May 28 2002

Updated KLineakConfig to work with version 0.3 of lineakd
Fixed up gui level error reporting to use qmessagebox instead
  of only displaying errors to stdout.
Moved the parsing code into a LineakParser class.
Moved a lot of c code to c++ code.

June 2 2002

Ugly hack to allow a user to start the program without having a ~/.lineak/lineak.conf
file. On startup it writes a minimal config file w/ the first keyboard in the map from the
parsing of the .def file. Then it opens this as the default keyboard. Users won't really
notice, or care IMHO as they can then go and select their own keyboard from the list
and set the commands.

June 7 2002

KlineakConfig now controls your lineakd process. You do not and should not start
lineakd before running KlineakConfig. It also now docks into your kicker panel.
There is a right mouse button menu that allows you to start, stop and restart lineakd.
It also picks up a local lineakkb.def file in ~/.lineak/ and will add any
keyboards defined there to the configuration choices automatically. On program exit
it allows you to have the ability to keep lineakd running in the background or
to stop it. I recommend stopping it if you plan on running klineakconfig again
b/c the detection of a running lineakd process is not very precise. It essentially
greps for lineakd being run by your userid and tried to kill it.

June 15 2002

KlineakConfig now has a dialog to allow you to define your own keyboard and easy
access keys by entering the name of the key, and then pushing the key right after.
You can view, modify and delete the key mappings to your hearts delight. It does limit
itself to attempting to grab something that really is an easy access key, but this is
a best guess. Some of the parsing code has been reorganized to allow the loading of
new definition files. The PID code has not yet been updated, but that's in the works
(it uses /proc to get the pid information).

August 16 2002

Applied fixes from Evan Teran for proper c++ compliance. His email goes
something like this:
############
hey, i was trying to make an ebuild for gentoo linux for klineakconfig and i
noticed some problems with current sources, i figured i'd give ya a heads up
:)

anyway, firstly you accidentally included a config.cache file from your
computer which tends to cause mysterious errors when compiling on another
machine with alternate config...might be a good idea to remove that.

also, my system uses gcc 3.2 and it picked up some errors in source, nothing
major. | Mostely it dies on the fact that you included the default params in
both the class definition and the implementation. | like this:

----------------------------------------------
klineakconfig.h:

virtual bool reparseDef(QString parsefile = QString::null);

klineakconfig.cpp:

bool KlineakConfig::reparseDef(QString parsefile = QString::null){
----------------------------------------------

the ISO standards says it should only be in the definition like this:

----------------------------------------------
klineakconfig.h:

virtual bool reparseDef(QString parsefile = QString::null);

klineakconfig.cpp:

bool KlineakConfig::reparseDef(QString parsefilel){
----------------------------------------------

modern compilers think you are trying to redefine the functions..which is bad
:(

anyway, after some minor modifcations i got it to compile, i can send you a
patch (once i figure out how to make one) if ya like :)

love this program! | looking forward to next version :)
##########
Thanks for the fixes! I'll make sure to be ISO compliant in all future code.
(err, well, *more* compliant anyways).

October 8 2002

Fixed gcc3.2 compile problems.
Fixed compile problems for distributions that do not have qt-designer compiled with KDE widget support.
Fixed up bug which caused klineakconfig to write an empty config file.
Fixed a bug which caused klineakconfig to lock up.
Added the ability to handle quotes in the commands.
Made running kkeyled and lineakd after exit configurable in the preferences.

Fixed lots of bugs.

October 10 2002

Bugfix release. Keyboard submission did not properly attach the definition
file. Instead there was a bug which caused it to create an attachment with
only the path to the definition file instead.

In the 0.5 release I neglected to increment the version number to 0.5. Now
I've set the version number to 0.5.1