IRC log of #zope3-dev for Saturday, 2005-06-18

*** projekt01 has quit IRC00:18
*** tvon has quit IRC00:20
*** bskahan has joined #zope3-dev00:50
*** SureshZ has left #zope3-dev00:50
*** philiKON has quit IRC00:51
srichterJ1m: what is the latest way to make a release?00:57
srichterI cannot build a package with my old script anymore00:57
srichterdid you remove the X?00:57
J1mYes00:58
J1mLook at what I checked in in releases00:58
J1mI checkd in a "test" script00:58
srichterok, this is the same I had as a shell script ;-)00:58
J1mNote that it assumes that that zpkgtools checkout is a sibling to the zope checkout00:59
srichterlike my setup :-)00:59
J1mcool00:59
J1mand it uses a local zope map that has relative path references in it.01:00
J1mThis wasn't possible until a couple of weeks ago01:00
srichterok01:01
srichterJ1m: I get an error:01:06
srichtertar xozf Zope-0.0.0.tgz01:06
srichter./configure --prefix `pwd`/z01:06
srichter./configure: line 24: syntax error near unexpected token `<<<'01:06
srichter./configure: line 24: `<<<<<<< .mine'01:06
srichtersomeone did not clean up something (this is a new pure checkout01:06
srichteroh ok, it might be zpckgtools, which I modified before01:07
srichtermy bad01:07
*** rk_ has joined #zope3-dev01:22
*** tvon has joined #zope3-dev01:30
*** xerophyte has left #zope3-dev01:48
*** rk_ has quit IRC01:54
*** bskahan has quit IRC02:01
J1msrichter, how's it going?02:20
*** J1m has quit IRC02:21
*** bskahan has joined #zope3-dev02:34
*** projekt01 has joined #zope3-dev03:02
*** yota has quit IRC03:04
*** dagnachew has joined #zope3-dev03:13
*** bskahan has quit IRC03:30
*** tarek has quit IRC03:42
*** projekt01 has quit IRC03:45
*** stub has joined #zope3-dev03:50
* dagnachew is away: supper !03:50
*** dagnachew has quit IRC05:49
*** stub has quit IRC06:30
*** bradb is now known as bradb-afk07:02
*** strichter has joined #zope3-dev09:03
*** srichter has quit IRC09:03
*** SureshZ has joined #zope3-dev09:24
*** SureshZ has left #zope3-dev09:37
*** timte has joined #zope3-dev09:54
*** philiKON has joined #zope3-dev10:42
*** philiKON has joined #zope3-dev10:51
*** jhauser has quit IRC11:04
*** philiKON has quit IRC11:10
*** timte has quit IRC12:16
*** yota has joined #zope3-dev13:13
*** projekt01 has joined #zope3-dev13:23
*** debugger has joined #zope3-dev13:28
*** andrew_m has joined #zope3-dev13:36
*** philiKON has joined #zope3-dev14:10
andrew_mi'm having some problems with logging out of zope3. does anyone know if that's a zope, browser or configuration issue?14:22
VladDraccould be a browser issue - some browsers have difficulty clearing the http authentitacion14:23
VladDrac..cation14:23
* VladDrac uses ff webdev extensions to be sure14:23
andrew_mthanks - is that something that fixes the problem e.g. on firefox?14:25
VladDracyeah (ff = firefox)14:25
VladDracit has a 'clear http authentication' option14:25
andrew_mi see.. i'll have a look at that, thanks14:27
*** philiKON has quit IRC14:31
andrew_mVladDrac: hmm.. before i spend my whole weekend searching for that extension ;), any hints where to find that?14:36
VladDraclet's see14:36
*** J1m has joined #zope3-dev14:37
VladDracvery annoying, firefox extensions urls now force you to a mozilla firefox upgrade page14:38
VladDrachttp://chrispederick.com/work/firefox/webdeveloper/14:38
andrew_moke, thanks.. was looking on the official firefox extensions page14:39
andrew_mVladDrac: do you make your users install that when they use firefox?14:40
VladDracno14:43
VladDracI use it as a developer14:43
*** bskahan has joined #zope3-dev14:47
andrew_mic.. i guess for the user the only way to log out then is to close the browser for now14:47
VladDracyup14:47
VladDracor switch to some sort of cookie based authentication14:47
andrew_mhmyeah..14:48
J1mlike session auth which is provided by zope.app.authentication14:48
andrew_mJ1m: ok, that sounds like an option to try for me14:48
*** strichter is now known as srichter14:53
*** ChanServ sets mode: +o srichter14:53
projekt01Jim, do you have time to answer a short question?14:54
srichterandrew_m: HTTP Authentication does not have the concept of logout14:54
