| *** pcardune has quit IRC | 00:03 | |
| benji | if indented == demented, then I applaud jodok's turning of J1m's psycopathy against him | 00:06 |
|---|---|---|
| J1m | :) | 00:07 |
| jodok | :) | 00:07 |
| J1m | You can't turn it against me. | 00:07 |
| J1m | It only entertains me. | 00:07 |
| benji | I forgot, you're imune to sanity. | 00:07 |
| J1m | That's why benji will never make me stop. | 00:08 |
| *** tarek has joined #zope3-dev | 00:08 | |
| * benji cups his hand to his ear to hear J1m's distant laugh of "mwa! ha! ha!" | 00:09 | |
| *** greenman has quit IRC | 00:09 | |
| *** pcardune has joined #zope3-dev | 00:18 | |
| *** rocky is now known as rocky|away | 00:19 | |
| *** pelle_ has joined #zope3-dev | 00:24 | |
| *** pelle_ has quit IRC | 00:24 | |
| *** dobee has quit IRC | 00:32 | |
| *** cbcunc has quit IRC | 00:34 | |
| *** b52lap has quit IRC | 00:39 | |
| *** b52lap has joined #zope3-dev | 00:40 | |
| *** benji has quit IRC | 00:41 | |
| *** greenman has joined #zope3-dev | 00:54 | |
| *** jodok has quit IRC | 01:00 | |
| *** lucielejard has quit IRC | 01:03 | |
| *** yota has quit IRC | 01:06 | |
| *** dunny has joined #zope3-dev | 01:08 | |
| rocky|away | does anyone know ... is there anything zope3 does to ensure that datetime.datetime.now() returns a datetime instance with a proper tzinfo set ? | 01:12 |
| J1m | no | 01:12 |
| J1m | I don't think there is anything it can do. | 01:12 |
| rocky|away | because i'm looking at some code in someone's formlib form that does time_zone.normalize(datetime.datetime.now()) but i don't possibly see how that would work | 01:13 |
| * J1m wants to switch to star dates | 01:13 | |
| *** derek|laptop has joined #zope3-dev | 01:13 | |
| rocky|away | since datetime.datetime.now() doesn't have tzinfo | 01:13 |
| *** derek|laptop is now known as derek_richardson | 01:15 | |
| srichter | rocky|away: no this will probably not work; the rule of thumb is to always store dates in UTC | 01:16 |
| *** faassen has quit IRC | 01:16 | |
| rocky|away | srichter: the code in question is in lovely.tag.browser.tag | 01:16 |
| srichter | then my suspicion is that time_zone.normalize does exactly that | 01:17 |
| * J1m wonders if time_zone.normalize should fail or at least warn if it gets a datetime wo a time zone | 01:17 | |
| srichter | it probably looks at the server timezone and converts the datetime to utc | 01:17 |
| J1m | That sounds excessively magic. | 01:17 |
| rocky|away | well, time_zone is trying to be looked up via idatetime.ITZInfo adapter lookup, and if that fails it simply uses pytz.UTC | 01:18 |
| rocky|away | pytz.UTC.normalize() barfs on any datetime that doesn't have a tzinfo | 01:18 |
| rocky|away | and... datetime.datetime.now() does not have any timezone | 01:18 |
| J1m | That's good. | 01:18 |
| rocky|away | right, but makes me wonder how this code ever worked | 01:18 |
| srichter | it probably doesn't or someone provided an adapter that does some magic liek I said | 01:19 |
| rocky|away | hmmm | 01:20 |
| *** jfroche has joined #zope3-dev | 01:20 | |
| *** J1m has quit IRC | 01:23 | |
| *** jinty has quit IRC | 01:52 | |
| *** derek_richardson is now known as derek|home | 02:02 | |
| *** pcardune has quit IRC | 02:20 | |
| *** redir has joined #zope3-dev | 02:27 | |
| *** pcardune has joined #zope3-dev | 02:41 | |
| *** Jell-O-Fishi has joined #zope3-dev | 02:46 | |
| *** lurker has quit IRC | 02:55 | |
| *** natea_ has joined #zope3-dev | 03:02 | |
| *** redir has quit IRC | 03:02 | |
| *** redir has joined #zope3-dev | 03:05 | |
| *** jukart has joined #zope3-dev | 03:12 | |
| *** whit has quit IRC | 03:21 | |
| *** alecm has quit IRC | 03:27 | |
| *** ksmith99 has quit IRC | 03:28 | |
| *** benji has joined #zope3-dev | 03:38 | |
| *** jinty has joined #zope3-dev | 03:44 | |
| *** ignas has quit IRC | 03:57 | |
| *** redir has quit IRC | 03:59 | |
| *** jfroche has quit IRC | 04:04 | |
| *** jukart has quit IRC | 04:06 | |
| *** derek|home has quit IRC | 04:07 | |
| *** natea_ has quit IRC | 04:16 | |
| *** RaFromBRC has quit IRC | 04:20 | |
| *** azazel has quit IRC | 04:30 | |
| *** projekt01 has joined #zope3-dev | 05:00 | |
| *** pcardune has quit IRC | 05:01 | |
| *** alecm has joined #zope3-dev | 05:05 | |
| *** alga has quit IRC | 05:12 | |
| *** jinty has quit IRC | 05:14 | |
| *** alecm has quit IRC | 05:24 | |
| *** jinty has joined #zope3-dev | 05:24 | |
| *** alecm has joined #zope3-dev | 05:25 | |
| *** alecm has quit IRC | 05:38 | |
| *** greenman has quit IRC | 05:39 | |
| *** whit has joined #zope3-dev | 05:45 | |
| *** alecm has joined #zope3-dev | 05:55 | |
| *** greenman has joined #zope3-dev | 05:59 | |
| *** alecm has quit IRC | 06:13 | |
| *** alecm has joined #zope3-dev | 06:18 | |
| *** febb has joined #zope3-dev | 06:19 | |
| *** alecm has quit IRC | 06:23 | |
