mobilecoinfoundation / cookiecutters Goto Github PK

One issue is, when you cookiecutter out a repo like this and cargo publish it for the first time, you will be the only owner, and the only one who is authorized to put it on crates.io.

Then, if such a person leaves the company, it may be impossible for us to publish any later versions of it on crates.io if we can't get a hold of them.

You can see owners on crates.io pages like here: https://crates.io/crates/aligned-array

For instance, with aligned-array, I made sure I added mobilecoinofficial/eng as a co-owner so that if I get hit by a bus, the show can go on.

It would be nice if something in the cookiecutter template would prompt people to do this, otherwise a lot of people will likely forget. I almost forgot this with mc-rand just now.

One issue I also ran into with mc-rand just now is:

chris@chris-ThinkPad-X1-Extreme-Gen-3:~/mobilecoinofficial/rand$ cargo owner --add github:mobilecoinfoundation:coredev
    Updating crates.io index
error: failed to invite owners to crate `mc-rand` on registry at https://crates.io

Caused by:
  the remote server responded with an error: only members of a team or organization owners can add it as an owner

So I am not actually authorized to add coredev to repos that I create -- it might be that James or Nick or someone has to be in the loop for this process. Or I guess I can give mobilecoinofficial/eng ownership over repos in mobilecoinfoundation?

This looks good to move forward providing a ci run for the other queued PRs to leverage.

I'm a bit behind my current tasks, but I would like to find time to:

• Add poetry for venv for consistent local runs
• Convert the test to pytest, for more narrowed test failures
• Output to https://docs.pytest.org/en/7.1.x/how-to/tmp_path.html#the-tmp-path-fixture
• utilize https://docs.python.org/3/library/filecmp.html with a deep comparison to
• I need to find it, but I'm pretty sure there was a pytest plugin to overwrite expecteds so one could pass a flag during the test runs and regenerate the output, similar to what the bash script is always doing.

For example I don't think the github_user is very robust right now, but I don't think we should have a full directory tree per test case, while a temp dir we could focus on the contents codeowners file

Originally posted by @nick-mobilecoin in #24 (review)

• #27
• #28
• #29
• #30

Examples should be updated when CI detects they need to be (it's still recommended to do this yourself, but this is like formatting/linting: the repo should do that for you if you forget)

We decided to approach our "test output" from a different perspective and supercede #25. Instead, we're going to treat the "expected output" as "examples" instead. Thinking about it this way changes what we want to do for goals.

• Keep the default-config example in the repo
• #39
• #40
• #41

most of our repos don't have wikis

Usually I put #![deny(missing_docs)] in all my crate's lib.rs, and I'd like to encourage other people to do this.

It occurs to me that maybe I shouldn't put this in every crate's source code, and just put it in CI instead

It would be nice if one of the questions is if it's intended for mobilecoinfoundation or mobilecoinofficial org

this is not good when the cargo lock files are being checked in, like with a new attested service