2006-07-05  Steven G. Johnson  <stevenj@ab-initio>

	* harminv.1: documented phase flip with -n

	* NEWS, harminv-main.c: added -n to NEWS, flipped phase for -n

	* COPYRIGHT, NEWS, acx_blas.m4, acx_lapack.m4, check.h,
	configure.ac, harminv-int.h, harminv-main.c, harminv.1, harminv.c,
	harminv.h, sines.c: version bump

	* harminv-main.c: fixed sign of phase in output; thanks to Andrew
	Norton for the bug report

2005-12-06  Steven G. Johnson  <stevenj@ab-initio>

	* harminv-main.c, harminv.1: added -n option to flip freq sign

2005-09-02  Steven G. Johnson  <stevenj@ab-initio>

	* acx_blas.m4, acx_lapack.m4: updated

	* NEWS, check.h, configure.ac, harminv-int.h, harminv-main.c,
	harminv.1, harminv.c, harminv.h, sines.c: version, copyright year
	bump

	* harminv-main.c, harminv.1, harminv.c: switched back to old Wall
	& Neuheuser eigenvalue routine, to remove singular null space

2004-05-26  Steven G. Johnson  <stevenj@ab-initio>

	* harminv.1: whoops, 4 == 6? thanks to Gareth McCaughan

2004-05-20  Steven G. Johnson  <stevenj@ab-initio>

	* NEWS, harminv.1: whoops

	* NEWS, configure.ac: version bump

	* harminv-int.h, harminv-main.c: minor

	* README, harminv-main.c, harminv.1: use at most 300 modes by
	default

	* README: see also -d

	* NEWS: ack

	* harminv.1: ehh, let the user decide on the best filter

	* harminv.1: don't recommend -e 1000, but -a 1e-4 often works well
	it seems

	* harminv-int.h: complex<long double> seems to be problematic with
	g++ 2.95

	* NEWS, configure.ac: version bump

	* harminv.c: NPOW correction to iterative computation of z^m

	* README, configure.ac, harminv-int.h, harminv-main.c, harminv.1,
	harminv.c: use long double types, spectral density instead of nf,
	ZGGEVX option

2004-05-19  Steven G. Johnson  <stevenj@ab-initio>

	* harminv.c: minor fixes, noticed on Solaris

	* NEWS, configure.ac: bumped version

	* harminv-main.c: -F uses |f|

	* harminv-main.c, harminv.1: added -F option

	* harminv.c: fixed CHECK strings

	* harminv-main.c: recognize -A

	* harminv.c: whoops, get_omega was wrong

	* README, configure.ac, harminv-main.c, harminv.c, harminv.h:
	consistent use of freq abbreviation

	* harminv.c: stupid clog workaround

2004-05-18  Steven G. Johnson  <stevenj@ab-initio>

	* configure.ac, harminv-int.h, harminv.c: more stable calculation
	of amplitude for |u| not close to unity; cache G's for U0

	* configure.ac: --enable-debug option

	* harminv-main.c: truth in advertising

2004-05-17  Steven G. Johnson  <stevenj@ab-initio>

	* configure.ac: beta version

	* Makefile.am, configure.ac: use proper aclocal inclusion of .m4
	files

	* README: typo

	* NEWS: updated

	* harminv-main.c: SOLVE_ONCE_ONLY preproc. flag

	* sines.c: sines -R

	* harminv-main.c, sines.c: handle RE-IMi format

	* harminv.c: comment re: zdotu

	* harminv.c: removed obsolete ref. to harminv_zdotu

	* harminv.1: note that -Q uses |Q|

	* configure.ac: version bump

	* README, harminv-main.c, harminv.1, harminv.c, harminv.h: -Q,
	get_Q, documentation

	* harminv-main.c, harminv.c: pass -1/-2 for initialize/finish in
	mode_ok

	* harminv-main.c: typo

	* harminv-main.c, harminv.c: specify_omega, allow commas

	* check.h, harminv-int.h, harminv-main.c, harminv.c, harminv.h:
	added mode_ok option (currently not recommended)

	* harminv-int.h, harminv.c, harminv.h: get_amplitudes,
	get_error,get_omega

	* harminv-int.h, harminv-main.c, harminv.c: use LAPACK ZGGEV,
	corrected error calculation, thresholds, ...