| *** pcardune has joined #zope3-dev | 06:25 | |
| *** whit has quit IRC | 06:29 | |
| *** febb has quit IRC | 06:37 | |
| *** febb has joined #zope3-dev | 06:37 | |
| *** febb has quit IRC | 06:40 | |
| *** febb has joined #zope3-dev | 06:40 | |
| *** febb has quit IRC | 06:41 | |
| *** oferw has joined #zope3-dev | 06:42 | |
| *** romanofski has joined #zope3-dev | 06:44 | |
| *** febb has joined #zope3-dev | 06:50 | |
| *** febb has joined #zope3-dev | 07:00 | |
| *** b52laptop has joined #zope3-dev | 07:02 | |
| *** greenman has quit IRC | 07:07 | |
| *** greenman has joined #zope3-dev | 07:08 | |
| *** b52lap has quit IRC | 07:18 | |
| *** febb_ has joined #zope3-dev | 07:24 | |
| *** febb has left #zope3-dev | 07:24 | |
| *** febb has quit IRC | 07:24 | |
| *** febb_ has quit IRC | 07:24 | |
| *** febb__ has joined #zope3-dev | 07:24 | |
| *** febb_ has joined #zope3-dev | 07:24 | |
| *** febb_ is now known as febb | 07:25 | |
| *** febb has joined #zope3-dev | 07:26 | |
| *** febb_ has joined #zope3-dev | 07:28 | |
| *** alecm has joined #zope3-dev | 07:29 | |
| *** alecm has quit IRC | 07:36 | |
| *** afd_ has joined #zope3-dev | 07:41 | |
| *** pbugni[afk] has joined #zope3-dev | 07:42 | |
| *** greenman has quit IRC | 07:52 | |
| *** pcardune has quit IRC | 07:59 | |
| *** oferw has quit IRC | 08:00 | |
| *** jukart has joined #zope3-dev | 08:00 | |
| *** dunny has quit IRC | 08:01 | |
| *** pghoratiu has joined #zope3-dev | 08:05 | |
| *** jukart has quit IRC | 08:19 | |
| *** pbugni[afk] is now known as pbugni | 08:19 | |
| *** naro has joined #zope3-dev | 08:28 | |
| *** projekt01 has quit IRC | 08:30 | |
| *** zagy has joined #zope3-dev | 08:32 | |
| *** pbugni is now known as pbugni[afk] | 08:38 | |
| *** baijum has joined #zope3-dev | 08:38 | |
| *** alecm has joined #zope3-dev | 08:41 | |
| *** jukart has joined #zope3-dev | 08:44 | |
| *** lisppaste6 has quit IRC | 08:59 | |
| *** yota has joined #zope3-dev | 09:08 | |
| *** timte has joined #zope3-dev | 09:18 | |
| *** dobee has joined #zope3-dev | 09:19 | |
| *** febb has quit IRC | 09:19 | |
| *** jodok has joined #zope3-dev | 09:19 | |
| *** pelle_ has joined #zope3-dev | 09:21 | |
| *** philiKON_ has joined #zope3-dev | 09:38 | |
| *** tiredbones has joined #zope3-dev | 09:39 | |
| *** sorin has joined #zope3-dev | 09:42 | |
| *** sorin is now known as sorindregan | 09:42 | |
| *** wrobel has joined #zope3-dev | 09:47 | |
| *** philiKON has quit IRC | 09:54 | |
| *** afd_ has quit IRC | 10:01 | |
| *** alecm has quit IRC | 10:03 | |
| *** yvl has joined #zope3-dev | 10:10 | |
| *** pcardune has joined #zope3-dev | 10:23 | |
| *** zagy has quit IRC | 10:40 | |
| *** ccomb has joined #zope3-dev | 10:44 | |
| *** zagy has joined #zope3-dev | 10:52 | |
| *** Henri__ has joined #zope3-dev | 11:08 | |
| *** sm is now known as sm-zzz | 11:10 | |
| *** lzdych has joined #zope3-dev | 11:14 | |
| *** afd_ has joined #zope3-dev | 11:17 | |
| *** Henri__ has left #zope3-dev | 11:31 | |
| *** baijum has quit IRC | 11:39 | |
| *** jodok has quit IRC | 11:43 | |
| *** pbugni[afk] has quit IRC | 11:47 | |
| *** goschtl has joined #zope3-dev | 11:56 | |
| *** baijum has joined #zope3-dev | 11:57 | |
| *** bigkevmcd has joined #zope3-dev | 11:59 | |
| *** jikanter has joined #zope3-dev | 12:29 | |
| *** wreutz has joined #zope3-dev | 12:43 | |
| *** Aiste has quit IRC | 12:43 | |
| *** philiKON_ has quit IRC | 12:44 | |
| *** philiKON has joined #zope3-dev | 12:44 | |
| *** pcardune has quit IRC | 13:00 | |
| *** jukart has quit IRC | 13:22 | |
| *** jukart has joined #zope3-dev | 13:24 | |
| *** greenman has joined #zope3-dev | 13:37 | |
| *** ka-jod has joined #zope3-dev | 13:45 | |
| *** ignas has joined #zope3-dev | 13:45 | |
| *** mkerrin has joined #zope3-dev | 13:54 | |
| *** rocky|away is now known as rocky | 14:11 | |
| *** salfield has joined #zope3-dev | 14:17 | |
| *** sorindregan has quit IRC | 14:25 | |
| *** benji has quit IRC | 14:27 | |
| *** sorin has joined #zope3-dev | 14:33 | |
| *** sorin is now known as sorindregan | 14:33 | |
| *** J1m has joined #zope3-dev | 14:37 | |
| *** whit has joined #zope3-dev | 14:37 | |
| *** BjornT has quit IRC | 14:40 | |
| *** alga has joined #zope3-dev | 14:42 | |
| *** greenman has quit IRC | 14:45 | |
| *** BjornT has joined #zope3-dev | 14:45 | |
| *** ghendi has joined #zope3-dev | 14:51 | |
| *** greenman has joined #zope3-dev | 14:53 | |
| *** Aiste has joined #zope3-dev | 14:57 | |
| *** Jell-O-Fishi has quit IRC | 14:58 | |
| rocky | does anyone know if zope provides the ability to guess what the currently-running-server's timezone is? | 15:00 |
| rocky | perhaps via pytz or something | 15:00 |
| rocky | hmm... DateTime seems to be able to figure it out... hmm | 15:02 |
| rocky | bah forget it, didn't realize time.tzname gave it to me | 15:03 |
