Correct "toolchain list" command verbose option

[BeniCheni]

Correct the verbose (or v) option of toolchain list to display full path of toolchains. Updated tests.

current observation - only displays user's rustup directory

→ rustup toolchain list -v                                                     
stable-x86_64-apple-darwin (default)	/Users/bchen/.rustup
nightly-2019-09-25-x86_64-apple-darwin	/Users/bchen/.rustup
nightly-x86_64-apple-darwin	/Users/bchen/.rustup

fixed version

→ rustup toolchain list -v                     
stable-x86_64-apple-darwin (default)	/Users/bchen/.rustup/toolchains/stable-x86_64-apple-darwin/
nightly-2019-09-25-x86_64-apple-darwin	/Users/bchen/.rustup/toolchains/nightly-2019-09-25-x86_64-apple-darwin/
nightly-x86_64-apple-darwin	/Users/bchen/.rustup/toolchains/nightly-x86_64-apple-darwin/

[kinnison]

You have some strange indentation which needs fixing in your test.

[kinnison]

Not bad at all for a first pass. Please simplify as directed and undo one of your changes, and ensure that the code is properly formatted so it passes the CI.

This is a distinct improvement in output.

[BeniCheni]

Rather than adding an extra function for the delimiter?

I'd rather leave the expansion of m.is_present("verbose") to within a local toolchain_list()

The -v/--verbose thing doesn't really need checking here, pick one and test that, no need for both.

@kinnison, updated per suggestions, ready for your next convenience to review again. Thanks!

[kinnison]

This is looking very good, but I had an idea about linked toolchains (see comments)

[BeniCheni]

It might be cute to also handle symlinked toolchains (rustup toolchain link) by reading the symlink and reporting the target rather than the location in ~/.rustup/toolchains

I'd prefer if this function took toolchain as &str even though in theory it's the only consumer of the String produced in the calling function -- just to futureproof things a little.

Greeting, @kinnison! Sorry about the delay. Been under the "water" w. life & work in last couple of days, but now I'm back and pushed an update to the PR.

Like your suggestion of symlink a lot to take the feature further! Updated the feature & added suggested test. Also updated to use &str instead of String for the toolchain param of the print_toolchain_path function.

Ready for your next convenience to review again. Thanks!

[kinnison]

Almost there, one little bogon to resolve...

[BeniCheni]

Almost there, one little bogon to resolve...

@kinnison, updated to address the bogon, :) Thanks.