What do you want fixed in 2.4?

[EricL]

While work proceeds on the C++ version, I'm going to be spending a little bit of time on 2.4 fixing bugs which will probably culminate in a 2.4B download at some point.  I'd like to know what you want fixed.  If there is something in 2.4 that bugs you, a crashing problem perhaps or some functionality that doesn't work quite right, please post it here and I will add it to the work list.  What I'm really looking for is an indication as to how important it is to you relative to other problems in 2.4 so I can prioritize.

Note that this is not the forum for full bug reports.  You don't have to give me tons of details here (but please referance the posting in the appropriate bug reporting forum if there is one).  This is just the place to post one sentence things of the sort "If this only worked in 2.4 I would buy you a case of your favorite beer and move off of 2.37.6 forever..."

[PurpleYouko]

Go for it  :D

[Griz]

ok Eric ...
 
now that you've got it a bit more stable I'll go back to
playing with it and see what else comes up.
maybe we can explore a same new leagues ...
or rerun them for 2.4 seeing as bots act differently with
the new physics ...
so a new 'pecking order' will probably result.
plus folks have come up with new bots to enter
into the fray.

I do like the physics of 2.4 ...
much more realistic movements ...
so it will be fun to have it working as well as 2.37.6
 
fixes so far:

1 Console tweaked ... for dead bots.
2 DNA activaction windows working
3 overflow errors for venom fixed
4 Leagues ... now loads leagues and InstaWin fixed.
5 error.sim save uncommented in compiled version

??? what else?

just trying to keep track so I can document what was done ...
and how,  on the wiki.

onward

[Testlund]

Well, EricL...you should find some postings in the bug report I've made about things I think whould be very nice to have fixed, like ALL settings in the GUI stays put as I want them! ...and day/night cycle working correctly, for instance when I save and quit and then restart and load the save, the day/night cycle shouldn't start over from the beginning. It should work EXACTLY as in 2.37 versions. Grr! ....and I also whould like ALL my settings to be saved in the settings file! As it is now when I load a settings file I still have to manually change back some settings because they get reset to some default values I don't like. I want LOW brownian motion! AND...some bots sink to the bottom with no gravity! Well, I've reported about all this before in the bug reports. I hope it's still there. Seeing as the C++ version seems to take forever I think it whould be a great idea if you could try and solve the problems with 2.4  B)

[Griz]

yes.
I forgot ... the settings don't all reload correctly.
will see if I can narrow that down to some specifics.
I think I did comment before on gravity ...
that even the lowest setting is too high for the subtle differences I would like
to see. would be nice to have the range be a magnitude of 10 lower on the
low end of the scale. perhaps a scaling factor for the current range:
* 0.1, * 1, * 10 or something.

[Griz]

ah yes!
the 'king bot' thing.
would be wonderful to have a way for the user to define what criteria
are used in the selection of 'best bot'.
I experimented with uncommenting and using Num's InvestedEnergy
thing that he had set up ... and it seemed to work for me.
It calculates the energy invested in all of a bot's offspring ...
adding up their collective nrg and body.
but that is heavily weighted then to how many offspring a bot has.
what I would like to be able to do is select and scale some other
inputs from within the program ...
such as: # of mutations, # of kills, age, or #offspring/age ...
for the bot itself ... to determine it's Value for being King bot ...
and having it's DNA saved.
the # of mutations being taken into account ...
to keep from saving a bot that has not mutated from the one you
have already save 5 times before.
I think Bots worked out something on this at one time ...
or at least rambled on about it ...
I think he had some good ideas even if unable to articulate them. ;)
seems his version of 2.37, available on the FTP, did something
with it. at the time I had no clue what he was talking about ...
so will try to relocate those threads, and run his version again.
not sure if he took the time to comment his code, tho. ;)

sorry I don't have a lot of time to devote right now ...
will play with it when I can.

[Testlund]

Ok, looks like the project is starting to move again. Just found a fix for the day/night cycle that Eric have posted. Will try it out.

[Old Henk]

The league option could use some serious streamlining (I daresay rework), in my opinion.

Henk

[Numsgil]

Leagues are honestly a mess.  It would be some serious work to rework them to be cleaner.  I offer that not as discouragement but as plain fact that should be considered before starting,

[Elite]

My Symbioticus doesn't work in 2.4  :(  The 'head' just spins round like it isn't connected to a tie at all, and the bot acts kinda blind and lame, and it slides everywhere randomly, otherwise it doesn't move, and the sharing doesn't work, and it doesn't shoot.

In fact, none of Symbioticus works in 2.4

Try to see what I mean. The newest version is attached.

[EricL]

Can you upload a settings file for me?  I'm sure you know this, but the settings in 2.4 have a huge impact on the way bots behave.  Some bots suited for one environment won't appear to work at all in another (like zero momentum).  A settings file would help me figure out whether there are bugs or just version differences.

I have your bot up and running in 2.4 and it seems to do stuff but I have not taken the time to check to see if it's the right stuff...

-E

[EricL]

Okay, found this.  *.multi is never getting set in 2.4.  I still have to do some research to figure out the best fix, but this is the cause of most if not all of the disfunction of Symbioticus (and probably most MBs) in 2.4.  Symbioticus uses *.multi all through the DNA and the fact that *.multi is always zero is the reason it never shoots, aims wrong, etc....

[Elite]

That would explain it then. Good luck on fixing it.

Great work you're doing here Eric - kudos to you!  :D

Sugestion number (2)

*.thisgene doesn't seem to work properly in 2.4. Viruses that use *.thisgene to propagate behave very weirdly indeed. Can you take a look.

Default settings. A generic *.thisgene virus is attached. I helps to put the virus' energy to 30000 since they die pretty quickly othewise.

See, the virus doesn't work, however if you put him really close to a veg you'll start to see vague flashes of cyan. Most odd.

[EricL]

Thanks Elite.  I'm having fun!

Regarding the virus bot, not sure exactly what I should be seeing.  What I see in my version of 2.4 (has all my fixes) is that if the bot is close enough to another veggy (or another generic virus bot) it shoots the virus into the other bot sucessfully.  The other bot then starts executing the gene (or more than one copy of the gene, depending upon how many cycles have passed and how many times the virus has been shot into it) and it starts shooting the virus back.  The number of genes in both bots starts to rise at an ever increasing rate as does the energy they consume per cycle and at somewhere around 50 genes, they die as they exhaust their energy executing all those virus genes.

I've not played with viruses yet, so I may not understand certain nuances.  Should I be seeing something differnt?

Also, I have a buddy version with the .multi fix I can email you if you want to try it.  I also changed it so .....deep breath..... numties actually gets updated the cycle after a tie is shot, not two cycles after.  This change really makes me nervous, since it is changing the way things have worked for a long time but I think its the right behavour and it seems to test okay with the bots I've tried.  I can always back it out if soemone hates it.  All the tref sysvars still don't update until the n+2 cycle though I think I know how to "fix" that too without screwing up the order of operations for everything else.  Check out this thread for more  info on this.

-E

[Numsgil]

Commenting on the numties thing...

I just remembered something PY said a long, long time ago (before I joined :P).  bots that do *.tiepres .deltie store to get rid of tie feeders will still give away a full cycle to the tie feeding bot to feed from it.

So while it isn't necessarily a "feature", it does impact certain viable feeding and defensive strategies.

Maybe we should add some sort of ability that determines the cycle delay that it takes bots to register new ties.