Guillaume Potier

I'm seing this when calling dmesg: ``` [ 963.574597] process '/cgHI5jEa/dummy' started with executable stack [ 5546.850740] show_signal_msg: 147 callbacks suppressed [ 5546.850744] rocksdb:low[160415]: segfault at 556cdc038e00 ip 0000556cdc038e00 sp...

I can reproduce it in a much much more consistent way when I hit ctrc-c just after rpc ready message: ``` # ./target/release/forest --chain=calibnet --target-peer-count 50 --encrypt-keystore false 2022-08-08T15:50:31.281Z WARN...

> our compile-time version string is already comparable, maybe it can be used? > > https://github.com/ChainSafe/forest/blob/main/forest/build.rs#L27 Yes we can use similar code than for `version`. I think that for some...

My comment was more like indicative of what have done other node implementations @ChainSafe. What's important here would be more to have the feature described by @lemmih. It's something that...

> What do the version numbers mean? > Why are they different from forest --version? This build version is supposed to be the version of the binary (client, miner or...

> What is BuildType/NodeType? Was this part of Lotus? Can we remove it? Those are similar of `BuildType`/`NodeType` in Lotus. `NodeType` tracks the kind of binary that exposes the RPC-api:...

> Is there a reason to keep the module? If so, add the reason to the comments. If not, remove the module. I think a little bit of clean up...

Also we shouldn't have two ways of computing forest version string. Lotus uses `build.UserVersion()` for both `lotus -v` and initing its `APIVersion.Version` string.

> Excellent investigation. Could you create the PR for cleaning up the module? Sure.

@lemmih Is this PR ok for merging? What's left to be done?