J1mprojekt01, sure14:55
projekt01What's the best way to fire a event in the method directlyProvide()14:55
srichterandrew_m: you can only rechallenge, but the browser keeps the old settings, unless the new ones are accepted14:55
projekt01;-)14:55
projekt01Replace the interface module?14:55
J1muh, yes14:56
J1msrichter, how goes the release?14:56
SteveAprojekt01: provide your own API that fires the event you want, and calls the zope.interface.diectlyProvides()14:56
andrew_msrichter: ok, that's good to know14:56
srichterI have started last evening and the functional test seems mysterios to me. I do have a plan for the import thingy though14:57
srichterprojekt01: I agree with SteveA14:57
projekt01But if I use a method hook, all other packages don't use this and I can't exactly control this part14:58
srichterprojekt01: right, but it is very unlikely that you are in general interested in knowing who got a directlyProvided interface15:00
projekt01Other packages = 3rd party packages15:00
srichterfor example, I doubt that you are interested in the skin the request has (which is a directlyProvides() call)15:00
projekt01I'm interested to control the "directly provided" interface order15:01
SteveAthat's like saying "i want to hook into all method calls"15:01
J1myou definately don't want to change the order implicitly15:01
SteveAit's hooking into a low-level interface that is used for very many things15:01
srichterbut you are only interested in doing this in a particular context, such as your specific content types or so15:02
projekt01But nobody can say which order this "directly provided" interface really have, two objects of the same type can have different orders15:02
J1mdirectlyProvides completely specifies the order by the order of its arguments.15:03
J1mThat is explicit.15:03
J1mIt would be evil to change that.15:03
J1mIf I say: directlyProvides(o, I1, I2)15:04
J1mIt would be bad if something switched the order of I1 and I2 behind the scenes.15:04
projekt01Yes, but if I directlyProvide(IA, [Iother]) and directlyProvide(IB, [Iothers])  or directlyProvide(IB, [Iothers]) and then directlyProvide(IA, [Iother]) I get a different order15:04
projekt01First I get (IA, IB) and second I get (IB, IA)15:05
projekt01Sorry, forgot the Iothers15:05
J1mI find that a bit hard to believe, since directlyProvides replaces the directly provided interfaces.15:06
projekt01If I directlyProvide a Interface I append it to the directlyProvided interface (each time) and this will produce different orders15:07
projekt01If we directlyProvide different Interfaces15:07
projekt01The interface order depends on the order you call the different directlyProvide at all and not how you call the method,15:09
projekt01Because if you call directlyProvide you don't take care on other interfaces orders (normaly) where the object allready provides15:10
J1mno, directlyProvides does not append15:10
J1m>>> o = C()15:10
J1m>>> directlyProvides(o, I1, I3)15:10
J1m>>> list(providedBy(o))15:10
J1m[<InterfaceClass __main__.I1>, <InterfaceClass __main__.I3>]15:10
J1m>>> directlyProvides(o, I2, I3)15:10
J1m>>> list(providedBy(o))15:10
J1m[<InterfaceClass __main__.I2>, <InterfaceClass __main__.I3>]15:10
J1m>>> o = C()15:10
J1m>>> directlyProvides(o, I2, I3)15:10
J1m>>> directlyProvides(o, I1, I3)15:11
J1m>>> list(providedBy(o))15:11
J1m[<InterfaceClass __main__.I1>, <InterfaceClass __main__.I3>]15:11
J1m>>>15:11
projekt01The method does not append but if we use the directlyProvide we put the additional interface in front or at the end (normaly)15:13
projekt01We don't take care on orders15:13
projekt01And this is like append, it depends what you do first15:13
projekt01And what next and last15:13
J1mThe directly provided interfaces get combined with the class implemented interfaces.15:14
J1mThis happens in a fixed way15:14
J1mit has no dependence on ordr of directlyProvides calls.15:14
srichterthat's because all these pieces of the framework do not care about the order, since the directly provided interfaces are usually unrelated to each other15:14
J1msrichter, that's often not true.15:15
