Nic Ferrier wrote:
But it's difficult to change your mind if you say "prove that logging
is the most widely used logging available and then I'll think about it
but I don't use it anyway because I wrote my own".
Well, that's not what I said. What I said was:
1. I'm not convinced that a logger handler should be included in
mod_python just because it's a standard module. A lot of packages out
there that have logging facilities don't necessarily use logger. Even
though it's part of the Python distribtion, that doesn't make it
uniquitous in its use. mod_python exposes apache's logging facility in
using interface that apache provides. I think that's the goal of
mod_python, to (1) provide access to apache's API and (2) provide
dispatching for handlers.
2. My counterargument to the logging interface as being "essential" is
that I neither use the logging module nor want to use it. I tried to
use it, but the implementation required a lot of overhead (mostly due to
threading issues and access to the right req object). YMMV, but my
experience led me to create a thinner wrapper around apache.log_error
that was simply easier to use and maintain. That really has nothing to
do with whether or not
In terms of changing my mind, that has to do with whether or not I
believe it should be part of the core mod_python distribution. And I
still think it should not. I still think it falls under the umbrella of
contributed code, just as if it was an implementation that utilized
cgitb or SimpleXMLRPCServer, both of which are standard modules and
arguably pretty darn useful in mod_python development.
Please note that I also disagree with including the publisher and psp
handlers as part of the core distribution, but, well, we can't turn back
the clock on that ;-) And besides, I'm probably one of the few here who
don't use them, making my opinion on that matter a minority view.
Unfortunately I have to argue from the position that logging is
officially part of the python language.
Respectfully, I don't think that has anything to do with it. There's a
lot that can be done with only the official parts of the Python language
that aren't included in mod_python but left as an option for the user to
implement. Recently a bid to include a WSGI implementation for
mod_python was shot down, which is a Python PEP and something I thought
would have been pretty neat to include. Regardless of Grisha's personal
feelings on that (and I have a pretty good idea of what they are ;-)), I
think it was ultimately the right decision for mod_python.
Nick
