Author Topic: Virus insertion  (Read 1923 times)

Offline Anonomous Guest Person

  • Bot Builder
  • **
  • Posts: 85
    • View Profile
Virus insertion
« on: October 19, 2006, 11:04:42 AM »
For some reason, every single virus I've looked at in the victim's DNA seems to be inserted BEFORE a stop as opposed to AFTER a stop (making it, from cond to stop, technically apart of another gene)

For theoretical example, say you have a virus like:
cond
start
0 .aimdx store
0 .aimsx store
stop

And it gets inserted into a bot like:
cond
*.nrg 1000 >
start
15 .repro store
100 .strbody store
15 .aimdx store
stop

Then the result seems to become:
cond
*.nrg 1000 >
start
15 .repro store
100 .strbody store
15 .aimdx store
cond
start
0 .aimdx store
0 .aimsx store
stop
stop

This bug occurs in v2.42.8 by the way.

Offline EricL

  • Administrator
  • Bot God
  • *****
  • Posts: 2266
    • View Profile
Virus insertion
« Reply #1 on: October 19, 2006, 12:07:42 PM »
This would not surprise me.  Viruses are an area that have been on the "verify it works right" list for the 2.4 fork for a long time, but I have not managed to get to it.  There have been some fixes to the core gene parsing routines in the 2.4 code fork however which could have broken this, so an insertion problem like this would not surprise me in the least.

Do you have a sim with a bot that exhibits this?  It need not be fancy or clean, just something I can use to excercise the code and step through it in the debugger.  Thanks.
Many beers....

Offline EricL

  • Administrator
  • Bot God
  • *****
  • Posts: 2266
    • View Profile
Virus insertion
« Reply #2 on: October 19, 2006, 12:23:52 PM »
Okay, I think I found this via code inspection.  Looks like AddGene wasn't adding 1 to what was being returned by NextStop(), which returns the position of the next stop codone in a genome.  Hopefully, I can release a buddy drop later today.  A sim which demonstrates the bug would still be nice so I can verify the fix.  Thanks.
Many beers....

Offline Anonomous Guest Person

  • Bot Builder
  • **
  • Posts: 85
    • View Profile
Virus insertion
« Reply #3 on: October 19, 2006, 05:39:24 PM »
Quote from: EricL
Okay, I think I found this via code inspection.  Looks like AddGene wasn't adding 1 to what was being returned by NextStop(), which returns the position of the next stop codone in a genome.  Hopefully, I can release a buddy drop later today.  A sim which demonstrates the bug would still be nice so I can verify the fix.  Thanks.

Thanks!

I might be uploading this wrong or something, but here.

Offline EricL

  • Administrator
  • Bot God
  • *****
  • Posts: 2266
    • View Profile
Virus insertion
« Reply #4 on: October 19, 2006, 07:03:56 PM »
Thanks.  Your sim posted fine.  It was what I thought it was.  Stay tuned for a buddy.
Many beers....

Offline Anonomous Guest Person

  • Bot Builder
  • **
  • Posts: 85
    • View Profile
Virus insertion
« Reply #5 on: October 19, 2006, 09:06:48 PM »
No problem. In fact, thanks for fixing that!

Though, out of curiosity, what do you mean by "buddy"?

Offline EricL

  • Administrator
  • Bot God
  • *****
  • Posts: 2266
    • View Profile
Virus insertion
« Reply #6 on: October 19, 2006, 09:27:36 PM »
Quote from: Anonomous Guest Person
Though, out of curiosity, what do you mean by "buddy"?
A "buddy" drop is an industry term for a "between version", iterim release of software that is usally released to one or more "buddies" for testing purposes.  Buddy drops in DB are lettered versions such as 2.42.8a, which is now available in the Changes for 2.42.9 topic and contains the fix for the virus issue.  Although buddies are generally stable releases, which may actually be "better" than the last official release, they are termed buddies becuase 1) a feature intended for ther next release may not be complete 2) I don't do a complete test pass before releasing them like I try to do for "real" releases and 3) I don't post the source code.
« Last Edit: October 19, 2006, 09:31:33 PM by EricL »
Many beers....