by Gnther Montag 

V. 0.7.2  17.2.2005 
some fixes, crashbreakers, endless-loop-unloopers, 
trained the thread-starter in psychotherapeutic self-relief
to prevent crash from chaos-clicking....

V. 0.7.1 15.2.2005 
ADDITIONAL SOUND DRIVER ACCESS FOR ALSA API
Should be helpful if hf ran only with -n
(no-mmap) option before, as the ALSA API gives better mmap support. Many
soundcards CAN do mmap(), although of the errors in/by the (OSS) drivers.
If you want to use ALSA, change the audio device from /dev/dsp to hw:0.0 in
/etc/hf.conf.
PORTECHO 
Test program for TCP socket. 
TCP socket improved, should now really be universal,
means, not only for F6FBB, but for any program that can be accessed via
TCP. (Don't launch the missiles!) 
(The F6FBB-special code is now only active if you use port 6300.)
RE-WRITTEN HOWTO 
now structured, clarified, especially for OSS/ALSA problems, F6FBB config. 
BUGFIXES... 
Compiler error because now in hfkernel c and c++ (MT63) is
linked, the linker must now be g++, or, gcc with -lstdc++. 
The crash with hfterm hanging in an endless loop cleared by a counter for
hfterm's polls for the hfapp socket, exits at a max value of 1000. 
This helps on my old box, not on my new (with SuSE 9.2).
The crash at start is gone if I do not let hfkernel run in an xterm-e.
Some more compiler warnings cleared. 
Auto-update of ~/hfterm.rc by hf script in case of program update.
Beautified output of paccalc, which calculates pactor quality per SNR.
Fixed bug in dcf77rx, summertime and wintertime was  changed. I is really
wintertime now!

V.0.7
20.1.2005
MT63:
Pawel Jalocha`s MT63 included as new mode, can be run from hfterm,
plus the complete mt63 package as console programs mt63rx- -tx and -trx,
plus experimental tools and calibration tool ratecal1. 
SQUELCH:
A squelch extra for fsk and mt63, triggered by the fft routines in 
spectrum.c (so it works after the first call of the spectrum).
Improves rx quality, makes the mailbox worth testing also for non-arq modes!
BEACON:
Mailbox and beacon combined now possible, mailbox-beacon text.
ALSA IN PREPARATION:
Dummy files alsa.c and alsa.h prepared, for alsa driver to come.
MAILBOX:
Improved configuration on F6FBB and documentation on this in the howto.
Example file. Instruction for a binary hack to change greeting message.
BUGFIXES:
reffreq and dxf77rx: they gave the "cpumhz" in Hz instead of MHz 
and wrote that also into config file, now o.k.
With the new font carriage-returns were printed
as "hook", a workaround for this now replaces it by spaces.
Backspaces now delete backwards correctly in Pactor and also MT63.
Autotx /rx and automatic QRT functions written new, clearer, many fixes.
Better names and comments for functions in mailbox.c, 
to make it clearer for others who want to use them in other programs.
Arranged directory structure in hfkernel completely new for clarity,
Partially re-sorted and renamed headers, cleared confused include wiring. 
Removed kdrv (kernel modules) and kernel (user interface to
kernel modules) completely because obsolete (old hardware only)
and not used, can not maintain it.

V.0.6.5
15.11.04
Sorry, 0.6.4 crashed after some minutes,
bug in automatic partial save of rx file,
overflow of rx window buffer.
Beautified font in gtk windows (now fixed width).

