The local devnet broke for a bit because of the L2 rewrite. It's still broken actually in terms of the rewards claim and it's something I'd like to work on but has been side lined to attend some other issues re: the launch of the public network. I've been sitting on these changes for a few weeks now and they're starting to get bit-rot so now's a good time to check it in.
Use designated initialisers to initialise network_config. There were some variables, can't remember which ones that were pulling mainnet configuration but being used in test networks that were caught with this.
I typically do static builds, always, and CURL was currently being built without SSL which meant that testing stagenet with a Arbitrum provider using https was failing. I added SSL to the build, curiously though this didn't seem to pull in a dynamic dependency on OpenSSL not sure why. Also had to omit zstd because we don't build that yet, but we should eventually ...
Make local devnet use its own hardforks for testing purposes
Move L2Tracker::SAFE_BLOCKS into network config (local devnet wants this very small so that we can test L2 transactions easily. All nodes in the local devnet are connected to the same RPC provider anyway).
Correctly use get_block_height instead of top_block.height which is the same kind of bug with the uninitialised block.hash bug fixed last week
Fix the BLSSigner tests because we now exclusively use verifyMsg and signMsg which matches the audited cryptography in the contracts. This meant the expected test results change.
The local devnet broke for a bit because of the L2 rewrite. It's still broken actually in terms of the rewards claim and it's something I'd like to work on but has been side lined to attend some other issues re: the launch of the public network. I've been sitting on these changes for a few weeks now and they're starting to get bit-rot so now's a good time to check it in.
httpswas failing. I added SSL to the build, curiously though this didn't seem to pull in a dynamic dependency on OpenSSL not sure why. Also had to omit zstd because we don't build that yet, but we should eventually ...get_block_heightinstead oftop_block.heightwhich is the same kind of bug with the uninitialisedblock.hashbug fixed last weekverifyMsgandsignMsgwhich matches the audited cryptography in the contracts. This meant the expected test results change.