[GitHub] [spark] dgd-contributor commented on a change in pull request #33459: [SPARK-36229][SQL] conv() inconsistently handles invalid strings with more than 64 invalid characters and return wrong value on overflow

[GitBox]

dgd-contributor commented on a change in pull request #33459:
URL: https://github.com/apache/spark/pull/33459#discussion_r677489071



##########
File path: 
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/util/NumberConverter.scala
##########
@@ -49,12 +49,14 @@ object NumberConverter {
    */
   private def encode(radix: Int, fromPos: Int, value: Array[Byte]): Long = {
     var v: Long = 0L
-    val bound = java.lang.Long.divideUnsigned(-1 - radix, radix) // Possible 
overflow once
+    // bound will always positive since radix >= 2
+    val bound = java.lang.Long.divideUnsigned(-1 - radix, radix)
     var i = fromPos
     while (i < value.length && value(i) >= 0) {
-      if (v >= bound) {

Review comment:
       I think the root of this misunderstanding problem is we treat a long 
value as an unsigned long value. I think we should introduce a new class of 
unsigned long value to provide the needed function such as comparing, checking 
for overflow...




-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org