J1mOften you do care about the order.15:15
J1mSometimes, one would like to directlyProvide an interface and have it take a lower precedence than existing interfaces.15:15
srichterbut I meant that the various packages that might directly provide an interface on the request, for example, do not care about the order15:16
J1mThere are similar issues with the interface combinations that happen with decorators.15:16
J1msrichter, usually that is true, yes.15:16
srichterfor example, the skin code does not care about the order of the other interfaces; it only cares about finding the skin interface15:16
* J1m expects SteveA to say he wishes nothing depended on interface order. 15:16
projekt01srichter, if the don't force using different adapters yes, otherwise no15:17
projekt01SteveA, hm, yes and no, it could be a interesting pattern if we can take really care on that part.15:18
J1mprojekt01, I could see having a flavor of directlyProvides that put some directly-provided interfaces after the implemented ones.15:18
J1mI certainly wouldn't do anything about this using events.15:18
SteveAi wonder about having an option to say "adapt, provided it isn't ambiguous"15:19
SteveAi've needed this for some use of getting adapters to determine authorization15:19
SteveAwhere having ambiguity may mean the programmer has not realized something, and made a mistake15:19
J1mIt's never ambiguous :)15:19
SteveAfor the whole system, yes15:20
SteveAfrom the partial knowledge of one developer, it often is15:20
SteveAor rather, it is easy to get the wrong idea15:20
SteveAwhich is not good, where security is concerned15:20
SteveAnow, you could say "don't use adapters for security then" ;-)15:20
J1m:)15:20
SteveAbut, in this case, my infrastructure raises an exception when there is such ambiguity15:21
J1mHow do you define ambiguity?15:21
SteveAin this particular case, when one name is in more than one interfaces provided by an object, where one of those interfaces does not extend the other15:22
SteveAi'm looking up an adapter that is going to say who is authorized to do 'permission' on that name15:22
SteveAoh, one thing i've come across a few times that seems to be missing from zope.interface15:23
J1mThat seems like a valid strategy15:23
J1mThat seems like a intresting strategy15:23
J1mThat seems like a interesting strategy15:23
projekt01SteveA, I don't understand, what have a name of a interface to do with that? You mean attribute name?15:24
SteveAis to be able to say not IFoo(object) or getAdapter(IFoo, object), but getAdapterSpecificallyFor(IFoo, object, specific_iterface=IThingObjectHas)15:24
SteveAor otherwise, getAdapterFactory(IThing, IFoo)15:24
SteveAalthough, i'd also have to say how many dependents i have, and the name15:24
SteveAprojekt01: by 'a name in an interface', i'm talking about an attribute name15:25
SteveAthat is, an attribute that is described by that interface15:25
J1mSteveA, you want adapter super IOW.15:25
SteveAor, described within the interface, if you prefer15:25
* SteveA tries to remember exactly what super does15:26
J1mI plan a redesign of the adapter registry that would allow super operations.15:26
projekt01SteveA, Ok, I see, it sounds interesing15:26
J1mIt is useful in inheritence for looking up a method that a method overrides.15:26
SteveAi'm not sure i want adapter super15:26
J1mIn the adapter case, it would be used by an adapter to call the adapter it overrised.15:27
SteveAbut i'm not sure i don't15:27
SteveAwhen you say getAdapter(IFoo, obj), then all interfaces provided by obj are considered, in a particular order15:27
J1mI wanted to do this for 3.1, but had too much customer work.15:27
SteveAwhat I want to do is to look up IFoo for a particular interface that I choose15:27
SteveAright now, i get around it by making an object that provides the one interface i'm interested in15:28
J1mYou can do that now in 2 steps.15:28
SteveAand say IFoo(object_that_provides_IBar_only)15:28
SteveAdarn customers, always getting in the way of writing software...15:28
J1mYou can get the adapter registry and call lookup on it.15:28
J1mYeah, they're a real pain sometimes15:29
