[stats-dev] Debugging Flags

Benjamin Gavin virtual_olympus at yahoo.com
Tue Nov 25 14:13:21 EST 2003


OK,
  I'm starting to look at the debug stuff on the stats pages.  What we
need to come up with is a fairly exhaustive list of the things we would
like to debug, how we would like the information displayed, etc.  My
current thoughts are that we would implement a bit-flag style system so
that you could easily turn on/off the debugging information that you were
interested in.  The flags we current have are:

DEFINE('DEBUG_TIME', 1);
DEFINE('DEBUG_TRACE', 2);
DEFINE('DEBUG_PROFILE', 4);
DEFINE('DEBUG_SHOW_QUERY', 8);

I can see some additional things that might be useful, for example:

DEBUG_TIME_QUERY
DEBUG_DISABLE_CACHE (when/if caching is implemented)

Currently, all these things are defined in config.inc.  The general
consensus is that we should place all the debugging pieces into debug.inc
and possibly include that in config.inc.

Another current problem that we have is in the construction of URLs.  In
order to implement debugging through the URL, we'd need to go in and
modify the application in hundreds of places where URLs are generated. 
Instead, I would prefer that we go with a cookie-based approach with a
entry-point page which handles setting the cookie to the appropriate value
(maybe a form with checkboxes)?  This would allow us to go to
http://stats-<blah>.distributed.net/debug.php and be presented with a page
which allows us to select the debugging features that we would like to
enable for the current session.  Then any pages we hit after that point
would pick up the debugging information from the cookie and use that to
initialize the debugging infrastructure.

Thoughts, comments, etc?

Thanks!
Ben [TheJet]


=====
/******************************/
Benjamin Gavin
virtual.olympus software
ben at virtual-olympus.com
http://www.virtual-olympus.com/
/******************************/

__________________________________
Do you Yahoo!?
Free Pop-Up Blocker - Get it now
http://companion.yahoo.com/


More information about the stats-dev mailing list