Comment # 11
on bug 10978
from Michael Mann
(In reply to Evan Huus from comment #9)
> Oh, I think I understand what's going on. The compression format is
> basically saying "these n bytes are the same as these other n bytes you've
> already seen at offset X" except we haven't processed X bytes yet so we're
> not validating that offset and are getting uninitialized bytes (instead of
> out-of-bounds access) because of the circular buffer thing.
I was trying to possibly find justifcation for your "easy" fix of just zeroing
out the bytes, but didn't have any luck. Based on your anaylsis does it make
sense to wait until we've seen offset X for decompression? Almost like this
needs 2 passes to work?
You are receiving this mail because:
- You are watching all bug changes.