hartig commented on PR #2501:
URL: https://github.com/apache/jena/pull/2501#issuecomment-2309447553

   Thanks Andy! Let me know if you need my help with anything.
   
   There is one thing remaining from the discussion above: [You mentioned the 
following performance-related 
concerns](https://github.com/apache/jena/pull/2501#issuecomment-2156169828) 
regarding the `CDTAwareParserProfile` introduced in this PR.
   > There is one place that needs investigation that I have found so far which 
is the mixes this - `CDTAwareParserProfile` is on the critical path for all 
parsing whether CDT's are used or not.
   >
   > The subclass quickly tests whether to invoke the superclass but parsing is 
sensitive to the JIT ptimizer. Parsing runs at about one/two microseconds per 
triple which isn't that many instructions and the new may
   change what optimizations the JIT performs.
   >
   > It may well make no measurable difference because the optimizer nowadays 
has ways to implement virtual method calls as direct function calls. The only 
way to find out is to try.
   
   Related to this, [I had a few 
proposals](https://github.com/apache/jena/pull/2501#issuecomment-2185172583), 
which I am repeating here again:
   
   Would it be an option to move the code of `CDTAwareParserProfile` into 
`ParserProfileStd` and, then, remove `CDTAwareParserProfile` altogether?
   
   Another option may be to enable users to switch off support for CDT literals 
(e.g., via an argument in `ModLangParse`, similar to `--strict`). If switched 
off, `RiotLib` would create a `ParserProfileStd` instance instead of a 
`CDTAwareParserProfile` instance.
   
   Yet another option may also be to combine both of these ideas. What do you 
think?


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: pr-unsubscr...@jena.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: pr-unsubscr...@jena.apache.org
For additional commands, e-mail: pr-h...@jena.apache.org

Reply via email to