Fri Feb 14 16:53:42 CET 2003 * hannes: bugfix
could not factorize x2+xy+y2 in Fp(a)[x,y], a2+a+1=0
added tests in factorize(f,a) (isPurePoly(f,a))
added sort option to FpFactorizeUnivariateCZ and to NTL factor routines
(NTLconvert.cc cf_factor.cc fac_cantzass.cc)
Thu Nov 7 11:29:33 MET 2002 Hans Schoenemann <hannes@mathematik.uni-kl.de>
* interface to NTL
* many fixes
Wed Aug 22 16:25:16 MET DST 2001 Hans Schoenemann <hannes@mathematik.uni-kl.de>
* Version 1.3c: some utility functions for libfac
* search for main var in factorize (experimental)
2000-09-02 Olaf Bachmann <obachman@mathematik.uni-kl.de>
* configure.in, etc: support for omalloc
2000-08-04 Olaf Bachmann <obachman@mathematik.uni-kl.de>
* ffops.cc: initialize ff_prime to 0, so that first call to
ff_setprime always initializes ff_invtab.
Tue Feb 16 19:00:03 1999 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* GNUmakefile.in (basefactorysrc): `imm.cc' added
(ftestm4src): `divrem.m4' added
(distfiles): `factory.cfg' added
* factory.cfg: new file
* GNUmakefile.in (.PHONY): dependencies `doxy', `doxyclean' added
(doxy, doxyclean): new targets
(mostlyclean): dependency `doxyclean' added
* GNUmakefile.in (doxysrc, doxyincl): new variables
(doxyfiles): new variable
* GNUmakefile.in (ftestm4src): `sqrfree.m4' added
* GNUmakefile.in (ftestsrc, ftestincl): ntl-stuff added
(ftestm4src): `gcd.ntl.m4', `revert.m4' added
Wed Sep 23 19:05:47 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* fac_ezgcd.cc (ezgcd_specialcase): bug fix (???). See comments
in the file for more information.
Tue Jul 7 18:34:39 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* GNUmakefile.in (ftestm4src): `extgcd.m4' added to distribution
Wed Jul 1 10:57:53 1998 <pohl@FUECHSE>
* int_intdiv.cc (modulocoeff): extra call added for `mpz_mod_ui()'
with void return value. Wrapped by #define `__MWERKS__'.
Tue Jun 30 16:38:41 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_algorithm.cc: doc fixes
* cf_algorithm.cc (divides): bug fix. Works for zero f and g.
More heuristic added.
* cf_algorithm.cc (psr, psq, psqr): bug fix. Work for `x' with
level lower than `level(f)' or `level(g)'.
* cf_algorithm.h: doc fixes
* ftmpl_inst.cc (tmax, tmin(Var, Var)): instantiations added
* GNUmakefile.in (ftestm4src): `divides.m4' added to distribution
Mon Jun 29 12:41:40 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* templates/ftmpl_functions.h: doc fixes
* GNUmakefile.in (basefactorysrc): `cf_inline.cc' and
`cf_iter_inline.cc' added to distribution
* canonicalform.h (class CanonicalForm): friend declarations for
`operator -', `operator +', `operator *', `operator /',
`operator %', `div', `mod' replaced by ordinary declarations
* canonicalform.cc, canonicalform.h (CanonicalForm,
~CanonicalForm, isOne, isZero, operator =): definitions of
methods moved to `cf_inline.cc'. Declarations marked as
`CF_INLINE' or `CF_NO_INLINE', resp.
(operator -, operator +, operator *, operator /, operator %,
div, mod): definitions of operators moved to `cf_inline.cc'.
Declarations marked as `CF_INLINE' or `CF_NO_INLINE', resp.
* canonicalform.h (CF_INLINE, CF_NO_INLINE): new #defines.
`cf_iter_inline.cc' (conditionally) #included.
* cf_iter.cc, cf_iter.h (operator++, coeff, hasTerms, exp):
definitions moved to `cf_iter_inline.cc'. Declarations marked
as `CF_INLINE' or `CF_NO_INLINE', resp.
* cf_iter.h (CF_INLINE, CF_NO_INLINE): new #defines.
`cf_iter_inline.cc' (conditionally) #included.
* configure.in (enable_cf_inline): new `--enable'-option
* config.h.in (CF_USE_INLINE): new #define
* cf_inline.cc (CanonicalForm(Variable, int))): assertion added
* cf_inline.cc (isOne, isZero, operator -(CF)): order of immediate
checks changed
* cf_inline.cc: new file
* cf_iter_inline.cc (operator ++, coeff, exp): assertions fixed
* cf_iter_inline.cc: new file
Fri Jun 26 12:24:34 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* int_cf.cc, int_cf.h (isOne, isZero): pure virtual methods
implemented. Declarations adapted.
* int_rat.cc, int_int.cc, int_rat.h, int_int.h, int_poly.h
(isZero, isOne): methods removed. Declarations adapted.
* int_pp.cc (isOne, isZero): slightly speeded up. Doc fixes.
* imm.h (imm_isone, imm_isone_p, imm_isone_gf): doc fixes
* imm.h (imm_iszero, imm_iszero_p, imm_iszero_gf): doc fixes
* int_poly.cc, int_pp.cc, int_rat.cc (neg): doc fixes
* int_int.cc (neg): doc fixes. Slightly speeded up.
* imm.h (imm_neg, imm_neg_p, imm_neg_gf): doc fixes
Mon Jun 15 16:04:26 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* configure.in: doc fixes
* configure.in: explicit path name generated slightly rewritten to
work on HP's. Absolute path names are generated already by
Makefile to keep `configure.in' more readable.
Fri Jun 12 16:28:48 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* int_intdiv.cc: `canonicalform.h' included
* gfops.cc, int_int.cc, readcf.y: order of includes fixed
* timing.h: bug fix. Test on `WINNT' negated.
Tue Jun 9 11:52:45 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* configure.in (ac_save_path): renamed to `$save_path'. All
references changed
(expl_datadir, expl_gftabledir): renamed to `$explicit_datadir',
`$explicit_gftabledir', resp. All references changed.
* configure.in: checks for gmp.h and libgmp.a search in
`$includedir' and `$libdir', resp.
* configure.in: new section added to get a safer explicit path
expansion
* configure.in: `$gftabledir' is not longer handled in a special
way for Singular
Mon Jun 8 17:26:55 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* singext.h: doc fix
* gfops.cc, singext.h (cf_get_table): use `feFopen()' to load
GF(q) tables if built with Singular. Declaration added to
`singext.h' (by Hannes).
* configure.in: calls to `AC_PATH_PROG' ``fixed''. Necessary
because of some old autoconf version (by Olaf)
* ftmpl_inst.cc: test on `WINNT' disabled with gcc
Wed Jun 03 16:47:54 1998 <pohl@FUECHSE>
* ftmpl_factor.h (template <class T> class Factor):
`operator == ( const Factor<T> &, const Factor<T> &)'
removed as friend and added as extra operator
Wed Jun 03 14:47:41 1998 <pohl@FUECHSE>
* timing.h, readcf.y, int_poly.cc, imm.h: test on WINNT
disabled with gcc
Mon May 11 11:36:22 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* config.h.in (INT64): unused declarations for MetroWerks removed
* imm.cc (imm_mul): wrapped by `#ifdef __MWERKS__'
* ffops.h (ff_bignorm): definition of `ff_bignorm' removed for
MetroWerks
(ff_mul): made a declaration for MetroWerks. Defined in
`ffops.cc'.
* imm.cc (imm_mul): new file for MetroWerks
* imm.h (MINIMMEDIATELL, MAXIMMEDIATELL): definition of long long
constants removed for mac
(imm_mul): made a declaration for MetroWerks. Defined in
`imm.cc'.
