IRC log of #zope3-dev for Friday, 2005-04-08

*** srichter has quit IRC00:10
*** Damascene has quit IRC00:13
*** mohsen has joined #zope3-dev00:40
*** hazmat has joined #zope3-dev00:43
*** philiKON has quit IRC00:54
*** J1m has quit IRC01:05
*** philiKON has joined #zope3-dev01:32
*** niemeyer has quit IRC01:35
*** ignas has quit IRC01:44
*** alga has quit IRC01:48
*** bskahan has joined #zope3-dev01:50
*** Damascene has joined #zope3-dev01:50
*** Dunebuggy has left #zope3-dev01:50
*** tarek_ has quit IRC01:53
*** d2m has quit IRC02:14
*** Theuni has quit IRC02:26
*** mohsen has quit IRC02:47
*** niemeyer has joined #zope3-dev03:05
*** hazmat has quit IRC03:27
*** philiKON has quit IRC03:37
*** RaFromBRC is now known as RaFromBRC|afk04:37
*** garrett-smith has quit IRC04:37
*** stub has joined #zope3-dev04:39
*** _projekt01 has left #zope3-dev04:44
*** tav|offl1ne has joined #zope3-dev05:09
*** stub has quit IRC05:15
*** tav|offline has quit IRC05:15
*** stub has joined #zope3-dev05:17
*** hazmat has joined #zope3-dev05:19
*** RaFromBRC has joined #zope3-dev05:20
*** `anthony has quit IRC05:24
*** `anthony has joined #zope3-dev05:42
*** RaFromBRC|afk has quit IRC05:43
*** bskahan has quit IRC06:00
*** MiUlEr has joined #zope3-dev06:32
*** hazmat has quit IRC07:03
*** hazmat has joined #zope3-dev07:31
*** niemeyer has quit IRC07:34
*** `anthony has quit IRC07:38
*** Damascene has quit IRC08:25
*** Damascene has joined #zope3-dev08:27
*** nimfa_ has joined #zope3-dev09:07
*** hdima has joined #zope3-dev09:14
*** hazmat has quit IRC09:19
*** `anthony has joined #zope3-dev09:21
*** Aiste|lunch has quit IRC09:24
*** hazmat has joined #zope3-dev09:25
*** `anthony has quit IRC09:26
*** AJC has joined #zope3-dev09:31
*** `anthony has joined #zope3-dev09:38
*** sashav_ has joined #zope3-dev09:42
*** nimfa_ has quit IRC09:51
*** sashav_ is now known as sashav09:53
*** Aiste has joined #zope3-dev10:04
*** hazmat has quit IRC10:08
*** `anthony has quit IRC10:30
*** ChickenBrain has left #zope3-dev10:35
*** Aiste has quit IRC10:48
*** Aiste has joined #zope3-dev11:02
*** lunatik has joined #zope3-dev11:04
*** `anthony has joined #zope3-dev11:47
*** `anthony has quit IRC11:52
*** projekt01 has joined #zope3-dev12:18
*** zagy has quit IRC12:24
*** J1m has joined #zope3-dev12:36
*** `anthony has joined #zope3-dev12:48
*** efge has joined #zope3-dev12:57
*** oferw has joined #zope3-dev13:27
*** oferw has quit IRC13:31
*** regebro has joined #zope3-dev13:39
*** bskahan has joined #zope3-dev13:46
*** SteveA has quit IRC14:10
*** d2m has joined #zope3-dev14:25
*** J1m has quit IRC14:50
*** SteveA has joined #zope3-dev14:59
*** SteveA has quit IRC15:02
*** AJC has quit IRC15:03
*** SteveA has joined #zope3-dev15:03
*** faassen has joined #zope3-dev15:08
*** FarcePest has joined #zope3-dev15:40
*** ignas has joined #zope3-dev15:40
*** gintas has joined #zope3-dev15:44
*** srichter has joined #zope3-dev15:52
*** ChanServ sets mode: +o srichter15:52
*** alga has joined #zope3-dev15:54
*** niemeyer has joined #zope3-dev16:02
*** tonico has quit IRC16:34
*** sashav has quit IRC16:35
*** tonico has joined #zope3-dev16:35
*** tonico has quit IRC16:43
*** J1m has joined #zope3-dev17:00
*** hdima has quit IRC17:11
*** shapr has joined #zope3-dev17:21
shaprI'd like to see the difference, can you point it out?17:22
shaprJ1m: do you have a list of 'basic abstractions' ?17:24
shaprI've been collecting such a thing lately, in hopes of coming up with a design for something that uses a lot less state than Zope.17:25
*** Theuni has joined #zope3-dev17:30
*** shapr has left #zope3-dev17:32
srichterJ1m: I really do not like having the ZODB code as externals :-) "svn stat" gives too much output now17:33
J1msrichter, me neither17:33
srichtera chance we can do it the old way again?17:34
*** Theuni has quit IRC17:35
J1msrichter, I'd like to stick with this for now.17:36
J1mI think we should give the other way a chance.17:36
srichterI have worked 2 days with it and I cannot stand it; it makes it more complicated to see what changed17:37
srichterit definitely decreases my productivity17:37
srichter(just some feedback from me)17:37
gintassrichter, cd into the directory you are working in before invoking 'svn st'17:37
srichterbut I do not want to do that17:38
srichterbecause I also oftemn make changes to doc and package-includes17:38