J1mThe lookup fuction takes interface specifications15:29
J1mIt returns a factory, which you can then call.15:29
J1mSo, it should be straightforward to implement an api like you want.15:29
SteveAokay, i'll take a look.  i'll need to see how names fit it too.15:29
SteveAthanks15:30
SteveAhmm15:30
SteveAsuper opertaions + events sounds a bit like partial adapters15:30
J1muh15:31
J1mI supose super could be used to implement partial adapters.15:31
J1mNot sure how events fit in.15:31
projekt01SteveA, J1m, srichter, thanks for the hints, I have to leave now, nice weekend and see you next time15:32
srichterbye15:32
*** J1m has quit IRC15:36
*** projekt01 has quit IRC15:52
*** mikka__ has joined #zope3-dev16:05
*** timte has joined #zope3-dev16:10
*** MiUlEr has quit IRC16:18
*** bskahan has quit IRC16:43
*** nederhoed has joined #zope3-dev16:50
*** anguenot has joined #zope3-dev16:51
*** mikka__ has quit IRC17:17
*** Aiste has quit IRC17:18
*** ignas has joined #zope3-dev17:35
*** garrett-smith has joined #zope3-dev18:51
*** garrett-smith has left #zope3-dev18:51
*** ignas has quit IRC18:51
nederhoedhow can I access the standardmacros from a page template that I add using the management interface?19:01
nederhoed<html19:02
nederhoed  metal:use-macro="views/standard_macros/view">19:02
nederhoedresults in the following error19:02
nederhoed Module zope.pagetemplate.pagetemplate, line 118, in pt_render19:03
nederhoedWarning: Macro expansion failed19:03
nederhoedWarning: exceptions.KeyError: 'views'19:03
nederhoedanyone?19:03
*** projekt01 has joined #zope3-dev19:15
srichteruse context/@@standard_macros19:33
srichterand also, this only works for filesystem-based development19:33
srichterno TTW supported19:33
*** SureshZ has joined #zope3-dev19:56
*** anguenot has left #zope3-dev20:22
nederhoedaha20:23
nederhoedmerci20:23
nederhoedyes! that worked, thank you20:26
*** bskahan has joined #zope3-dev21:15
*** J1m has joined #zope3-dev21:49
J1msrichter, ayt?21:50
srichterJ1m: yep21:50
J1mI just sent you a note.21:51
srichterI am sending the release notes right now21:51
J1mThere is now a separate release for windows.21:51
srichterok, that's what I thought21:51
srichterso this will be the Windows source release that Tim will use21:51
J1myes21:52
J1mIt includes some extra files to support running as a service.21:52
srichterbtw, we really need to share the *.cfg files between Zope-Win and Zope, so that I will not forget to updat them21:52
srichterdoes SVN support links?21:52
J1msvn has externals, but I'm told that you can't use them with files.21:53
srichterok21:53
J1mMost config files are shared.21:53
J1mZope-win is very small and builds on Zope21:53
srichterok, so I guess I have to live with that for now21:53
srichterbut PUBLICATION.cfg, for example, is duplicated21:53
J1mThere's not much to live with21:53
J1mThat is the only one.21:53
srichterah, since you are here; I think we can remove ReleaseSupport now, right?21:54
srichterok, good to know21:54
J1mwhat's that?21:54
srichterit is a directory in the Zope repository that we used for maps to create releases21:54
J1mNo, I don't think we are ready to get rid of it.21:55
srichterreally?21:55
srichterI built the release totally without it21:55
srichteryou basically copied everything interesting into releases21:55
J1mYes, but there are ather releases supported there,21:55
J1mYes, for Zope 3.21:56
J1mMaybe we should get rid of the zope3.map from there.21:56
srichter(except for a couple of map files I created to package zwiki, bugtracker, psycopgda and mysqlda)21:56
srichteroh, maybe I can create similar releases directories for those packages21:56
srichterand for zwiki and bugtracker I simply add them to Zope's21:57
J1mBut I suspect that other releases use the zope3.map21:57
srichtermmh, maybe21:57
srichterWe leave it there, until I release those products for 3.121:57
J1mI think fred uses it for zope.interface.21:58
srichterI see21:58
srichterthat could move to Zope3/releases as well21:58
J1mwell, I guess the zope,interface release could use the map that's already there.21:59
srichteryep21:59
J1mCongrats on the release!22:02
