Tier Nolan [ARCHIVE] on Nostr: π Original date posted:2015-06-22 π Original message:On Mon, Jun 22, 2015 at ...
π
Original date posted:2015-06-22
π Original message:On Mon, Jun 22, 2015 at 8:10 PM, Martin Schwarz <martin.schwarz at gmail.com>
wrote:
> Gavin,
>
> in 2022 your proposal (BIP as well as code) crosses the 32MB maximum
> message size limit. In order to avoid deployment of code that
> deterministically
> fails fatally in 2022, I'd propose to stop the doublings at 32MB for now
> and fix
> the message size limit in the mean time.
There is an exception in the code for "block" messages.
https://github.com/gavinandresen/bitcoinxt/commit/c81898ec46e4962daf975e352931b848026fdc34#diff-7ec3c68a81efff79b6ca22ac1f1eabbaR3548
This means 2MB limit for all other messages. "block" messages are limited
to the max block size for 2 hours into the future.
I think setting it to a week into the future might be better, since it is
only a DOS protection. This would guarantee that message sizes are
reasonable. The size check would still be done anyway.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/attachments/20150622/2457767e/attachment.html>
π Original message:On Mon, Jun 22, 2015 at 8:10 PM, Martin Schwarz <martin.schwarz at gmail.com>
wrote:
> Gavin,
>
> in 2022 your proposal (BIP as well as code) crosses the 32MB maximum
> message size limit. In order to avoid deployment of code that
> deterministically
> fails fatally in 2022, I'd propose to stop the doublings at 32MB for now
> and fix
> the message size limit in the mean time.
There is an exception in the code for "block" messages.
https://github.com/gavinandresen/bitcoinxt/commit/c81898ec46e4962daf975e352931b848026fdc34#diff-7ec3c68a81efff79b6ca22ac1f1eabbaR3548
This means 2MB limit for all other messages. "block" messages are limited
to the max block size for 2 hours into the future.
I think setting it to a week into the future might be better, since it is
only a DOS protection. This would guarantee that message sizes are
reasonable. The size check would still be done anyway.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/attachments/20150622/2457767e/attachment.html>