rog-o-matic for debian testing (wheezy)

[songbird]

picks up where:

http://code.google.com/p/rogomatic/ r23

leaves off.

an updated version to run on GNU/Debian Linux
testing (wheezy in development). i could not
find any version that would compile. many
compile oddities and language troubles remain.
i almost gave up because of how bad this old
code is and the various bugs, but then again
i really did want to see it run... so here it
is.

many bugs still, but at least it does work
with a version of rogue 5.4.4 slightly
modified to fix one trouble and a few grammar
errors (also provided):

http://www.anthive.com/rog/rog.html

please, no criticism, it's summer now and i
have little time to work on this until next
winter, but would like people to tinker with
it if they like.

if you want the full history of changes get
the version that includes the git commits.

have fun, peace, etc.

songbird

[songbird]

songbird wrote:

not sure if this would work on an
ubuntu system or not, but someone can
always give it a try. :)

the project is still a learning project
for me and is going ok, slowly at times when
i'm busy with gardening or other projects,
but still ambling along in that i test
newer compiler versions (gcc-4.7) with it
and recent releases of other related code
(make, autotools, curses, lex).


...
> http://www.anthive.com/rog/rog.html

rogomatic-r1.0.0 picks up where another
project stopped. i've posted several steps
in getting to r1.0.0 and a full history in
git of even the other svn version. i'm new
enough to git that don't be surprised by
oddities in there...

i recently took some time learn a little
of GNU autotools and removed the framewerk
build environment. something i've wanted
to do for a long time now. this makes it
easier for others to compile as now they
only need a system with "sh" (for ./configure)
"make", a C compiler and "curses".

i haven't fixed any more bugs in the code
though. the most annoying bug at the moment
is the scanner for the screen codes isn't
working 100%, but it doesn't actually affect
the play. sometime when i get a chance this
year i'll redo the scanner in lex/flex.

peace, have fun, etc...


songbird

[songbird]

songbird wrote:

...

- Improved weapon handling in Rog-O-Matic to not cause
problems with cursed weapons. Previously, at times a
wield would fail part way through and then if the weapon
letter happened to be a command letter it would send that
to Rogue as a command instead of cancelling the command
once it detected the wield didn't fail. This was causing
a lot of other error messages to appear. I need make
similar changes to the armor handling (not yet done).

- Put the usleep delay feature I added a few versions
ago under the control of Rog-O-Matic's '-w' (watched) option
(makes testing go faster if it can run full blast).

- Updated both Rogue and Rog-O-Matic to use the SEED
environment variable (if set) to make it easier to replicate
runs.

Get the LATEST at:

http://www.anthive.com/rog/rog.html


- Major revision of the debugging tools/scripts. Many
rewritten, a few added, directory structure redone.

- This is not an upward compatible change, so the location
is changed too. These allow me to run rogomatic/rogue in a
replicable way and to turn on debugging tracing, and run
things in series with learning on and off. Also included
a script to compare two series runs of Rog-O-Matic/Rogue --
helps to check if the changes I'm making actually improve
things or fix the bug. Also was fun to run it comparing
no-learning to learning to see how much of a difference the
genetic algorithm actually makes. (I had a lot of fun
writing these scripts but now that they are more mature and
useful I can actually get back to fixing Rog-O-Matic and
Rogue, now that I can replicate runs that hit errors at the
deeper levels). Also, added a script to select a previous
run from where they are stored so it can be started up
again.

at:

http://www.anthive.com/rog/other/things.html


Cheers,


songbird