I would love to come up with a confidential test case.  I will see what
I can put together (I will have to scrub the data thoroughly - it is
cellphone traffic records).

To be fair, it does take a LOT of data to trigger this message.  My
project had been in production for over 2 years before first exhibiting
a DTM shortage.  More recently, I have heard elsewhere that the DTM
limitation shouldn't be an issue anymore, but that doesn't change the
fact that this problem was threatening the schedule of my project until
I did that change.

Could you point to any sections of code that are supposed to make this
happen?  I would like to have a look!



tlj

Timothy Jones - Sr. Systems Engineer, Development 
(813) 273-4743 - Syniverse Technologies 
The success of most things depends upon knowing how long it will take to
succeed.  
--Baron de Montesquieu, political thinker


-----Original Message-----
From: Joseph Kesselman [mailto:[EMAIL PROTECTED] 
Sent: Monday, February 06, 2006 10:15 AM
To: Timothy Jones
Cc: xalan-j-users@xml.apache.org
Subject: Re: No more DTM IDs


The quoted comment is somewhat outdated. Documents can now overflow from
one DTM ID to another, so the IDENT_DTM_NODE_BITS value is more a matter
of
efficiency than of actual document size limitation.

However: With current code, I'm surprised that you're actually managing
to
run out of DTM IDs unless you are using huge numbers of document() calls
or
doing extremely deep recursion of certain kinds of nested RTF
construction.
If you can come up with a reasonable-sized non-confidential testcase
which
demonstrates what you're doing to provoke the problem, it might be worth
posting that to Bugzilla so we can consider whether there's something we
need to do about it.


Regarding "why does it have to be a compile-time setting?" -- I believe
some Java compilers have fully inlined static-final compile-time
constants.
Hence, I would be concerned about making this reconfigurable at runtime;
I'd want to see performance tests on a number of platforms before I'd be
really comfortable with changing it to be a system property.

And as mentioned above, I find myself wondering whether this is a matter
of
patching a symptom rather than fixing the actual disease.

______________________________________
Joe Kesselman -- Beware of Blueshift!
"The world changed profoundly and unpredictably the day Tim Berners Lee
got bitten by a radioactive spider." -- Rafe Culpin, in r.m.filk

Reply via email to