What about leading and trailing whitespace around the $() operator?

[anki-code]

The original question from laloch:

What about leading and trailing whitespace around the $() operator? Can you do mkdir -p /tmp/$(whoami)/garbage?

[anki-code]

Hi @laloch! Thank you for the great question!

Can you do mkdir -p /tmp/$(whoami)/garbage?

At this time in xonsh 0.9.24 this example is not working and I've tested the @() operator behavior:

echo 123@('456')
#123456

echo 123@(['456','789'])
# 123456 123789

Python substitution operator expands the prefix and duplicate it. This behavior looks good IMHO.

What about $() and XEP-2 proposal?

If we install xep-2 fork and try the similar example we can found that XEP-2 logic improves the behavior comparing with current xonsh:

mkdir -p /tmp/dir_test && cd /tmp/dir_test

# xonsh
mkdir -p here/@($(whoami))/garbage
tree here
#here
#└── pc\012
#    └── garbage

# 🔴 Unwanted 'pc'$'\n' directory.

# xonsh-xep-2
# pip install -U git+https://github.com/anki-code/xonsh-xep-2
rm -rf here
mkdir -p here/@($(whoami))/garbage
tree here
#here
#└── pc
#    └── garbage

# 🟢 Everything as expected.

As result my answer at this time will be: mkdir -p /tmp/$(whoami)/garbage should do the same as mkdir -p /tmp/@($(whoami))/garbage along with XEP-2 where $() is iterable object.

Does it make sense?

[daniel-shimon]

Maybe this should be part of the XEP? (make /tmp/$(whoami)/garbage expand like /tmp/@($(whoami))/garbage)

[anki-code]

Yeah, I think it will be good. I'm waiting here for @laloch feedback as from author of the question. If it works for him I'll be happy to add the suggestion to the XEP-2.