ignasgintas, that increases the risk of forgetting to commit tests, or some minor changes ...17:38
srichterand if I go into the dir I tend to forget them17:38
srichteralso, I run the tests from the root17:39
gintasyou can work around the problem17:39
gintascheck out ZODB and other modules externally17:39
gintasthen in your checkout's Zope3/src create symlinks to those checkouts17:39
ignassvn st | grep -v external ? :)17:39
gintassvn st will still be noisy, but it will be quick because it won't recurse into the modules17:40
srichterbut I do not want to do that. That is all more work, so I have more overhead which makes me less productive17:40
srichterthe point is it is more confusing or more work than before17:40
srichterand I do not understand why17:40
gintasas far as I understand previously changes would get lost because they were committed to the wrong repository17:41
J1mThe best way to work around this is probably to create a script that runs svn status and filters the output.17:41
gintasconceptually I like the new model better17:41
srichterJ1m: yep, that would at least only require extra work once; but then I have to remember a new command (I know, I am lazy!)17:42
ignassvn st | grep -v external | grep -v X | grep -v ^$:)17:42
ignasthe :) is optional17:42
gintasgrep -v X ?17:43
ignas^X eoulf br better i guess17:43
gintassrichter, my first suggestion was moot, I was thinking about svn up (which takes a while until all the externals are checked) rather than svn st17:44
srichteranyway, I just wanted to raise my voice that I do not like this stuff17:45
J1mI think we should try it a bit longer.17:45
srichterI wish there would be an SVN option to turn off status checking on externals17:45
J1mIf we decode to switch back to copies, we should *not* expect Tim to maintain the copy.17:46
srichterI agree17:46
J1mIt should be *our* responsibility to synchronize w ZODB.17:46
*** SteveA_ has joined #zope3-dev17:49
*** gintas has quit IRC17:57
faassensrichter: I looked around for such a switch, but apparently the svn developers said that would be bad. I wasn't clear on why. :)17:59
*** lunatik has left #zope3-dev18:01
faassenMarius already sent something to the mailing list about this.18:04
faassenand there's also an issue, though according to Marius it conflates too many issues.18:04
faassenOkay, 'checkout', 'export', 'update', and 'status', now accept a --ignore-externals option as of r12832.  Changing this issue summary to track just the stickiness aspect now.18:04
faassenapparently it's coming to some version of svn.18:04
faassenthat was a message from jan 23.18:05
*** tonico has joined #zope3-dev18:05
faassenso it looks like help is on the way, in some unknown version of svn.18:09
*** zagy has joined #zope3-dev18:09
faassenhopefully 1.2, out april/may.18:09
faassenno, even mar/april.18:09
faassenso just about now. :)18:09
faassenyes, changelog confirms it's going to be in 1.218:11
*** SteveA has quit IRC18:11
faassenthey already shipped a rc1 of 1.2 yesterday.18:12
*** nimfa has joined #zope3-dev18:14
*** Aiste has quit IRC18:18
*** nimfa is now known as Aiste18:19
efgeI want to store for the duration of a transaction a list of persistent objects, can I rely on id(ob) not changing under me ? So that I can use id(ob) as a key in a dict18:21
SteveA_id(ob) will not change for an objec18:26
SteveA_you might want to use _p_oid though18:26
SteveA_if they are persistent objects18:26
SteveA_because if they are collected and then reloaded from the connection18:26
SteveA_then the id will change18:26
SteveA_because it will be a new object18:26
SteveA_why not just store a reference to the objects themselves in the list?18:26
*** niemeyer has quit IRC18:27
SteveA_garbage collected18:27
efgefor speed I'd prefer to use a dict18:27
efgeto check for presence faster18:27
SteveA_in schooltool, there's a PersistentObjectsSet (or similar)18:27
efge(some objects may be added several times, I want to keep only one)18:27
SteveA_that is a set, the "setness" based on the _p_oid18:27
SteveA_you could use a dict18:28
SteveA_keys _p_oid, values objects18:28
efgeok I'll use _p_oid then, thanks18:28
SteveA_when you say "persistent"... you mean 1st class persistent, right?18:28
efge_p_oid is set as soon as my ob is stored as subobject of an object that has a _p_jar, right ?18:29