	* mandelshtam/CQZ.f, mandelshtam/fdm1d.f,
	mandelshtam/fdm1d_multiwindow.f, mandelshtam/readme-multi.txt,
	mandelshtam/test.f: added multi-window code

	* mandelshtam/Makefile, mandelshtam/fd_qz99.f,
	mandelshtam/generate.f, mandelshtam/in_qz, mandelshtam/readme.txt:
	added mandelshtam code (not for distribution)

2004-05-15  Steven G. Johnson  <stevenj@ab-initio>

	* harminv-main.c, harminv.1: added SORT_Q option

	* NEWS: wrap

	* NEWS: formatting

	* NEWS: space

	* NEWS, configure.ac: version bump

	* harminv.c: require one additional c[]...forcing data beyond end
	of array to zero degraded the accuracy

	* NEWS: blah

	* NEWS: updated

	* harminv.1, COPYRIGHT: copyright year updates

	* configure.ac, harminv.c: whoops

	* Makefile.am: whoops, distribute acx_* files

	* Makefile.am, README, blas_wrap.f, harminv.c: don't bother with
	blas_wrap.f

2004-05-14  Steven G. Johnson  <stevenj@ab-initio>

	* NEWS, README: tweaks

	* ChangeLog, mkdist.sh: whoops

	* AUTHORS, ChangeLog: GNU-lly correct

	* harminv.1: minor

	* NEWS: news

	* ChangeLog: created automatically by mkdist.sh

	* check.h, harminv-int.h, harminv-main.c, harminv.c, harminv.h,
	sines.c: copyright year update

	* README: linking info

	* harminv.1: document -s

	* Makefile.am, blas_wrap.f, harminv-main.c, harminv.c, harminv.h:
	added sorting, BLAS dot products & norms, ...

	* sines.c: added -s

	* README: blah

	* README: more explanation of method

	* harminv.c: added newer reference

	* Makefile.am, README, autogen.sh, configure.ac, harminv.1,
	harminv.c, harminv.pc.in, mkdist.sh: updated comments

	* Makefile.am, configure.ac, copyright.h, harminv-int.h,
	harminv-main.c, harminv.c, harminv.h, sines.c: separate library
	and executable

2004-03-13  Steven G. Johnson  <stevenj@ab-initio>

	* configure.ac, harminv.c: minor fixes

2002-11-04  Steven G. Johnson  <stevenj@ab-initio>

	* COPYING, ChangeLog, Makefile.am, Makefile.in, NEWS, acx_blas.m4,
	acx_lapack.m4, config.guess, config.sub, configure.ac,
	configure.in, harminv-main.c, harminv.c, install-sh, sines.c:
	automake-ized

2002-03-28  Steven G. Johnson  <stevenj@ab-initio>

	* sines.c: add noise option -N

2002-03-18  Steven G. Johnson  <stevenj@ab-initio>

	* harminv-main.c: bug fix in -f parsing

2001-10-28  Steven G. Johnson  <stevenj@ab-initio>

	* configure.in: slightly updated complex tests

2000-06-06  Steven G. Johnson  <stevenj@ab-initio>

	* sines.c: bug fix: default # timesteps should be NPERIODS
	periods, not 1.

	* README: added README

2000-06-05  Steven G. Johnson  <stevenj@ab-initio>

	* harminv.1: note that inputs are whitespace-separated

	* harminv.1: many improvements

	* configure.in: cray has CGEEV, not ZGEEV

	* configure.in, harminv.c: more cray fixes (sigh)

	* configure.in, harminv.h: Cray doesn't have carg

	* sines.c: don't divide by zero, not all machines support 1/0 =
	inf

	* configure.in: check for I constant

	* harminv.c: make F77_FUNC slightly more general

	* harminv.c: whoops, ref was correct

	* harminv.c: fixed reference

	* harminv.c: disable LAPACK "optimal workspace" computation, as
	LAPACK seems to be buggy sometimes.

	* Makefile.in: 'make dist' needs to run autoheader.

	* configure.in, harminv-main.c, harminv.c: cleaned up warnings

	* sines.c: fixed compilation error under C

2000-06-04  Steven G. Johnson  <stevenj@ab-initio>

	* sines.c: explained meaning of Im[freq]

	* Makefile.in, sines.c: added sines test program

	* harminv.1: weakened description of error output.

	* harminv-main.c: match column ordering in man page

	* harminv-main.c, harminv.1: added man page

	* Makefile.in, configure.in, harminv-main.c, harminv.c, harminv.h:
	added C++ complex<double> support

	* harminv-main.c: sign-related bugfixes

	* harminv-main.c: bug fix

	* COPYING, COPYRIGHT, check.h, copyright.h, harminv-main.c,
	harminv.c, harminv.h: added copyright info

	* Makefile, Makefile.in, config.guess, config.sub, configure.in,
	harminv-main.c, harminv.h, install-sh: autoconfiscated

	* Makefile, check.h, harminv-main.c, harminv.c, harminv.h: New
	file.

	* Makefile, check.h, harminv-main.c, harminv.c, harminv.h: created
	repository

