*** bskahan has quit IRC | 00:10 | |
*** BjornT has quit IRC | 00:11 | |
*** bradb has quit IRC | 00:17 | |
*** jdz_ has quit IRC | 00:18 | |
*** mgedmin has quit IRC | 00:22 | |
*** srichter has quit IRC | 00:23 | |
*** bskahan has joined #zope3-dev | 00:45 | |
*** GaryPoster has quit IRC | 00:49 | |
*** Aiste has quit IRC | 00:54 | |
*** Theuni has quit IRC | 00:56 | |
*** hazmat has quit IRC | 00:59 | |
*** deo|food is now known as deo | 00:59 | |
*** srichter has joined #zope3-dev | 01:08 | |
*** ChanServ sets mode: +o srichter | 01:09 | |
*** BjornT has joined #zope3-dev | 01:13 | |
*** jhauser_ has joined #zope3-dev | 01:20 | |
*** alga has quit IRC | 01:25 | |
*** stub has joined #zope3-dev | 01:36 | |
*** jhauser has quit IRC | 01:36 | |
*** philiKON has joined #zope3-dev | 01:36 | |
*** philiKON has quit IRC | 01:42 | |
*** benji_york has quit IRC | 01:53 | |
*** niemeyer has quit IRC | 01:58 | |
*** hazmat has joined #zope3-dev | 02:03 | |
*** tarek__ has quit IRC | 02:12 | |
*** tarek_ has joined #zope3-dev | 02:19 | |
*** ignas has quit IRC | 02:22 | |
*** ignas_ is now known as ignas | 02:22 | |
*** ignas_ has joined #zope3-dev | 02:22 | |
*** J1m has quit IRC | 02:32 | |
*** hazmat has quit IRC | 02:38 | |
*** hazmat has joined #zope3-dev | 03:03 | |
*** WebMaven_ has quit IRC | 03:16 | |
*** GaryPoster has joined #zope3-dev | 03:27 | |
*** GaryPoster has quit IRC | 03:30 | |
*** GaryPoster has joined #zope3-dev | 03:55 | |
*** GaryPoster has quit IRC | 04:06 | |
*** BjornT has quit IRC | 05:31 | |
*** bradb has joined #zope3-dev | 06:25 | |
*** srichter has quit IRC | 06:30 | |
*** srichter has joined #zope3-dev | 07:24 | |
*** bskahan has quit IRC | 07:51 | |
*** d2m has quit IRC | 08:15 | |
*** Fosten has joined #zope3-dev | 08:43 | |
*** Fosten has left #zope3-dev | 08:43 | |
*** d2m has joined #zope3-dev | 08:48 | |
*** `anthony has quit IRC | 08:50 | |
*** hdima has joined #zope3-dev | 09:23 | |
*** Theuni has joined #zope3-dev | 10:07 | |
*** sashav has joined #zope3-dev | 10:14 | |
*** tarek_ has quit IRC | 10:16 | |
*** Aiste has joined #zope3-dev | 10:40 | |
*** tav|offline has quit IRC | 10:48 | |
*** tav|offline has joined #zope3-dev | 10:49 | |
projekt01 | Is anybody using the PAU and has a setup with sites and subsites? | 11:12 |
---|---|---|
projekt01 | I think there is now way to login to a subsite I the first site denies the zope.View. | 11:13 |
projekt01 | The first site is allways offering the login even if I point to the subsites URL | 11:14 |
projekt01 | Is this correct? | 11:14 |
projekt01 | It's not the PAU which forces this, it's the ZopePublication. This means it's the default way. | 11:16 |
projekt01 | My question, what can I do if I like to offer a login directly an the subsite? | 11:17 |
*** MalcolmC has joined #zope3-dev | 11:25 | |
*** Aiste has quit IRC | 11:31 | |
*** SteveA has quit IRC | 11:41 | |
*** stub has left #zope3-dev | 11:43 | |
*** SteveA has joined #zope3-dev | 11:44 | |
*** Aiste has joined #zope3-dev | 11:51 | |
*** __gotcha has joined #zope3-dev | 11:54 | |
*** __gotchaway has quit IRC | 11:55 | |
*** Aiste has quit IRC | 11:59 | |
*** sashav has quit IRC | 12:06 | |
*** dlk has joined #zope3-dev | 12:09 | |
*** ignas has quit IRC | 12:27 | |
*** ignas has joined #zope3-dev | 12:28 | |
*** ignas has quit IRC | 12:28 | |
*** ignas_ is now known as ignas | 12:28 | |
*** ignas_ has joined #zope3-dev | 12:28 | |
*** ChanServ sets mode: +o srichter | 12:36 | |
*** j-w has joined #zope3-dev | 12:38 | |
*** watzo has joined #zope3-dev | 12:38 | |
*** admp has joined #zope3-dev | 12:46 | |
*** tarek_ has joined #zope3-dev | 13:03 | |
*** SteveA has quit IRC | 13:08 | |
*** SteveA has joined #zope3-dev | 13:17 | |
*** sashav has joined #zope3-dev | 13:34 | |
*** watzo has quit IRC | 13:34 | |
*** bradb has quit IRC | 13:34 | |
*** tvon has quit IRC | 13:34 | |
*** watzo has joined #zope3-dev | 13:36 | |
*** bradb has joined #zope3-dev | 13:36 | |
*** tvon has joined #zope3-dev | 13:36 | |
*** admp has quit IRC | 13:40 | |
*** bradb has quit IRC | 13:40 | |
*** Aiste has joined #zope3-dev | 13:43 | |
*** mgedmin has joined #zope3-dev | 13:45 | |
*** sashav_ has joined #zope3-dev | 13:46 | |
*** dlk has left #zope3-dev | 13:51 | |
*** sashav has quit IRC | 13:53 | |
*** BjornT has joined #zope3-dev | 13:56 | |
*** bradb has joined #zope3-dev | 13:56 | |
*** MrTopf has joined #zope3-dev | 14:08 | |
MrTopf | hi | 14:08 |
srichter | MrTopf: hi, long time, no see :-) | 14:27 |
srichter | I will be back online in 30-40 mins | 14:27 |
*** srichter has quit IRC | 14:27 | |
*** __gotcha has quit IRC | 14:49 | |
*** SteveA__ has joined #zope3-dev | 15:01 | |
*** SteveA__ is now known as SteveA_ | 15:02 | |
*** SteveA has quit IRC | 15:02 | |
*** SteveA_ is now known as SteveA | 15:02 | |
*** srichter has joined #zope3-dev | 15:04 | |
*** ChanServ sets mode: +o srichter | 15:06 | |
*** `anthony has joined #zope3-dev | 15:06 | |
*** sashav_ has quit IRC | 15:07 | |
*** stub has joined #zope3-dev | 15:11 | |
projekt01 | srichter, Do you have time to answer a conceptual question? | 15:15 |
srichter | yep | 15:15 |
projekt01 | I will write a mail to the zope3-dev list | 15:16 |
projekt01 | It's about traverser and subsites. | 15:16 |
srichter | ok | 15:16 |
projekt01 | Should arrive | 15:18 |
ignas | what is the common speed of processing contributors agreement ? | 15:19 |
ignas | i have sent one like 2 weeks ago ... | 15:20 |
projekt01 | ignas, I think jim is very bussy, Did you send it him directly? | 15:20 |
ignas | by email | 15:20 |
ignas | yes | 15:20 |
projekt01 | Oh, did you send it to jim at zope dot com | 15:21 |
ignas | jim ar zope org | 15:21 |
ignas | s/ar/at | 15:21 |
projekt01 | Try at zope.com | 15:21 |
srichter | yeah. zope.com is the right onwe | 15:22 |
srichter | he will do it almost immediately, so it only should take a couple of hours, once he is in the office in about an hour | 15:22 |
ignas | won't he be angry about geting the scan twice ? (it's heavy) | 15:22 |
srichter | naeh | 15:23 |
srichter | do a lower res scan | 15:23 |
srichter | or make it a gray scale image, which should make it small | 15:23 |
ignas | srichter, it's indexed png already | 15:24 |
ignas | just sent it so it's too late for pngcrush | 15:25 |
*** niemeyer has joined #zope3-dev | 15:31 | |
*** deo has quit IRC | 15:33 | |
*** stub has left #zope3-dev | 15:52 | |
*** SureshZ has joined #zope3-dev | 15:57 | |
*** J1m has joined #zope3-dev | 16:01 | |
*** SteveA__ has joined #zope3-dev | 16:21 | |
projekt01 | srichter, did you understand the problem I described in the mail? | 16:22 |
srichter | I have not looked; I thought you changed your mind | 16:22 |
srichter | hold on | 16:23 |
projekt01 | Ok, thanks | 16:23 |
*** vlado has joined #zope3-dev | 16:25 | |
*** mkerrin has joined #zope3-dev | 16:26 | |
srichter | projekt01: your analysis is very good; I just do not have an answer | 16:27 |
srichter | J1m is the security expert | 16:27 |
projekt01 | There are many other problems in the situation right now | 16:28 |
srichter | mmh, this is really tricky too | 16:28 |
projekt01 | We have a Allow/Deny concept whcih doesn't work with subsites | 16:28 |