* fac_cantzass.cc (CantorZassenhausFactorFFGF,
CantorZassenhausFactorExt): new local variables `firstFactor',
`secondFactor' introduced to fix order of evaluation of
recursive factorizations
Mon Apr 20 16:34:14 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* .cvsignore (stamp-h): entry re-added. After all, its
necessary.
* configure.in (MKINSTALLDIRS, MAKEHEADER): instead of using an
absolute path, use relativ path and our own variables
`FACTORY_MKINSTALLDIRS' and `FACTORY_MAKEHEADER'. Do not use
cache to get the variables' values.
* GNUmakefile.in (MKINSTALLDIRS, MAKEHEADER): use the values from
`FACTORY_MKINSTALLDIRS' and `FACTORY_MAKEHEADER', resp.
Tue Apr 14 14:05:47 1998 Wilfred Pohl <pohl@mathematik.uni-kl.de>
* fac_univar.cc (UnivariateQuadraticLift): type for `log()' must
be float, hence constant `2' replaced by `2.0'
* config.h.in (INT64): new typedef for Metroworks compiler
Mon Apr 6 14:06:49 MET DST 1998 Jens Schmidt
* int_intdiv.cc:
***** merge from branch `factory-gcd' to main trunk
* ftmpl_inst.cc: instantiations for `tabs( int )' added
* fac_univar.cc (UnivariateQuadraticLift, UnivariateLinearLift):
call to `div( CF, CF )' replaced by call to `operator / ( CF, CF )'
Mon Mar 23 17:00:00 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* GNUmakefile.in (ftestm4src): references `$(ftestm4develsrc)'
instead of `$(ftestm4src)' from development `GNUmakefile' to
avoid problems when building distributions
Tue Mar 17 16:57:10 MET 1998 Jens Schmidt
* int_int.cc, int_int.h, int_poly.cc, int_poly.h, int_intdiv.cc,
* ChangeLog:
***** merge from branch `factory-gcd' to main trunk
* GNUmakefile.in:
***** merge from branch `factory-gcd' to main trunk
* timing.h, examples/application.cc:
***** merge from branch `factory-gcd' to main trunk
Tue Mar 17 14:22:56 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
**************** <<< changes in branch `factory-gcd' ****************
* int_poly.cc (dividecoeff, divremcoeff): use `divideTermList()'
to do their work
(divideTermList): new function copied from former
`divTermList()'. Declaration added.
* int_poly.cc (divTermList): uses `div()' to divide coefficients
* int_int.cc (dividesame, dividecoeff, divsame, divcoeff,
modulosame, modulocoeff, modsame, modcoeff, divremsame,
divremcoeff, divremsamet, divremcoefft): methods moved to
int_intdiv.cc. Slightly speeded up.
* int_int.h (normalizeMPI, uiNormalizeMPI, normalizeMyself,
uiNormalizeMyself): new `InternalIntger' methods. Declarations
added.
* int_int.h (class InternalInteger): static member `initialized' removed
* int_intdiv.cc: new file
* GNUmakefile.in (basefactorysrc): int_intdiv.cc added to
distribution
* timing.h (TIMING_END_AND_PRINT): added for windows
(TIMING_DEFINE_PRINTPROTO, TIMING_PRINT): macros common to all
platforms extracted from `#ifdef' directive
* timing.h (TIMING_RESET): new macro
Fri Mar 13 17:05:10 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* examples/application.cc (main): cosmetic changes
* canonicalform.cc, canonicalform.h, cf_algorithm.cc, cf_factor.cc,
* cf_map.cc, cf_map.h, fac_ezgcd.cc, fac_multivar.cc,
* fac_univar.cc, fac_util.cc, fac_util.h, ffops.cc, ftmpl_array.h,
* ftmpl_factor.h, ftmpl_functions.h, ftmpl_list.h, ftmpl_matrix.h,
* sm_sparsemod.cc, sm_util.cc, sm_util.h, ftest/ChangeLog,
* ftest/commonden.m4, ftest/norm.m4, templates/ftmpl_array.cc,
* templates/ftmpl_factor.cc, templates/ftmpl_functions.h,
* templates/ftmpl_list.cc, templates/ftmpl_matrix.cc, ChangeLog,
* GNUmakefile.in, cf_algorithm.h, cf_gcd.cc, factory.template,
* ftmpl_inst.cc:
***** merge from main trunk to branch `factory-gcd'
**************** >>> changes in branch `factory-gcd' ****************
Thu Mar 12 12:21:33 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_gcd.cc (maxnorm): function removed. All references replaced
by `maxNorm()'.
* fac_univar.cc (norm): function removed. All references replaced
by `euclideanNorm()'.
* fac_util.cc (maxCoeff): function removed. Declarations
adapted. All references replaced by `maxNorm()'.
* cf_algorithm.cc (divides): slightly speeded up.
* canonicalform.h (sign): new function
* canonicalform.cc (blcm): new function. Declaration added.
Wed Mar 11 10:39:33 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_algorithm.cc (cden, common_den): renamed to
`internalBCommonDen()' and `bCommonDen()', resp. Declarations
adapted. All references changed.
* cf_algorithm.h (abs): function `abs()' moved from
`canonicalform.h' to `cf_algorithm.h'. All referring files
include `cf_algorithm.h'.
* cf_algorithm.cc (maxNorm, euclideanNorm): new functions.
Declarations added.
* cf_algorithm.cc (internalBCommonDen): uses `blcm()' instead of
`lcm()'
* GNUmakefile.in (ftestm4src): `norm.m4' and `commonden.m4' added
to distribution
* templates/ftmpl_functions.h (tabs): new function
* ffops.cc (ff_setprime): uses `memset()' to clear inversion
table
* ftest/ChangeLog, ftest/ftest_io.cc, ftest/ftest_io.h,
* ftest/ftest_util.cc, ftest/ftest_util.h, ftest/ftest_util.m4,
* ftest/gcd.fex, ftest/runfex:
***** merge from branch `factory-gcd' to main trunk
Tue Mar 10 15:44:06 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* factory.template, ftmpl_inst.cc, templates/ftmpl_array.cc,
templates/ftmpl_factor.cc, templates/ftmpl_list.cc,
templates/ftmp_matrix.cc: #include directives for mac fixed
* ftmpl_array.h, ftmpl_factor.h, ftmpl_functions.h, ftmpl_list.h,
ftmpl_matrix.h: new files. #include of `templates/ftmpl_xxx.h'
changed to include of `ftmpl_xxx.h'.
* GNUmakefile.in (basefactoryincl): new header files added to
distribution
Fri Feb 27 10:24:22 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
**************** <<< changes in branch `factory-gcd' ****************
* var_intglobal.cc, var_intglobal.h: files removed (do not belong
to this branch!)
* cf_primetab.h: file removed
* cf_defs.h (SW_QUOTIENT, SW_FAC_USE_BIG_PRIMES): switches removed
* cf_linsys.cc (linearSystemSolve, solve, bound determinant2):
incorrect respective unused code removed
* cf_char.cc (setCharacteristic): adapted to new prime generation
scheme
* cf_char.cc (setCharacteristic): previous adaption slightly
changed. The largest small prime is now looked up only once.
* cf_gcd.cc (gcd_poly_univar0): adapted to new prime generation
scheme
* fac_ezgcd.cc (findBound): adapted to new prime generation
scheme
* cf_linsys.cc (determinant): adapted to prime generation
scheme
* sm_sparsemod.cc (internalSparsemod): adapted to new prime
generation scheme
* fac_univar.cc (choosePrimes): adapted to new prime generation
scheme
* sm_sparsemod.cc (sparsemod): debug output added
* fac_univar.cc (cf2double): function removed
* fac_univar.cc (UnivariateQuadraticLift, UnivariateLinearLift):
argument `Gamma' removed. All callers changed.
* fac_univar.cc (UnivariateQuadraticLift, UnivariateLinearLift): call to
`div( CF, CF )' replaced by call to `operator / ( CF, CF )'
Thu Feb 26 17:58:31 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* gcd_pmultiv.h (class CFPrimitiveMultivariateGCD): renamed to
`CFPrimitiveMultivariateGcd'. All references changed.
(class CFSubResGCD): renamed to `CFSubResGcd'. All references
changed.
* gcd_pmultiv.h (class CFPrimitiveMultivariateGCD): does no longer
inherit virtual from `CFPrimitiveGcd'
(class CFSubResGcd): does no longer inherit from
`CFPrimitiveUnivariateGCD'
* gcd_pmultiv.h (class CFSubResGcd): new method `clone()'
* gcd_pmultiv.h: new file
* gcd_pmultiv.cc (CFSubResGcd::execute): bug fix. Superfluous
break removed.
(execute): bug fix. Assertion fixed.
* gcd_pmultiv.cc: new file
**************** >>> changes in branch `factory-gcd' ****************
Fri Feb 20 17:41:23 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* GNUmakefile.in (installcf, installmem): `ranlib' libraries after
installing them for SUN machines
Tue Feb 10 10:47:04 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_switches.h: doc fix
Mon Feb 9 15:57:07 1998 Ruediger Stobbe <rstobbe@de.oracle.com>
* cf_factor.cc (factorize): factorize now handles characteristic
0 correct, e.g. rational coefficients and SW_RATIONAL = ON.
Mon Feb 9 11:40:41 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_primetab.cc: new file
Thu Feb 5 16:53:21 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* configure.in (enable_timing, enable_debugoutput): extra timing
and debugoutput magic removed
* GNUmakefile.in (@timingtargets@, @debouttargets@,
@debtimingtargets@): extra timing and debugoutput magic removed
Tue Feb 3 14:03:00 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* GNUmakefile.in (ftestm4src): `bextgcd.m4' and `bgcd.m4'
removed. References variable `ftestm4src' instead.
Mon Feb 2 09:55:16 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_chinese.cc (chineseRemainder( const CanonicalForm & ...)):
completely rewritten
Fri Jan 30 15:06:26 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_chinese.cc (chineseRemainder): debug output added
* cf_chinese.cc (chineseRemainder( const CFArray & ...)): slightly
speeded up
* cf_gcd.cc (iextgcd, igcd): functions removed. All references
replaced by `bextgcd()' and `bgcd()', resp.
* canonicalform.h (iextgcd): declaration removed
Thu Jan 22 10:38:34 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* int_cf.cc, int_int.cc int_rat.cc (bgcdsame, bgcdcoeff,
bextgcdsame, bextgcdcoeff): new methods. Declarations added.
* GNUmakefile.in (ftestm4src): `bgcd.m4' and `bextgcd.m4' added to
distribution
* canonicalform.h (mvar): bug fix. Declared as `Variable mvar
(...)'.
* canonicalform.cc (bgcd, bextgcd): new functions. Declarations
added.
Fri Jan 16 09:08:17 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_switches.h: doc fix
* cf_algorithm.h (divides): declaration moved from canonicalform.h
to cf_algorithm.h
* cf_switches.cc: doc fix
Tue Jan 13 14:56:42 1998 Olaf Bachmann <obachman@mathematik.uni-kl.de>
* configure.in: changed order of library checks (-lm before
-lgmp); check for atof in -lm check, to work around a bug on
some linux machines.
Tue Jan 6 11:01:23 1998 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* .cvsignore (stamp-h): entry removed
Wed Dec 17 09:56:25 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* canonicalform.cc (operator ==, operator !=): rewritten and
speeded up
(operator >): slightly speeded up
* int_cf.h (class InternalCF): `sign()' declared as const
* int_poly.cc (comparesame): doc and assertion fix
* int_int.cc (sqrt, ilog2): assertion fixes
* int_pp.cc (comparecoeff): assertion fix
* int_rat.cc (comparesame): doc fix
* int_cf.cc (InternalCF::sign): method made pure virtual.
Declaration adapted.
* int_cf.cc (taildegree): bug fix. Did not check whether CO is
zero.
* imm.h (imm_cmp): doc fix
* canonicalform.cc (CanonicalForm::gcd): method removed.
Declaration adapted.
* canonicalform.cc (degree, taildegree): one more time slightly
speeded up. Hopefully for the last time.
Fri Dec 12 10:06:33 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* imm.h (imm_cmp_p, imm_cmp_gf): comparisons fulfill law of
trichitomy
* readcf.y (yylex): parser reads generator from GF(q) correctly
Tue Dec 9 10:03:07 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* GNUmakefile.in (.PHONY): dependencies `$(srcdir)/winnt/factory.h'
and `$(srcdir)/winnt/factoryconf.h' added
* winnt/factory.h: automatically created from `factory.template'
using `makeheader'
* winnt/factoryconf.h: automatically created from
`factoryconf.template' using `makeheader'
* GNUmakefile.in (dist): dependencies `$(srcdir)/winnt/factory.h'
and `$(srcdir)/winnt/factoryconf.h' added
($(srcdir)/winnt/factory.h): new target
($(srcdir)/winnt/factoryconf.h): new target
* GNUmakefile.in (distntfile): new variable
(distfiles): `$distntfiles' added to distribution
(dist): creates directory `winnt/'
(distfiles): `bin/makeheader.pl' added to distribution
* config.h.in (INT64): new define
* bin/makeheader.pl: new file
* int_poly.cc (comparecoeff): unused argument `acoeff' removed
Mon Dec 8 18:56:39 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* INSTALL, README: doc fixes for the Windows NT port
Thu Nov 27 21:00:00 1997 Ruediger Stobbe <rstobbe@de.oracle.com>
* winnt: new directory containing configuration files for
Microsoft Windows NT port using MS Visual C++ 5.0
* winnt/INSTALL.nt: new file that describes the make process on
Windows NT
* winnt/nt.mak: Makefile for Windoes NT
* winnt/config.h: preconfigured file for Windows NT
* winnt/factoryconf.h: for the first time this file is a hand-made
version of factoryconf.h for Windows NT
* winnt/factory.h: for the first time this file is a hand-made
version of factory.h for Windows NT
* config.h.in: new typedef INT64 for 64 bit integer type of various
platforms (currently 'long long int' for gcc and '__int64' for
MS VC++)
* readcf.y: malloc.h and memory.h are included in the case of WINNT
to have prototypes for alloca and memcpy
* ffops.h: now uses INT64 instead of long long int
* ffops.cc: now uses INT64 instead of long long int
* imm.h: now uses INT64 instead of long long int
* int_poly.cc: includes strstrea.h instead of strstream.h in the
case of WINNT (a typical Microsoft 8.3 problem - its a pity)
* cf_random.cc: the way const data members were used was not
conforming to the C++ standard. Now there is a version of
RandomGenerator that should run the same way on gcc as well as on
CodeWarrior (mac) and MS VC++ (nt) and every other compiler
conforming to the C++ standard
* fac_iterfor.cc: fixed assertion that refers to a variable that
was not in the scope. Wondering why gcc compiled it
* fac_sqrfree.cc: isSqrFreeZ now returns a bool, as declared in
the prototype. Again wondering about gcc's behaviour
* sm_sparsemod.cc: array declaration of varf, varg and schnitt on
on the heap instead on the stack because the array size is not
a constant. delete [] is called at a suitable place
* fac_cantzass.h: MS VC++ does not like a default value for a
parameter that is a reference (in some sense this is good behaviour)
Therefore there are now no default values for the parameter list
of FpFactorizeUnivariateCZ
* fac_cantzass.cc: No reference parameters for FpFactorizeUnivariateCZ
* cf_factor.cc: FpFactorizeUnivariateCZ is called with full parameter
list
* fac_univar.cc: FpFactorizeUnivariateCZ is called with full parameter
list
* ftmpl_inst.cc: MS VC++ does not like to explicitly instantiate
template functions - tmax and tmin are therefore not instatiated in
the case of WINNT
Anyway it is not clear if this file is needed when using MS VC++
since the handling of templates in MS VC++ done in the spirit of
AT&T's Cfront compilers which do a more automatic job
* templates/ftmpl_list.h: MS VC++ wants a forward declaration of
template class List<T> and template class ListIterator<T>. This
should be no problem for gcc
* timing.h: changes in the case of WINNT
Fri Nov 21 15:25:27 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* canonicalform.h (class CanonicalForm): inline method hasGcd()
removed.
