To further clarify, what I did to prepare this test is:
mkdir -p test/a/b
cd test
echo > a/c
On Sun, Nov 25, 2018 at 11:11 AM Fernando Santagata <
nando.santag...@gmail.com> wrote:
> Here's output of 'a/b'.IO.d from the REPL:
>
> > 'a/b'.IO.d
> True
>
> On Sun, Nov 25, 2018 at 1:52 AM Timo Paulssen <t...@wakelift.de> wrote:
>
>> The dir method gives you entries in the directory you pass. If you don't
>> pass a test it'll use the default test which is none(".", ".."), i.e.
>> "anything except . and ..".
>>
>> I'm not sure why using { .IO.d } as the test would not give you b,
>> though. Can you check what "a/b".IO.d outputs? Maybe that can give us a
>> clue.
>>
>> HTH
>> - Timo
>> On 24/11/2018 22:18, Fernando Santagata wrote:
>>
>> Hello,
>>
>> I think that I don't understand how the 'test' argument of 'dir' works.
>> I have a directory 'a', which contains a subdirectory 'b' and a file 'c';
>> I want to select only the subdirectories of 'a'.
>>
>> Using the REPL I tried to ask the content of the directory 'a':
>>
>> > my @dirs = dir('a')
>> ["a/c".IO "a/b".IO]
>> > my @dirs = dir('a', test => { .IO.d })
>> ["a/.".IO "a/..".IO]
>> Why omitting the test the code returns the right list, while adding the
>> test it returns just '.' and '..'?
>>
>> If I do the same thing for the top level directory '.' the behavior is
>> different:
>>
>> > my @dirs = dir('.', test => { .IO.d })
>> [".".IO "a".IO "..".IO]
>>
>> Now I can see the directory 'a'.
>> If I descend a level, doing a 'cd a', the behavior is consistent with
>> what I see at the previous level.
>> I'm confused.
>>
>> I'm using version 2018.10.
>>
>> --
>> Fernando Santagata
>>
>>
>
> --
> Fernando Santagata
>
--
Fernando Santagata