projekt01 | Once you denied you get restricted the the parent site's PAU login concept | 16:29 |
projekt01 | An hang arround there | 16:29 |
srichter | J1m: have you looked at Roger's traversal question mail? It raises a very interesting security issue; I do not know the answer | 16:29 |
projekt01 | But your login info is below | 16:29 |
projekt01 | Yu never reach the subsites Pau again | 16:29 |
mgedmin | projekt01, I have a real-world metaphor for your situation | 16:29 |
mgedmin | imagine a building with a locked door | 16:29 |
mgedmin | imagine an office inside the building, with another locked door | 16:29 |
mgedmin | to get into the office, you need two keys | 16:30 |
mgedmin | in your example, you need to log into the outer site, and then into the inner site | 16:30 |
srichter | mgedmin: very good anology | 16:30 |
mgedmin | I think you can grant the necessary permissions to make the outer site traversable by anonymous users | 16:30 |
mgedmin | (leave the main building door unlocked) | 16:30 |
J1m | Yes | 16:31 |
J1m | This is right | 16:31 |
mgedmin | I think you should be able to grant those permissions while denying anonymous users view access to the outer site | 16:31 |
J1m | You can grant access to the traversal adapters. | 16:32 |
J1m | That should be enough. | 16:32 |
projekt01 | Which traversal adapter? All? | 16:33 |
J1m | Jusr the publication traversal adapters. | 16:33 |
projekt01 | In the core? | 16:33 |
* J1m looking it up | 16:34 | |
projekt01 | Why not use trusted ITraversal adapters? | 16:34 |
J1m | <adapter | 16:35 |
J1m | for="zope.app.container.interfaces.IItemContainer | 16:35 |
J1m | zope.publisher.interfaces.browser.IBrowserRequest" | 16:35 |
J1m | provides="zope.publisher.interfaces.browser.IBrowserPublisher" | 16:35 |
J1m | factory="zope.app.container.traversal.ItemTraverser" | 16:35 |
J1m | permission="zope.Public" | 16:35 |
J1m | trusted='1' | 16:35 |
J1m | /> | 16:35 |
J1m | <adapter | 16:35 |
J1m | for="zope.app.container.interfaces.ISimpleReadContainer | 16:35 |
J1m | zope.publisher.interfaces.browser.IBrowserRequest" | 16:35 |
J1m | provides="zope.publisher.interfaces.browser.IBrowserPublisher" | 16:35 |
J1m | factory="zope.app.container.traversal.ItemTraverser" | 16:35 |
J1m | permission="zope.Public" | 16:35 |
J1m | trusted='1' | 16:35 |
J1m | /> | 16:35 |
J1m | You do this in an overrides zcml file for your application. | 16:35 |
J1m | You don't modify the zope configuration. | 16:35 |
projekt01 | Can we change this later in the core? | 16:36 |
projekt01 | If it's working | 16:36 |
J1m | It works, we use it in our projects. | 16:36 |
mgedmin | projekt01, I imagine that there can be a security requirement that anonymous users must not distinguish between URLs that do not exist and URLs that exist but are password protected | 16:36 |
mgedmin | I don't know any actual use cases, though | 16:37 |
projekt01 | I'm sure nobody is using subsites. Right? | 16:37 |
J1m | projekt01, I'd have to ponder whether this should be a default. | 16:38 |
projekt01 | Right now we have a two step login with subsites. Is this the case we like to have. (It's also OK). For me it's just important to understand the concept. | 16:39 |
J1m | No, we should not have a 2-step login. | 16:39 |
projekt01 | Like mgedmin says you need all keys for the building | 16:39 |
projekt01 | J1m, this is the case we have now. | 16:40 |
J1m | I'll note, that eventually, I want to change the publication process so that a system can have multiple rots. | 16:40 |
J1m | projekt01, not with those adapter declarations. | 16:40 |
