Hi Gerhard,
I'm posting this separately under the same heading because it's probably
less relevant.
While analysing the store-janitor code I found some strange things in
the way Store-janitorImpl.freeMemory() increments the store index.
I assume that each successive call to freeMemory() is intended to free
an item from the next store, starting again when the last store is
reached. If there are 3 stores then the sequence for index should be:
0,1,2,0,1,2,0,1,2,0,1,2,... Is this correct? (three different indexes)
When there are 3 stores my debug log reports the following sequence:
Index before: -1,0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,....(four different
indexes repeat)
Freeing at: -1,0,1,2,*,0,1,2,*,0,1,2,*,0,1,2,....
(corrected): 0,0,1,2,0,0,1,2,0,0,1,2,0,0,1,2,....(store 0 is more
frequent)
Index after: 0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3,....
'*' is where debug message says "Starting from the beginning" and from
the code you can see that it actually does a freeing at 0.
if (this.getLogger().isDebugEnabled() == true) {
this.getLogger().debug("Starting from the beginning");
}
this.resetIndex();
((Store)this.getStoreList().get(0)).free();
this.setIndex(0);
(corrected) is what actually gets freed, taking into account the above
and the following false debug message on the first attempt.
if(this.getIndex() == -1) {
if (this.getLogger().isDebugEnabled() == true) {
this.getLogger().debug("Freeing at index=" +
this.getIndex());
}
((Store)this.getStoreList().get(0)).free();
this.setIndex(0);
} else {
It says it is "Freeing at index=-1" but then actually frees at index 0.
Conclusion 1: Debug printout is a bit misleading.
Conclusion 2: Items are removed from store 0 twice as often as from
stores 1 and 2.
Have I got this right? Is it intentional or a mistake?
Regards,
Peter.
P.S. The debug printout is on my initial posting under this heading.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]
