----------------------------------------------------------------
BEFORE YOU POST, search the faq at <http://java.apache.org/faq/>
WHEN YOU POST, include all relevant version numbers, log files,
and configuration files.  Don't make us guess your problem!!!
----------------------------------------------------------------

Dear Nick

I greatly appreciate your detailed response to my problems. 
Unfortunately you don't address the issue which worries me the 
most.  

My main question is whether there is a built-in way of protecting 
oneself from a user who refreshes his browser many many times?

(I should have left all the other questions for other threads.)

I'm contemplating running a time check on the session object and 
forcing the user to wait 5-10 seconds before using the same servlet 
twice, but -- is this necessary? advisable? or even valid?  Am I 
somehow circumventing default protections?


> Are you intializing all your locally-scoped variables? 

I can't claim not to have missed one or two.  But -- why do you ask?  
The site is quite comfortably serving a few hundred user's per day.  
So I don't see how your question is relevant.  I'm busy confirming all 
the details so I can meet future demand.  (In January it has to ramp 
up to 10s of thousands per day.)  Perhaps, your question is 
relevant, but I don't understand how.


> You should never synchronize "until desired consistency is
> achieved". You should always know why you are synchronizing.
> Remember when you say "synchronize", you are implicity saying
> "synchronize(this)": the entire object is getting locked, not just
> the method. 

The crude reality is that this is a VERY rush job.  We are currently 
doing user testing.  In future we will examine the 'sychronization' 
issues one by one.  I was concerned that by sort of forcing a single 
thread model I was causing unexpected side effects.  If so, I don't 
really understand how that would produce the behaviour I'm seeing.


> Don't forget to commit every jdbc transaction, even "select"s:
> because jdbc considers even reads can trigger writes. 

Interesting point, but MySQL has no 'commit'!  What I'm doing is 
freeing my connections back to the dbConnectionBroker pool.  Do I 
need to satisfy JDBC as well??  I hadn't thought of that.


> And are you using a "finally" clause to ensure you are committing
> and closing your cursor? 

Yes, absolutely, hence my severe concern over what I'm seeing!   If I 
hit <ctrl-r> a half a dozen times then at least four out of six times, the 
code never executes as far as the 'finally' -- it just ceases to 
execute........ somewhere!

Sure hope someone out there can help me, it's got me pretty 
worried.

Rgds, 

Hasan



  Hasan
  ~~~~~                 HASAN BRAMWELL

Address: Casilla 17-17-1004, Quito, Ecuador                   Tel : (593) (2) 372-748
Internet: [EMAIL PROTECTED]                                      or : (593) (9) 722-221


--
--------------------------------------------------------------
Please read the FAQ! <http://java.apache.org/faq/>
To subscribe:        [EMAIL PROTECTED]
To unsubscribe:      [EMAIL PROTECTED]
Archives and Other:  <http://java.apache.org/main/mail.html>
Problems?:           [EMAIL PROTECTED]

Reply via email to