J1m | And only because you haven't given everyone view in the root | 16:40 |
J1m | If we had multiple roots, then each site would be it's own root and you wouldn't have to traverse a top site to get to it. | 16:41 |
J1m | I think School Tool did something like this. | 16:41 |
J1m | The new publication object on my Bobo branch would allow this. | 16:41 |
projekt01 | Not to the root, I denied the zope.View permission on the first site. | 16:42 |
projekt01 | J1m, I don't think your proposal will work, the method _maybePlacefullyAuthenticate in ZopePublication tries also to authenticate on each ISite | 16:42 |
J1m | Yeah, so what | 16:43 |
J1m | On the top site, authentication will fail and an unauthenticated user will be returned. | 16:43 |
projekt01 | This forces also the login on the first site | 16:43 |
mgedmin | perhaps there should be two different permissions -- zope.View and zope.Traverse? | 16:44 |
mgedmin | then projekt01 could just allow traversal for the outer site but disallow views | 16:44 |
projekt01 | Yup, I was thinking about this today:-) | 16:44 |
*** SteveA has quit IRC | 16:44 | |
projekt01 | I think travers and view are different parts | 16:44 |
projekt01 | J1m, perhaps we can put the concept of the method _maybePlacefullyAuthenticate into a event. This can be overriden. | 16:46 |
*** regebro has joined #zope3-dev | 16:46 | |
*** tvon has quit IRC | 16:46 | |
projekt01 | J1m, perhaps the BeforeTraverseEvent and AfterTravereEvent is the right place to do _maybePlacefullyAuthenticate() | 16:47 |
*** regebro has quit IRC | 16:52 | |
*** BjornT has quit IRC | 16:52 | |
J1m | projekt01, yes, eventually it will be done in an event that fires just before an object is traversed. | 16:55 |
J1m | That is, effectively, what happens now. | 16:55 |
J1m | The main reason to move it out to an event will be to make the publisher leaner. | 16:55 |
projekt01 | How can I get rid of the method _maybePlacefullyAuthenticate() right now? | 16:57 |
J1m | why do you want to? | 16:57 |
J1m | You can't without replacing the publication object. | 16:57 |
projekt01 | It forces a login on the site and stop the traverser travers to the subsite. | 16:57 |
J1m | But I don't recommend that. | 16:57 |
J1m | No, it doesn't. | 16:57 |
J1m | It tries to authenticate. | 16:58 |
*** regebro has joined #zope3-dev | 16:58 | |
J1m | It does not issue a challenge unless you try to acess something that is not public. | 16:58 |
J1m | ]If you make the traversers public, as I suggested, then you won't have a problem. | 16:58 |
J1m | If you make the traversers public, as I suggested, then you won't have a problem. | 16:59 |
J1m | Alternatively, you can grant zope.View for everyone in the root site and deny it in the subsite. | 16:59 |
projekt01 | Our site is more then just a login hook. I can't give zope.View permissions there. | 17:00 |
J1m | OK, then make the traverses public. | 17:00 |
mgedmin | J1m, the other day I was thinking about noninheritable grants | 17:00 |
mgedmin | you can't just say "I want zope.View for this container, but not for objects contained within it" | 17:00 |
J1m | Or you can replace the publication object if you insist. | 17:01 |
J1m | mgedmin, in our customer projects, our grants are mostly non-acquireable. | 17:01 |
mgedmin | how do you do that? | 17:01 |
mgedmin | a different security policy? | 17:01 |
J1m | (That can only be acquired by objects that can't have their own grants.) | 17:01 |
J1m | Yes, we are using a different security policy. | 17:02 |
mgedmin | ah | 17:02 |
