Add support for excluding functions from coverage with attribute

[benjaminfuchs]

PR Summary

Fixes #2268

This PR adds support for excluding specific code constructs from coverage analysis using [ExcludeFromCodeCoverageAttribute()].

Changes:

• New CoverageLocationVisitor: Skips AST nodes in script blocks marked with [ExcludeFromCodeCoverageAttribute()].
• Get-CommandsInFile updated: Replaced predicate-based filtering with CoverageLocationVisitor.
• Updated Tests: Ensures nodes with [ExcludeFromCodeCoverageAttribute()] are skipped.
  Verifies inclusion of unmarked nodes and relevant control flow statements.

PR Checklist

• PR has meaningful title
• Summary describes changes
• PR is ready to be merged
  • If not, use the arrow next to Create Pull Request to mark it as a draft. PR can be marked Ready for review when it's ready.
• Tests are added/update (if required)
• Documentation is updated/added (if required)

[benjaminfuchs]

If the approach is valid, I'll refine the tests and add proper documentation. Feedback is welcome!

[fflaten]

Thank you for looking into this! It works as intended, though I've suggested some improvements.

While reviewing I realized we could make this more efficient, see comment about FindAll. So I'd like to discuss an alternative approach. See draft PR #2598 which should produce the same results.

[fflaten]

FindAll will trigger this for commands in nested scriptblocks even though a parent is excluded. Not ideal to reverse (check parents) on every step.

$ast = {
    1 | ForEach-Object {
        [System.Diagnostics.CodeAnalysis.ExcludeFromCodeCoverageAttribute()]
        param($a, $b)
        $a + $b
        2 | ForEach-Object {
            3 | ForEach-Object {
                continue
             }
        }
    }
}.Ast
$ast.FindAll($predicate, $searchNestedScriptBlocks)

# Processing .. output added in IsExcludedByAttribute

Processing 1
Processing ForEach-Object {
        [System.Diagnostics.CodeAnalysis.ExcludeFromCodeCoverageAttribute()]
        param($a, $b)
        $a + $b
        2 | ForEach-Object {
            3 | ForEach-Object {
                continue
             }
        }
    }
Processing $a + $b
Processing 2
Processing ForEach-Object {
            3 | ForEach-Object {
                continue
             }
        }
Processing 3
Processing ForEach-Object {
                continue
             }
Processing continue

Expression Redirections Extent
---------- ------------ ------                                                                                                                                      
1          {}           1                                                                                                                                           
           {}           ForEach-Object {… 

We could avoid this by rewriting to use an AstVisitor, stopping early at a ScriptBlockAst where the attribute is set.

[benjaminfuchs]

@fflaten Thanks for the feedback! I agree with your point about improving efficiency by replacing FindAll with Visit. The approach in draft PR #2598 looks good to me. My check for the attribute also seems a bit over-engineered. If you'd like, I can take over your implementation and finalize the PR with tests and documentation updates, or let me know how you'd like to proceed.

[fflaten]

Looks like you've simplified yours as well. Feel free the use and complete the Visitor-based code. Was waiting on @nohwnd's thoughts but personally I'd prefer it as it's more extensible.

[nohwnd]

Will try to have a look today in the evening.

[nohwnd]

It would be nice to have a more high-level api test here, similar to the "Single class using '" tests above. Where we enter and leave the coverage and get the results.

[nohwnd]

Add some short info about what this class does if possible. The original predicate is somehow (at least for me) easier to grasp what the intent is.

[nohwnd]

And possibly also why we prefer it over the original predicate.

[nohwnd]

Great job here,thank you! Some small nitpicks on the code :)

[fflaten]

Thanks. @benjaminfuchs Let me know if you'd like me to address any of the comments.

[nohwnd]

I can also address them.

…

-j

________________________________ Od: Frode Flaten ***@***.***> Odesláno: Tuesday, January 14, 2025 8:54:18 PM Komu: pester/Pester ***@***.***> Kopie: Jakub Jareš ***@***.***>; Mention ***@***.***> Předmět: Re: [pester/Pester] Add support for excluding functions from coverage with attribute (PR #2593) Thanks. @benjaminfuchs<https://github.com/benjaminfuchs> Let me know if you'd like me to address any of the comments. — Reply to this email directly, view it on GitHub<#2593 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABLYLYK76PWF4DVMDWV7PSL2KVTOVAVCNFSM6AAAAABUAW2RJ2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKOJQHE4DEMZRG4>. You are receiving this because you were mentioned.Message ID: ***@***.***>