* int_poly.cc (coeff): explicit cast to CanonicalForm added
(coeff): cosmetic change
* int_rat.cc (comparecoeff): bug fix. Did not calculate dummy
correctly.
(comparecoeff): assertion moved into `if' branch
* int_pp.cc (comparesame): assertion added
* int_int.cc (comparesame): assertion added
* int_cf.cc (coeff): explicit cast to CanonicalForm added
* int_cf.cc (comparecoeff): function made pure virtual.
Declarations adapted.
* int_cf.h (class InternalCF): new pure virtual function
comparecoeff()
* imm.h: doc fix
* GNUmakefile.in (ftestm4src): fbinops.m4 added to distribution
Thu Nov 20 16:21:51 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* imm.h: functions rereordered
Wed Nov 19 12:21:27 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* gfops.h: functions reordered
* imm.h (imm_sign): doc fix
* imm.h: functions reordered
Thu Nov 13 09:10:53 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_chinese.cc (chineseRemainder): `const CF' in arguments
changed to `const CF &'. Declaration adapted.
* configure.in: doc fix
* GNUmakefile.in (ftestall): new target
(.PHONY): dependencies `ftestall' `ftestclean' `ftestdistclean'
added
* GNUmakefile.in (%.d): dependency generation uses temporary files
(clean): removes these temporary files
* GNUmakefile.in ($(ftestexec)): dependecy `factoryconf.h' added
Wed Nov 12 18:08:51 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* configure.in: check for `GetOpt.h' removed
Thu Nov 6 16:43:17 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* configure.in: check for `libg++.a' commented out
Wed Nov 5 14:48:46 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* GNUmakefile.in (conftest): new target
($(ftestexec), gengftables.o): dependency `conftest' added
* GNUmakefile.in (gengftables.o): some dependencies removed
(gengftables): dependencies slightly changed
* configure.in: exits with error on absence of libgmp.a or libm.a
* configure.in: check for `unistd.h' added
* configure.in: check for `libg++.a' added
Wed Oct 29 12:41:04 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* configure.in: check for `ln -s' added
* configure.in: checks for header files necessary for the FTE
added
* GNUmakefile.in (LN_S): new variable
* GNUmakefile.in ($(ftestexec)): new target
* GNUmakefile.in (ftestsrc, ftestincl, ftestm4src, ftestdistfiles,
ftestexec): new variables
(distsrc, distincl, distfiles, dist): FTE added to distribution
(ftestclean, ftestdistclean): new targets
(clean, distclean): dependencies ftestclean and ftestdistclean,
resp., added
* GNUmakefile.in (ftest/GNUmakefile): new target
* GNUmakefile.in (gengftables): rule slightly simplified
* test_install.cc (main): calls to `operator<<()' replaced by
calls to `printf()'
* configure.in (M4FLAGS): new output variable
* configure.in (M4): check for `m4' added
* configure.in: creates `ftest/GNUmakefile'
Tue Oct 28 14:46:08 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
**************** Factory Version 1.3b released ****************
* NEWS: updated to 1.3b
* README: doc fix. Updated to 1.3b.
* INSTALL: doc fix. Updated to 1.3b.
* configure.in (factory_version): updated to 1.3b
* examples/GNUmakefile: doc fix
* configure.in (factory_configuration): new variable
* config.h.in (FACTORYCONFIGURATION): new #define
* configure.in (ac_cv_shift): set to `yes' in case of cross
compiling
* examples/application.cc (main): check on existence of GF(q)
tables removed
Mon Oct 27 15:14:19 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* examples, examples/GNUmakefile, examples/application.cc,
examples/gcd.cc, examples/factorize.cc: new files
* GNUmakefile.in (distfiles, dist): `examples/' and `examples/*'
added to distribution
* canonicalform.cc (operator >, operator <): bug fix. Calls
`comparecoeff()' in the right way.
Fri Oct 24 18:44:37 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* GNUmakefile.in (TMPLCXXFLAGS): set to `$(LIBCXXFLAGS)'
* cf_gcd.cc (gcd): bug fix. Handles polynomials with rational
coefficients correctly
Thu Oct 23 14:13:42 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* ftmpl_inst.cc: implicit instantiations of `ListItem<T>' changed
to explicit instantiations
* factory.template, factoryconf.template, gengftables.cc,
ftmpl_inst.cc: doc fix
* cf_globals.cc (factoryConfiguration): new external variable.
Declaration added.
* test_install.cc (main): prints factoryConfiguration
Wed Oct 22 15:41:30 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* configure.in: doc fixes
* canonicalform.cc (operator ()): speeded up
* cf_random.cc (factoryrandom): returns a random number without
calculating any remainder if n equals zero
* ffops.cc (ff_setprime): does not clear table of primes if the
prime number used last time equals new one
Tue Oct 21 13:24:48 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* test_install.cc: new file
* GNUmakefile.in (distsrc): distribution file test_install.cc
added
(installtest, test_install, iftmpl_inst.o, test_install.o): new
targets
(clean): removes test_install
(.PHONY): phony target installtest added
Wed Oct 15 16:31:53 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* canonicalform.cc (taildegree): slightly speeded up
* canonicalform.cc (operator []): assertion added. Bug fix (did
not work for immediates).
* canonicalform.cc (den, deriv, operator []): explicit casts to
CanonicalForm added
(sqrt): explicit calls to CFFactory::basic removed
Tue Oct 14 16:47:28 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* canonicalform.cc (mvar, degree): slightly speeded up
Fri Oct 10 10:19:07 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* gfops.cc (gf_get_table): casts to 'char *' added to avoid
warnings
* int_pp.cc: superfluous arguments removed
* int_int.cc: doc fix
* int_cf.cc: doc fix
* int_rat.cc: doc fix
* int_cf.cc (den): returns 'CFFactory::basic( 1 )' instead of
'genOne()'
* canonicalform.cc (LC, degree( v ), tailcoeff, deriv): slightly
speeded up
* canonicalform.cc (Lc): new method. Declaration added.
* int_cf.cc (Lc): new method. Declaration added.
* int_poly.cc (Lc): new method. Declaration added.
Thu Oct 9 10:20:56 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_algorithm.cc (divides): moved from canonicalform.cc to
cf_algorithm.cc. Declaration moved, too. In all files which
refer to this function #include of cf_algorithm.h added.
(divides): doc fix
* canonicalform.cc (divides): moved to cf_algorithm.cc.
Declaration moved, too.
* canonicalform.h (Lc): new function
Wed Oct 1 14:21:53 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_random.h: (factoryrandom): n defaults to zero
* cf_random.cc (factoryrandom): checks for n = zero added
Tue Sep 30 10:22:43 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* GNUmakefile.in (%.d): order of dependencies changed
Mon Sep 29 08:11:42 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* GNUmakefile.in (%.d): `factoryconf.h' added as dependency
* canonicalform.h: doc fix
* cf_ops.cc (size( CF ), size( CF, Var )): new functions.
Declarations added.
