Darwinbots Forum
Code center => Bugs and fixes => Topic started by: Testlund on July 09, 2006, 03:34:47 AM
-
I'm not sure if this could be a bug. Check out the screenshot. Big cyan dots flooting around inside 3 bots. This screenshot is from my evosim.
-
Here's another one. A big cyan dot attached to a really tiny bot. This one kept reproducing dispite it allmost had no energy! This screenshot is from a sim with Animal Minimalis and Alga Minimalis.
-
Here's an attached save of the sim.
-
I couldn't find this in your sim. It only has about a dozen bots in it and tons of veggies.
A cyan dot is a virus shot impacting. That is probably what you are seeing. Shot impacts have radius 20, so if your bots are super small and you are zoomed in, they will seem large by comparison to the bot, whose radius may be smaller. Shot impact dots should only last one cycle though. Note that the point of impact where the impact circle is displayed may not appear on the edge or inside the bot. This is because the impact may have occurred early in the cycle and the bot may have moved since then. Bots are displayed where they are at the end of the cycle. Shot collisions can happen anytime during the cycle. This difference can be pronouced for small raidus bots, so shot imapct dots can oftne appear outside the bot.
-
Ok, check out this save instead. It's from my evosim. Somewhere near the upper left corner there are some bots wich have made ties together. Two of them have cyan dots in them, but there are no virus dots entering from anywhere. The cyan dots just keeps on flooting around inside the bots. Also some of the bots eye dots are located inside the bots instead of sitting at the edge of the cell wall, where they should normally be.
-
I have noticed this as well ...
and often have cyan dots attached to a bot ...
as if tied, but with no tie visible ...
moving along with it as if it were a shadow
for as long as the bot lives.
trying to click on it to see what it is ...
results in the bot being selected instead.
so it's like a blue ghost ...
seems friendly ...
I think I'll call them 'Caspers'.
they seem to 'haunt' certain bots more than others ...
as I notice them with my CW 'circling' bots all the time.
attached is a sim which seems to generate a lot of them ...
there are at least 3 here ...
bottom center, another just to the left of that one ...
and another in the upper right.
checking out your new incoming energy thang ...
seems to work just fine.
I do have a question on energy input tho ...
will do it in another post.
-
Okay, found this. When I added the new shot impact dots in 2.42.7, I missed the case where the creation of a virus shot fails because the gene being referenced is beyond the end of the DNA. In Testlund's sim, bot 37 (and presumably other bots which display similar persistant dots) is attempting, over and over, to create a virus shot using gene number 10661, which is of course, beyond it's 11 genes. The creation of the shot fails as expected, but in the process of almost creating and then failing the creation of the virus shot, an erronous virus shot impact dot is displayed at the point of shot origin, over and over each cycle, as the bot trys to create a new virus each cycle.
Any bot trying and failing to create a virus will exhibit a dot for that cycle.
Totally harmless. Fixed in 2.42.8.
Also some of the bots eye dots are located inside the bots instead of sitting at the edge of the cell wall, where they should normally be.
This is an artifact of your monitor/ graphics adapter resolution/dimensions. The edge of the bots are displayed as perfect ciricles on the screen (because the underlying VB circle command is aware of your monitor resolution and adjusts accordingly and it is used to display the bot edge) but unless you are using monitor resolution dimensions with an aspect ratio of exactly 4:3 (e.g. 800X600) the bot's eye will be displayed on the true, actual, underlying edge which will not coincide with the displayed edge (the eye will trace an ellipse). If you watch the eye, it will touch the bot edge either at the top and bottom or on the sides, depending on whether you are using short and fat or tall and skinny monitor dimensions respectivly, then go inside the bot the rest of the time, tracing an elipse. Shot impacts will exhibit the same issue. Some shots will appear to go inside the edge and yet not impact. In reality, this is just an artifact of the way the edge is displayed on monitors with aspect ratios that do not exactly match the field dimensions - the shot did not actually go inside the bot, it just looks that way. The ellipse that the eye traces is the true bot edge and shot impacts only occur if they go inside that.
Don't worry, the underlying physics of everything is fine. It's just an artifact of the way bot edges are displayed using the VB6 circle command.
Some day I may implement custom field dimensions and this will get fixed as a side effect of that, but not today.
-
Large numbers in .vshoot prevent the virus from firing. Is there an upper limit on .vshoot?
Also, sometimes virus shots aren't visible for some reason
-
Large numbers in .vshoot prevent the virus from firing. Is there an upper limit on .vshoot?
I need to spend some hours on the virus code to better understand it. The code in this area is total spagettii and probably rife with problems. In fact, I think I just caught one with how costs are applied for firing a virus....
From a quick look, syntanticly, 32000 is the legal limit in vshoot of course. Semantically, the nrg of the resulting shot (which impacts the shot's range) is clipped at 32000/20 or 1600 so using values abouve 1600 in vshoot equates to 1600.
I could find nothing related to vhsoot that would prevent the virus from firing, but if the virus is longer than 10 base pairs, then the virus shot is stored for a number of cycles equal to twice the number of base pairs in the virus. This could be what you are seeing.
Note that this is just how the code appears to work today. It may not be how it is supposed to work.
Also, sometimes virus shots aren't visible for some reason
Don't know about this one. If you have a sim which demonstrates this, I'll take a look. As above, perhaps what you are seeing is the delay in actually firing the virus...
-
On a related note, what kind of shot do the green blobs represent?
-
On a related note, what kind of shot do the green blobs represent?
FlashColor(1) = vbBlack ' Hit with memory shot
FlashColor(-1 + 10) = vbRed ' Hit with Nrg feeding shot
FlashColor(-2 + 10) = vbWhite ' Hit with Nrg Shot
FlashColor(-3 + 10) = vbBlue ' Hit with venom shot
FlashColor(-4 + 10) = vbGreen ' Hit with waste shot
FlashColor(-5 + 10) = vbYellow ' Hit with poison Shot
FlashColor(-6 + 10) = vbMagenta ' Hit with body feeding shot
FlashColor(-7 + 10) = vbCyan ' Hit with virus shot
-
I like your signature Eric
-
I like your signature Eric
I think you and PY are the only ones who can truly understand...
-
You're probably right there Eric.
On a very literal note
I once spent a week or two in northern Italy trying to rewrite a manual for an Italian made Gas Chromatography Autosampler that had originally been written in Italian and then badly translated.
I can testify that actually reading Italian is about the hardest thing I ever tried to do.