Author Topic: Bugs in 2.42.8f  (Read 6935 times)

Offline Elite

  • Bot Overlord
  • ****
  • Posts: 532
    • View Profile
Bugs in 2.42.8f
« Reply #15 on: November 10, 2006, 02:45:10 AM »
WOW! That's a huge change

My batterybots now work indistinguishably from how they were meant to work. PY's inchworm now works much more compatably. Exen Hexagonis manages to form hexagons (but fails quite a lot - exactly as it did in 2.37.6)! MBs can turn, fix angles, set lengths, feed, set memory locations through the tie. Perfect

Antbots work perfectly too (nothing to do with ties, but I just noticed)

Yay, well done  

I've attached a work in progress. A 2.4-only 2-bot MB that moves through changing tie lengths. Needs a little work, but here's what I've got so far ...
« Last Edit: November 10, 2006, 02:45:40 AM by Elite »

Offline Elite

  • Bot Overlord
  • ****
  • Posts: 532
    • View Profile
Bugs in 2.42.8f
« Reply #16 on: November 10, 2006, 11:23:25 AM »
Got another bug

Take the sim I've got below, and load it
Select one of the tied bots, and move it around a little (it may have something to do with the vision grid too, then the bot has something straight ahead in it's eye5). It should crash with "Error. Invalid procedure call or argument"

Offline EricL

  • Administrator
  • Bot God
  • *****
  • Posts: 2266
    • View Profile
Bugs in 2.42.8f
« Reply #17 on: November 10, 2006, 01:46:14 PM »
Yea, I caught this one myself last night.  Basically, now that angular momentum is working, if you fling the bot with the focus around enough by a tie, the combination of it's movement and the angular momentum transfer of flinging it around can push the internal .aim calculation beyond 2Pi, which causes a crash when the circle indicating the bot focus is drawn since the VB Circle method has problems with some arguments > 2Pi and there was insufficient protection in the focus ring drawning routine to deal with it.  Fixed in the next buddy/release post 2.42.8h.
Many beers....