* fac_univar.cc: doc fix
* cf_chinese.cc: doc fix
* cf_resultant.cc: doc fix
* cf_gcd.cc (maxnorm): spurious variable `h' removed
* cf_gcd.cc (balance): slightly speeded up
Fri Sep 26 12:54:41 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* assert.h: prints the function now where the assrtion occured if
we are translating with gcc
Thu Sep 25 13:51:01 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_gcd.cc (chinesePoly): function removed
(gcd_poly_univar0): slightly speeded up
* cf_gcd.cc (lcm): bug fix. Returns zero now if f or g equals
zero.
* cf_gcd.cc (gcd_poly): does not call sparsemod on univariate
polynomials an longer
* cf_gcd.cc (vcontent): assertion added.
* cf_gcd.cc (content( CF, Variable )): slightly speeded up.
Assertion added.
Wed Sep 24 12:35:40 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* canonicalform.cc (deriv): speeded up
Tue Sep 23 17:24:24 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* int_int.cc: doc fix
* canonicalform.cc: doc fix
Fri Sep 12 17:25:14 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_gcd.cc (isqrt): function removed
* canonicalform.cc: doc fix
Wed Sep 10 15:36:15 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* int_cf.cc (InternalCF::sign): assertion added
* int_rat.cc (InternalRational::sign): call to mpz_cmp_si()
replaced by call to mpz_sgn()
* int_pp.cc (InternalPrimePower::sign): call to mpz_cmp_si()
replaced by call to mpz_sgn()
* int_int.cc (InternalInteger::sign): call to mpz_cmp_si()
replaced by call to mpz_sgn()
* int_pp.cc, int_int.cc, int_rat.cc, int_poly.cc, int_cf.cc,
canonicalform.cc (sign): doc fix
* imm.h (imm_sign): bug fix. imm_sign() returned 0 for sign(1)
when calculating in GF(q).
* imm.h (imm_sign): bug bug fix
* gfops.h (gf_sign): doc fix
* configure.in (expl_datadir): initialization fixed. If
$expl_datadir is relative we prepend the current working
directory now.
* cf_gcd.cc (isqrt): function removed
Tue Sep 9 09:06:17 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* README: doc fix
* INSTALL: doc fix
* cf_globals.h: doc fix
* cf_globals.cc: #include fix, doc fix
* cf_primes.cc, fac_cantzass.cc, int_pp.cc, int_rat.cc: #include
fix
* cf_factor.cc: #include fix
(factorize): cf_glob_switches.isOn() replaced by isOn() for
sake of uniformity
* cf_ops.cc (swapvar_between): arguments to operator * swapped
* cf_ops.cc (swapvar_between, swapvar_rec): functions reordered
and declarations removed
* cf_ops.cc (swapvar_between1, swapvar_rec1, swapvar1): new
experimental functions, to be removed in future
* cf_chinese.cc: doc fix
* fac_multihensel.cc: #include fix
* fac_berlekamp.cc: #include fix
* cf_util.cc: #include fix
* ffops.h: #include fix
Mon Sep 8 12:39:09 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_algorithm.cc (psr, psq, psqr): assertions added
* cf_algorithm.cc (psq, psqr): return correct values now if deg(f)
< deg(g)
(psr, psq, psqr): return correct values now if f = 0
* cf_algorithm.cc (cden): recursion terminates now for f in base
domain instead for f in coefficient domain
* cf_map.h: doc fix
* cf_map.cc: doc fix
* cf_algorithm.h: doc fix
* cf_switches.h: doc fix
* cf_switches.cc: doc fix
Thu Sep 4 10:19:25 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_ops.cc: (replacevar, replacevar_between): new functions
* canonicalform.h (replacevar): new declaration
* cf_algorithm.cc: doc fix
* cf_resultant.cc: doc fix
* cf_algorithm.cc: new file
(psr, psq, psqr, cden, common_den): functions moved from
cf_ops.cc to cf_algorithm.cc
* cf_ops.cc (psr, psq, psqr, cden, common_den): functions moved
to cf_algorithm.cc
* cf_algorithm.h: declarations of functions from cf_algorithm.cc
moved from canonicalform.h to cf_algorithm.h
* canonicalform.h: declarations of functions from cf_algorithm.cc
moved to cf_algorithm.h. In all files which refer to these
functions #include of cf_algorithm.h added.
* GNUmakefile.in (basefactorysrc): cf_algorithm.cc added to
distribution
* cf_ops.cc (apply): bug fix. In case f.inCoeffDomain() result is
now initialized to f.
Mon Sep 1 10:47:41 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* canonicalform.h (resultant): declaration moved to cf_algorithm.h
* cf_algorithm.h (resultant): declaration moved from
canonicalform.h to cf_algorithm.h
* cf_ops.cc (resultant): function moved to cf_resultant.cc
* cf_resultant.cc (trivialResultant): new function
* cf_resultant.cc (resultant): moved from cf_ops.cc to
cf_resultant.cc. Completely rewritten.
* cf_ops.cc (psr): returns f now if deg(f) < deg(g)
* cf_resultant.cc (subResChain): new function
* cf_algorithm.h (subResChain): new declaration
* cf_factor.cc (factorize( CanonicalFormn, bool ), sqrFree):
default argument initializer from definition removed
Fri Aug 29 09:10:28 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* canonicalform.cc (degree( Variable )): slightly speeded up
(deriv( Variable )): slightly speeded up
* canonicalform.h (deriv): doc fix
* canonicalform.cc (deriv( Variable )): assertion added
(deriv( Variable )): bug fix
* cf_algorithm.h: declarations of functions from cf_linsys.cc
moved from canonicalform.h to cf_algorithm.h
* canonicalform.h: declarations of functions from cf_linsys.cc
moved to cf_algorithm.h
* cf_factor.cc (sqrFree( CF )): removed
(sqrFree( CF, bool )): default argument initializer added
(factorize( CF, bool )): default argument initializer added
* cf_algorithm.h (sqrFree( CF )): declaration removed
(sqrFree( CF, bool )): default argument initializer added to
declaration
(factorize( CF, bool )): default argument initializer added to
declaration
* GNUmakefile.in (basefactorysrc): cf_resultant.cc added to
distribution
* cf_algorithm.h: declarations from cf_factor.h moved to
cf_algorithm.h
* cf_factor.h: declarations moved to cf_algorithm.h. cf_factor.h
removed. All #include statements changed.
* GNUmakefile.in (basefactoryincl): cf_factor.h removed from
distribution
* canonicalform.h: declarations of functions from cf_factor.cc
removed. In all files which refer to these functions #include
of cf_algorithm.h added.
* cf_algorithm.h: new file
* cf_algorithm.h: declarations from cf_chinese.h moved to
cf_algorithm.h
* cf_chinese.h: declarations moved to cf_algorithm.h.
cf_chinese.h removed. All #include statements changed.
* GNUmakefile.in (basefactoryincl): cf_chinese.h removed from
distribution. cf_algorithm.h added to distribution.
Thu Aug 28 09:21:15 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_map.h: superfluous '#include "templates/ftmpl_functions.h"'
removed
Mon Aug 4 16:57:27 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* canonicalform.cc (sqrt): doc fix
Fri Aug 1 11:06:58 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_map.cc (operator << ( ostream& s, const CFMap & m )):
rewritten
* configure.in: checks for compiler characteristic are done with C
instead of C++
Thu Jul 31 14:59:07 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* fac_univar.cc (kBound): bound fixed
* canonicalform.h (sqrt): doc fix
* canonicalform.cc (sqrt): immediate case completely rewritten
Wed Jul 30 09:40:08 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_linsys.cc (bound): off by one errors fixed
(linearSystemSolve): off by one errors fixed
* cf_linsys.cc (bound): debug output added
* cf_linsys.cc (bound): returns now bound + 1
* canonicalform.cc (sqrt): sqrt(0) works now
(sqrt): assertion added
(sqrt): 'int h' changed to 'long long h'. Casts added.
