babashka-sci-dev

borkdude 2023-10-10T10:18:52.436269Z

@eval2020 When running the fs tests locally on my machine I'm getting:

$ bb test

Running tests in #{"test"}

Testing babashka.fs-test

FAIL in (glob-test) (fs_test.clj:107)
:hidden option should be disabled by default should be enabled (when not provided) when pattern starts with a dot
expected: (= (quote (".gitignore")) (map normalize (fs/glob "." ".git*")))
  actual: (not (= (".gitignore") (".gitignore" ".git")))
Any ideas?

borkdude 2023-10-10T13:45:30.386909Z

Can anyone run fs's tests locally using bb test with the most recent commits?

eval2020 2023-10-10T13:51:22.360999Z

checking…

eval2020 2023-10-10T13:55:19.670189Z

glob does not return folders for me - is this OSX specific? Can provide a patch for this pattern to be more specific.

borkdude 2023-10-10T13:57:02.626789Z

are you on macos?

eval2020 2023-10-10T13:57:07.391789Z

yes

borkdude 2023-10-10T13:57:10.324659Z

m1?

eval2020 2023-10-10T13:57:15.266099Z

yep

borkdude 2023-10-10T13:57:37.988229Z

.git is not a folder in my case, it is

$ cat .git
gitdir: ../.git/modules/fs

borkdude 2023-10-10T13:57:45.385739Z

because of the submodules

borkdude 2023-10-10T13:57:59.068779Z

you are likely running it as a non-submodule of babashka

borkdude 2023-10-10T13:58:01.475399Z

which explains the issue

borkdude 2023-10-10T13:58:06.007749Z

ok, Il

borkdude 2023-10-10T13:58:11.946309Z

I'll just change the pattern myself

👍 1
eval2020 2023-10-10T13:58:13.560109Z

no submodules here indeed

borkdude 2023-10-10T13:58:54.470939Z

done

borkdude 2023-10-10T14:00:01.114939Z

I have some doubts about the auto-hidden thing since on Windows files are hidden via an attribute, not by convention

borkdude 2023-10-10T14:00:25.398319Z

Maybe fs should just treat all dot-starting files as hidden, or maybe not...

borkdude 2023-10-10T14:00:55.034659Z

anyway for glob/match it probably made sense what we did

eval2020 2023-10-10T14:01:20.278119Z

so (glob "." ".some") on Win already yielded dotfiles then? Which makes the code now more portable?

eval2020 2023-10-10T14:01:42.351739Z

> anyway for glob/match it probably made sense what we did ah you mean the same

borkdude 2023-10-10T13:08:36.232619Z

https://github.com/babashka/neil/issues/191