jas0501 Lv 1
The latest release still contains an ungraceful crash when removing bands during wiggle. The program just ends.
The latest release still contains an ungraceful crash when removing bands during wiggle. The program just ends.
We have never found a way to reliably reproduce this. If someone can help us with that, please do.
It would help if we knew how the keyboard command is parsed by the program. Debugging the problem is difficult from the gamer perspective, specifically because of the inconsistent nature of the problem.
Q1) Is polling used to accept keyboard input or is it interrupt driven?
Q2) When the call is made to remove the band, how is the wiggle function suspended during the removal?. Is it sequential in nature, ie, is the interpolation for the 0th thru Nth terms completed prior to execution of the band removal, or is the function immediate, asynchronous, it splits a full iteration into two parts, pre-removal and post?
If the second case, where the removal is asynchronous, this may be the problem, as a divide by zero or similar may be encountered on the following iteration due to a possible unexpected discontinuity at or near the transition index, and if unhandled could cause problems,or possibly, on second thought, the pointer to the indexing may be lost due to breaking the "chain" in midstride.
I'm just not sure how you're doing things, so it's kinda difficult to judge.
I hope this has some sense to it for you.
Keep the dialog open. if we can get more info on the methods you use, we can put all of our brains together and solve it
What OS are you using to try to reproduce this?
It's pretty much easy to do in Windows XP.
How long in units of seconds does a full iteration (0th thru Nth term) take to complete? To qualify the question, how long does it take for a 120 residue protein?
An intermittent bug can't be diagnosed to find a specific sequence that causes it. That said it occurs often when I type
W -> R -> W
with bands present. The may intersect with the fact that key events are not handled properly. Sometime they are lost.
As an alternative to finding the bug the why not change the implementation of the "R" key during a wiggle to do
Stop
Remove
Wiggle.
As it is, the successful remove of the band does not immediately take effect, either as a bug, or due to the queue of event not yet displayed.
The "R" key implementataion as Stop;Remove bands;Wiggle may also be close to what the uses desires.
We weren't able to reproduce it, not on Linux or Windows.
I can certainly reproduce it at will in WinXP
Is there some debug that can be turned on to capture more detailed info that what's in the error log?
Well, please try to post detailed steps for us to reproduce the bug. You can't debug the problem remotely, so just do your best in helping us reproduce it.
I'm not how many more ways we can describe this.
Throw on some bands, press 'W', wait for the wiggle to slow down, press "R"
BAM - crash to desktop. I just did it 3 times in 5 minutes.