J1mThanks!22:02
srichterthank you22:03
srichterI know you did a lot of work there too22:03
srichterbtw, I decided not to create a branch yet and keep the feature freeze for while longer22:03
J1mwhy?22:03
srichterthis way Julien, Dimitry and other bug fixers have an easier time22:03
J1mNow that you mention it, I was going to complain about that.22:04
srichterLOL22:04
srichterwe have the tag22:04
J1mhave you announced a feature freeze?22:04
srichteryeah, we did that a long time back22:04
srichterI will send another message in a moment22:04
J1mwell, you should announce one now.22:04
J1mEspecially since we don't have a branch.22:04
srichterdone22:10
bskahansrichter, J1m: very nice22:10
srichterJ1m: does a news item take a while to appear on the front page after it is published?22:14
J1mI have no idea.22:14
srichterok, I'll just wait a few mins then22:14
srichterJ1m: will Zope-win also run on Linux?22:18
J1myes22:19
srichterok, then I can test it22:19
J1myes22:19
srichterJ1m: I saw that you decided not to include versioncontrol22:20
J1myes, or wfmc22:21
srichterright, I figured you wanted to wait with wfmc22:21
J1mI don't know that anyone but us is interested in using versioncontrol.22:21
srichterright22:22
srichterI think people are interested, but they don't like the approach22:22
srichterFlorent said that CPS does it another way which he likes better22:22
srichterand wants to use for Z3 ECM22:23
J1mI like it better too, if I understand it correctly.22:23
srichterI did not follow the discussion22:24
J1mWas there a discussion?22:25
srichteron the z3labs list there was a brief one22:26
*** bskahan has quit IRC22:33
srichterarggggggggghhhhhhhhhh, I hate zope.org22:52
srichterit never does what I want22:52
srichterJ1m: even though I have published everyhting in my new Software release Products/Zope3 is broken22:53
srichterdo you know why?22:53
srichterespecially since this works: http://www.zope.org/Products/Zope3/swpackage_view22:55
*** tvon has quit IRC22:59
srichterd2m: are you around?23:03
srichterd2m_: ping23:03
d2msrichter: hi23:04
srichterd2m: are you on the Web team?23:04
d2myes23:04
srichterd2m: Because you are the only one in the #zope-web channel23:04
srichterI really need your help23:05
d2m;)23:05
srichterfor some reason http://www.zope.org/Products/Zope3/ does not work23:05
srichterand I just sent an announcement providing this URL23:05
srichtersome priv issue23:05
srichteron the other hand this works: http://www.zope.org/Products/Zope3/swpackage_view23:05
srichtershould I just create an index_html that forwards to this view?23:06
d2mok, must be a 'private' object somewhere -- i'll check the folder23:06
srichterI thought I published everything23:06
srichterit stopped working after I added the beta 1 Windows file23:06
srichteralso, my news item did not make it to the news list23:07
*** dagnachew has joined #zope3-dev23:09
d2msrichter: checking....23:10
srichterd2m: thanks a lot!23:10
srichterd2m: quick in dirty is better than slow and correct, right now :-)23:10
d2msrichter:  http://www.zope.org/Products/Zope3/ works for me now, http://www.zope.org/News/ and http://www.zope.org/news.rss too23:13
srichterok, what did you do?23:14
srichterdoes it just work?23:14
srichteryipee, for me too :-)23:14
srichterd2m: so you gotta tell me what was wrong23:15
srichterbecause it is release season I have to play this game a couple of times in the next weeks23:15
d2msrichter: no idea, i just hit it a bit :)23:15
srichterooooohhhhhhhh, no!23:15
srichterits the zope.org ghost again :-)23:16
d2msrichter: well, the NewsItem had no EffectiveDate set23:16
srichterahh, I see, good to know that you need this23:16
srichtersame for the other objects?23:16
d2mthe NewsItem should show up on frontpage within the next minutes23:16
srichtercool23:17
srichterthanks a whole lot!23:18
d2msrichter: what is the difference between the 2 release files ?23:33
*** gsbain has joined #zope3-dev23:41
srichterthe windows one installs a fix, so that the release will work with Windows services23:41
d2mi see, ok -- great news btw !23:42

Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!