[Mlir-commits] [mlir] Fix printing of `mlirUniformQuantizedSubChannelTypeGetNumBlockSizes` in 32-bit machine. (PR #133763)

[Kevin Gleason]

GleasonK wrote:

LGTM. My main question was - should we just be using int64_t instead of intptr_t for the return type, but it looks like the way you currently have it is more in line with the existing "get num" APIs like `intptr_t mlirUniformQuantizedPerAxisTypeGetNumDims(...)` and the test

```
fprintf(stderr, "num dims: %" PRIdPTR "\n",
          mlirUniformQuantizedPerAxisTypeGetNumDims(perAxis));
```



https://github.com/llvm/llvm-project/pull/133763