Fred Hebert

> because the next_state function receives the execution result, it means the initial generation of commands needs to happen side by side with the execution, because otherwise we don't have...

The initial thread also had mentions of maybe generating an ID every time the host name is `@localhost` since it does not provide a lot of information. Any views on...

A lot of the environment set up for that file can likely be skipped if you run the compile job from within a visual studio terminal since that’s what sets...

`backoff:increment(1)` already exists. It just does it by a bitshift rather than exponentiation since the default exponentiation function in Erlang works on floats and handling for timers is probably a...

The tests won't give you good results, since they're tests that willingly break and fail so that I can manually check the output that results. I willingly did not check...

I believe there is no way to fix this. The CT hooks just have disjoint calls for `init_per_`, and test cases themselves [carry no information about the groups they belong...

Sounds workable; might make sense to identify which request returns which results with a Ref, allowing people to parallelize calls.

Looks like a good idea. A test may be tricky to put in place but worth it (maybe behind a version/dynamic switch). It seems like Dialyzer is failing on OTP-25....

You may want to use https://github.com/Supersonido/rebar_mix as a plugin, which also provides hooks to consolidate protocols to get them included in a release.

IIRC the issue is probably that we moved _checkout directories' build directory to `?DEFAULT_CHECKOUTS_OUT_DIR` avoid the symlinking issue of prior times: https://github.com/erlang/rebar3/blob/048412ed4593e19097f4fa91747593aac6706afb/apps/rebar/src/rebar_app_utils.erl#L256-L261 This was done as part of https://github.com/erlang/rebar3/pull/2276 and...