[llvm] [X86][DAGCombiner][SelectionDAG] - Fold Zext Build Vector to Bitcast of widen Build Vector (PR #135010)

[Simon Pilgrim via llvm-commits]

RKSimon wrote:

> @RKSimon @phoebewang zext to bitcast is not beneficial for other targets. Bitcast is blocking there folding. We can take our optimization to X86isellowering and do the needful there and will handle all the regressing test cases for X86. Should it sound good?

Sorry I missed this - will take a look

https://github.com/llvm/llvm-project/pull/135010