2005-07-19  Martin Maechler  <maechler@stat.math.ethz.ch>

	* DESCRIPTION (Version): 1.2-2

	* R/fracdiff.R (fracdiff): use .C(), no longer .Fortran()
	* src/fdcore.c (fracdf):   dito

	* Calling: update

	* tests/sim-ex.R: if(FALSE) library(*, lib="..MM..")

2005-07-06  Martin Maechler  <maechler@stat.math.ethz.ch>

	* src/fdhess.c (hesspq_): move "inline" declarations to beginning
	of loop: against warning "ISO C89 forbids mixed declarations and code"

2005-07-02  Martin Maechler  <maechler@stat.math.ethz.ch>

	* DESCRIPTION (Version): 1.2-1

	* src/fdsim.c (fdsim): finally found "off by 1" indexing bug
	  {introduced only in 1.1-2, two weeks ago} which accessed s[0]
	  and hence sometimes gave huge garbage initially.

	* tests/sim-ex.R: new: for fracdiff.sim() bug search
		now also a speed test.


2005-06-30  Martin Maechler  <maechler@stat.math.ethz.ch>

	* R/fracdiff.R (fracdiff.sim): add 'n.start', 'rand.gen', etc;
	  similar as in arima.sim.
	 NOTA BENE: changes the default fracdiff.sim() result as soon as
	 ---------  p + q >= 1 !

2005-06-29  Martin Maechler  <maechler@stat.math.ethz.ch>

	* DESCRIPTION (Version): 1.2-0	{never released}

	* src/fdcore.c: using FD_EXTERNAL and including all the
	``common block'' declarations:

	* src/mach_comm.h: all these are new, and included by
	* src/maux_comm.h: the *.c files that need them.
	* src/tols_comm.h:
	* src/gamm_comm.h:
	* src/hess_comm.h:

	* README: added several general notes

	* R/fracdiff.R (fracdiff): .C("fdhpq"): 'x' is not neeeded

	* src/fdcore.c (fdcom): move  fdcom() {Common Block Initialization}
	* src/fdhess.c (fdcom): from fdhess.c to fdcore.c

2005-06-17  Martin Maechler  <maechler@stat.math.ethz.ch>

	* src/Makevars : drop the non-portable "-O3 -Wall" flags


2005-06-17  Martin Maechler  <maechler@stat.math.ethz.ch>

	* DESCRIPTION (Version): 1.1-2
	* DESCRIPTION (Date): 2004-10-02 --- never released ---> now to CRAN

	* tests/ex-Vinod.R: add the example that failed (memory error)

2004-10-02  Martin Maechler  <maechler@stat.math.ethz.ch>

	* R/fracdiff.R (fracdiff): new integer work array passed to C.

	* src/fdcore.c (fracdf_): etc: new  iw[] integer work array.

	* README: explain the "cast" bug in pqopt_()'s call to lmder1_()...

	* src/fracdiff.h: new for common declarations

2004-09-18  Martin Maechler  <maechler@stat.math.ethz.ch>

	* src/fdcore.c: now translated from fortran
	* src/fdgam.c:	using f2c, my "f2c-clean"
	* src/fdhess.c: and lots of manual cleaning.
	* src/fdmin.c:
	* src/fdsim.c:

2004-04-29  Martin Maechler  <maechler@stat.math.ethz.ch>

	* R/fracdiff.R (fracdiff): make 'lenw' compatible to check in
	src/fdcore.f; {also make "info = 1" message more informative}

	* src/fdcore.f (fracdf): in the case of too small workspace,
	return the desired size to R.

2004-01-12  Martin Maechler  <maechler@stat.math.ethz.ch>

	* DESCRIPTION (Version): 1.1-1

	* tests/ex.R: changed version; less precision - portable?
	* tests/ex.Rout.save:

2004-01-07  Martin Maechler  <maechler@stat.math.ethz.ch>

	* man/fracdiff.Rd: explain `h' argument and its default;
	    now list all components in \value{};  also 'dtol':
	* R/fracdiff.R: dtol < 0 now does work as the docs have always said.

	* src/Makevars (PKG_LIBS): new;	 use BLAS_LIBS.

2003-12-29  Martin Maechler  <maechler@stat.math.ethz.ch>

	* DESCRIPTION (Version): 1.1-0 -> released to CRAN

	* tests/ex.R: new - first regression tests

	* R/fracdiff.R (fracdiff): found *the* bug!! : use result$w !

	* src/fdmin.f:	declared all; checked with "implicit none"
	* src/fdhess.f: <ditto>
	* src/fdcore.f: <ditto>

	* src/fdgam.f (d9lgmc): fixed typo: s/d9gmlc/d9lgmc/
				checked all with "implicit none"

	* DESCRIPTION (Maintainer): Martin Maechler (was ORPHANED)
	* DESCRIPTION (License): GPL

