Jim Hester

I am not sure what the objection to `dependencies = c("soft", "Config/Needs/website"))` is. Without the soft this is essentially equivalent to `install_github("r-lib/pkgdown", dependencies = FALSE)`, e.g. it would only install...

so the way `TRUE` works is it gives you the 1st order soft dependencies and the 2nd order hard dependencies. soft gives you all of the recursive soft dependencies even...

I can't reproduce this, and this error simply occurs when the endpoint returns HTML or XML instead of JSON, likely due to an error page. What package you were you...

This is because one of the dependencies is from GitHub, and remotes is trying to check its remote hash. We should probably tryCatch it so we don't have an error...

We sort of have this idea in `pak`, there is `cran::[email protected]`, though on `cran::` is supported at the moment.

I realize what you propose is an additional constraint on top of this, which I think makes sense.

It looks to me like you are missing the protocol, are you passing it to `install_git()`?

My main worry with making the GitLab remote more complex is our team doesn't use GitLab, so it is possible this will break in the future without us realizing it....

I guess we should strip them, though it would then break updating packages later. However if you still set the `GITLAB_PAT` when you run `update_packages()` would the update work?

To do this we would have to change that line to remove the environment restriction and also `inherits = TRUE`. This would mean that if there was another function called...