Repository navigation
Get-ChildItem: wildcard matching behavior differs between -Path and -Include parameters when using escape characters #3724
Description
Note:
Get-Itemappears not to be affected, onlyGet-ChildItem- Problem exists on all supported platforms.
Wildcard patterns support ` as the escape character to represent literal ` instances in the filename and to escape the following metacharacters: * ? [ ]
Steps to reproduce
# Create a file that contains a literal backtick and a literal substring that would be a valid
# sub-pattern if interpreted as a wildcard expression.
New-Item -Type File 'a`[b]'
# Match the newly created file using a wildcard expression with suitable escaping.
# The following 2 command should be equivalent, but currently aren't.
Get-ChildItem * -Include 'a```[b`]'
Get-ChildItem -Path 'a```[b`]'Expected behavior
Both Get-ChildItem commands should match the newly created file and output information about it.
Actual behavior
The 2nd command unexpectedly fails:
Get-ChildItem : Cannot find path '/Users/jdoe/a```[b`]' because it does not exist.
...
Again, note that Get-Item does not exhibit this problem.
Environment data
PowerShell Core v6.0.0-alpha (v6.0.0-alpha.18) on macOS 10.12.4
PowerShell Core v6.0.0-alpha (v6.0.0-alpha.18) on Ubuntu 16.04.1 LTS
PowerShell Core v6.0.0-alpha (v6.0.0-alpha.18) on Microsoft Windows 10 Pro (64-bit; v10.0.14393)
Activity
iSazonov commentedon May 8, 2017
Related #3414
Is it only macOS issue?
mklement0 commentedon May 8, 2017
Ilya (@iSazonov): No, it affects all supported platforms - I've updated the original post to make that clear.
mklement0 commentedon Sep 5, 2019
I don't know what fixed it, but the problem no longer occurs as of PowerShell Core 7.0.0-preview.3
iSazonov commentedon Sep 5, 2019
Michael Klement (@mklement0) Should we add a test to exclude regression or we already have it?
mklement0 commentedon Sep 5, 2019
Ilya (@iSazonov) Providing such tests sounds like a good idea, given that we don't know why the problem went away and given that there'll be a v2 filesystem provider.
I haven't explicitly checked if we already have such tests, but I'd be surprised if we did, given that this problem was seemingly never tackled as such.
On a quick meta note: I know you've pinged me on a few other issues, but the bigger ones I haven't had the time and energy to respond to, but I generally appreciate your systematic efforts to reach out to others (who may care / have something to contribute) in the discussions.
iSazonov commentedon Sep 5, 2019
Michael Klement (@mklement0) I do not expect an immediate reaction to pings. I just hope you get some interesting links for thought. Your reviews is always great!
I will add a test for the issue. Thanks!
iSazonov commentedon Sep 5, 2019
Perhaps kwkam could reference a PR where the issue was fixed.