Closed makkarpov closed 3 months ago
@makkarpov I think we should most likely swap it before we call writeToken
. Do you want to do a PR ?
Sure, I can do the PR.
You mean wrap in the SwappedByteBuf
for the second time? It should not be a big deal in performance terms (since tokens are short), but construction like SwappedByteBuf(SwappedByteBuf(x))
look awkward. Shouldn't we just remove allocateNativeOrder
from the allocation of mintTokenBuffer
? Byte order of this buffer does not seem to be relevant anywhere.
Sounds good
Version: ~0.0.64.Final~ 0.0.65.Final
QuicTokenHandler
receives swapped byte buffer inwriteToken
, but a regular buffer invalidateToken
.This could give you a lot of debugging time figuring out why none of the
readInt
,writeInt
and similar functions works correctly.