| *** edgordon has quit IRC | 15:06 | |
| *** benji has joined #zope3-dev | 15:10 | |
| *** edgordon has joined #zope3-dev | 15:15 | |
| *** __mac__ has joined #zope3-dev | 15:19 | |
| *** niemeyer has joined #zope3-dev | 15:20 | |
| *** xbeanx has joined #zope3-dev | 15:21 | |
| *** mgedmin has joined #zope3-dev | 15:25 | |
| romanofski | moin | 15:27 |
| Theuni | hmm. | 15:30 |
| Theuni | philiKON: ping | 15:30 |
| Theuni | I'm wondering about using the pluggable authentication utility as a global utility. | 15:30 |
| philiKON | pong | 15:30 |
| Theuni | I stumbled over something that I'm not sure is intentional. | 15:30 |
| philiKON | um, ok | 15:31 |
| Theuni | the beforeTraversal hook needs an IAuthentication utility that *must* return an unauthenticated principal | 15:31 |
| *** ghendi has quit IRC | 15:31 | |
| Theuni | However, the interface says it *may* return None. | 15:31 |
| Theuni | PAU always returns None for the unauthenticatedPrincipal. | 15:31 |
| Theuni | In a normal case this ends up using the unauthenticated principal from an earlier IAuthentication utility. | 15:32 |
| Theuni | However, if it's a root utility, the request.principal becomes none and zope.app.exception dies. | 15:32 |
| Theuni | So the question I have is: | 15:32 |
| *** lisppaste6 has joined #zope3-dev | 15:32 | |
| Theuni | Can request.principal be None instead of an unauthenticated principal and should other code cope with this or should the first IAuthentication utility always return an unauthenticated principal. | 15:33 |
| philiKON | i'm puzzled by the fact that the interface permits None | 15:34 |
| Theuni | That's correct though. | 15:34 |
| philiKON | i think the PAU should try to return the unauthenticated principal if possible | 15:34 |
| Theuni | It allows an IAuthentication utility to not return a principal which results in the traversal to use the previous principal. | 15:34 |
| philiKON | s/the/an/ | 15:34 |
| philiKON | i see | 15:34 |
| philiKON | ok, that makes sense | 15:34 |
| Theuni | If it does return an unauthenticated principal, then this cascade doesn't work. | 15:34 |
| Theuni | IMHO it should be configurable to some extent. | 15:35 |
| philiKON | well, i guess a RootPAU needs to special-case this for the unauthenticated principal | 15:35 |
| Theuni | In the case that it is a root utility it definitely doesn't make sense. | 15:35 |
| Theuni | Thought so. | 15:35 |
| * Theuni takes a look at IRequest though | 15:35 | |
| Theuni | Hmm. IRequest doesn't tell whether the principal can be None. | 15:36 |
| Theuni | Most code assumes that it never is None. | 15:36 |
| philiKON | i think it shouldn't be None | 15:36 |
| Theuni | k. | 15:36 |
| Theuni | Well. Then we need two things: | 15:36 |
| Theuni | a) a rootPAU and a slightly different interfaces for authentication utilities that work at the root. | 15:37 |
| *** ghendi has joined #zope3-dev | 15:37 | |
| philiKON | (i don't see why we need another interface) | 15:37 |
| Theuni | Well. Right. We don't. | 15:37 |
| Theuni | The publication treats the None case as 'Unauthorized'. | 15:38 |
| Theuni | Which is *very* interesting | 15:38 |
| Theuni | because this will always die | 15:38 |
| Theuni | because Unauthorize expects the principal not to be none. | 15:38 |
| Theuni | but it may be raisen before the principal is ever set | 15:38 |
| philiKON | hugh | 15:38 |
| Theuni | sooooo ... there are legal cases where request.principal is None | 15:38 |
| Theuni | or we forbid that a root authentication utility can return None as the unauthenticated principal. | 15:39 |
| philiKON | i think that's a sensible interpretation | 15:39 |
| philiKON | which means that i guess we do need another interface | 15:39 |
| Theuni | Hmm. Looking at that code also tells me that the transaction must begin much earlier | 15:39 |
| Theuni | philiKON: another reason why this sounds right, is that we have to places asking for authentication utilities | 15:40 |
| Theuni | a) beforeTraversal and b) _maybePlaceful...() | 15:40 |
| Theuni | and they have slightly different semantics. | 15:40 |
| philiKON | yup | 15:40 |
| philiKON | well, not slightly | 15:40 |
| philiKON | beforeTraversal happens once | 15:40 |
| philiKON | before everything | 15:40 |
| Theuni | Here is another comment: raise Unauthorized # If there's no default principal | 15:40 |
| philiKON | maybePlaceful happens after each traversal step | 15:40 |
| Theuni | i meant different semantics for a single call ;) | 15:41 |
| Theuni | how they work with iauthentication | 15:41 |
| philiKON | i see | 15:41 |
| * Theuni detects: Nobody every created a different root authentication utility or they didn't care. | 15:41 | |
| * philiKON assumes the former | 15:42 | |
| Theuni | I wonder whether we can get the zodbless patch into Zope 3.4 | 15:42 |
| Theuni | we have two days to go for a merge on the trunk | 15:42 |
