axpy
(LinBox): Compute the negative of the inverse of the constant
coefficient first, then multiply through by that
(Inverse): Rename MatrixDomain _MD to VectorDomain _VD
* linbox/field/large-modular.h (LinBox): Rename ParamModular to LargeModular
* linbox/field/Makefile.am (include_HEADERS): Rename param-modular.h to
large-modular.h
* linbox/field/vector-domain.[Ch]: Rename references to MatrixDomain to
VectorDomain
* linbox/field/Makefile.am (include_HEADERS): Rename matrix-domain.* to
vector-domain.*
* linbox/field/matrix-domain.h (LinBox): Add scalar-vector multiplication
* linbox/field/matrix-domain.C (LinBox): Use FieldAXPY for dot products
(MatrixDomain::axpy): Remember to increment i in the last loop
(LinBox): Use axpy and axpyin when possible
(LinBox): Specialize everything using MatrixDomainType rather than
MatrixDomainSimpleType
* linbox/randiter/large-modular.h: Rename ParamModularRandIter to
LargeModularRandIter
* linbox/randiter/Makefile.am (include_HEADERS): Rename param-modular.h to
large-modular.h
* tests/test-*.C: Update
* tests/.cvsignore: Update
* tests/test-vector-domain.C: Rename references to MatrixDomain to
VectorDomain
* tests/Makefile.am (test_vector_domain_SOURCES): Rename
test-matrix-domain to test-vector-domain
* tests/test-matrix-domain.C (testSparseAXPY, testDenseAXPY): Add these
tests
2002-03-11 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/matrix-domain.C (LinBox): Remove precondition checks for sparse
axpy
(LinBox): Check that i has not gone past the end of y in the loop
2002-03-06 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/matrix-domain.h (LinBox): Remove definitions that derive from
the field class; they are not necessary right now
2002-02-28 Bradford Hovinen <hovinen@cis.udel.edu>
* TODO: Added notes here
2002-02-28 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/field/matrix-domain.h (LinBox): Remove extraneous template
specialization
* linbox/field/matrix-domain.C (LinBox): Use _F.init to initialize field
elements
* linbox/field/matrix-domain.h: #include vector-traits.h
* tests/test-common.h (printSparseSeqVector): Implement
* tests/test-matrix-domain.C: Added this test
2002-02-27 Bradford Hovinen <hovinen@cis.udel.edu>
* doc/Makefile.am (SUBDIRS): Add $({DOCPLUSPLUS|TEX}_SUBDIRS)
2002-02-27 Bradford Hovinen <hovinen@cis.udel.edu>
* configure.in: Added doc/{html|tex}/Makefile; added call to check
for Doc++; Add a check for latex and only build latex
documentation if that is present
(TEX_SUBDIRS): Make sure this gets included even if Doc++ isn't found
* examples/example.C (main): Comment alignment fix
* linbox/*/*.dxx: Update paths
* Added missing *-wrappers.dxx
* linbox/*/*.dxx: Update paths; add missing .dxx files
* linbox/field/matrix-domain.h (LinBox): Add specialization for SparseSequence
vectors
2002-02-20 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/Makefile.am (noinst_PROGRAMS): Use $(TESTS)
2002-02-13 Bradford Hovinen <hovinen@cis.udel.edu>
* HACKING: Update information on directory structure
* configure.in: Properly quote error message strings
Use exit status 1 rather than -1 on error
* HACKING: Added section with checklists for adding header files
and tests to Linbox.