V.0.6.4
9.11.04
Calibration programs:
reffreq now with -n option.
Tested the -n options for reffreq and dcf77rx:
With both, the (mostly needed) 
sound_corr is quite the same and could be useful,
the others vary.
Improved output of dcf77rx and reffreq,
They now write to the config file /etc/hf.conf themselves!
Unified names of the 3 correction factors.
Many and important bugfixes:
Shift is now indicated correctly in spectru m.
Fixed Password error on mailbox connect, 
hanging pactor connect on mailbox disconnect.
(Seems bug of F6FBB:
In my mailbox_test_port() function I test read, write and error 
with a select() call. When F6FBB disconnects, it gives no err (OOB) msg,
but the select says, there is something to read, but no read possible!)
Fixed ugly gtk errors removed (came from writing to not-yet-existing
gtk-texts), ugly segmentation fault on exit of hfterm.
rpm was too fat, contained the (not needed) hfkernel_...'s, removed them
from dist.
Again thanks a lot to the mailinglist friends, especially Dave Hambly! 

V.0.6.3
16.10.04
Improved and debugged the hf start-script very much, 
it uses dialog now, it checks configuration and gives info for newbie users.
Mailbox test function: Simulates remote user by writing into the
hfterm rx window.
Renamed hfkernel_user -> hfkernel
Took hfkernel_link and hfkernel_kernel out of /usr/bin
to make rpm smaller and because not experimented with at the moment.
Improved rpm, manpages now in the rpm!
Clarified doc, explained common errors in short.
Bugfixes...
Thanks a lot to the mailinglist friends for bugreports 
although I could not yet fix them all...

V.0.6.2
26.09.04
more bugfixes, e.g. in makefiles
specfiles improved, options for SuSE, Debian
new manpage man hf
startscript improved

V.0.6.1
13.09.2004
Some bugfixes
with patch by Chris Vine (thanks!):
 - characters longer than 1 line removed, gave errors with gcc >= 3.3
 - some missing includes
corrected some mistakes (hard copies of cp or ginstall) in makefiles
docs and examples now in /usr/share/hf
and - depending on distribution - in /usr/share/doc/hf
or /usr/share/doc/packages/hf.
specfile prepared in doc dir, please try rpm 
Pactor mailbox list (thanks Jost) included in doc and help
tx in color

V.0.6
30.07.2004
a lot done, so 0.6..
Main things:
- THE AUTOTX AND AUTORX FUNCTIONS!
- THE BEACON!
- THE MAILBOX!	
- The experimental no-mmap code (option -n)
works bad for ESS 1938 (sounds good, but all some seconds xruns in receive),
works quite good for old soundblaster 16, receives a bit and sends a bit in pactor
and rtty, so hopa also the other modes.
For halfduplex it is not yet implemented,
I did not try it with calibration. But hope is there.
- THE HF startscript that can do all, 
made hfkernel_user suid root, to be started simply by user.
- Configuration file /etc/hf.conf.
-h (force halfduplex) option.
- hfkernel:
xrun endless loop crash brake count, exits after 1300 xruns.
- Doc: 
you read it! simplified, put hfterm.txt and hfkernel.txt into 
the main readme, which is now called howto, simpler now. 
- Updated manpages, but will not update them regularly.
- Lots of debug. Maybe made some more bugs.
- dcf77: 
solved a little "year 2000 problem" in the output, now it's 2004, 
is this a record?
clarified output, reduced verbose levels from 8 to 2, 
now also output without -v, rotating rod, system time set function, 
built a LF converter (circuit in doc),
tested dcf77rx program which is sensitive to atmospheric noise, but good.
- RTTY: 
diddle.
- Pactor:
In pactor rx removed the workaround for bad crc, did not do good.
- hfterm: 
shortcuts with strg instead of shift, does not disturb while
writing to the tx window any more.
Descrambled status display output. 