| philiKON | it's a bit convoluted | 15:42 |
| philiKON | but in principle i'm +1 | 15:42 |
| philiKON | we should get jinty to merge it | 15:42 |
| Theuni | I looked at the patch and it seems reasonable. However, we should write tests before merging it. | 15:43 |
| Theuni | I think I'll do that now. | 15:43 |
| Theuni | I could really use that. | 15:43 |
| Theuni | Hmm. beta for tomorrow. I better get started. | 15:44 |
| Theuni | I wonder whether J1m took a look at the patch already. | 15:45 |
| * J1m catches up ... | 15:45 | |
| Theuni | :) | 15:45 |
| Theuni | two issues at hand: one was the semantics of IAuthentication as a global utility, the other was the zodbless patch that jinty provided. | 15:46 |
| *** salfield has quit IRC | 15:47 | |
| *** philiKON has quit IRC | 15:47 | |
| J1m | I think any major patch for 3.4 would be really silly. | 15:48 |
| J1m | The goal of 3.4 is to continue the past releases because we haven't quite got out act tohether with eggs. | 15:48 |
| J1m | The goal of 3.4 is to continue the past releases because we haven't quite got our act together with eggs. | 15:49 |
| J1m | Theuni, request.principal should not be None. | 15:49 |
| J1m | imo, that would be a bug. | 15:49 |
| Theuni | Hmm. | 15:49 |
| J1m | zodb-less zope should just be a package/buildout that gets releases like anything else. | 15:50 |
| J1m | I don't see why it should be tied to 3.4 in any way. | 15:50 |
| *** philiKON has joined #zope3-dev | 15:50 | |
| Theuni | Ok. Let's talk about the request first. | 15:50 |
| J1m | I'm really out of touch with our eggification efforts. :/ | 15:51 |
| Theuni | I derive the following todos for me for getting request.principal right: | 15:51 |
| J1m | I hope to get caught up a DZUG if not sooner. :) | 15:51 |
| Theuni | The global/root authentication utility must not return None as the unauthenticated principal. | 15:51 |
| Theuni | Actually that seems like the only todo. | 15:51 |
| J1m | BTW, this is configurable through the publication object. | 15:52 |
| J1m | right | 15:52 |
| Theuni | As a side note: I think transaction.begin() should be done immediately at the beginning of beforeTraversal() | 15:52 |
| J1m | or, alternatively, the publication could choose to provide one if it doesn't. | 15:52 |
| J1m | Or even look up an independent factory. | 15:52 |
| Theuni | Right. | 15:52 |
| Theuni | Dunno what's better. | 15:52 |
| J1m | I wish we had made the publisher look more like the bobo publisher. | 15:53 |
| J1m | too late now. | 15:53 |
| J1m | I've been working on some monitoring code that would probably have been cleaber if I had some more hook points. | 15:53 |
| J1m | I suggest decoupling this issue from the 3.4 release. | 15:54 |
| J1m | I'm not sure you coupled it though. | 15:54 |
| J1m | So perhaps ignore the previous remark. :) | 15:54 |
| Theuni | :) | 15:54 |
| Theuni | Alright. I can ignore zodbless for now too. Would have been nice. | 15:55 |
| Theuni | I'm undecided for which way to go with the unauthenticated principal though | 15:55 |
| J1m | Sure, it still is, just don't tie it to 3.4. | 15:55 |
| Theuni | :) | 15:55 |
| Theuni | The patch is intrusive so decoupling it from the core is hard ... | 15:55 |
| J1m | It would be great to see some kind of release of zodb-less zope based on the new way. | 15:55 |
| Theuni | but we don't have to merge it now. | 15:56 |
| J1m | Then the patch is very wrong. | 15:56 |
| J1m | Bad patch Bad bad patch. | 15:56 |
| J1m | :) | 15:56 |
| Theuni | :) | 15:56 |
| J1m | All you should need to to replace the publication object. | 15:56 |
| Theuni | ah. he extended the existing publication instead of replacing. | 15:56 |
| J1m | if anything else is needed, then there is probably a bug somewhere else. | 15:57 |
| J1m | A more bobonic publication would probably support this better. :) | 15:57 |
| * Theuni points the discussion back to the question of the principal. | 15:57 | |
| *** oferw has joined #zope3-dev | 15:58 | |
| * Theuni gets a coffee. if nobody objects he then will introduce rootauthentication utilities. | 15:59 | |
| *** greenman has quit IRC | 16:06 | |
| ignas | rootauthentication utilities? | 16:13 |
| Theuni | new interface: IRootAuthentication. Requires that unauthenticatedPrincipal never returns None | 16:14 |
| J1m | I don't think that is needed. | 16:15 |
| Theuni | That's why I was trying to ask which way to implement the principal issue. | 16:15 |
| Theuni | s/implement/get rid of/ | 16:16 |
| J1m | You have lots of choices. | 16:16 |
| Theuni | Sure. That was one of those I offered. Nobody objected. ;) | 16:16 |
| J1m | - Change the publication to use an unauthenticated user factory | 16:16 |
| J1m | - Change pau to return an unauthenticated principle. | 16:17 |
| J1m | - Don't use a pau as a root auth util :) | 16:17 |
