v0.6.4

Bug fixes

• Fix busy looping in connect #992

v0.6.3

include a fix where nodes start to go eat all the CPU for network activity, utilising the executor of the network with contained and managed CPU instead of the DefaultGlobal executor.

House Keeping

• network: Get rid of DefaultExecutor #985
• Add block search by stake pool #960

v0.6.2

This release contains mainly changes in the way we handle connections (see link below). It fixes issues regarding keeping connections open for too long resulting in bloating the node

House keeping

• Rework pending client connections #981

v0.6.1

bug fix

• treat the network blocks just like every other blocks even at bootstrap #949

v0.6.0

house keeping

• make the trusted peer list selection more random to avoid bloating the rendezvous points (#946 )
• provide some network info via new REST end point (#846 #939 )
• dropped releasing binaries for FreeBSD and NetBSD (none here were using them and they were adding time to maintain, however contribution to bring that back are welcome #933 )
• actually start purging the leader logs for stake pool operator, reducing a lot the memory consumption per log (#937)
• put an hard limit on incoming p2p connections (#909)

Security

• added cargo audit to our CI so CVE are directly reported at every PR (#929 )

Bug fixes

• fix nodes that would not sync at all because of no publicly available addresses ( #926 )
• make nodes fail right away at start up if node cannot listen on the given port number (#911 )
• drop early client connections that still in pending state, preventing potential DoS (#940)

Breaking changes

For nodes who want to be publicly reachable and advertise themselves as rendezvous point (aka trusted peer), they will now need to generate a secret (jcli key generate --type ed25519) and gives the associated public key to their users (jcli key to-public).

For nodes who need to connect to the rendezvous points, they will need to add the public id of the trusted peer along the IP address.

Node running with versions 0.5.* or older won't be compatible with nodes running from 0.6 and future versions.

v0.6.0-rc1+1

v0.6.0-rc1+1

v0.6.0-rc1

Release candidate for v0.6.0, not for production yet

house keeping

• provide some network info via new REST end point (#846 #939 )
• dropped releasing binaries for FreeBSD and NetBSD (none here were using them and they were adding time to maintain, however contribution to bring that back are welcome #933 )
• actually start purging the leader logs for stake pool operator, reducing a lot the memory consumption per log (#937)
• put an hard limit on incoming p2p connections (#909)

Security

• added cargo audit to our CI so CVE are directly reported at every PR (#929 )

Bug fixes

• fix nodes that would not sync at all because of no publicly available addresses ( #926 )
• make nodes fail right away at start up if node cannot listen on the given port number (#911 )
• drop early client connections that still in pending state, preventing potential DoS (#940)

Breaking changes

For nodes who want to be publicly reachable and advertise themselves as rendezvous point (aka trusted peer), they will now need to generate a secret (jcli key generate --type ed25519) and gives the associated public key to their users (jcli key to-public).

For nodes who need to connect to the rendezvous points, they will need to add the public id of the trusted peer along the IP address.

Node running with versions 0.5.* or older won't be compatible with nodes running from 0.6 and future versions.

v0.5.6+lock

v0.5.6+lock

v0.5.6

house keeping

• improve the REST's get settings data (#887 , thanks @rinor )
• some network logs are no longer reported at the higher default levels when there are successful error handling on invalid/incompatible connections #885

bug and fixes

• REST's get next-id now returns 40x instead of 500 on invalid block id #888
• lot of p2p network fixes and improvements #899 #889 #901 #896
• connection limits #890

v0.5.5

House keeping

• some comfort logging regarding block creation/update (#880 )
• remove resource limitation (2MiB max) of the node (#874 )
• stop the node if a panic/exception happens in the system (#873 ): this will prevent us from not detecting when a component of the node has failed.

Bug fixes

• less panic: #876 #850 #864 #870

full change logs