Resync with recent change to DelayedArray

Use getDefaultBlockLength(type(x)) instead of
DelayedArray:::get_default_block_maxlength(type(x)).

Author: hpages

DESCRIPTION

@@ -1,5 +1,5 @@
 Package: minfi
-Version: 1.27.5
+Version: 1.27.6
 Title: Analyze Illumina Infinium DNA methylation arrays
 Description: Tools to analyze & visualize Illumina Infinium methylation arrays.
 Authors@R: c(person(c("Kasper", "Daniel"), "Hansen", role = c("cre", "aut"),
@@ -55,7 +55,7 @@ Imports:  S4Vectors,
           grDevices,
           graphics,
           utils,
-          DelayedArray (>= 0.7.17),
+          DelayedArray (>= 0.7.27),
           HDF5Array,
           BiocParallel
 Collate:  AllGenerics.R

R/DelayedArray_utils.R

@@ -18,20 +18,18 @@ setMethod("dimnames", "arrayRealizationSink", function(x) {
 # Helper functions for setting up ArrayGrid instances --------------------------
 
 colGrid <- function(x) {
-    block_maxlen <- max(nrow(x),
-                         DelayedArray:::get_default_block_maxlength(type(x)))
-    makeRegularArrayGridOfCappedLengthViewports(dim(x), block_maxlen,
+    block_len <- max(nrow(x), getDefaultBlockLength(type(x)))
+    makeRegularArrayGridOfCappedLengthViewports(dim(x), block_len,
                                                 "first-dim-grows-first")
 }
 
 rowGrid <- function(x) {
-    block_maxlen <- max(ncol(x),
-                         DelayedArray:::get_default_block_maxlength(type(x)))
+    block_len <- max(ncol(x), getDefaultBlockLength(type(x)))
     ## Is the "hypercube" shape really intentional? Or should the
     ## "last-dim-grows-first" shape be used instead? (Like in
     ## DelayedMatrixStats:::block_APPLY() when MARGIN == 1L.)
     ## [H.P. -- 06/29/2018]
-    makeRegularArrayGridOfCappedLengthViewports(dim(x), block_maxlen,
+    makeRegularArrayGridOfCappedLengthViewports(dim(x), block_len,
                                                 "hypercube")
 }