Hi Stack,
Trace.java calls random.nextLong() to generate SpanId, and its value
will be negative 50% of the time. The bug can be easily reproduced with the
two test cases I have added in the patch.
Currently MilliSpanDeserializer is not consumed within htrace project.
However if anyone (internal or external) needs to use the deserializer,
this would be a blocking issue.
Thanks,
Long Zhou
On Wed, Dec 31, 2014 at 10:56 AM, Stack <[email protected]> wrote:
> On Wed, Dec 31, 2014 at 10:48 AM, Long Zhou <[email protected]> wrote:
>
> > Can we include fix for HTRACE-40 in this release?
> >
> > HTRACE-40 <https://issues.apache.org/jira/browse/HTRACE-40>:
> > MilliSpanDeserializer throws NumberFormatException when Span has negative
> > Long value
> >
>
> Does a negative number make any sense? You think this a critical issue?
> St.Ack
>
