This is a bug in javac
starting with JDK 9 (which made some changes regarding string concatenation, which I suspect is part of the problem), as confirmed by the team javac
with bug id JDK-8204322 . If you look at the corresponding bytecode for the line:
array[i++%size] += i + " ";
IS:
21: aload_2
22: iload_3
23: iinc 3, 1
26: iload_1
27: irem
28: aload_2
29: iload_3
30: iinc 3, 1
33: iload_1
34: irem
35: aaload
36: iload_3
37: invokedynamic #5, 0 // makeConcatWithConstants:(Ljava/lang/String;I)Ljava/lang/String;
42: aastore
Where the last aaload
is the actual load from the matrix. However, the part
21: aload_2 // load the array reference
22: iload_3 // load 'i'
23: iinc 3, 1 // increment 'i' (doesn't affect the loaded value)
26: iload_1 // load 'size'
27: irem // compute the remainder
Which roughly corresponds to the expression array[i++%size]
(minus the actual load and store), it’s in there twice. This is incorrect, as the specification says in jls-15.26.2 :
A compound assignment expression of the shape
E1 op= E2
equivalent toE1 = (T) ((E1) op (E2))
where is itT
is the kind ofE1
, except thatE1
it is evaluated only once.
So, for the expression array[i++%size] += i + " ";
the part array[i++%size]
it only needs to be evaluated once. But it is rated twice (once for the cargo and once for the store).
So yes, this is a bug.
Some updates:
The bug was fixed in JDK 11 and there will be a back-port to JDK 10 (but not JDK 9, since no longer receives public updates ).
Aleksey Shipilev mentions on the JBS page (and @DidierL in the comments here):
Solution: compile with
-XDstringConcat=inline
This will return to use StringBuilder
to do the concatenation and does not have the bug.