Open rabbibotton opened 8 months ago
This is solved unmasking (if required) the payload before using fast-write-sequence each time. The reason this works I suspect has to do with optimizations used to copy data to the buffer but I do not have time to investigate. I will try and submit a pull request later today.
Any news on the time frame to merge this fix?
Any payload that uses more than one frame (ie, :continue with large payloads) fails.
The buffer information is not being stored properly with fast-write-sequence properly. So that finish-output-buffer returns a correct sized vector but the data is garbage. So most likely the issue is in fast-io
A quick test to confirm I used a string to accumulate the buffer and everything worked.