mgedmin | but security policy is global -- you can't switch to a different one in a site | 17:02 |
*** SureshZ has quit IRC | 17:02 | |
J1m | right, not currently. | 17:02 |
mgedmin | that means, if you want multiple apps to interoperate in a single zope instance, you have to use the default security policy | 17:02 |
J1m | You have to use the same security policy, yes. | 17:03 |
J1m | (you don't have to use the default one.) | 17:03 |
J1m | Eventually, I'd like to change this. | 17:03 |
*** bskahan has joined #zope3-dev | 17:03 | |
J1m | You could actually change this yourself now. | 17:03 |
J1m | The security policy is really just an interaction factory. | 17:04 |
J1m | You could create a security policy that looks up an interaction factory. | 17:04 |
projekt01 | Is it not a better solution to Deny in the root by default and Allow in a site as the default concept. | 17:05 |
J1m | That depends on your application. | 17:06 |
projekt01 | Why? Which usecase can't be done with this pattern | 17:06 |
*** tvon has joined #zope3-dev | 17:06 | |
*** hdima has quit IRC | 17:06 | |
regebro | Q: Ehm, OK, so I have an addform, right. But I want one of the fields to have a different default depending on what object it is created in. How do I do that? | 17:06 |
regebro | Something in update() perhaps? | 17:07 |
J1m | You can provide a class that overrides _setUpWidgets. | 17:08 |
regebro | OK, I'll check into that. | 17:08 |
J1m | Call the base than call setRenderedValue on the desired widget. | 17:09 |
J1m | projekt01, maybe I want the top site to be public and use subsites for member functions. | 17:10 |
*** tvon has quit IRC | 17:10 | |
*** tvon has joined #zope3-dev | 17:10 | |
projekt01 | Jim, this can be done in the principals.zcml with the right configuration, right? | 17:11 |
*** SureshZ has joined #zope3-dev | 17:11 | |
J1m | Can what be done? | 17:11 |
projekt01 | Give zope.View permission to Unauthenticated principals. | 17:12 |
projekt01 | Then you can access the top site by default | 17:13 |
*** tvon has quit IRC | 17:15 | |
projekt01 | I see your solution with the trusted adapter, but I'm scary about to use a non standard way. If we whould change the default way (Deny by default and Allow optional) it's much more clear what to do. | 17:15 |
*** tvon has joined #zope3-dev | 17:15 | |
J1m | Right, you could give zope.View to Unauthorized. I suggested that earlier. That is an option. | 17:16 |
J1m | Zope 3 is meant to be pluggable. I see nothing scary about overriding component registrations as long as you understand what's going on. | 17:17 |
J1m | I'll note that, in our app, we want to allow traversal to any object, even if we have to traverse objects that a user doesn't have other access to. | 17:18 |
projekt01 | Yup | 17:18 |
projekt01 | Not just in our app. Right now the zope server offers a 2 step login for subsites. | 17:19 |
projekt01 | This is the default concept right now. And I see you don't like that. Right? | 17:19 |
J1m | That depends on how it is configures. | 17:20 |
J1m | If you grant View to the unauthenticated user or Everybody, you won't get a two-step login. | 17:21 |
projekt01 | Yes everything is changable, but that's the default if you stat the server, and you have to register trusted adapters in a override.zcml. | 17:21 |
projekt01 | I'm wondering if this can be done form scripters? | 17:21 |
J1m | No, this can't be done by scriptors today. | 17:21 |
J1m | Gotta go | 17:21 |