SteveA_so, you might need to use connection.add18:29
SteveA__p_jar is a connection, of course18:29
SteveA_connection.add() was added for exactly this reason18:29
efge(thing is, I'm asking this here but this is for Zope 2 and ZODB 3.2 ;-)18:30
SteveA_this requires zodb3.318:30
efgeadd() ? yes I think I saw that, and it's messier in ZODB 3.218:30
SteveA_i doubt it has been backported18:31
efgeok to hell with speed, I'll use a simple list of aq_base(ob)18:31
efgeand optimize later :)18:31
SteveA_whatever aq_base() means ;-)18:32
SteveA_you can use a dict of id(aq_base(ob)) : aq_base(ob)18:32
SteveA_so, the ob won't be collection18:32
SteveA_so, the ob won't be collected18:32
SteveA_and that gives you a speedy lookup too18:33
SteveA_seeing as it is all in one transaction18:33
efgeactually they have no __hash__ so I can't18:33
efgeah misread sorry18:33
efgeso id() is ok then ?18:33
SteveA_provided you keep a reference to the object18:34
efgeI do, yes18:34
SteveA_which of course you have to do anyway to look it up18:34
SteveA_as you can't look up an object by id, except from C18:34
efgeand anyway I want to stored the wrapped version too18:34
*** zagy has quit IRC18:37
*** bskahan is now known as bska|lunch18:39
*** zagy has joined #zope3-dev18:41
*** hazmat has joined #zope3-dev18:52
*** regebro has quit IRC18:57
*** mkerrin has joined #zope3-dev19:05
srichterfaassen: thanks, that will definitely help; I hope it can be a svnrc option as well19:09
*** hazmat has quit IRC19:09
*** bska|lunch is now known as bskahan19:22
*** tvon has joined #zope3-dev19:26
*** niemeyer has joined #zope3-dev19:27
*** Aiste has quit IRC19:38
*** bradb has quit IRC19:50
*** stub has quit IRC19:50
*** hazmat has joined #zope3-dev19:55
*** regebro has joined #zope3-dev20:16
*** bskahan has quit IRC20:21
*** faassen has left #zope3-dev20:25
*** bskahan has joined #zope3-dev20:26
*** tvon has quit IRC20:38
*** tvon has joined #zope3-dev20:41
srichterJ1m: do you have a moment?20:48
srichterI think I found the problem with the test failure20:48
srichterfor my preference code to work, I have to manually (as in the Python code) register a checker function:20:49
srichterdefineChecker(PreferenceGroup, PreferenceGroupChecker)20:49
srichterhowever, in, a cleanup mfunction is registered with the test cleanup registry that cleans up the entire checker registry20:50
srichterwhich kills my security declaration20:50
srichterso, now I do not know what to do....20:50
J1mWhy don't you define the checker in zcml?20:50
srichterdo we have a directive that let's me specify a checker function for a particular class?20:50
srichterdo we have a directive for this?20:51
J1mDo you need a custom checker?20:51
srichterI need it to be dynamically generated, yes20:51
srichterbecause attributes will be available based on the schema a preference group provides20:51
J1mDo you generate classes?20:53
J1manyway, you could use __Security_checker__20:53
srichterno, but instances provide different types of attributes20:53
J1muse __Security_checker__20:53
srichterdoes this accept a function or a checker class?20:54
J1mIt is a checker instance.20:54
srichterwhat is the value __Security_checker__ should have?20:54
srichterok, I guess I could do this in the __init__20:55
srichteror whenever the schema changes, I change the checker as well20:55
srichterJ1m: btw, there are several failures when running tests with -N 220:57
srichter(some are in apidoc, but some are soemwhere else too20:57
J1mThat doesn't surprise me20:58
*** bradb has joined #zope3-dev20:58
srichterJ1m: btw, I fixed the persistent module code. it works again and I fixed a bug in src/zodbcode (I hope it was not copied)20:59
J1mCool. I still don't want to include that in 3.1. :)21:00
srichterof course21:00
*** MacYET has joined #zope3-dev21:04
*** bskahan has quit IRC21:11
*** MacYET has left #zope3-dev21:24
*** efge has left #zope3-dev21:28
*** mkerrin has quit IRC21:35
*** th1a has quit IRC21:45
*** ignas has quit IRC21:59
*** tvon has quit IRC22:15
*** th1a has joined #zope3-dev22:21
*** tvon has joined #zope3-dev22:37
*** tvon has quit IRC22:48
*** tvon has joined #zope3-dev23:00
*** tvon has quit IRC23:26
*** tvon has joined #zope3-dev23:26
*** tvon has quit IRC23:33
*** regebro has quit IRC23:36

Generated by 2.15.1 by Marius Gedminas - find it at!