This file describes changes in the cvec package.

2.5.6 (2016-11-08)
  - Fix a bug reported by Jan De Beule which resulted in cvec incorrectly
    converting certain GAP matrices to cvec matrices.

2.5.5 (2016-03-08)
  - Maintenance release

2.5.4 (2015-12-18)
  - Some functions accept slightly more general input. For example, it
    is now in some cases acceptable to use finite field elements like
    Z(4)^0 (which is equal, but not identical, to Z(2)^0) as input when
    cvec is expecting elements of GF(2).
  - Fix EmptySemiEchelonBasis: its documentation said it expects a
    sample vector, but the implementation expected a sample matrix. Now
    both are possible, and both are documented.
  - Fix some errors in the manual.
  - Move website to http://gap-packages.github.io/cvec/
  - Stop using GenericHashFunc_C (formerly provided by the orb package)

2.5.3 (2014-11-08)
  - Tweak the build system to avoid potential issues when the source
    code gets repackages by third parties (e.g. by the GAP team, when
    creating their big package archives)

2.5.2 (2014-09-24)
  - Fix some minor issues in the documentation
  - Remove some leftover GAP 4.4 compatibility code
  - Update Max Neunhoeffer's contact data
  - Use AutoDoc to generate title page of the manual from PackageInfo.g
  - Add Max Horn as a maintainer

2.5.1 (2014-04-04)
  - Previous release had an invalid release date in
    PackageInfo.g (2014-04-31 instead of 2014-03-31).
  - Merge HISTORY and CHANGES files.

2.5 (2014-03-31)
  - Add methods for \^ for cvecs and Frobenius automorphisms
  - Add methods for \^ for cmats and Frobenius automorphisms
  - Fix a bug in unpickling semi echelonised bases
  - Update and tweak the package build system
  - Move package repository and website to GitHub

2.4 (2012-09-14)
  - Fix an ELMS_LIST bug for CMats reported by Felix Noeske

2.3 (2012-07-13)
  - Fix CMat bug: used to not ShallowCopy its first argument
  - Fixed DuplicateFreeList and Filtered bug for CMats.

2.2 (2012-05-28)
  - Compile manual against GAP 4.5.4.
  - Add script "configure_like_gap".
  - Adjust README and installation instructions.

2.1 (2012-02-03)
  - Fix a serious bug in MinimalPolynomialMC for matrices over finite
    fields. WARNING: Wrong results were produced by version 2.0!

2.0 (2012-01-31)
  - Improve implementation of MinimalPolynomialOfMatrixMC and use it
    as a method for MinimalPolynomial for row list matrices.
    Document it for the first time.
  - Completely change build setup to use automake/autoconf/libtool as
    done in the IO package.

1.8 (2011-03-24)
  - Slightly improved minimal polynomial routine by suggestion of
    referee.
  - Use C-Hashfunction of orb 3.0 if it is there.
  - Many optimisations for GF2
  - Faster ScalarProduct for GF2
  - Fast EntryOfMatrixProduct for prime fields
  - Fixed configure script to work with GAP 4.5 without breaking 4.4.
  - Fixed a bug with zero length matrices introduced after 1.7
  - Added a Length method for semi echelonised bases
  - Documented the Vectors operation for semi echelonised bases
  - Put revision 402 as version 1.8 on the web

1.7 (2007-09-26)
  - Introduce filters for cmats and cvecs to distinguish cases of q=p
    and q <= MAXSIZE_GF_INTERNAL.
  - Kernel method for cmat row access
  - Operation ScalarProductsRows with kernel method for prime fields
    used for condensation
  - Fixed infinite recursion in matrix multiplication because of
    Winograd method.
  - Fixed problem of polynomial factorisation over "DefaultField"
    instead of BaseDomain in Minimal and Characteristic polynomial
    routines.
  - New method for SumIntersectionMat for cmats. 
  - Use constructors for new objects in new interface, implement them.
  - Change all addresses to St Andrews.
  - Require GAP >= 4.4.9
  - Tune vector matrices interface and polish implementation.
  - Added a pickling/unpickling method for semi echelon basis.
  - Quite a few little changes to the vector/matrix interface
    after discussions in Braunschweig.
  - Make TraceMat compatible with GAP4R4.
  - Added matobjplist.g* from the dev library for compatibility with 4r4.
  - Put revision 324 as version 1.7 on the web.

1.6 (2007-04-02)
  - Improve MinimalPolynomialMC, change calling convention.
  - Fixed a dangerous bug in matrix multiplication using the
    memory-saving Winograd trick.
  - Better Value method for CMats (do nothing a human would not also do)
  - Put AddMatrix and MultMatrix into interface and implement it for cmats
  - Put revision 298 as version 1.6 on the web.

1.5 (2007-02-07)
  - Improve usage of grease, in particular for level 1 for 2 < q < 257
  - Improve FILL_GREASE_TAB by using trick coming from analysis: we only
    have to multiply a few times by the primitive root, not by powers of it.
  - Fixed a serious bug in SemiEchelonBasisMutableTX which occurred if
    the first row of the matrix was zero.
  - New operation CompanionMatrix (not yet documented)
  - GreaseCalibration for different machines (not yet documented)
  - Winograd trick for multiplication, corresponding calibration function
  - Put copyright and licensing information in all files.
  - Add version number to archive name, fixed some things in PackageInfo.g.
  - Put revision 280 as version 1.5 on the web.

1.4 (2006-11-29)
  - Adjust to new buffered I/O.
  - New operation ChangeBaseDomain with methods for cvecs and cmats.
  - Put revision 212 as version 1.4 on the web.

1.3 (2006-09-01)
  - change nearly everything to the new vector/matrix interface
  - implement new Intersection functionality no longer using Zassenhaus
  - allow for more than one package with a C-part
  - added test/polynomials.g to check characteristic and minimal polynomials
  - added Memory for cvecs and cmats and to interface

1.2 (2006-08-18)
  - ?

1.01 (2006-06-08)
  - ?

1.0 (2005-06-28)
  - Initial release
