Improve pipeline indentation handling in UseConsistentIndentation rule

[liamjpeters]

PR Summary

• Two small improvements in the UseConsistentIndentation rule's handing of pipeline indentation. Flagged in this issue and issue comment.

  1. Rule never finds the ends of nested pipelines, as it had a fast-return when it reached a pipeline who's end position was beyond the current tokens line. It would always find the outer/parent pipeline (which contains the nested pipeline and spans past it) and return early, empty-handed.

     This would lead to some cases of "runaway" indents.

     PR sorts the pipelines, in place, by their end line number and column number. This allows the nested pipeline to be found first (as they end before their parent pipeline).

  2. Rule was using a single counter to keep track of pipeline indentation and then removing it all when it found a pipeline end.

     There were some edge cases that were, even with the first issue resolved, not quite right.

     PR adds in tracking of each pipeline's indentation contribution separately.

• Small improvements in the UseConsistentIndentation rule's handling of indentation where opening braces are on the same line. Previously, adjacent openers like .foreach({, @(@{, or ([PSCustomObject]@{ each incremented the indentation level, producing double or triple indentation. Now only the last unclosed opener on a line affects indentation, fixing common patterns like .foreach({...}) and $list.Add([PSCustomObject]@{...}).

Copilot generated (and I reviewed) a bunch of tests to cover these cases. It's really getting quite good at generating tests now 🤖

Fixes #1680
Fixes #1933
Fixes #2159

Potentially also Fixes the main issue raised in #1168

PR Checklist

• PR has a meaningful title
  • Use the present tense and imperative mood when describing your changes
• Summarized changes
• Change is not breaking
• Make sure all .cs, .ps1 and .psm1 files have the correct copyright header
• Make sure you've added a new test if existing tests do not effectively test the code changed and/or updated documentation
• This PR is ready to merge and is not Work in Progress.
  • If the PR is work in progress, please add the prefix WIP: to the beginning of the title and remove the prefix when the PR is ready.

[andyleejordan]

Nice work — this hits #1933, #1680, and #2159 with one coherent change.

The two structural moves are exactly the ones the bugs called for:

• Replacing the flat currentIndentationLevelIncreaseDueToPipelines counter with a Dictionary<PipelineAst, int> so each pipeline's contribution is unwound only when that pipeline ends — that's the actual fix for the runaway-indent-on-nested-pipelines bug.
• The end-position sort on pipelineAsts so inner pipelines are matched before outer ones (and the comment explaining why MatchingPipelineAstEnd's early-break needs that ordering) is a nice clarity improvement.

The unification of paren/curly opener handling under a single openerSkippedIndentation stack with HasUnclosedOpenerBeforeLineEnd is a clean refactor of the previous lParenSkippedIndentation-only logic, and the depth-relative scan (return false on our own closer, true on a nested unclosed opener at line end) is correct for the multiple-openers-on-a-line cases. All 90 tests pass locally.

LGTM — approving.

— Drafted by Copilot (Claude Opus 4.7)