On Wed, Jul 29, 2009 at 2:24 PM, Michael Bayer<[email protected]> wrote:
>
> Jon Nelson wrote:
>> I assume that the purpose of this is to save performance?
>
> correct.
>
>> I note that the way much of the logging (esp. in pool.py) seems to be
>> written is like this:
>>
>> somelogger.log("String ..." % (var1,var2,var3))
>> and of course, regardless of the loglevel specified this always
>> interpolates the variables - a fairly significant performance penalty
>> for sure.
>
> well no, because all the log statements are underneath conditionals and
> are not evaluated otherwise.
Right. I knew that but *should* have said that /in absence of the
conditionals/ you will *always* pay the performance penalty of
interpolation. Even if we start out with loglevel=INFO but the log
statement is, say, DEBUG we still pay the interpolation penalty (for
things like repr or %r, easily far more expensive than the method
overhead).
>> Shouldn't the style be:
>>
>> somelogger("String ...", var1,var2,var3)
>>
>> which prevents interpolation unless the loglevel is actually
>> sufficient?
>
> yes, it should. patches are welcome, but again not a big deal.
>
> Would that not obviate the need for _should_log_info (and
>> _should_log_debug) ??
>
> unlikely. The conditionals save an enormous number of method calls into
> the logging package in any case. Python seems to incur an inordinate
> amount of overhead for method calls.
>
>> I'm willing to supply a patch that changes the format of the calls to
>> the logging system if that's useful.
>
> sure ! if you are curious, try running the tests test_zoomark and
> test_zoomark_orm to see the method call overhead added by even null
> logging.
OK. Sweet. Assuming I have time, of course.
--
Jon
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"sqlalchemy" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/sqlalchemy?hl=en
-~----------~----~----~----~------~----~------~--~---
