Post-Traversal Hook (+Post-Publication Hook)
Hi, I'm cleaning up the alternative BaseRequest for Zope 2.7 right now which implements a post traversal hook (which will go into 2.8) and discovered a small issue I would like to get some feedback from. Originally the patch called all registered methods after the user has been determined, but before eventually unauthorized is thrown. Therefore the SecurityManager already could deliver the user object, but it is not available within the REQUEST as the well known "AUTHENTICATED_USER". Anyway, does anybody see a special reason to call the hooks a) before unauthorized b) after complete successful authorization I see that there could be two hooks make sense in there, but can't decide for a single one. (My tummy tells me to go for the one after complete authentication, so we have the complete REQUEST environment as in opposition to site access rules.) Also Dario Lopez-Kaesten came up with a request for a hook that is called after the publication. Is there anybody objecting such a "post publication hook"? Otherwise I would go for - implementing post-traversal variant b - implementing a post-publication hook Cheers, Christian PS: Looks a bit like I should write a proposal on this ...? -- Christian Theune, gocept gmbh & co. kg http://www.gocept.com - ct@gocept.com fon: 03496 3099112 fax: 03496 3099118 mobile: 0179 7808366
what does publication mean in this context? Robert Am Freitag, 17. Oktober 2003 11:05 schrieb Christian Theune:
Hi,
I'm cleaning up the alternative BaseRequest for Zope 2.7 right now which implements a post traversal hook (which will go into 2.8) and discovered a small issue I would like to get some feedback from.
Originally the patch called all registered methods after the user has been determined, but before eventually unauthorized is thrown. Therefore the SecurityManager already could deliver the user object, but it is not available within the REQUEST as the well known "AUTHENTICATED_USER".
Anyway, does anybody see a special reason to call the hooks
a) before unauthorized b) after complete successful authorization
I see that there could be two hooks make sense in there, but can't decide for a single one. (My tummy tells me to go for the one after complete authentication, so we have the complete REQUEST environment as in opposition to site access rules.)
Also Dario Lopez-Kaesten came up with a request for a hook that is called after the publication. Is there anybody objecting such a "post publication hook"? Otherwise I would go for
- implementing post-traversal variant b - implementing a post-publication hook
Cheers, Christian
PS: Looks a bit like I should write a proposal on this ...?
-- mit freundlichen Grüssen Robert Rottermann www.redCOR.ch
Am Fr, 2003-10-17 um 13.38 schrieb robert:
what does publication mean in this context?
I'm referring to the point in time after the method that was addressed by the request (e.g. index_html or __call__) was callen but hasn't been given back yet. So after the result of the request is known. Cheers, Christian -- Christian Theune, gocept gmbh & co. kg http://www.gocept.com - ct@gocept.com fon: 03496 3099112 fax: 03496 3099118 mobile: 0179 7808366
On Fri, Oct 17, 2003 at 01:46:08PM +0200, Christian Theune wrote:
Am Fr, 2003-10-17 um 13.38 schrieb robert:
what does publication mean in this context?
I'm referring to the point in time after the method that was addressed by the request (e.g. index_html or __call__) was callen but hasn't been given back yet. So after the result of the request is known.
hmm... the only issue that occurs to me is that the method could take a long time to complete, whereas pre-publication should generally be very quick. Don't know if that matters except when debugging in a hurry. -- Paul Winkler http://www.slinkp.com Look! Up in the sky! It's THE NOISOME BATMAN! (random hero from isometric.spaceninja.com)
On Fri, Oct 17, 2003 at 12:15:07PM -0400, Paul Winkler wrote:
On Fri, Oct 17, 2003 at 01:46:08PM +0200, Christian Theune wrote:
Am Fr, 2003-10-17 um 13.38 schrieb robert:
what does publication mean in this context?
I'm referring to the point in time after the method that was addressed by the request (e.g. index_html or __call__) was callen but hasn't been given back yet. So after the result of the request is known.
hmm... the only issue that occurs to me is that the method could take a long time to complete, whereas pre-publication should generally be very quick. Don't know if that matters except when debugging in a hurry.
duh, never mind, I didn't really read the end of your first message, in which you clearly propose two hooks, one before and one after publication. Carry on :-) -- Paul Winkler http://www.slinkp.com Look! Up in the sky! It's FANTASTIC BURRITO MAFIOSO! (random hero from isometric.spaceninja.com)
participants (3)
-
Christian Theune -
Paul Winkler -
robert