* cf_defs.h (SW_RATIONAL, SW_QUOTIENT, SW_SYMMETRIC_FF,
SW_BERLEKAMP, SW_FAC_USE_BIG_PRIMES, SW_FAC_QUADRATICLIFT,
SW_USE_EZGCD, SW_USE_SPARSEMOD): #defines replaced by 'const
int's
* canonicalform.cc (fillVarsRec, getNumVars, getVars): moved to
cf_ops.cc
* cf_ops.cc (fillVarsRec, getNumVars, getVars): moved from
canonicalform.cc to cf_ops.cc
* cf_ops.cc: doc fix
Thu Jul 24 12:44:46 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* GNUmakefile.in (distfiles): bin/folding.el and bin/fold-docu.el
added
Wed Jul 23 09:45:35 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_map.h (class CFMap): references to List<CanonicalForm>
replaced by CFList
* cf_map.cc (CFMap( CFList )): 'P.append()' replaced by 'P.insert()'
* cf_chinese.cc (chineseRemainder( CF, ... )): operator % replaced
by function mod()
* README: doc fix
Tue Jul 22 11:04:30 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_map.cc (CFList, CFListIterator, MPListIterator): references to
List<CanonicalForm>, ListIterator<CanonicalForm>, and
ListIterator<MapPair> replaced by their respective typedefs
* canonicalform.cc (initCanonicalForm): initialization of switches
removed since they are initialized by constructor
* bin/folding.el: new file
* bin/fold-docu.el: new file
* bin/fold-docu.el (fold-docu): works now for all fold marks
* README: doc fix
* cf_switches.h (CFSwitchesMax): new constant
* cf_switches.h (class CFSwitches): number of switches replaced by
references to 'const int CFSwitchesMax'
* cf_switches.cc (CFSwitches): number of switches replaced by
references to 'const int CFSwitchesMax'
**************** Factory Version 1.3a released ****************
* configure.in: doc fix
* INSTALL: doc fix
* NEWS: doc fix
* README: doc fix
Mon Jul 21 12:58:04 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* sm_sparsemod.cc (internalSparsemod): sparsemod() renamed to
internalSparsemod()
(sparsemod): new function which is a simple wrapper around
internalSparsemod()
* GNUmakefile.in (distfiles): NEWS and INSTALL added to
distribution
* configure.in (factory_version): version number incremented
* INSTALL: version number incremented
* README: version number incremented
* configure.in: doc fix
* GNUmakefile.in (dist): changed suffix of distribution suffix
(distclean): changed suffix of distribution suffix
* NEWS: doc fix
* INSTALL: doc fix
* INSTALL: new file
* README: doc fix
* GNUmakefile.in (installgftables): bug fixes undone
(installgftables): completely rewritten
Sat Jul 19 08:57:34 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* GNUmakefile.in (installgftables): bug fix
* GNUmakefile.in (dist): output is more beautyful now
* cf_gcd.cc (gcd_poly): call to sparsemod() added
* cf_ops.cc (resultant): assertion added
(resultant): new variable flipFactor
(resultant): handling of trivial cases fixed
(resultant): some other minor bug fixes
* cf_ops.cc (resultant): doc fix
* README: doc fix
* NEWS: doc fix
* GNUmakefile.in (uninstallgftables): brute force ('rm -rf')
replaced by a safer variant
Thu Jul 17 11:05:04 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* sm_util.cc: doc fix
(interpol): doc fix
* sm_util.h (REArray): bug fix
* sm_sparsemod.cc: new file
* GNUmakefile.in (basefactoryincl): files sm_util.h and
sm_sparsemod.h added
(basefactorysrc): files sm_util.cc and sm_sparsemod.cc added
* NEWS: new file
Wed Jul 16 11:35:06 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* .cvsignore (config.cache): new entry
* canonicalform.cc (CanonicalForm::ilog2): new method
* canonicalform.h (class CanonicalForm): new declaration of ilog2
(ilog2): new function
* int_cf.cc (InternalCF::ilog2): new method
* int_cf.h (class InternalCF): new declaration of ilog2
* int_int.cc (InternalInteger::ilog2): new method
* int_int.h (class InternalInteger): new declaration of ilog2
* fac_univar.cc (norm, dnorm): function dnorm replaced by norm
(kBound): uses ilog2() and norm() now
Tue Jul 15 14:47:53 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* sm_util.h (REArray): new typedef. All references to
Array<REvaluation> changed to REArray.
* sm_sparsemod.h: new file
Mon Jul 14 10:00:59 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* sm_util.h: doc fix
* sm_util.h (dinterpol): bug fix of declaration
* sm_util.cc: doc fix
* sm_util.cc: new file
* sm_util.h: new file
* int_cf.h (OBJDEL): spurious macro OBJDEL removed
* README: new file
* GNUmakefile.in (distfiles): README added to distribution
* canonicalform.cc (initCanonicalForm): initialization of
SW_USE_SPARSEMOD added
* canonicalform.cc (CanonicalForm::degree( Variable )): doc fix
* configure.in: doc fix
* configure.in (libmem, headerfactory, factoryconf): output
variables removed
* configure.in (templatedir): initialization fixed
Fri Jul 11 09:48:26 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* canonicalform.cc (CanonicalForm::degree): doc fix
* GNUmakefile.in: dependencies checked
* GNUmakefile.in: doc fix
* GNUmakefile.in (.PHONY): phony targets updated
* GNUmakefile.in (clean): core and *~ will not be removed any
longer
* GNUmakefile.in (depend): dependencies config.h and factoryconf.h
removed
* GNUmakefile.in (dist): 'chmod 777' removed
* GNUmakefile.in (gengftables): dependency on 'all' replaced by
dependency on $(libfactory) and libcfmem.a
* GNUmakefile.in (gengftables.o): dependency cf_util.h removed
(gengftables.o): dependencies factoryconf.h and config.h added
* GNUmakefile.in (installcf, installgftables): bug fix
(uninstallcf): bug fix
* GNUmakefile.in (libmem, headerfactory, factoryconf): variables
removed. All references changed (to 'libcfmem.a', 'factory.h',
and 'factoryconf.h', resp.).
* GNUmakefile.in (realmaintainer-clean): target removed
* GNUmakefile.in (templatesrc, templateincl): file
ftmpl_functions.h moved from $(templateincl) to $(templatesrc)
* GNUmakefile.in (TESTLDFLAGS): flag '-L.' moved to target
gengftables
(gengftables): '-L.' moved from $(TESTLDFLAGS) to command
* GNUmakefile.in (TESTCXXFLAGS): new variable
(gengftables.o): references to $(LIBCXXFLAGS) replaced by
references to $(TESTCXXFLAGS)
Tue Jul 1 14:26:29 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* GNUmakefile.in (dist): use ln instead of cp to build
distribution
* canonicalform.cc (initCanonicalForm, mmInit): initialize memory
managment if compiled with Singular. New declaration of
mmInit().
* cf_random.cc (RandomGenerator): definitions of 'const long int'
in class for mac removed. Instead, they are initialized in the
constructor.
