On Mon, Apr 2, 2012 at 6:04 PM, Matthew Toseland
<toad at amphibian.dyndns.org> wrote:
> On Saturday 31 Mar 2012 04:03:10 Zlatin Balevsky wrote:
>> >>
>> >> In 'log.info("Random message: %s", obj.toString())' evaluating
>> >> 'obj.toString()' was what caused issues, not recycling it, or so I
>> >> believe to remember.
>> >
>> > You don't need to call obj.toString() before calling log() - just pass the
>> > object itself. Then the only overheads are:
>> > 1) Calling the function, including passing the arguments - which hopefully
>> > will be optimised to be on some stack (but the code might not have been
>> > JITted yet).
>> > 2) Looking up whether it needs to be logged.
>> >
>>
>> 3) Autoboxing because you cannot pass a primitive if the argument is
>> Object without creating a <? extends Number>. ?Small ranges of
>> Char/Short/Integer/Long values are cached, anything outside those will
>> end up creating garbage if the shouldLog predicate evaluate true even
>> once.
>
> Will it still autobox them if they are on the stack, and never used?
If the isDebuggable predicate never evaluates true, then it's not a
problem. But if it evaluates true even once for the lifetime of the
loaded class it will create these objects. It's a quirk of current
hotspot jvm which will not affect production deployments but will
affect the developer: turning on debug logging for one class will
cause all primitives in all log statements to be instantiated. I had
some sample code earlier in the thread.
