/azp run

@ArmaanMcleod Rethinking related issue I believe right fix could be removing backtick from SpecialChars. I add a comment there. Could you please try this in new PR?

@ArmaanMcleod Rethinking related issue I believe right fix could be removing backtick from SpecialChars. I add a comment there. Could you please try this in new PR?

@iSazonov Could do that but current fix is not breaking behaviour since backtick handled separately in Escape() method. If there is a decision to allow this then this is the simplest fix.

@ArmaanMcleod Discussion in the issue was too short, did not found root of the issue. I think their assumption about Escape() is wrong. So I would like to see the test results after removing backtick from SpecialChars so that we can investigate further.

@ArmaanMcleod Discussion in the issue was too short, did not found root of the issue. I think their assumption about Escape() is wrong. So I would like to see the test results after removing backtick from SpecialChars so that we can investigate further.

@iSazonov Yep done. Lets see test results and see.

I see only our updated xUnit test failed. Could you please fix tests.
It confirms we are on right way.

I see only our updated xUnit test failed. Could you please fix tests. It confirms we are on right way.

@iSazonov Yep will do. I also think it is worth bring back Pester tests and doing -like comparison with escaped result, that will cover the match comparison which was highlighted in the issue.

E.g. We need a test which validates below works:

'a`b' -like 'a`b'
True

I think with a change like this the more tests the better to cover all grounds.

@mklement0 Could you please play with the build and try to find side (break something) effects?

@iSazonov, I haven't really followed the discussion, so I'm hoping @IISResetMe can take a look.

/azp run

Forgive me if I'm missing something obvious here, but this change appears to break IsMatch in the same way that Escape is broken - that certainly wasn't my intention with #16306

This behavior:

PS ~> 'abc`def' -like 'abc``def'
True
PS ~> 'abc`def' -like "abc````def"
True
PS ~> [WildcardPattern]::new('abc``def').IsMatch('abc`def')
True

... is entirely correct, I'm not suggesting we change that at all - having to use the same escape character for both the literal syntax and the pattern is potentially confusing, and I'm sure we can do a better job of explaining it in the docs, but it's entirely expected.

Instead of breaking IsMatch, WildcardPattern.Escape() should escape literal `'s:

PS ~> [WildcardPattern]::Escape('a`b')   # correct/desired behavior
a``b
PS ~> [WildcardPattern]::Escape('a`b')   # incorrect/current behavior
a`b

If we make that change instead, then this will start working as expected:

PS ~> 'abd`def' -like [WildcardPattern]::Escape('abc`def')  # desired behavior
True

@IISResetMe Thanks!

I see

This says that WildcardPattern explicitly process escaping in Pattern. So the PR is not right way.

Thanks @IISResetMe @iSazonov

Yeah had the same conclusion and just got confused. I will close this PR since its not the correct fix.

We can start again in another PR. Happy for you to take over if you have the right fix in mind. If not i will have a look based on our findings when i have time 🙂

📣 Hey @ArmaanMcleod, how did we do? We would love to hear your feedback with the link below! 🗣️

🔗 https://aka.ms/PSRepoFeedback

Microsoft Forms