| Theuni | The first choice seems ok, the second would kill existing behaviour that people rely on. | 16:17 |
| Theuni | The third would make me abandon my existing work ;) | 16:17 |
| *** meatballhat has joined #zope3-dev | 16:18 | |
| Theuni | I think using a factory is probably the cleanest solution. | 16:18 |
| J1m | sounds like you prefer door #1. | 16:18 |
| *** meatballhat_ has joined #zope3-dev | 16:18 | |
| Theuni | ack. | 16:18 |
| *** meatballhat has quit IRC | 16:18 | |
| *** Vq^ has quit IRC | 16:20 | |
| rocky | dobee: ping | 16:23 |
| dobee | rocky: pong | 16:26 |
| Theuni | J1m: If I got those factories right, then an IUnauthenticatedPrincipalFactory would be adapting a request or nothing. Which one makes more sense to you? | 16:27 |
| *** meatballhat_ is now known as meatballhat | 16:28 | |
| J1m | I expect it to be a utility that is called without arguments. | 16:30 |
| Theuni | Ah. Ok. That works as well of course. | 16:30 |
| J1m | I suppose it could also be an adapter that adapts nothing. :) | 16:30 |
| Theuni | isn't that what a utility is? | 16:30 |
| J1m | No. | 16:31 |
| J1m | absolutely not. | 16:31 |
| rocky | dobee: looking through lovely.tag which you seemd to have a hand in... have you noticed that on the edit form for tags that if you say only add 1 new tag than the previous tags are added again to the tag frequency counter? | 16:31 |
| philiKON | a utility is pre-instantiated | 16:31 |
| J1m | A utility is a component that is looked up by interface. | 16:31 |
| philiKON | and an adapter isn't? | 16:31 |
| rocky | lol | 16:31 |
| * rocky wonders what kind of funky adapters J1m writes | 16:31 | |
| Theuni | I was looking at the zope.app.authentication code and they're using various forms of adaptation so I was trying to extrapolate a pattern ... ;) | 16:31 |
| J1m | an adapter is an object that is computed to satisfy an inteface. | 16:31 |
| J1m | Adaptation always involved computation. | 16:32 |
| J1m | even if the computation leads to returning an existing object. | 16:32 |
| rocky | btw anyone here read pje's pep 3124 for py3k dealing with interfaces/adapters/overloading/aspects ? | 16:32 |
| J1m | Utilities are always preexisting objects. | 16:32 |
| J1m | sigh | 16:33 |
| rocky | J1m: that's a very good distinction btw, never really looked at it that way before | 16:33 |
| benji | IOW, utilities are just arbitrary things that can be looked up by interface and name. | 16:34 |
| *** jodok has joined #zope3-dev | 16:34 | |
| Theuni | Hmm. The default unauthenticated principal is an interesting thing. | 16:35 |
| ignas | hmm, who if anyone is maintaining ZODB egg? | 16:36 |
| * rocky thought pje's pep on interfaces was pretty interesting -- a scaled down version of zope3 stuff | 16:36 | |
| J1m | ignas, I am | 16:37 |
| ignas | J1m: if you install ZODB as an egg in some specific directory (not your system python path) | 16:37 |
| ignas | runzeo for example - works | 16:38 |
| ignas | mkzeoinstance | 16:38 |
| ignas | works as well | 16:38 |
| ignas | but if you mkzeoinstance | 16:38 |
| ignas | the zeo scripts (zeoctl) in the new bin directory | 16:38 |
| ignas | only knows where to find the ZODB egg | 16:40 |
| ignas | and can't find other eggs | 16:40 |
| ignas | maybe it would be possible to add some parameter like "EGG_PATH" to the script | 16:40 |
| * ignas has no idea, but i will have similar problems with schooltool | 16:40 | |
| ignas | and would like to see how these should be fixed ;) | 16:41 |
| J1m | Isn't schooltool moving to zc.buildout? | 16:41 |
| Theuni | J1m: Any idea what the default unauthenticated principal could use as an id? | 16:41 |
| ignas | J1m: no, not at the moment | 16:42 |
| J1m | Theuni, probably something based on it's module __name__. | 16:42 |
| J1m | ignas, ah, I though it was. | 16:42 |
| Theuni | k | 16:43 |
| J1m | ignas, I have as low an opinion of mkzeoinstance as I do of mkzopeinstance. :) | 16:43 |
| J1m | You might submit a bug report. | 16:43 |
| J1m | I would deal with the report in whatever way involves the least work for me, which is likely to be something along the lines of "mkzeoinstance only works if ZODB is installed into site-packages.". | 16:44 |
| ignas | :) | 16:45 |
| dobee | rocky: hm, dunno - well that's a bug in this case i guess | 16:47 |
| J1m | If it is easy to fix this otherwise, I might be willing to. | 16:47 |
| J1m | ignas, you might consider using buildout just to install your zeo server. | 16:48 |
| ignas | hmm | 16:49 |
| ignas | the question is - whether i want to have schooltool as an egg or not | 16:49 |
| ignas | the upside of this being that any user can install schooltool-SVN easily | 16:50 |
| ignas | the downside - i have yet to see an application deployed as an egg | 16:50 |
| *** baijum has quit IRC | 16:52 | |
| *** alecm has joined #zope3-dev | 16:59 | |
