<< return to Pixycam.com

Pixy 2.1 python free(): invalid next size (fast) error

Pixy Software
#1

I’m encountering this error when running a python script on a raspberry pi 5, and it instantly aborts the process without even printing a stacktrace or something. The weirdest thing is it happens at seemingly random intervals, not when is request the ccc blocks. I couldn’t find any examples of others encountering this problem, what could be causing this?
The pixy-relevant portion of the script (basically just detetcting whenever the robot sees a red or a green object):

import pixy
pixy.init()

…

def getColor(attempts:int=5)->int:
    i=0
    rt=-1
    while rt==-1 and i<attempts:
        blocks=pixy.BlockArray(1)
        n=pixy.ccc_get_blocks(100,blocks)
        log("detected %s blocks"%n,INFO)
        for i in range(n):
            b:pixy.Block=blocks[i]
            log("detected at %s %s, size %s %s with signature: %s"%(b.m_x,b.m_y,b.m_width,b.m_height,b.m_signature),INFO)
        if n>1:
            beep_parallel()
            log("multiple blocks detected!",WARNING)
        if n==0:
            beep_parallel()
            log("no blocks detected!",WARNING)
            rt=-1
        else:            
            b1:pixy.Block=blocks[0]
            rt= GREEN if b1.m_signature==1 else RED
        i+=1
        sleep(0.1)
    if rt==-1:
        log("did not find object!",ERROR)
        rt=GREEN
    log("Detected final color %s"%("GREEN" if rt==GREEN else "RED"),INFO)
    return rt

I tried to provide the right context, but if something else is needed let me know!
Thanks,
Csaba

#2

Hello,
Can you reproduce the abort running the unchanged Pixy Python examples?

Edward

#3

Hi, in the meantime I figured it out, it was an error entirely on my part, but I will share what I did wrong in case somebody makes the same mistake.
The problem is with this line of code:

blocks=pixy.BlockArray(1)

And this line:

pixy.ccc_get_blocks(100,blocks)

So I create a BlockArray with a size of one, but put up to 100 blocks of data in it :sweat_smile:
So this is actually a super simple fix.
P.S.:
While searching what could be causing my problem, I noticed that it only crashes at the fourth time I do this, which still seems a little weird (how doesn’t it crash the first time?).

#4

Hello,
Nice work finding the issue :slight_smile: Python manages memory in an inscrutable way. I’m guessing the overwriting only starts being a problem after several calls to ccc_get_blocks().

Edward