On Jul 29, 8:19 am, Malcolm Tredinnick <[EMAIL PROTECTED]> wrote: > On Fri, 2007-07-27 at 05:11 +0000, SmileyChris wrote: > > Sorry bout the unfinished second half of that last message, I was just > > thinking through type and I meant to delete it before posting. > > > On Jul 27, 4:16 pm, Graham Dumpleton <[EMAIL PROTECTED]> > > wrote: > > > The other problem is that although with the better WSGI adapters > > > SCRIPT_NAME is provided correctly set for you automatically, in > > > mod_python one cannot actually deduce SCRIPT_NAME automatically with > > > the way the adapter works and so you push onto the user the need to > > > define the mount point manually a second time. > > > I get that mod_python is a bit dumb, but my change isn't changing > > anything for mod_python users. It only changes the behaviour of the > > wsgi request handler. > > > > What it comes down to is there are a few sides issues that really need > > > to be looked at at the same time and the simplistic change may not be > > > adequate to cover these or not result in the same behaviour on all > > > hosting solutions. > > > Granted, the one thing I hadn't put too much consideration into is url > > `reverse` method. At the moment, it is totally decoupled from request, > > so it can't figure out SCRIPT_NAME. Perhaps request could get a > > SCRIPT_NAME aware reverse method (which would still work without it)? > > > Are there any other issues I'm missing? > > Coming in a bit late to this thread because I'm catching up on a back > log today. Reverse URL construction (especially w.r.t things like HTTP > redirects) is the big missing part. I was talking about this with a > couple of people at OSCON, including discussing with Jacob how we can > incorporate this so that the necessary backwards incompatibility changes > aren't too painful. I've got a similar patch in one of my branches > locally that I'll compare with your patch Chris and grab any bits I've > missed. It'll be a couple of days before I'm all set up to test this > correctly, since I'm travelling and reinstalled my laptop shortly before > leaving, so I need to configure Apache, etc, first. > > Haven't worked out exactly how to acommodate things like mod_python not > giving the full SCRIPT_NAME yet, but it's a relatively minor issue since > it's just a fact of life we need to work around. Something will present > itself. > > We do need to fix this, so thanks for the patch so far. We might as well > fix the whole thing at once so that it's easier to move an existing > setup to a different URL root without having to edit all the code > (letting Apache handle the SCRIPT_NAME portion). Initially there will be > a settings variable that will preserve backwards-compat and we'll break > it a bit before 1.0 after everybody's had time to adjust their code. So > it will be relatively painless.
FWIW, I know I said that SCRIPT_NAME can't be calculated in mod_python. Well, that isn't strictly true. If you are using mod_python 3.3.1 and have used the Location directive then it is probably possible by using req.hlist.location. I didn't mention this before because the fact it doesn't work with older versions of mod_python nor when Location is not being used sort of limits its usefulness. Didn't seem to make sense to therefore depend on it. Graham --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Django developers" 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/django-developers?hl=en -~----------~----~----~----~------~----~------~--~---
