#programming glitches

[jjeinn-tae]

...

OK, this is by far the weirdest bug I’ve ever encountered since high school, where for whatever reason a “cout <<” line didn’t want to respond, so I just had to repeat the line, completely baffling my teacher and myself.

This is a similar caliber, my crit code rolls a random number between 0 and 100, if it’s below your crit chance, you successfully crit.  That part works, I’ve verified that.  Yet, then I pass that crit flag to Fire(), and it doesn’t touch it until it checks to see if it needs to do the crit multiplier, or destroy the crit trail.  And 100% of the time, whether the crit was successful or not in the previous part, crit is set to false here.  I don’t touch the variable aside for the initial assigning, passing and the 3 checks (1 for the actual, 2 for the debugging to confirm that yes, the bit is swapping!) I don’t get how this is even possible, there’s nowhere the code goes in between, and even if it did, it’d be out of scope!

And so, I am just staring at this, see the glaring (but minor) inefficiency in the code up there (I check for crit every update instead of every time it fires) and miraculously the bug is fixed.

what

[jjeinn-tae]

Infinite fear!

So, in coding up my little rat moods, I accidentally made it so that their biggest fear (the scariest thing around) could be themselves.  Oops.  That makes their fear rating infinity, as they are no distance from themselves. So then they indefinitely try to run away from themselves.

I feel bad for them, poor rats.

[jjeinn-tae]

what

...I...  Huh.

First of all, I fixed the “squishy” wall collisions before fixing wall collisions outside of tile 0,0.  Which was a surprise, but there were errors all over in this code (I guess that’s what I get for doing it while kind of incoherent!).

It didn’t take much longer for me to fix the outside of 0,0 collisions, except wait...  No, it’s only collisions directly on the diagonal from 0,0.  I...  Don’t know how that can be, but OK, that’s really weird.

[jjeinn-tae]

Have you ever fixed a bug in your code, that only activated in a very specific instance, and then looked at it only to be confused how it *ever* worked in the first place in the normal instances?

My “unload panel” function checked the .pnl file (just my little UI file) to see which GUI elements to delete.  I found out there was an error pretty much immediately, as I’d cycle through the subpanels and the messages from the previous would stick around until I went through one full time.

I finally went to go back and fix it as part of my tuneups on function, and realized that the unload panel function forgets to open the file...  Um, what.  How did it ever start working on the second cycle?

(spoilers, the setup of my animation tool left the previous panel file in memory, so the only one that stuck around was the default “point” one, until you cycled to it again.  Loading an animation, and thus clearing the .pnl from memory was the culprit.  But like, I just had it load the file as it *should*.)

[jjeinn-tae]

No more silk-bladed swords!  Still crashing, but it seems unrelated to rendering.  So, this will be more difficult to track down.

[jjeinn-tae]

Oops, rats currently gain infinite speed along both axes when near their target of search.  That was a divide by zero.

[jjeinn-tae]

Ah yes, working perfectly as intended.  Surely my bow can fire these titanic arrows, and their glacially slow pace creating a wall of death is just the best balance decision I could ever make

[awkwardsonic06photos]

You had ONE JOB Shadow.