V.0.5
24.01.2004
a lot done, so 0.5.
should have become CHRISTMAS EDITION but too late..
my wife left me already... (only for holiday, this was fun only)
hfkernel:
variable "brake" counts to "brakepoint" (70) and exits when too many xruns.
Xrun and endless loops had caused crases with some soundcards / drivers.
(mmap problem).
fixed assembler 'bugs' (l1/user/refclock.c and util/filtbench.c)
with the help of Micha, via SuSE-programming mailing list, tnx!
(before the asm's had compiled with old, not with new gcc 3.2, so
I had outcommented them in my despair....)
Special important is the first one, now I hope time correction 
with RDTSC runs with half duplex soundcards.
pactor.c: more debug output, workaround for bad crc's in memory arq 
(which had interrupted connections).
A SPECIAL NEW PACTOR-ONLY STANDBY MODE IMPROVES PACTOR MONITORING !!!
In pactor rx a workaround for bad crc, it prints instead of bad crc now with warning.
hfterm:
rx window scrolls now, and many other graphic beautifications improvements!
I think it looks better now.
changed layout, fixtext now right side.
Fixtexts: Can now also parse last log entry, name, qth, rst, logged call etc.!
File-on-demand-store mode, automatic storage of rx to prevent window overflow (and crash),
now the text freezes before storing, i hope this prevents storage of rubbish, 
which had made it appear sometimes as a binary file before.
menu option to send file other than fixtext. Beacon prepared.
Start of a fix in hfterm/src/main.c: edit_backspace. 
Now it does as if it sends backspaces in
pactor (it deletes backwards in rx window), but still does not send. 
monitor: less output, better automatic deletion, hope less crashing.
Status display: removed rubbish output in RTTY mode.
scripts:
New startscript hfsudo, because knoppix needs sth like this.
combined startscripts now begin with "xterm -e" so that you see
hfkernel running in background window on X-screen.
hft starts hfterm, it is called by the 3 hfkernel-starters.
help:
examplefiles for fixtexts, brags now in /usr/share/doc/packages/hf/hf-examplefiles.
Much more, i forgot...

V.0.4.3
11.11.2003
hfkernel:
Fixed another self-made bug in l1/user/oss.c:
hdx_driver obuf mmap, restored as it was, now it works, sorry!
(It had been one of my desparate trials in making it
run for ALSA)
scripts:
In hfkernel_user -h the option -i also appears now,
it had been there before. (invert PTT)
Improved main start script,
now the cpu Mhz rate is fetched automagically from dmesg.
hfterm:
Little bug in hfterm/src/main.c in versionstring, removed \n.
Logbook: replaced "delete all" (nonsense!!) by "archivate log"!
general:
Changed some file names, seems now clearer and shorter.
Changed directory settings, adapted to FHS.
errlog now in /var/log.
Doc now in /usr/share/doc/packages/hf.
Binaries and startscripts now in /usr/bin.
Personal directory now hf instead of hftermfiles.
Improved documents.

V.0.4.2
4.10.2003
IMPORTANT BUGFIXES !!!
fixed bugs in l1/user/oss.c:
the PROT_WRITE (see above) in mmapping the soundcard
was WRONG, it made crash in endless loop.
Found the reason for segfaults: 
it was a memcpy to a not-allocated region,
both in hdx_driver and fdx_driver.
Wonder why it had worked before with OSS!
Now it runs with ALSA 0.9.7 in my machine.

V.0.4.1
9.9.2003
begin to care for this ChangeLog
begin with three-parts-version numbers
removed startkern_pl in source directory, not used any more
l1/user/oss.c line 249: fixed bug in calculation of tx-delay
l1/user/oss.c line 337: added '| PROT_WRITE ' in mmap instruction
    hope this helps overcoming program stop for ALSA OSS emulation
    in my machine with ALSA 0.9.0 it ends up with xrun after this change
    ALSA OSS emu is at the momend very hard worked at
    see www.alsa.org for docu and mailing list

V.0.4
about March-June 2003
many little bugs I do not remember
so it compiles now also with newer compiler
in the hfterm graphic interface I brought most of the functions to work
like log, parameters storing, help.
Fixed stack overrun in monitor function by limiting contents of the GtkText
many other things
rewrote documentation, Translation German / English

V.0.3a
about 2001-2002
by Axel
many ideas and new functions for the graphic interface
thanks for inspiring me!

V.0.2a
(by Axel)
graphic refresh & many good ideas!!!
thanks!!!

V.0.1
(has been made by Tom Sailer)
the main work!
See his docs, man hfkernel! 
Thanks !!!!