Darwinbots Forum

Code center => Bugs and fixes => Topic started by: Welwordion on May 02, 2006, 05:22:35 PM

Title: birthtie counted with numties?
Post by: Welwordion on May 02, 2006, 05:22:35 PM

I thought numties is not meant to count the birthtie? but it seems to do so (using newest version)
(you also should check multi)

Title: birthtie counted with numties?
Post by: Numsgil on May 03, 2006, 01:47:11 PM

The short answer is that I don't know if it's supposed to count it or not.

Title: birthtie counted with numties?
Post by: EricL on May 03, 2006, 02:01:21 PM

Both numties and multi relfect all ties for all bots, tier or tie-ee, all the time in the current code, including the birth tie, even during the cycle in whcih the tie is created.  The code updates them for both bots at the time the tie is created so even if the second bot has yet to go that cycle, they are updated by the time it comes to him.

Any reason why this should not be the case?

Title: birthtie counted with numties?
Post by: Welwordion on May 03, 2006, 04:00:16 PM

Well if you try to write conditions concering the number of ties this can be a hindrance, for example instead of shooting a tie in a certain cycle which can lead to a failed MB construction if you miss, its more sure to check if you already have the right number of ties and act accordingly.
Also multi is designed to find out if a bot is a multibot so he can activate sharenrg etc, and a bot is no multibot if he has just a birthtie. (maybe just modyfing multi is enough, so you know if you only have one tie if its a birthtie or not)

Title: birthtie counted with numties?
Post by: PurpleYouko on May 03, 2006, 05:20:31 PM

A bot should never count as a multi-bot with a birth tie.
.multi is only set to a value of one on the cycle in which a tie hardens.
That is unless the code has been messed up in 2.4  

Title: birthtie counted with numties?
Post by: EricL on May 03, 2006, 06:25:23 PM

Quote from: PurpleYouko

A bot should never count as a multi-bot with a birth tie.
.multi is only set to a value of one on the cycle in which a tie hardens.
That is unless the code has been messed up in 2.4  

The code has been messed up then.  I was not aware of this nuance.  I'll fix in in 2.42.4.

Title: birthtie counted with numties?
Post by: Endy on May 04, 2006, 03:34:52 AM

It's always been that way... If a bot is already sharing, it can start sharing with offspring. Kind of oddball, doesn't happen too much so you don't really notice it until babies start being eaten their parent.

Probably needs fixing. We could also just decrease the amount of time the birthtie lasts below the limit for hardening. Not 100% sure, but that should fix the problem. Maybe also add in a couple of different readbacks, one for hardties and a seperate one for soft ties.

Title: birthtie counted with numties?
Post by: PurpleYouko on May 04, 2006, 11:58:04 AM

Yes I have seen birth ties turn yellow indicating sharing. However I regularly use the *.multi sysvar to detect a hardened tie and have never come across an instance where it shows a true value too soon.

Then again to be fair, since Hunter 2.3 figured out an easy way to delete a birth tie, I pretty much always use some sort of birth tie removal gene in every bot so perhaps this bug has always been there.

Either way it needs to be squished good.

Title: birthtie counted with numties?
Post by: EricL on May 04, 2006, 12:03:18 PM

I broke multi in 2.42.3.  I do penenace.  I will fix it in 2.42.4.