Jeremy Evans

Looking at RFC7230: `field-name` is `token`, which is VCHAR (0x21-0x7e) - `"(),/:;?@[\]{}` : `/\A[-!#-'*+.0-9@-Z\\^-z|]+\z/` `field-value` is any number of SP (0x20), HTAB (0x9), VCHAR (0x21-0x7e), obs-text (\x80-0xff): `/[ \t!-~\x80-\xff]*/n` You...

I think adding something like this to the README is acceptable, I just didn't think it should be in SPEC. Here is my recommendation for modifications to the proposed language....

I think the code example should index into the environment hash. Otherwise, it looks fine.

Tried these instructions and did not run into any major issues. I was able to install postgres, restore a backup of my asdf-postgres database to the mise-postgres database, and bundle...

> IMO, I would not ship the `timeout` prefixing code in the general case for now, I think that's pretty invasive. > > > This is not a replacement for...

> Because of the first commit, authentication tests are failing with the following exception. @jeremyevans, how can I fix this issue? Architecturally, is the first commit incorrect? For catch use...

Looks like E2E failed as it spent over 90 minutes trying to download boot images. That doesn't seem related to this change.

You mention the use of OpenBSD pledge/unveil, but `File.writable?` does not work for that: ``` $ ruby -r unveil -e 'Pledge.unveil("."=>"r"); p File.writable?("."); File.write("./a", "")' true -e:1:in `write': Permission denied...

I'll check with other OpenBSD developers regarding whether `access(2)` should respect `unveil(2)` limitations.

@KJTsanaktsidis I checked and the unveil/access issue was considered a bug. A fix has just been committed for it. Thanks for asking about it.