| *** whit has quit IRC | 17:01 | |
| *** meatballhat has quit IRC | 17:03 | |
| *** meatballhat has joined #zope3-dev | 17:04 | |
| *** whit has joined #zope3-dev | 17:04 | |
| *** pghoratiu has left #zope3-dev | 17:05 | |
| *** alga_ has joined #zope3-dev | 17:07 | |
| J1m | ignas, we deploy applications using eggs all the time. | 17:12 |
| ignas | J1m: not using eggs, but packaged as eggs | 17:12 |
| J1m | And we've been deploying applications using buildout for years. | 17:12 |
| ignas | like easy_install schooltool | 17:12 |
| J1m | eggs aren't really for deploying applications. | 17:13 |
| ignas | and then - it magically works | 17:13 |
| ignas | hmm, why not? | 17:13 |
| ignas | zeo with mkzeoinstance seems to work ... | 17:13 |
| J1m | ZODB, and even mkzeoinst should work just fine with easy_install. | 17:13 |
| J1m | as long as you do things the way easy_install wants you to. | 17:14 |
| J1m | buildout gives you more flexibility than easy_install. | 17:14 |
| J1m | We (zc) don't like installing into system Python. | 17:14 |
| *** sorindregan has quit IRC | 17:14 | |
| benji | and as long as you don't want to do anything easy_install doesn't handle (e.g., set up cron jobs) | 17:14 |
| J1m | So we have a buildout recipe for ZEO that: | 17:15 |
| J1m | - sets up configuratiopn files | 17:15 |
| J1m | - sets up system rc scripts | 17:15 |
| J1m | - sets up a pack cron job | 17:15 |
| J1m | - sets up log rotation | 17:15 |
| J1m | It is slightly red-hat specific. | 17:15 |
| J1m | I'm sure it would be easy, if not trivial, to make it work with debian and other distros. | 17:16 |
| J1m | And, unlike mkzeiinst, it actually puts files where system admins want them. :) | 17:16 |
| ignas | admins will be installing schooltol through apt-get install ... | 17:17 |
| ignas | but we need an easy way for people to try out schooltool | 17:17 |
| ignas | buildout does not seem easy ;) | 17:17 |
| J1m | It seems as easy to me as a regular Zope checkout. | 17:17 |
| J1m | python2.4 bootstrap.py | 17:18 |
| J1m | bin/buildout | 17:18 |
| J1m | If you provide a complete buildout.cfg for them, that's all they have to do. | 17:18 |
| J1m | And, of course, you could provide a wrapper script that automates those 2 commands. | 17:19 |
| J1m | Heck, you could even provide a make file. :) | 17:20 |
| benji | boo, hiss | 17:20 |
| * Theuni waves autoconf at benji | 17:20 | |
| ignas | we have a makefile ;) | 17:21 |
| benji | I acually don't mind autoconf, but that's mostly because I don't have to use it. ;) | 17:21 |
| ignas | it works | 17:21 |
| ignas | without buildout even | 17:21 |
| ignas | but it still seems too complicated at the moment | 17:21 |
| ignas | for users | 17:21 |
| ignas | ;) | 17:21 |
| benji | and make doesn't work to well on Windows | 17:21 |
| *** jinty has joined #zope3-dev | 17:21 | |
| ignas | J1m: schooltool users do not map with Zope3 users that much :/ | 17:21 |
| J1m | You can wrap whatever UI you want around buildout. | 17:22 |
| Theuni | argh. | 17:22 |
| ignas | the last couple of users were only competent enough to follow a list of precise steps | 17:22 |
| * Theuni must open his eyes wider. | 17:22 | |
| J1m | There's no need to expose buildout to your users if you don't want to. | 17:22 |
| ignas | one of them managed to break his svn checkout 3 times | 17:23 |
| ignas | by doing "nothing" ;) | 17:23 |
| * J1m is tired of arguing with ignas about this. | 17:23 | |
| ignas | J1m: i understand that I can use buildout to solve my problem, but i am not sure it is the best solution | 17:24 |
| *** alga has quit IRC | 17:25 | |
| *** __mac__ has quit IRC | 17:29 | |
| rocky | Theuni: you wouldn't happen to have an example code/zcml for setting up a traversal adapter in zope 2.9 would you? | 17:29 |
| * Theuni shows his empty pockets | 17:30 | |
| *** Aiste_ has joined #zope3-dev | 17:31 | |
| rocky | hehe ... i'm asking you because i see an irc log online (via google) that shows a converation between you and philiKON describing your experiences setting up traverse stuff (he explained difference between ITraversable and IPublishTraverse) | 17:31 |
| *** yvl has quit IRC | 17:31 | |
| *** Aiste has quit IRC | 17:32 | |
| Theuni | that's probably not in zope 2.9 ;) | 17:32 |
| rocky | http://zope3.pov.lt/irclogs/%23zope3-dev.2006-07-16.log.html | 17:33 |
| rocky | The|unii try to hook into traversal with five in Zope 2.9.4 | 17:33 |
| rocky | that's me quoting you :) | 17:33 |
| Theuni | darn | 17:38 |
| * Theuni runs | 17:38 | |
| timte | rocky: I tried that too, with no luck. I think you need to do it the old bobo traverse way. It's fixed in five 1.5 I think. | 17:38 |
| Theuni | that's almost a year ago | 17:38 |
| Theuni | what the hell was i doing there?!? | 17:38 |
| rocky | lol @ Theuni | 17:38 |
| rocky | timte: right now i'm looking to see how FiveTraversable does stuff | 17:39 |