* cf_random.h (factoryrandom, factoryseed): declarations marked as
public
Mon Jun 30 15:38:08 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* GNUmakefile.in (LIBCFLAGS, LIBCXXFLAGS, TMPLCXXFLAGS): option
-Wall replaced by references to $(WARNFLAGS)
(WARNFLAGS): new variable
* imm.h (MINIMMEDIATELL, MAXIMMEDIATELL): initialization for mac
added which does not like a 'LL' suffix
* templates/ftmpl_matrix.cc: include paths for mac added
* templates/ftmpl_list.cc: include paths for mac added
* templates/ftmpl_factor.cc: include paths for mac added
* templates/ftmpl_array.cc: include paths for mac added
* factory.template: include paths for mac added
* canonicalform.h: include paths for mac added
* ftmpl_inst.cc: include paths for mac added
* cf_gcd.cc: include paths for mac added
* cf_map.cc: include paths for mac added
* cf_map.h: include paths for mac added
* fac_ezgcd.cc: include paths for mac added
* configure: checked into RCS
* readcf.cc: checked into RCS
Fri Jun 27 10:33:29 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* bin/makeheader: checked into RCS
* bin/install-sh: checked into RCS
* bin/mkinstalldirs: checked into RCS
Thu Jun 26 16:06:39 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* GNUmakefile.in: doc fix
Wed Jun 25 10:02:51 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* configure.in (enable_gmp): bug fix
(CPPFLAGS, LDFLAGS): bug fix
Tue Jun 24 11:15:25 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* configure.in (with_builtingmp): command line option
--with-builtingmp removed. All references to with_builtingmp
and some additional code to deal with the command line options
--with-gmp and --with-builtin-gmp removed, too.
* configure.in (enable_assertions): disabled by default
* configure.in (enable_gmp): with_gmp and with_builtin_gmp are set
to '$(includedir)' if command line option --enable-gmp is given
* configure.in: fixes in configure's messages
* configure.in: doc fix
Mon Jun 23 12:44:50 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* config.h.in: doc fix
* factory.template: doc fix
* factoryconf.template: doc fix
Fri Jun 20 12:09:00 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* templates/: all files prefixed with 'ftmpl_'. All #include
statements changed.
* GNUmakefile.in (templatesrc, templateincl): all files prefixed
with 'ftmpl_'
* ftmpl_inst.cc, tmpl_inst.cc: tmpl_inst.cc renamed to
ftmpl_inst.cc
* GNUmakefile.in: all references to tmpl_inst replaced by
references to ftmpl_inst
Thu Jun 19 14:33:17 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* config.h.in: doc fix
* factoryconf.template: doc fix
* headers of almost all source files fixed
Thu Jun 12 13:12:49 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* cf_globals.cc: doc fix
Fri Jun 6 09:38:05 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* configure.in: some stupid warnings removed if "dangerous"
combinations of with's/enable's are specified on commandline.
Some other fixes in configure's messages.
* configure.in (enable_memdebug): code removed which sets
enable_memdebug to 'normal' since this is really never tested
* configure.in (with_gmp, with_builtingmp): new command line
options. Rest of configure.in adapted to new scheme of
specifying where gmp lives.
* configure.in: doc fix
* GNUmakefile.in (distsrc, clean): gengftables added to
distribution
(distfiles): ChangeLog added to distribution
Thu Jun 5 15:03:48 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
* configure.in (enable_debugoutput, enable_timing): dependency on
with_Singular removed so '--with-Singular --enable-debugoutput'
and '--with-Singular --enable-timing' are possible
* GNUmakefile.in (gengftables.o): check added which makes make
complain if factory was not configured in the right way
* configure.in: checks on sizeof(int) == sizeof(void *) == 4
removed
* debug.h: dependency on #define NDEBUG removed
* factoryconf.template: doc fix
* config.h.in: added to distribution as source file. config.h.in
is not created from acconfig.h any longer.
* acconfig.h: removed from distribution
* GNUmakefile.in: ($(srcdir)/config.h.in, $(srcdir)/stamp-h.in,
realmaintainerclean, dist, distfiles): changed to reflect the
changes described above
* config.h.in (INCL_CONFIG_H): new #define to prevent multiple
inclusion of config.h
* GNUmakefile.in ($(srcdir)/configure): make will not complain
about a missing autoconf any longer
* canonicalform.cc (operator>>): returns 0 again if #define
SINGULAR and #define NOSTREAMIO are switched on
Before Mon Jun 4 1997 Jens Schmidt <schmidt@mathematik.uni-kl.de>
This is a list of most of the changes I made after I took the
responsibility for factory development and before I started to
update the ChangeLog regularily. It is unlikely to be complete.
Besides minor changes at "source code level" (bug fixes, new
features) which are not really visible to the user in general
there are quite a lot of changes at "organization level" (aka
preprocessor level) and at "distribution level" (configure,
Makefile).
First, the changes at distribution level:
* bin: some useful/necessary scripts added to distribution
* install-sh: moved to bin
* ChangeLog: added to distribution
* configure.in: totally rewritten
* config.h.in, conf.h.in: conf.h.in replaced by a totally
rewritten config.h.in. Furthermore, config.h is now explicitly
included in allmost every source file of the distribution.
* factoryconf.template: new file. factoyconf.h is included instead
of config.h in all source files which are not translated but
installed (e.g. templates).
* factory.template, factory.h: factory.h replaced by
factory.template since makeheader is now included in
the distribution
* factory.template: rewritten to new makeheader syntax
* factory.template, singfactory.h: Singular stuff moved from
singfactory.h to factory.template. Wrapped by #ifdef
SINGULAR.
* singfactory.h: removed from distribution
* gengftables.cc: code for generating gftables added to
distribution
* GNUmakefile.in, Makefile.in: Makefile.in replaced by a totally
rewritten GNUmakefile.in
* INSTALL: removed from distribution since not up to date
* memman.h, mmalloc.c, mmallocs.c, mmheap.c, mmprivate.h,
mmutil.c, mmallocb.c, mmblock.c, mminit.c, mmspec.c: new memory
manager added to distribution
* mpext.cc, mpext.h, mpfactory.h: removed from distribution. No
more MP support.
* config.h.in (MULTIPROTOCOL): #define MULTIPROTOCOL removed
* int_int.h (class InternalInteger): declarations dependent on
#define MULTIPROTOCOL removed
* int_rat.h (class InternalRational): declarations dependent on
#define MULTIPROTOCOL removed
* readcf.cc: added to distribution since not all users have
bison available
Now for the changes at organization level:
If we refer to the "DEBOUT macros", the "ASSERT macros", or the
"TIMING macros" we mean one (or more) of the macros defined in
debug.h, assert.h, and timing.h, resp.
* assert.h (NOASSERT, NDEBUG): #define NDEBUG to switch off
assertions replaced by #define NOASSERT. For sake of backward
compatibility NDEBUG works, too.
* assert.h (DEBINCLEVEL, DEBDECLEVEL, DEBOUTSL, DEBOUT, DEBOUTLN):
DEBOUT macros moved to debug.h
* assert.cc: renamed to debug.cc
* debug.h, debug.cc: new files. DEBOUT macros moved from
assert.h/assert.cc to debug.h/debug.cc. All source files which
need these macros include debug.h.
