Raymond Raymond <[EMAIL PROTECTED]> writes:
> Thanks for you answer Knut.
> In org.apache.derby.impl.store.raw.log.LogToFile.java,
> in method checkpointWithTran(), there is a piece of code
> like following:
>
> synchronized (this){
> // has someone else found a problem in the raw store?
> if (corrupt != null){
> throw
> StandardException.newException(SQLState.LOG_STORE_CORRUPT, corrupt);
> }
> // if another checkpoint is in progress, don't do anything
> if (inCheckpoint == true)
> proceed = false;
> else
> inCheckpoint = true;
> approxLogLength = endPosition; // current end position
> // don't return from inside of a sync block
> }
> if (!proceed){
> return false;
> }
>
> see, someone did comment "don't return from inside of a sync block", I
> am not
> sure who wrote the comment. I am just wondering what potential problem it
> will case to return from the inside of a sync block.
Hm, that looks strange... I can't see that there's any problem with
returning from inside that synchronized block. Perhaps it is a
workaround for a JIT bug in an old JVM? I have seen strange pieces of
code other places as well with comments about bugs in certain VMs. It
could also be that the code has changed after the comment was written,
or the comment could have been based on a misunderstanding. But I'm
just guessing.
--
Knut Anders
