I suggest follow RI for this case. Because spec says nothing about NPE in
DataOutputStream(OutputStream) constructor, and following spec may cause
legacy applications broken. Furthermore, RI's behaviour is acceptable this
time. :)
On 8/8/06, Vladimir Ivanov <[EMAIL PROTECTED]> wrote:
I need a c
+1 to follow RI
Here are links to similar bugs:
http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=4010323
http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=4012133
AFAIU this is not a RI bug but a feature today.
On 8/8/06, Vladimir Ivanov <[EMAIL PROTECTED]> wrote:
On 8/8/06, Richard Liang
On 8/8/06, Richard Liang <[EMAIL PROTECTED]> wrote:
Vladimir Ivanov wrote:
> I need a consultation: should we follow RI or spec in the case of issue
> 1100?
>
I think we shall follow RI for this this problem. Because the behavior
of RI is reasonable though it seems that it is not compliant wit
Vladimir Ivanov wrote:
I need a consultation: should we follow RI or spec in the case of issue
1100?
I think we shall follow RI for this this problem. Because the behavior
of RI is reasonable though it seems that it is not compliant with Spec.
If we decide to throw NPE, maybe some use applic
I need a consultation: should we follow RI or spec in the case of issue
1100?
Thanks, Vladimir
On 8/8/06, Vladimir Ivanov (JIRA) <[EMAIL PROTECTED]> wrote:
[classlib][io] compatibility: Harmony method
DataOutputStream(null).writeBytes("") throws NPE while RI does not