* configure.in: Print error message if GMP was not found
* HACKING: Update text
(http): Added a section on having the right development tools and
added some information on autogen.sh/configure options
* CHANGED-INTERFACES: Update text
2002-02-09 Bradford Hovinen <hovinen@cis.udel.edu>
* COPYING: Copy over LGPL
2002-02-08 Bradford Hovinen <hovinen@cis.udel.edu>
* all: s/element_/Element_/g
2002-02-08 Saunders <saunders@udel.edu>
* examples/Makefile.am: change load order
2002-02-08 Bradford Hovinen <hovinen@cis.udel.edu>
* linbox/blackbox/inverse.h: Update #include vector-traits.h
(Inverse): Correctly instantiate minpoly
Include minpoly.h
(class Inverse): Make _z mutable
2002-02-08 Saunders <saunders@udel.edu>
* macros/gmp-check.m4: set LD_LIBRARY_PATH
2002-02-08 Bradford Hovinen <hovinen@cis.udel.edu>
* macros/*-check.m4: Unset *_{CFLAGS|LIBS} if the library was not found
Invoke found macro only if version check succeeded
Invoke not found macro if version check failed
2002-02-08 Saunders <saunders@udel.edu>
* tests/Makefile.am: change link order.
2002-02-08 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-inverse.C (main): Use prevprime(2^32) for default modulus
(main): Fix banner
* tests/test-diagonal.C (main): Use prevprime(2^32) for default modulus
* tests/test-sparse-matrix.C (main): Use 101 for default modulus
* tests/test-minpoly.C (main): Use prevprime(2^32) for default modulus
* tests/Makefile.am (TESTS): Added test test-inverse
2002-02-07 Bradford Hovinen <hovinen@cis.udel.edu>
* tests/test-sparse-matrix.C (main): Use GF(65521) by default
* tests/test-minpoly.C (main): Use GF(65521) by default
* tests/test-diagonal.C (main): Use GF(65521) by default
* tests/Makefile.am (TESTS): Added test test-diagonal
* tests/test-sparse-matrix.C: Don't #include minpoly.h
2002-02-04 Bradford Hovinen <hovinen@cis.udel.edu>
* configure.in (AC_OUTPUT): Add examples/data/Makefile
2002-02-04 Bradford Hovinen <hovinen@cis.udel.edu>
* configure.in (CFLAGS): Make sure all CFLAGS variables make it in
to compilations
(CXXFLAGS): s/CFLAGS/CXXFLAGS/
* Makefile.am (SUBDIRS): Put examples last in the list
* linbox/blackbox/sparse-matrix-aux.h (apply): Initialize *y_iter to 0 before
looping
* macros/ntl-check.m4: Correct check in error message
* macros/*.m4: Remove setting LD_LIBRARY_PATH, etc.
* tests/test-sparse-matrix.C (testNilpotentApply): Reset even with every
iteration
* tests/test-minpoly.C (testNilpotentMinpoly): Fix test of correctness
* tests/test-common.h (printPolynomial): Write 0 for the zero polynomial
(printPolynomial): Make sure 1 polynomial gets printed correctly
* tests/test-minpoly.C (testIdentityMinpoly): Swap coefficients in the
test
* tests/test-minpoly.C (main): Update banner
* tests/test-common.h (applyPoly): Implement; copy from
linbox/blackbox/inverse.h
* tests/test-sparse-matrix.C: Move minpoly tests to test-minpoly
* tests/Makefile.am (TESTS): Add test test-minpoly
2002-02-03 Bradford Hovinen <hovinen@cis.udel.edu>
* */*.{C|h|inl}: Change Element to element, Integer to integer
* */*.{C|h|inl}: Updated header comments to reflect correct
filename locations
* */.cvsignore: Added/updated .cvsignore files
* configure.in: Add AC_PROG_CC, AC_PROG_RANLIB for building static
libraries
Add AM_PROG_LIBTOOL
* */Makefile.am: Added infrastructure for building static library
liblinbox.a so that we can avoid symbol collisions
2001-12-31 Bradford Hovinen <hovinen@cis.udel.edu>
* HACKING: Added documentation on coding style, fleshed out
algorithms a bit
distrib
>
Fedora
>
18
>
x86_64
>
media
>
updates
>
by-pkgid
>
a352fda7bb79791386fbf14df9be0f15
>
files
>
10