* assert.h (PVIRT_VOID, PVIRT_INTCF, PVIRT_BOOL, PVIRT_INT,
PVIRT_CHARCC): PVIRT_*-macros moved from int_cf.h to assert.h
* int_cf.h (PVIRT_VOID, PVIRT_INTCF, PVIRT_BOOL, PVIRT_INT,
PVIRT_CHARCC): PVIRT_*-macros moved to assert.h
* assert.h (__ASSERT, __STICKYASSERT): the preprocessor instead of
fprintf() pastes the message into the format string now
* assert.h (__ASSERT1, ASSERT1, STICKYASSERT1): new macros
* assert.h (__WARN, WARN, STICKYWARN): new macros
* canonicalform.cc (operator>>): dependency on #define SINGULAR
removed
* cf_defs.h (INCL_CF_DEFS_H, incl_cf_defs_H): #define
incl_cf_defs_H renamed to #define INCL_CF_DEFS_H so
cf_defs.h is included only once
* cf_factor.cc (factorize): dependency on #define MULTIFACTOR
removed
* cf_factory.cc (CFFactory::basic( int, int )): error message
changed to call of ASSERT macros
* fac_util.cc: (crossprod): error message changed to call of
ASSERT macros
* int_cf.cc: several error messages changed to calls of ASSERT
macros
* int_cf.h (InternalCF::InternalCF( InternalCF& )): error message
changed to call of ASSERT macros
* int_poly.cc (InternalPoly::InternalPoly( InternalPoly& )): error
message changed to call of ASSERT macros
* templates/array.cc (Array<T>::operator[], Array<T>::operator+=):
error messages changed to calls of ASSERT macros
* templates/list.cc (List<T>::getFirst, List<T>::getLast,
ListIterator<T>::getItem): error messages changed to calls of
ASSERT macros
* cf_gcd.cc: debug output changed to calls of DEBOUT macros
* cf_linsys.cc: debug output rewritten and changed to calls of
DEBOUT macros
* fac_berlekamp.cc (QPrintFF, QPrintGF): definition and calls of
QPrintFF(), QPrintGF() wrapped by #ifdef DEBUGOUTPUT
* fac_univar.cc: debug output rewritten
(hprint): calls to hprint() wrapped by calls to #define
DEBOUTHPRINT, which definition depends on #define DEBUGOUTPUT
(instead of NDEBUG, as before). All callers changed.
* cf_linsys.cc: superfluous #define TIMING removed
* fac_distrib.cc: superfluous #undef DEBUGOUTPUT removed
* fac_ezgcd.cc: some superfluous #defines (DEBUGOUTPUT, TIMING)
and #includes removed
* fac_multihensel.cc: some superfluous #defines (DEBUGOUTPUT, TIMING)
removed
* fac_multivar.cc: some superfluous #defines (DEBUGOUTPUT, TIMING)
removed
* fac_univar.cc: some superfluous #defines (DEBUGOUTPUT, TIMING)
and #includes removed
* config.h.in (GFTABLEDIR): new #define
* gfops.cc (gf_get_table): uses #define GFTABLEDIR to get path
to the GF(q) tables
* config.h.in (MDEBUG, NOASSERT, TIMING, DEBUGOUTPUT): new
#defines
* config.h.in (NOSTREAMIO): new #define NOSTREAMIO to switch off
stream IO. Operators <<, >>, and related stuff wrapped by
#ifndef NOSTREAMIO.
* debug.cc (deb_level, level_msg, deb_inc_level):
deb_inc_level()'s output made more pretty at time of first call
* debug.h (DEBOUT, DEBOUTLN): slightly modified so they print
"objects" instead of only one "msg" and one "object". All
callers changed.
* debug.h (DEBUGOUTPUT): #define DEBUGOUTPUT is switched off now
if #define NDEBUG is switched on
* fac_univar.cc (ZFactorizeUnivariate): timing stuff changed to
calls of TIMING macros
* gmpext.h: references to #define INCL_GCDEXT_H replaced by
references to #define INCL_GMPEXT_H
* initgmp.cc (initializeGMP): dependency on #define SINGULAR
removed
* mmallocb.c (MM_TEST, TEST): references to #define TEST
replaced by references to #define MM_TEST
* readcf.h: '#include <stdio.h>' replaced by '#include <iostream.h>'
* readcf.y: dependency on #define COMEAU removed
* templates/matrix.cc (ASSERT): definition of ASSERT macros
replaced by inclusion of factoryconf.h
* timing.h (HZ): new #define to calculate ticks per second in a
more portable way. All references changed.
* timing.h (TIMING_DEFINE_PRINT, TIMING_END_AND_PRINT): stream IO
changed to std IO
* timing.h (TIMING_END_AND_PRINT): new macro
Last not least, changes at source code level:
* canonicalform.cc (mapinto): fixed mapping from int to GF(q)
* cf_factor.cc (sqrFree( CanonicalForm, boolean )): new
function which sorts output if desired
* cf_factor.cc (sqrFree( CanonicalForm )): changed to call new
sqrFree()
* canonicalform.h, cf_factor.h (sqrFree( CanonicalForm, boolean )):
new declaration
* fac_sqrfree.cc (compareFactors, sortCFFList): new functions used
by new sqrFree()
* fac_sqrfee.h (sortCFFList): new declaration
* cf_defs.h (SW_USE_SPARSEMOD): new switch
* cf_globals.cc (factoryVersion): new external
* cf_globals.h (factoryVersion): new declaration
* config.h.in (FACTORYVERSION): new #define
* factory.template: '#include cf_globals.h' added because of
version string
* cf_linsys.cc (linearSystemSolve, determinant, determinant2):
typedef int_ptr added
(linearSystemSolve, determinant, determinant2): 'new (int*)[..]'
replaced by 'new int_ptr[..]'
* templates/matrix.h (Matrix<T>::T_ptr): new typedef
* templates/matrix.cc (Matrix<T>::Matrix( int, int ),
Matrix<T>::Matrix( Matrix<T> ), Matrix<T>::operator=): 'new (T*)[..]'
replaced by 'new T_ptr[..]'
* cf_iter.h: superfluous '#include <iostream.h>' removed
* cf_factor.h: superfluous '#include <iostream.h>' removed
* cf_reval.h: superfluous '#include <iostream.h>' removed
* memutil.c: superfluous '#include <stdio.h>' removed
* mmalloc.c: superfluous '#include <stdio.h>' removed
* cf_random.cc (RandomGenerator): new class
(RanGen): new static
(factoryrandom, factoryseed): calls to stdlib random generator
replaced by methods of class RandomGenerator
* cf_switches.h (RATIONAL): superfluous #define RATIONAL removed
* int_int.cc (IntInt): superfluous #define IntInt removed
* int_pp.cc (IntPP): superfluous #define IntPP removed
* fac_univar.cc (initHG, ZFactorizeUnivariate): STICKYWARNings
added because of some strange errors which occur in these
functions
* fac_univar.cc (kBound): cast to double added in calculation of
the bound
(UnivariateQuadraticLift): cast to double added in calculation
of no_iter
* fac_univar.cc (max_fp_fac, MAX_FP_FAC): '#define MAX_FP_FAC'
replaced by 'const int max_fp_fac'. All references changed.
* fac_univar.cc (ZFactorizeUnivariate): superfluous declaration of
'CFFList G' removed. Some trivial structural changes. Test on
'D != 0' (before deleting D) removed.
* gfops.h (gf_gf2ff, gf_isff): declarations marked as public for
Singular
* factory.template: '#include "gfops.h"' added
* gfops.cc (gf_get_table): gf_table[gf_q] is guaranteed to be 0
* gfops.cc (gf_get_table): rewritten. Major changes: copies
Singulars tables if these are present instead of reading from
disk. Uses standard IO instead of stream IO to read the tables.
* singext.h (nfCharQ, nfM1, nfMinPoly, nfPlus1Table): declarations
of Singular objects added to copy Singulars GF(q) tables
* gfops.cc (gf_get_table): length of temporary buffer replaced by
'const int gf_maxbuffer'
* gfops.cc (gf_isff): rewritten
* gfops.cc (gf_maxtable, MAXTABLE): '#define MAXTABLE' replaced by
'const int gf_maxtable'. All references changed.
* gfops.cc (gf_valid_combination): 'int m' replaced by references
to gf_maxtable
* gfops.cc (intVec2CF): new function
* gfops.h (gf_mipo): new declaration
* imm.h (class InternalCF): superfluous declaration of class
InternalCF removed
* imm.h (imm_intval): convertion from GF(q) to integer added
* int_int.cc (InternalInteger::genOne): Ruedigers last fryday
afternoon fix
* variable.cc (Variable::name): casts to int added
* variable.cc (Variable::Variable( char ),
Variable::Variable( int, char ), rootOf): 'delete ...' changed
to 'delete [] ...' where necessary