| *** romanofski has quit IRC | 17:40 | |
| Theuni | rocky: if your read through that complete log, then you should find what you need | 17:40 |
| Theuni | that's what my belly tells me | 17:40 |
| Theuni | and my belly also tells me to get an apple | 17:40 |
| rocky | hehe | 17:42 |
| *** mgedmin has quit IRC | 17:42 | |
| *** meatballhat has quit IRC | 17:42 | |
| *** mgedmin has joined #zope3-dev | 17:44 | |
| *** salfield has joined #zope3-dev | 17:44 | |
| *** meatballhat has joined #zope3-dev | 17:45 | |
| *** Aiste_ is now known as Aiste | 17:50 | |
| *** norro has joined #zope3-dev | 17:50 | |
| *** jinty has quit IRC | 17:50 | |
| *** oferw has quit IRC | 17:51 | |
| *** timte has quit IRC | 17:55 | |
| *** jbb666 has joined #zope3-dev | 18:16 | |
| Theuni | J1m: btw. did you agree that i can move the transaction.begin() before the root global authentication? | 18:23 |
| J1m | No, I was hoping that you would forget so I didn't have to look. | 18:24 |
| * J1m looks | 18:25 | |
| Theuni | :) | 18:25 |
| *** afd_ has quit IRC | 18:26 | |
| J1m | Theuni, that looks good to me. | 18:27 |
| Theuni | k | 18:27 |
| * Theuni scratches his head for testing that ... | 18:27 | |
| J1m | You could create an insane auth that modifies the database during authentication. :) | 18:28 |
| Theuni | :) | 18:28 |
| Theuni | Something along those lines. | 18:28 |
| Theuni | The reason I want to move it is because I have a utility that actually does talk to the database. ALthough it only reads it, I feel like it should be in the same transaction. | 18:29 |
| Theuni | What happens if transaction.begin() is called twice btw? | 18:29 |
| J1m | a begin implies and abort. | 18:29 |
| Theuni | k. | 18:29 |
| J1m | transaction boundaries are otherwise implicit. | 18:29 |
| J1m | a commit or about ends the current transaction and implicitly begins a new one. | 18:30 |
| Theuni | right. what's the rule on keeping references to objects from earlier transactions? | 18:30 |
| Theuni | they will just join the new transaction, right? | 18:31 |
| rocky | dobee: does the tagging engine in lovely.tag do any security checks? i mean when returning getItems() does it discern what items you're allowed to have access to? | 18:31 |
| J1m | yes, object references and transaction boundaries have nothing to do with each other. | 18:31 |
| dobee | rocky: no | 18:31 |
| *** zagy has quit IRC | 18:32 | |
| J1m | what you really want to avoid is keeping object refs after a connection has been closed. | 18:32 |
| Theuni | right. | 18:32 |
| Theuni | so the current situation would really only be annoying if you had a "crazy" authentication utility that does modify the objects during authentication | 18:32 |
| Theuni | otherwise the begin would just implicitly abort() the former transaction but the user object that was loaded will be fine. | 18:33 |
| J1m | Or if you had a left over uncommitted transaction from some previous activity, which should be very very unlikely. | 18:34 |
| J1m | yes | 18:34 |
| Theuni | ok. i'll put that in the tracker for now, my time is running out here. | 18:35 |
| Theuni | urks | 18:36 |
| Theuni | someone removed zope.index from the checkout | 18:36 |
| J1m | yes, fred sent a note about that. | 18:37 |
| J1m | There should be an external now. | 18:37 |
| *** goschtl has quit IRC | 18:38 | |
| Theuni | ah, maybe subversion fooled me | 18:39 |
| *** wreutz has quit IRC | 18:39 | |
| Theuni | jupp. it stumbled over the directory that was still there. | 18:39 |
| * Theuni hits subversion on the stack | 18:40 | |
| *** ignas has quit IRC | 18:42 | |
| * Theuni grumbles. test failures from zope.index. | 18:43 | |
| *** wreutz has joined #zope3-dev | 18:43 | |
| * J1m is sure tests pass on a clean checkout | 18:43 | |
| J1m | Not that I've tried. :) | 18:43 |
| Theuni | hrn. gotta look later. gotta run. | 18:43 |
| *** lzdych has quit IRC | 18:51 | |
| *** WebMaven is now known as fog_of_war | 18:57 | |
| *** fog_of_war is now known as WebMaven | 18:57 | |
| *** lucielejard has joined #zope3-dev | 19:01 | |
| *** norro has quit IRC | 19:02 | |
| *** wreutz has quit IRC | 19:03 | |
| andrew_m | how can i turn a _p_oid into an int? I assume the root object ID ('\x00\x00\x00\x00\x00\x00\x00\x01') is 1? | 19:11 |
| J1m | If this is just for visual inspection, use: | 19:12 |
| J1m | import zodb.utils | 19:12 |
| J1m | sorry | 19:13 |
| J1m | import ZODB.utils | 19:13 |
| J1m | print ZODB.utils.u64(oid) | 19:13 |
| J1m | You shouldn't use OIDs in application code. | 19:13 |
| J1m | They aren't required to be serialized integers. | 19:14 |
| andrew_m | i see, thanks. i might use object paths instead as IDs. | 19:17 |
| *** harobed has quit IRC | 19:21 | |
| *** jodok has quit IRC | 19:23 | |
| J1m | I hope you're aware of the intid framework, | 19:23 |
| andrew_m | J1m: hmm, no. i'm trying to use as little of zope as possible due to its complexity | 19:27 |
| J1m | Then why are you here? You'd rather solve a problem yourself that reuse an existing solution? | 19:28 |
| J1m | than | 19:28 |
| andrew_m | well, zope has some great, simple features (persistence, http server, object views) | 19:29 |
| ccomb | ...and intid :) | 19:30 |
| rocky | andrew_m: i would argue that zope3 the library/framework is much better than zope3 the app server :) | 19:30 |
| andrew_m | rocky: can be. but you need to invest a lot of time and effort to dig into it | 19:30 |
| ccomb | I did it, and it's worth doing it | 19:31 |
| andrew_m | after a year of zope3 development i find myself fading out more and more of zope depended stuff | 19:32 |
| *** ghendi_ has joined #zope3-dev | 19:32 | |
| *** roym has joined #zope3-dev | 19:35 | |
| ccomb | that's good news, zope itself is becoming less and less zope dependant | 19:36 |
| andrew_m | heh | 19:36 |
| *** ghendi_ has quit IRC | 19:36 | |
| *** sm-zzz is now known as sm | 19:49 | |
| *** RaFromBRC has joined #zope3-dev | 19:51 | |
| *** srichter has quit IRC | 19:52 | |
| *** ksmith99 has joined #zope3-dev | 19:52 | |
| *** ghendi has quit IRC | 19:53 | |
| *** roym has left #zope3-dev | 19:54 | |
| *** tiredbones has quit IRC | 19:58 | |
| *** jukart has quit IRC | 20:01 | |
| *** tiredbones has joined #zope3-dev | 20:12 | |
| *** tiredbones_ has joined #zope3-dev | 20:13 | |
| *** tiredbones has quit IRC | 20:22 | |
| *** jbb666 has quit IRC | 20:23 | |
| *** Aiste_ has joined #zope3-dev | 20:24 | |
| *** Aiste has quit IRC | 20:26 | |
| *** Aiste_ is now known as Aiste | 20:26 | |
| *** b52laptop has quit IRC | 20:32 | |
| *** lurker has joined #zope3-dev | 20:36 | |
| *** BjornT_ has joined #zope3-dev | 20:38 | |
| *** BjornT has quit IRC | 20:38 | |
| *** dobee has quit IRC | 20:39 | |
| *** jodok has joined #zope3-dev | 20:44 | |
| *** rocky is now known as rocky|away | 20:47 | |
| *** dobee has joined #zope3-dev | 20:52 | |
| *** rocky|away is now known as rocky | 20:55 | |
| *** jukart has joined #zope3-dev | 20:57 | |
| *** Aiste has quit IRC | 21:02 | |
| *** pcardune has joined #zope3-dev | 21:03 | |
| *** pbugni[afk] has joined #zope3-dev | 21:03 | |
| *** tarek has quit IRC | 21:09 | |
| *** pcardune has quit IRC | 21:13 | |
| *** ignas has joined #zope3-dev | 21:14 | |
| *** J1m has quit IRC | 21:14 | |
| *** febb has joined #zope3-dev | 21:14 | |
| *** ignas has quit IRC | 21:18 | |
| *** b52laptop has joined #zope3-dev | 21:19 | |
| *** pbugni[afk] is now known as pbugni | 21:32 | |
| *** lurker has quit IRC | 21:41 | |
| *** jukart has quit IRC | 21:43 | |
| *** dobee has quit IRC | 21:47 | |
| *** ChanServ sets mode: +o hazmat | 21:51 | |
| *** RaFromBRC has quit IRC | 21:51 | |
| *** jbb666 has joined #zope3-dev | 21:52 | |
| *** jodok has quit IRC | 21:52 | |
| *** jodok has joined #zope3-dev | 21:54 | |
| *** whit is now known as whit|lunch | 21:55 | |
| *** mgedmin has quit IRC | 21:56 | |
| *** lurker has joined #zope3-dev | 21:58 | |
| *** RaFromBRC has joined #zope3-dev | 22:01 | |
| *** mkerrin has quit IRC | 22:02 | |
| *** Jell-O-Fishi has joined #zope3-dev | 22:05 | |
| *** lurker_ has joined #zope3-dev | 22:09 | |
| *** lucielejard has quit IRC | 22:13 | |
| *** lucielejard has joined #zope3-dev | 22:14 | |
| *** lurker has quit IRC | 22:18 | |
| *** pbugni is now known as pbugni[afk] | 22:23 | |
| *** pcardune has joined #zope3-dev | 22:25 | |
| *** tarek has joined #zope3-dev | 22:36 | |
| *** rocky is now known as rocky|away | 22:40 | |
| *** ka-jod has quit IRC | 22:44 | |
| *** gumpa has joined #zope3-dev | 22:49 | |
| *** pcardune has joined #zope3-dev | 22:57 | |
| *** pcardune has quit IRC | 22:59 | |
| *** natea|away has joined #zope3-dev | 22:59 | |
| *** rocky|away is now known as rocky | 23:00 | |
| *** natea|away is now known as natea|nyc | 23:00 | |
| *** RaFromBRC is now known as RaFromBRC|lunch | 23:01 | |
| *** naro has quit IRC | 23:03 | |
| *** jfroche has joined #zope3-dev | 23:05 | |
| *** natea|nyc_ has joined #zope3-dev | 23:09 | |
| *** natea|nyc has quit IRC | 23:10 | |
| *** alga_ has quit IRC | 23:10 | |
| *** pcardune has joined #zope3-dev | 23:12 | |
| *** romanofski has joined #zope3-dev | 23:25 | |
| *** jbb666 has quit IRC | 23:29 | |
| *** lurker_ is now known as lurker | 23:34 | |
| *** pcardune has quit IRC | 23:36 | |
| *** natea|nyc_ has quit IRC | 23:39 | |
| *** whit|lunch has joined #zope3-dev | 23:44 | |
| *** whit|lunch is now known as whit | 23:49 | |
| *** harobed has joined #zope3-dev | 23:51 | |
| *** whitmo has joined #zope3-dev | 23:57 | |
| *** whit has quit IRC | 23:57 | |
| *** whitmo has quit IRC | 23:59 | |
| *** whit has joined #zope3-dev | 23:59 | |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!