address comments

Author: Peiming Liu

mlir/lib/Dialect/SparseTensor/IR/SparseTensorDialect.cpp

@@ -1256,7 +1256,7 @@ bool ConcatenateOp::needExtraSort() {
   bool allSameOrdered = llvm::all_of(getInputs(), [dstStt](Value op) {
     return getSparseTensorType(op).hasSameDimToLvl(dstStt);
   });
-  // TODO: When conDim != 0, as long as conDim corresponding to  the first level
+  // TODO: When conDim != 0, as long as conDim corresponding to the first level
   // in all input/output buffers, and all input/output buffers have the same
   // dimToLvl, the tmp COO buffer is still unnecessary (e.g, concatenate
   // CSC matrices along column).

mlir/lib/Dialect/SparseTensor/Transforms/SparseTensorRewriting.cpp

@@ -905,7 +905,7 @@ struct ConcatenateRewriter : public OpRewritePattern<ConcatenateOp> {
 
     ForeachOp foreachOp;
     for (Value input : op.getInputs()) {
-      // Build a for op for each input tensor to append new values into the
+      // Builds a for op for each input tensor to append new values into the
       // output tensor.
       foreachOp = rewriter.create<ForeachOp>(
           loc, input, iterArg ? ValueRange{iterArg} : ValueRange{},
@@ -914,7 +914,7 @@ struct ConcatenateRewriter : public OpRewritePattern<ConcatenateOp> {
             SmallVector<Value> dstLcvs(dstTp.getLvlRank());
             for (Dimension d = 0; d < dimRank; d++) {
               Value crd = dcvs[d];
-              // Transform coordinates for the concatenating dim.
+              // Transforms coordinates for the concatenating dim.
               if (d == conDim)
                 crd = builder.create<arith::AddIOp>(loc, crd, offset);
               // FIXME: `toStoredDim` is deprecated