Sync yarp branch with main

[andyw8]

This brings in the latest changes from main. There were few conflicts in DocumentSymbol due to #947 and #949, so I thought it best to have a review of how that was resolved.

[github-actions]

Benchmark results in seconds (slowest at top)

          textDocument/completion average: 0.273015 std_dev: 0.009705
          textDocument/diagnostic average: 0.065767 std_dev: 0.01489
          textDocument/definition average: 0.006715 std_dev: 0.003463
      textDocument/selectionRange average: 0.005266 std_dev: 0.000909
   textDocument/documentHighlight average: 0.002204 std_dev: 0.00061
               codeAction/resolve average: 0.001801 std_dev: 0.00043
               textDocument/hover average: 0.001728 std_dev: 0.000321
           textDocument/inlayHint average: 0.001648 std_dev: 0.000449
        textDocument/documentLink average: 0.001336 std_dev: 0.000472
 textDocument/semanticTokens/full average: 0.001315 std_dev: 0.000315
            textDocument/codeLens average: 0.001313 std_dev: 0.000263
      textDocument/documentSymbol average: 0.001311 std_dev: 0.000379
    textDocument/onTypeFormatting average: 0.001287 std_dev: 0.000419
          textDocument/codeAction average: 0.001249 std_dev: 0.000294
textDocument/semanticTokens/range average: 0.001233 std_dev: 0.000256
          textDocument/formatting average: 0.001228 std_dev: 0.00054
        textDocument/foldingRange average: 0.001198 std_dev: 0.000308


================================================================================
Comparison with main branch:

 textDocument/semanticTokens/full faster by 63.577 %
textDocument/semanticTokens/range faster by 47.359 %
      textDocument/documentSymbol faster by 65.215 %
        textDocument/foldingRange faster by 66.256 %
          textDocument/formatting unchanged
          textDocument/diagnostic unchanged
        textDocument/documentLink faster by 62.938 %
           textDocument/inlayHint unchanged
      textDocument/selectionRange faster by 16.12 %
   textDocument/documentHighlight faster by 42.01 %
               textDocument/hover unchanged
          textDocument/codeAction unchanged
    textDocument/onTypeFormatting unchanged
               codeAction/resolve unchanged
          textDocument/completion unchanged
            textDocument/codeLens faster by 65.032 %
          textDocument/definition unchanged


================================================================================
Missing benchmarks:

RubyLsp::Requests::ShowSyntaxTree

[andyw8]

Note: Had to mark as skip for now, as with other tests in this class.

[vinistock]

Did you use merge to bring the changes in from main? I see changes that are related to other commits (e.g.: bumping the version) present in this commit.

Our default is to always use rebase. Will the fact that the changes from other commits are merged here in this different commit create more conflicts when rebasing this branch?