Merge branch 'master' of github.com:das-developers/das2C

Author: cpiker

das2/array.c

@@ -595,7 +595,7 @@ const byte* DasAry_getIn(
 	if((nDim > pThis->nRank)||(nDim < 0)){
 		das_error(DASERR_ARRAY, "Rank %d array '%s' does not have an index "
 				     "number %d ", pThis->nRank, pThis->sId, pThis->nRank, nDim);
-		return 0;
+		return NULL;
 	}
 
 	/* Get index info item at this partial index, or the item pointer for

das2/array.h

@@ -472,7 +472,7 @@ DAS_API void dec_DasAry(DasAry* pThis);
  * @memberof DasAry
  */
 DAS_API byte* DasAry_disownElements(
-	DasAry* pThis, size_t* pLen, size_t* pOffest
+	DasAry* pThis, size_t* pLen, size_t* pOffset
 );
 
 /** Does this array own it's own memory?

das2/variable.c

@@ -720,8 +720,12 @@ DasAry* _DasVarAry_strideSubset(
 	}
 	size_t uRemain = 0;
 	const byte* pBaseRead = DasAry_getIn(
-		pThis->pAry, pThis->base.vt, nVarRank, base_idx, &uRemain
+		pThis->pAry, pThis->base.vt, DasAry_rank(pThis->pAry), base_idx, &uRemain
 	);
+	if(pBaseRead == NULL){
+		*pContinue = false;
+		return NULL;
+	}
 
 	/* make a variable stride from the array stride, note that the var_stride
 	 * may be degenerate and have offset changes of 0 */

das2/variable.md

@@ -1,4 +1,4 @@
-# Developer notes for DasVar_copy()
+# Developer notes for DasVar_subset() and _DasVar_strideSubset()
 
 DasAry is unique in that it allows all array dimensions to be completly ragged.
 This is rare.  Most major binary array implementations such as those in NumPy, 
@@ -7,14 +7,14 @@ This is reasonable as the majority of real world datasets can be accessed
 via strided indexing.  And those that can't are often fill-padded to so that
 strided indexing applies.
 
-Since strided arrays are common, the `DasVar_copy()` function switches over
+Since strided arrays are common, the `DasVar_subset()` function switches over
 to strided copies when the extraction region satisfies the strided condition.
 This allows for copy out in a loop that skips the slow top-down access function
-`DasAry_getAt()`.  In addition, of the requested subsection is actually a continous
-range of memory with no repeats, then a simple memcpy() is used to speed up the
-process further.
+`DasAry_getAt()`.  In addition, of the requested subsection is actually a 
+continous range of memory with no repeats, then a simple memcpy() is used to
+speed up the process further.
 
-To understand the internal implementation of DasVar_copy() a bit more, some
+To understand the internal implementation of DasVar_subset() a bit more, some
 background is useful.
 
 ## Strided Arrays
@@ -53,7 +53,7 @@ are added to read all values:
 ## Slicing alters iteration bounds
 
 To slice an array we hold one (or more) of the indices constant and then 
-iterate over the rest.  The equation is not changed, only the ineration
+iterate over the rest.  The equation is not changed, only the iteration
 indicies for one or more items.  For example to get a slice at j = 2 
 then: