| *** b52laptop has joined #zope3-dev | 00:02 | |
| *** RaFromBRC is now known as RaFromBRC|lunch | 00:19 | |
| *** maurits has quit IRC | 00:25 | |
| *** nathany has quit IRC | 00:27 | |
| *** dunny has joined #zope3-dev | 00:28 | |
| *** norro_ has quit IRC | 00:30 | |
| *** projekt01 has joined #zope3-dev | 00:37 | |
| *** scooterXL_ has joined #zope3-dev | 00:39 | |
| *** tarek has joined #zope3-dev | 00:43 | |
| *** jodok has quit IRC | 00:46 | |
| *** b52laptop has quit IRC | 00:47 | |
| *** b52laptop has joined #zope3-dev | 00:48 | |
| *** tarek has quit IRC | 00:56 | |
| *** rmarianski has quit IRC | 00:56 | |
| *** rmarianski has joined #zope3-dev | 00:57 | |
| *** J1m has quit IRC | 00:59 | |
| *** benji has quit IRC | 01:00 | |
| *** whit has quit IRC | 01:01 | |
| *** regebro has quit IRC | 01:04 | |
| *** RaFromBRC|lunch is now known as RaFromBRC | 01:06 | |
| *** lucielejard has quit IRC | 01:09 | |
| *** natea_ has quit IRC | 01:11 | |
| *** timte has quit IRC | 01:11 | |
| *** scooterXL_ has quit IRC | 01:17 | |
| *** jsadjohnson has quit IRC | 01:19 | |
| *** malthe has joined #zope3-dev | 01:19 | |
| *** rocky has quit IRC | 01:23 | |
| *** redir has quit IRC | 01:24 | |
| *** reco has joined #zope3-dev | 01:27 | |
| *** natea__ has joined #zope3-dev | 01:34 | |
| *** junkafarian has quit IRC | 01:34 | |
| *** reco_ has joined #zope3-dev | 01:35 | |
| *** junkafarian has joined #zope3-dev | 01:35 | |
| *** reco has quit IRC | 01:35 | |
| *** jamur2 has quit IRC | 01:36 | |
| *** lurkymclurkleton has quit IRC | 01:37 | |
| *** reco_ has quit IRC | 01:40 | |
| *** rocky has joined #zope3-dev | 01:47 | |
| *** sp0cksbeard has quit IRC | 01:48 | |
| *** norro has quit IRC | 02:00 | |
| *** harobed has quit IRC | 02:04 | |
| *** rcrafton has quit IRC | 02:14 | |
| *** natea__ has quit IRC | 02:29 | |
| *** rmarianski has quit IRC | 02:35 | |
| *** rcrafton has joined #zope3-dev | 02:53 | |
| *** deo has quit IRC | 03:06 | |
| *** greenman has quit IRC | 03:12 | |
| *** redir has joined #zope3-dev | 03:18 | |
| *** seletz has joined #zope3-dev | 03:18 | |
| *** seletz has quit IRC | 03:19 | |
| *** srichter has quit IRC | 03:19 | |
| *** b52laptop has quit IRC | 03:20 | |
| *** junkafarian has quit IRC | 03:41 | |
| *** Theuni has quit IRC | 03:45 | |
| *** pcardune has quit IRC | 03:47 | |
| *** rocky has quit IRC | 04:08 | |
| *** rcrafton has quit IRC | 04:36 | |
| *** greenman has joined #zope3-dev | 04:41 | |
| *** ktwilight has quit IRC | 04:58 | |
| *** ktwilight has joined #zope3-dev | 04:59 | |
| *** malthe has quit IRC | 05:00 | |
| *** projekt01 has quit IRC | 05:06 | |
| *** greenman has quit IRC | 05:23 | |
| *** salfield has quit IRC | 06:41 | |
| *** rocky has joined #zope3-dev | 06:47 | |
| *** redir has quit IRC | 06:54 | |
| *** baijum has joined #zope3-dev | 06:56 | |
| *** baijum has quit IRC | 07:12 | |
| *** RaFromBRC has quit IRC | 07:17 | |
| *** greenman has joined #zope3-dev | 07:26 | |
| *** greenman has quit IRC | 07:41 | |
| *** pcardune has joined #zope3-dev | 07:52 | |
| *** stub has joined #zope3-dev | 07:54 | |
| *** jukart has joined #zope3-dev | 08:17 | |
| *** Theuni has joined #zope3-dev | 08:33 | |
| *** afd__ has joined #zope3-dev | 08:38 | |
| *** baijum has joined #zope3-dev | 08:53 | |
| *** sm has quit IRC | 09:01 | |
| *** greenman has joined #zope3-dev | 09:05 | |
| *** kursor has joined #zope3-dev | 09:12 | |
| *** greenman has quit IRC | 09:13 | |
| *** dobee has joined #zope3-dev | 09:24 | |
| *** baijum has quit IRC | 09:28 | |
| *** timte has joined #zope3-dev | 09:34 | |
| *** timte has quit IRC | 09:44 | |
| *** sorindregan has joined #zope3-dev | 09:44 | |
| *** pcardune has quit IRC | 09:52 | |
| *** junkafarian has joined #zope3-dev | 09:59 | |
| *** baijum has joined #zope3-dev | 10:02 | |
| *** scooterXL has joined #zope3-dev | 10:05 | |
| *** jodok has joined #zope3-dev | 10:06 | |
| *** timte has joined #zope3-dev | 10:08 | |
| *** baijum has quit IRC | 10:10 | |
| *** tarek has joined #zope3-dev | 10:28 | |
| *** goschtl has joined #zope3-dev | 10:30 | |
| *** tarek has quit IRC | 10:34 | |
| *** kursor has quit IRC | 10:35 | |
| *** tarek has joined #zope3-dev | 10:36 | |
| *** MJ has joined #zope3-dev | 10:40 | |
| *** scooterXL has quit IRC | 10:41 | |
| *** philiKON has joined #zope3-dev | 10:49 | |
| *** harobed has joined #zope3-dev | 10:50 | |
| *** baijum has joined #zope3-dev | 11:09 | |
| *** scooterXL has joined #zope3-dev | 11:20 | |
| *** scooterXL has quit IRC | 11:23 | |
| *** baijum has quit IRC | 11:25 | |
| *** yotaff has joined #zope3-dev | 11:26 | |
| *** kursor has joined #zope3-dev | 11:35 | |
| *** maurits has joined #zope3-dev | 11:42 | |
| *** greenman has joined #zope3-dev | 11:53 | |
| *** norro has joined #zope3-dev | 11:53 | |
| *** mkerrin has joined #zope3-dev | 11:57 | |
| *** jayaraj has joined #zope3-dev | 12:18 | |
| *** b52laptop has joined #zope3-dev | 12:22 | |
| *** sunew has joined #zope3-dev | 12:27 | |
| *** sunew is now known as sunew|away | 12:30 | |
| *** salfield has joined #zope3-dev | 12:41 | |
| *** jayaraj has quit IRC | 12:41 | |
| *** jodok has quit IRC | 12:45 | |
| *** jodok has joined #zope3-dev | 12:46 | |
| *** kursor has quit IRC | 12:56 | |
| *** dobee has quit IRC | 13:26 | |
| *** zagy has joined #zope3-dev | 13:28 | |
| *** alga has quit IRC | 13:35 | |
| *** MJ is now known as MJ|lunch | 13:40 | |
| *** tarek has quit IRC | 13:51 | |
| *** tarek has joined #zope3-dev | 14:03 | |
| *** jsadjohnson has joined #zope3-dev | 14:08 | |
| *** projekt01 has joined #zope3-dev | 14:11 | |
| *** quodt has joined #zope3-dev | 14:17 | |
| *** whit has joined #zope3-dev | 14:27 | |
| *** b52laptop has quit IRC | 14:28 | |
| *** b52laptop has joined #zope3-dev | 14:28 | |
| *** MJ|lunch is now known as MJ | 14:30 | |
| *** dunny has quit IRC | 14:38 | |
| *** philiKON has quit IRC | 14:44 | |
| *** whit has quit IRC | 14:44 | |
| *** jukart has quit IRC | 14:45 | |
| *** djohnson has joined #zope3-dev | 14:47 | |
| djohnson | morning | 14:47 |
|---|---|---|
| *** redir has joined #zope3-dev | 14:49 | |
| *** baijum has joined #zope3-dev | 14:53 | |
| *** b52laptop has quit IRC | 14:59 | |
| *** b52laptop has joined #zope3-dev | 14:59 | |
| *** zagy has quit IRC | 15:08 | |
| *** benji has joined #zope3-dev | 15:12 | |
| *** redir has quit IRC | 15:15 | |
| *** stub has quit IRC | 15:18 | |
| *** J1m has joined #zope3-dev | 15:21 | |
| *** jodok has quit IRC | 15:32 | |
| *** projekt01_ has joined #zope3-dev | 15:35 | |
| *** baijum has quit IRC | 15:40 | |
| *** philiKON has joined #zope3-dev | 15:41 | |
| *** niemeyer has joined #zope3-dev | 15:43 | |
| *** whit has joined #zope3-dev | 15:45 | |
| *** afd__ has quit IRC | 15:45 | |
| *** projekt01 has quit IRC | 15:54 | |
| *** MattBowen|work has joined #zope3-dev | 15:55 | |
| *** lurkymclurkleton has joined #zope3-dev | 15:57 | |
| *** b52laptop has quit IRC | 15:59 | |
| *** b52laptop has joined #zope3-dev | 15:59 | |
| *** jsadjohnson has quit IRC | 15:59 | |
| *** jamur2 has joined #zope3-dev | 16:00 | |
| *** MJ has quit IRC | 16:09 | |
| *** MJ has joined #zope3-dev | 16:09 | |
| *** greenman has quit IRC | 16:15 | |
| philiKON | J1m: ping | 16:22 |
| J1m | in a meeting | 16:22 |
| philiKON | k | 16:22 |
| *** chammiya has joined #zope3-dev | 16:27 | |
| *** b52laptop has quit IRC | 16:27 | |
| *** b52laptop has joined #zope3-dev | 16:28 | |
| tarek | http://buildbot.zope.org does not exists anymore ? | 16:29 |
| tarek | I wanted to take a look at the css, but it seems to be gone | 16:29 |
| J1m | tarek, yes, it didn't survive the cvs.zope.org crash. | 16:30 |
| tarek | oh ok... | 16:30 |
| J1m | we didn't bother with it because benji didn't think anyone was using it. | 16:30 |
| philiKON | there's http://zopebuildbot.whq.gocept.com | 16:30 |
| philiKON | J1m: back from a meeting? | 16:31 |
| J1m | not yet? I'll let you know. :) | 16:31 |
| tarek | J1m, philiKON ok thanks guys | 16:31 |
| philiKON | J1m: thx :) | 16:31 |
| *** thruflo has joined #zope3-dev | 16:33 | |
| J1m | ok, philiKON, what's up? | 16:33 |
| Theuni | J1m: you pinged me a while ago. is that still current? | 16:34 |
| J1m | Theuni, no | 16:34 |
| Theuni | k | 16:34 |
| philiKON | J1m: i'd like to ask a few questions about zope.publisher.paste | 16:35 |
| J1m | sure | 16:35 |
| *** lucielejard has joined #zope3-dev | 16:35 | |
| benji | tarek: I contributed a version of the CSS to python.org's buildbots, you can look there | 16:35 |
| philiKON | J1m: so, basically, the application object is generic, the publication is "your application". | 16:36 |
| benji | http://www.python.org/dev/buildbot/all/ | 16:36 |
| J1m | Yes | 16:36 |
| philiKON | J1m: i suppose you do bootstrap configuration (zcml) and database setup in your publication's __init__ then? | 16:36 |
| J1m | If you want. | 16:36 |
| J1m | Did you notice my new zope.org key upload program? | 16:36 |
| philiKON | where do *you* do it? | 16:36 |
| philiKON | yes, i saw it :) | 16:36 |
| philiKON | it's a neat example | 16:37 |
| philiKON | but it doesn't have zodb nor zcml setup | 16:37 |
| J1m | Not that that is a model of how apps should be built, but for a one page app, its a readonable hack, especially since I was/am pressed for time. | 16:37 |
| philiKON | right. i would probably have done it as a pure wsgi app then :) | 16:38 |
| philiKON | but as said, it's a nice demo for this | 16:38 |
| tarek | benji, Ah ok thanks | 16:38 |
| J1m | Eventually, I'd like to create a publisher that provides a bobonic experience, where you put simple apps in modules. | 16:38 |
| philiKON | anyway, i wonder, when you use zope.publisher.paste, do you do zcml and zodb configuration at all, and if so, where? | 16:38 |
| J1m | well, I like bein able to rely on publisher semantics. | 16:38 |
| J1m | Had I written a pure wsgi app, then I'd have had to deal with the cgi module directly. | 16:39 |
| J1m | The publisher provides a fair bit of value, imo. | 16:39 |
| philiKON | sure | 16:39 |
| philiKON | that's true | 16:39 |
| philiKON | request + response | 16:39 |
| philiKON | anyway, not really my question | 16:39 |
| philiKON | i'm just wondering about a good pattern for doing zcml and zodb setup | 16:39 |
| J1m | I do want to implement a minimal publisher so as not to bring in lots of extra dependencies. | 16:39 |
| philiKON | i like the simplicity and flexibility of zope.publisher.paste | 16:39 |
| J1m | me too. :) It unties the gordian knot of zope.app.server etc. | 16:40 |
| philiKON | right | 16:40 |
| J1m | Of course, it provdes a base for othe frameworks. | 16:40 |
| philiKON | sure | 16:41 |
| J1m | BTW, I think, as I've stated for a long time, ZConfig has caused a lot of the grief. | 16:41 |
| J1m | Because of it's inflexible schema management. | 16:41 |
| philiKON | yeah | 16:42 |
| philiKON | being able to pass configuration around the way pastedeploy does it is nice | 16:42 |
| J1m | yes, up to a point. | 16:42 |
| philiKON | i suppose we'll do the zcml setup in the publication then | 16:42 |
| philiKON | it seems the logical place | 16:42 |
| J1m | It would be nice, imo, if the application had fill access to the configuration db. | 16:42 |
| J1m | The ways it's done, you almost have to call out to separate configuration files. | 16:43 |
| J1m | although you could have options who's values are ZConfig snippets. | 16:43 |
| philiKON | yuck | 16:43 |
| *** rcrafton has joined #zope3-dev | 16:43 | |
| * philiKON doesn't like zcml-in-buildout.cfg or zconfig-in-buildout.cfg | 16:43 | |
| philiKON | or paste.ini-in-buildout.cfg for that matter | 16:44 |
| philiKON | but i suppose that's a matter of taste and deployment method | 16:44 |
| J1m | so, for example, you could probably have a zodb option that had the ZConfig that defined the databases. | 16:44 |
| philiKON | yup | 16:44 |
| philiKON | was thinking about that too | 16:44 |
| J1m | I understand your reaction, although it obviously doesn;t bothe me as much as it does you. | 16:44 |
| J1m | I *really* like having everything in one place. | 16:44 |
| philiKON | i understand that of course | 16:45 |
| J1m | To me, that's worth micing syntaxes, which, fortunately, config-parser makes pretty easy. | 16:45 |
| *** sorindregan has quit IRC | 16:45 | |
| philiKON | and it doesn't bother me a lot because *i* know where to look for things and where to modify things | 16:45 |
| philiKON | othe rpeople don't | 16:45 |
| philiKON | sooooo many people start messing with parts | 16:45 |
| philiKON | i've seen it over and over and over | 16:45 |
| *** derek|laptop has quit IRC | 16:46 | |
| philiKON | anyway, not the discussion i want to get it into :) | 16:46 |
| J1m | Yeah, I guess that's an editcation issue. | 16:46 |
| *** derek|laptop has joined #zope3-dev | 16:46 | |
| J1m | Yeah, I guess that's an education issue. | 16:46 |
| philiKON | yup, in a way | 16:46 |
| philiKON | it's also an issue of what people are generally used to and what buildout does | 16:46 |
| J1m | I guess there should be a greater emphasis up front in the buildout docs. | 16:46 |
| J1m | There are lots of systems that manage configuration for people. | 16:47 |
| J1m | This is not a new concept. | 16:47 |
| philiKON | yeah. debian has quite a lot of configuration file generation as well | 16:47 |
| philiKON | which always confuses the heck out of me when i want to change soething once a year :) | 16:47 |
| J1m | It's the price of automation. | 16:47 |
| philiKON | yup | 16:47 |
| philiKON | anyway, thanks for the input on zope.publsiher.paste | 16:48 |
| J1m | It really made writing this key app a lot easier. | 16:48 |
| philiKON | we're considering using it for grok's wsgi story | 16:48 |
| J1m | I wish I had the bobonic flavor. | 16:48 |
| J1m | cool | 16:48 |
| philiKON | heh, right | 16:48 |
| J1m | we really need to do zope.app.publisher.paste. | 16:49 |
| J1m | I'm just soooo pressed for time. | 16:49 |
| J1m | But, back to simple apps. | 16:49 |
| philiKON | what would that flavour of ...paste do? | 16:49 |
| philiKON | just configure zope3's standard publication? | 16:49 |
| J1m | It would deal with the richer request-selection semantics there, like xml-rpc etc. | 16:50 |
| philiKON | ah, right | 16:50 |
| philiKON | maybe that'll come up as we deal with grok's wsgi story | 16:50 |
| J1m | for simple apps, I'd like to have something as simple as bobo used to be. | 16:50 |
| philiKON | grok is pretty simple :) | 16:50 |
| *** projekt01_ is now known as projekt01 | 16:50 | |
| J1m | and I want simple apps to start quickly enough that they could be delployed as cgi. | 16:51 |
| J1m | The key upload script really should be a cgi script. | 16:51 |
| *** sp0cksbeard has joined #zope3-dev | 16:51 | |
| J1m | Is there a cgi wsgi wrapper? | 16:52 |
| philiKON | yeah, i think so | 16:52 |
| philiKON | in wsgiref | 16:52 |
| J1m | certainly writing one would be simple enough. | 16:52 |
| philiKON | allows you to mount any wsgi app in a cgi server | 16:52 |
| *** AdamAtNCPC has joined #zope3-dev | 16:52 | |
| * philiKON goes back to sprinting | 16:52 | |
| philiKON | thanks again! | 16:53 |
| J1m | But the sample publisher in the test took about .6 seconds just to import on a not too slow machine. | 16:53 |
| philiKON | ouch | 16:53 |
| J1m | mainly, I assume because of all of the zope.publisher dependencies. | 16:53 |
| philiKON | regarding the bobonic flavour, are you planning on working with the repoze guys a bit on that? | 16:53 |
| J1m | not really | 16:54 |
| philiKON | i haven't really looked at this, but they do have some sort of primitive bobo-like thing | 16:54 |
| philiKON | i see | 16:54 |
| * philiKON doesn't really know much of the repoze internals anyway | 16:54 | |
| * philiKON should find the time to look at it | 16:54 | |
| J1m | I think their focus is very much on zope 2, which is fine. | 16:54 |
| J1m | I really want to do something smaller. | 16:55 |
| philiKON | i think chris wants to fit zope 3 into the picutre as well | 16:55 |
| *** aclark has joined #zope3-dev | 16:55 | |
| J1m | I think repoz is a bigger project than it needs to be. <shrug> | 16:55 |
| mcdonc | we'd like to use a z3 publisher | 16:55 |
| philiKON | there he is :) | 16:55 |
| philiKON | hey mcdonc | 16:55 |
| mcdonc | hey philiKON | 16:55 |
| mcdonc | it would be ideal if a newer publisher didn't delegate to the request for traversal; we'd likely just use it then | 16:56 |
| J1m | mcdonc, you don't have to delegate to the request for traversal. | 16:56 |
| J1m | You could do traversal yourself in the getApplication call or in the first TraverseName call. | 16:57 |
| J1m | The traveral data is available and manipulatable. | 16:58 |
| J1m | This was to allow traversers to consune multiple names. | 16:58 |
| *** AdamAtNCPC has left #zope3-dev | 16:58 | |
| J1m | consume | 16:58 |
| mcdonc | so what's the relationship between that and what i mentioned at http://www.gossamer-threads.com/lists/zope/dev/213361#213361 | 16:59 |
| mcdonc | (zope.publisher's "publish" function) | 16:59 |
| J1m | I didn't response there because I had forgotten that you can traverse the entire path in one go, if you wish. | 17:00 |
| J1m | respond | 17:00 |
| *** salfield has quit IRC | 17:01 | |
| J1m | The publication *can* do all the traversal if it wants. | 17:01 |
| J1m | I don't particularly want to make it do all the traversal. | 17:01 |
| *** b52laptop has quit IRC | 17:03 | |
| *** b52laptop has joined #zope3-dev | 17:03 | |
| mcdonc | maybe i'm missing it, but it seems like that top-level "publish" function is always used during publishing and is not pluggable? | 17:03 |
| mcdonc | one reason "obob" exists is because in a bare wsgi context we dont yet have a "request" | 17:04 |
| mcdonc | we only have the WSGI environment | 17:04 |
| mcdonc | and our "publication" is actually the thing that manufactures the request | 17:04 |
| mcdonc | but maybe i'm just off in the weeds, it's been a few weeks since i looked at it | 17:05 |
| J1m | right, it is not pluggable, but the request object has an API for manipulating the traversal stack. So, for example, a publication's getApplication method could use this API to consume the traversal stack, doing the traversal itself. The object it returns could then be the result of that traversal. | 17:05 |
| projekt01 | mcdonc, there is a pluggable traverser in z3c.traverser probably this could give you hints about the traversal step consume part | 17:07 |
| mcdonc | i'll need to see how it would be possible to use that publishing dance against something that implements z2's request iface | 17:09 |
| mcdonc | or maybe the publication will need to adapt it at each step | 17:10 |
| philiKON | mcdonc: see five.publication in my svn.zope.org sandbox | 17:21 |
| philiKON | mcdonc: it's a publication for zope.publisher that runs Zope2 | 17:21 |
| philiKON | mcdonc: well, tries to ;) | 17:21 |
| mcdonc | obob "helper" (akin to a publication) assumes there is no request, just the wsgi environ, which works pretty well for us because it means we just create whatever kind of request during helper setup using the environ based on the kind of app we're serving (z3 or z2). | 17:22 |
| *** dvschramm has joined #zope3-dev | 17:24 | |
| *** salfield has joined #zope3-dev | 17:24 | |
| *** jsadjohnson has joined #zope3-dev | 17:24 | |
| *** kursor has joined #zope3-dev | 17:27 | |
| *** afd_ has quit IRC | 17:28 | |
| *** goschtl has quit IRC | 17:31 | |
| *** afd_ has joined #zope3-dev | 17:31 | |
| *** goschtl has joined #zope3-dev | 17:31 | |
| *** MJ has quit IRC | 17:41 | |
| *** regebro has joined #zope3-dev | 17:43 | |
| *** rmarianski has joined #zope3-dev | 17:46 | |
| *** jsadjohnson has quit IRC | 17:51 | |
| *** sm has joined #zope3-dev | 17:51 | |
| *** pcardune has joined #zope3-dev | 17:52 | |
| *** kursor has quit IRC | 17:59 | |
| *** pcardune has quit IRC | 18:05 | |
| *** aclark is now known as aclark|away | 18:09 | |
| *** nathany has joined #zope3-dev | 18:11 | |
| *** zagy has joined #zope3-dev | 18:19 | |
| *** faassen has joined #zope3-dev | 18:24 | |
| J1m | Theuni, ayt? | 18:25 |
| *** b52lap has joined #zope3-dev | 18:30 | |
| *** b52laptop has quit IRC | 18:30 | |
| *** rcrafton has quit IRC | 18:34 | |
| *** fcorrea has joined #zope3-dev | 18:34 | |
| *** rcrafton has joined #zope3-dev | 18:34 | |
| *** djohnson has quit IRC | 18:38 | |
| *** projekt01 has quit IRC | 18:44 | |
| J1m | tlotze, ? | 18:46 |
| Theuni | here | 18:47 |
| J1m | how many zeo cache fixed did you commit? | 18:47 |
| Theuni | 'fixes'? | 18:47 |
| J1m | sorry yes | 18:47 |
| Theuni | hmm | 18:47 |
| J1m | I want to merge them to the 3.8 branch. | 18:48 |
| J1m | I know of 1. | 18:48 |
| Theuni | let me have a look at the checkin log | 18:48 |
| Theuni | one larger part was the test i added | 18:48 |
| Theuni | from the changelog i suspect: the fix for #98275 and the problem with objects larger then the total cache size | 18:50 |
| Theuni | i'm still having a look at the svn log though | 18:50 |
| *** b52lap has quit IRC | 18:51 | |
| Theuni | J1m: #181712 is not ported to 3.8 yet as the situation is different than on the turnk | 18:54 |
| Theuni | trunk | 18:54 |
| *** b52laptop has joined #zope3-dev | 18:54 | |
| J1m | Is that a launchpad #? | 18:55 |
| Theuni | yes | 18:55 |
| Theuni | issue 98275 was ported to 3.8 already | 18:55 |
| Theuni | and before that i didn't touch the cache afair | 18:56 |
| *** agroszer has joined #zope3-dev | 18:56 | |
| *** goschtl_ has joined #zope3-dev | 18:57 | |
| J1m | k, I'll port http://svn.zope.org/?rev=82895&view=rev | 18:57 |
| Theuni | yup | 18:57 |
| Theuni | it's the only obvious one i see | 18:57 |
| J1m | Do you know if there are any signal handlers in zope.server? | 18:59 |
| J1m | I don't see any. | 18:59 |
| * Theuni looks | 19:02 | |
| * J1m tries to figure out how to understand the zope 3 shutdown process. :) | 19:02 | |
| *** lucielejard has quit IRC | 19:02 | |
| Theuni | that's a meta activity! | 19:03 |
| *** djohnson has joined #zope3-dev | 19:03 | |
| J1m | yes | 19:03 |
| Theuni | this reminds me so of 2003 ... | 19:03 |
| *** goschtl has quit IRC | 19:04 | |
| Theuni | hmm | 19:06 |
| Theuni | thats threadedasync, right/ | 19:06 |
| Theuni | ? | 19:06 |
| *** quodt has quit IRC | 19:06 | |
| J1m | I think I'll just add a print to FileStorage.close and see if it's called. :) | 19:06 |
| Theuni | heh | 19:06 |
| J1m | threadedasync shouldn't have anything to do with signals. | 19:06 |
| Theuni | that's an area i'm not too familiar with | 19:06 |
| Theuni | i mean. the main loop i found is in zope.app.server | 19:07 |
| Theuni | and that comes from threadedasync | 19:07 |
| J1m | I'm 80% sure we don't shutdown cleanly. | 19:07 |
| J1m | If we don't, I've found a source of cache corruption. | 19:07 |
| Theuni | true | 19:08 |
| Theuni | probably the scan() method then should work a bit harder to detect this state | 19:08 |
| J1m | I doubt the format has enough redundant data for it to recover. | 19:09 |
| J1m | well, I suppose it could recover what it's already scanned and mark everything else as free. | 19:09 |
| Theuni | right | 19:09 |
| Theuni | detection would be my major concern | 19:09 |
| *** grahamperrin has joined #zope3-dev | 19:10 | |
| J1m | well, it detects corruption now. | 19:10 |
| Theuni | uh | 19:11 |
| Theuni | right | 19:11 |
| Theuni | :) | 19:11 |
| *** malthe has joined #zope3-dev | 19:11 | |
| Theuni | it might be nice if it detects it earlier | 19:11 |
| Theuni | or maybe we're talking about different symptoms | 19:12 |
| J1m | weird, I just tried to bzip a cache file and it only got a tiny bit smaller. | 19:12 |
| J1m | which does not compute. | 19:12 |
| Theuni | Do you think there might currently be a way to corrupt the cache, let scan succeed with faulty data and only fail much later? | 19:12 |
| J1m | No. I don't have many opinions yet. :) | 19:13 |
| Theuni | alright. | 19:13 |
| Theuni | what overall size was the cache file? | 19:13 |
| J1m | But I think it's possible that, if nothing else, when scan detects a problem, it could keep what it already scanned successfully and mark everything else as free. | 19:14 |
| Theuni | right | 19:14 |
| J1m | I had a 4GB cache file. | 19:14 |
| Theuni | then a tiny bit smaller isn't much :) | 19:14 |
| J1m | It compressed to about 3.9G. | 19:14 |
| J1m | It compressed to about 3.8G. | 19:14 |
| J1m | I was hoping for much more. | 19:15 |
| Theuni | yeah | 19:15 |
| Theuni | i would have thought that to be more too | 19:15 |
| Theuni | maybe pickles aren't that redundant ... | 19:15 |
| Theuni | but still | 19:15 |
| *** pcardune has joined #zope3-dev | 19:15 | |
| J1m | I suspect many images because we aren't using blobs yet. | 19:15 |
| Theuni | i don't see any signal handling in any zope.* packages | 19:17 |
| Theuni | zdrun does something | 19:17 |
| Theuni | but i'm not sure thats relevant | 19:17 |
| J1m | no | 19:17 |
| *** quodt has joined #zope3-dev | 19:18 | |
| *** J1m has quit IRC | 19:19 | |
| *** b52laptop has quit IRC | 19:27 | |
| *** sunew|away has quit IRC | 19:27 | |
| *** mkerrin has quit IRC | 19:28 | |
| *** goschtl_ has quit IRC | 19:32 | |
| *** goschtl has joined #zope3-dev | 19:32 | |
| *** dvschramm has quit IRC | 19:34 | |
| *** dvschramm has joined #zope3-dev | 19:35 | |
| *** afd_ has quit IRC | 19:39 | |
| *** J1m has joined #zope3-dev | 19:44 | |
| *** dobee has joined #zope3-dev | 19:46 | |
| *** dobee has quit IRC | 19:53 | |
| *** salfield has quit IRC | 19:55 | |
| *** dobee has joined #zope3-dev | 19:57 | |
| *** baijum has joined #zope3-dev | 20:00 | |
| *** grahamperrin has left #zope3-dev | 20:04 | |
| *** reco has joined #zope3-dev | 20:04 | |
| *** djohnson has quit IRC | 20:05 | |
| *** reco has quit IRC | 20:07 | |
| *** reco has joined #zope3-dev | 20:08 | |
| *** scooterXL has joined #zope3-dev | 20:12 | |
| *** RaFromBRC has joined #zope3-dev | 20:16 | |
| *** aclark|away is now known as aclark | 20:27 | |
| *** thruflo has quit IRC | 20:27 | |
| *** whit has quit IRC | 20:34 | |
| *** whit has joined #zope3-dev | 20:34 | |
| *** baijum has quit IRC | 20:45 | |
| *** jodok has joined #zope3-dev | 20:48 | |
| *** whit has quit IRC | 20:50 | |
| *** yotaff has quit IRC | 20:54 | |
| *** tarek has quit IRC | 21:01 | |
| *** harobed has quit IRC | 21:02 | |
| *** tarek has joined #zope3-dev | 21:03 | |
| *** ktwilight_ has joined #zope3-dev | 21:04 | |
| *** djohnson has joined #zope3-dev | 21:08 | |
| *** goschtl_ has joined #zope3-dev | 21:15 | |
| *** ktwilight has quit IRC | 21:17 | |
| *** goschtl has quit IRC | 21:22 | |
| Theuni | J1m: the ClassProvides descriptor seems to be broken for subclasses | 21:29 |
| Theuni | if class A provides an interface then a subclass of A will list __provides__ when you call `dir` but accessing __provides__ will raise an attributeerror | 21:29 |
| J1m | No, it works perfectly | 21:29 |
| Theuni | should we consider that a bug? | 21:29 |
| J1m | no | 21:30 |
| Theuni | interesting | 21:30 |
| Theuni | can you enlighten me why? | 21:30 |
| J1m | depending on how you interpret dir, dir is broken. | 21:30 |
| J1m | ClassProvides makes declarations for the class, **not** the subclasses. | 21:31 |
| *** scooterXL has quit IRC | 21:31 | |
| J1m | The descriptor it uses intentionally raises AttributeError if used on a subclass. | 21:31 |
| *** goschtl_ has quit IRC | 21:31 | |
| *** goschtl has joined #zope3-dev | 21:32 | |
| J1m | descriptors are allowed to raise AttributeError. Code that finds a descriptor and assumes that an attribute is available are Broken. | 21:32 |
| J1m | with a capital B. :) | 21:32 |
| Theuni | heh | 21:34 |
| J1m | Not everyone agrees with me regarding whether or not these declarations should be inherited. | 21:35 |
| Theuni | looks like python doesn't ;) | 21:36 |
| J1m | As Antonin Scalia would say, they're wrong. :) | 21:36 |
| Theuni | i read through the code and i think i understand the dance going on there | 21:37 |
| Theuni | so you think the dance of removing __providess_ from the locals and putting it on the class object should avoid inheritance? | 21:39 |
| J1m | I don't know exactly what you're saying, but no. | 21:40 |
| Theuni | hmm | 21:41 |
| J1m | The only way I know of to avoid inheritance is to have the descriptor complain if it is used on a subclass. | 21:41 |
| Theuni | ah | 21:41 |
| Theuni | reading the code I see that __provides__ is being moved out of the locals of the class and stuck onto the class object. i'm not quite sure why that dance happens. | 21:41 |
| Theuni | so you're saying there should be a way to declare that you don't want an attribute to be inherited? | 21:42 |
| J1m | I have no idea what the "locals of the class" is. | 21:42 |
| Theuni | the class' dict | 21:42 |
| J1m | Before discussing this with me, you need to learn about meta classes. | 21:42 |
| Theuni | i'm referring to _classProvides_advice | 21:43 |
| J1m | and how Python classes actually work. | 21:43 |
| Theuni | looks like it | 21:43 |
| * Theuni wanders off to read | 21:43 | |
| *** dunny has joined #zope3-dev | 21:50 | |
| *** djohnson has quit IRC | 21:53 | |
| *** scooterXL has joined #zope3-dev | 22:15 | |
| J1m | Theuni, lots of tests breal when I try to merge 82895. :( | 22:44 |
| *** agroszer_ has joined #zope3-dev | 22:44 | |
| J1m | oh, and I see why. | 22:46 |
| J1m | Gaaah | 22:46 |
| *** agroszer_ has quit IRC | 22:47 | |
| *** faassen has quit IRC | 22:48 | |
| Theuni | ? | 22:48 |
| J1m | I had to adjust the changes to take versions into account. | 22:48 |
| J1m | The trunk code has version support ripped out. | 22:48 |
| Theuni | right | 22:49 |
| * Theuni runs | 22:49 | |
| *** RaFromBRC is now known as RaFromBRC|brb | 22:49 | |
| *** dobee has quit IRC | 22:52 | |
| *** djohnson has joined #zope3-dev | 22:53 | |
| *** agroszer has quit IRC | 22:56 | |
| *** dobee has joined #zope3-dev | 22:56 | |
| *** dobee has quit IRC | 22:57 | |
| *** nathany has quit IRC | 22:59 | |
| *** derek|laptop has quit IRC | 22:59 | |
| *** lisppaste6 has quit IRC | 22:59 | |
| *** tlotze has quit IRC | 22:59 | |
| *** nathany has joined #zope3-dev | 22:59 | |
| *** derek|laptop has joined #zope3-dev | 22:59 | |
| *** lisppaste6 has joined #zope3-dev | 22:59 | |
| *** tlotze has joined #zope3-dev | 22:59 | |
| *** dobee has joined #zope3-dev | 23:02 | |
| *** RaFromBRC|brb is now known as RaFromBRC | 23:06 | |
| *** djohnson has quit IRC | 23:11 | |
| *** lisppaste6 has quit IRC | 23:14 | |
| *** lisppaste6 has joined #zope3-dev | 23:15 | |
| *** scooterXL has quit IRC | 23:20 | |
| *** harobed has joined #zope3-dev | 23:22 | |
| *** dobee has quit IRC | 23:23 | |
| *** BjornT has quit IRC | 23:47 | |
| *** greenman has joined #zope3-dev | 23:55 | |
| *** aclark is now known as aclark|away | 23:57 | |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!