*** J1m is now known as J1m|away | 17:22 | |
*** BjornT has joined #zope3-dev | 17:22 | |
*** mkerrin has quit IRC | 17:22 | |
*** tvon|x31 has joined #zope3-dev | 17:24 | |
*** tvon has quit IRC | 17:25 | |
*** zagy has quit IRC | 17:30 | |
*** mgedmin has quit IRC | 17:33 | |
*** zagy has joined #zope3-dev | 17:35 | |
*** tvon has joined #zope3-dev | 17:40 | |
*** tvon|x31 has quit IRC | 17:40 | |
*** tarek_ has quit IRC | 18:12 | |
*** tarek_ has joined #zope3-dev | 18:15 | |
*** SteveA__ is now known as SteveA | 18:18 | |
regebro | Well J1m|away is away, but he still rocks, and so does Zope3. The widget update thing works, and now I get different default days depending on which day I am viewing when I press "Add events". This calendar application will rock! (In a month or two). | 18:30 |
*** Theuni has quit IRC | 18:32 | |
*** zagy has quit IRC | 18:36 | |
*** regebro has quit IRC | 18:37 | |
*** Theuni has joined #zope3-dev | 18:37 | |
*** mgedmin has joined #zope3-dev | 18:38 | |
projekt01 | mgedmin, thanks for the building/key sample ;-) | 18:42 |
*** j-w has left #zope3-dev | 18:49 | |
mgedmin | I'm glad it was helpful | 18:55 |
projekt01 | What do you think, is the standard way right now OK, or should we workout another securitypolicy? I'm not sure what's Jim's ideas for this. | 18:56 |
mgedmin | it's a question that requires much thought | 18:57 |
mgedmin | I am not prepared to answer now | 18:57 |
projekt01 | Perhaps we should start a proposal with usecases and show different server configurations. | 18:58 |
mgedmin | yes | 18:58 |
projekt01 | Later we can find a good "everybody fit" solution | 18:58 |
mgedmin | perhaps the default security policy should use adapters | 18:58 |
mgedmin | ICheckPermission(object).hasPermission(principal, permission_id) or something | 18:59 |
mgedmin | I believe Steve uses something like that in Launchpad | 18:59 |
projekt01 | At least it should be replaceable. | 19:00 |
projekt01 | What's Launchpad? | 19:00 |
mgedmin | https://launchpad.ubuntu.com/index.html | 19:01 |
projekt01 | thanks | 19:02 |
*** SureshZ has left #zope3-dev | 19:04 | |
*** SureshZ has joined #zope3-dev | 19:07 | |
*** watzo has quit IRC | 19:10 | |
*** deo has joined #zope3-dev | 19:10 | |
*** tvon|x31 has joined #zope3-dev | 19:13 | |
*** tvon has quit IRC | 19:13 | |
*** bskahan has quit IRC | 19:18 | |
*** tarek_ has quit IRC | 19:24 | |
*** MrTopf has quit IRC | 19:29 | |
*** zagy has joined #zope3-dev | 19:40 | |
*** Aiste has quit IRC | 19:41 | |
*** Aiste has joined #zope3-dev | 19:44 | |
*** watzo has joined #zope3-dev | 19:59 | |
*** SureshZ has quit IRC | 20:08 | |
*** MalcolmC has quit IRC | 20:10 | |
*** mohsen has joined #zope3-dev | 20:27 | |
*** bskahan has joined #zope3-dev | 20:40 | |
*** vlado has quit IRC | 21:00 | |
*** sashav has joined #zope3-dev | 21:02 | |
*** ignas is now known as ignas|away | 21:16 | |
*** tvon|x31 has quit IRC | 21:49 | |
*** sashav has quit IRC | 21:56 | |
*** alga has joined #zope3-dev | 22:18 | |
*** mohsenX has joined #zope3-dev | 22:56 | |
*** jhauser_ has quit IRC | 23:10 | |
*** bradb has quit IRC | 23:11 | |
*** zagy has quit IRC | 23:12 | |
*** hazmat has quit IRC | 23:13 | |
*** bskahan has quit IRC | 23:16 | |
*** tvon has joined #zope3-dev | 23:19 | |
*** srichter has quit IRC | 23:26 | |
*** jhauser has joined #zope3-dev | 23:31 | |
*** bskahan has joined #zope3-dev | 23:33 | |
*** jhauser has joined #zope3-dev | 23:34 | |
*** hazmat has joined #zope3-dev | 23:37 | |
*** srichter has joined #zope3-dev | 23:40 | |
*** ChanServ sets mode: +o srichter | 23:41 | |
*** jhauser has quit IRC | 23:56 | |
*** bradb has joined #zope3-dev | 23:58 | |
*** mgedmin has quit IRC | 23:59 |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!