The method for decoding a frame has an invalid structure to analyze datastreams
The current method to decode frames searches for the first appearance of the characters S and T. After that, the decoder thinks that the subsequent frames are full bit sized, thus checks S and T immediately on the next frame, however, the bug happens whenever a frame have more bytes than expected, introducing a shift error on the next frames, causing a decoding error on the whole data stream.
For example, if we expect N bytes after Header (say S T), and we have this erroneous frame:
S T <N + 1 bytes>
The decoder will read S T <N Bytes>, and the next frame will have a different Header: BitN S
This is causing critical errors due to the loss of (a lot of) data.
Comments (2)
-
reporter -
reporter - changed status to resolved
changes to resolve
#5, using the new decoder we keep the indexes of good frames and we look dynamically along all the array all possible frames→ <<cset e85ec43b8ad1>>
- Log in to comment
Blocking bug detected, see issue
#5→ <<cset c32a4204b306>>