IRC log of #zope3-dev for Wednesday, 2005-10-12

*** niemeyer_ has quit IRC00:01
jennerhmm... strange, my /usr/lib/python2.4/site-packages/zope has no __init__.py00:07
*** jinty has quit IRC00:08
jennerso obviously `python -c "from import IContainer"` fails00:09
* jenner switches to debian zope3 package00:16
jennersheesh, I can't get the configure.zcml right after I've moved the code from to photo.py00:53
*** zbir_ has quit IRC00:56
*** efge has quit IRC01:04
*** efge has joined #zope3-dev01:05
*** efge has joined #zope3-dev01:07
*** efge has quit IRC01:26
*** anguenot has quit IRC01:28
jennerphiliKON: ayt?01:40
srichterhe mentioned earlier that he went away, probably to bed ;-)01:41
*** bradb has left #zope3-dev01:42
projekt01srichter, what's the state of viewlet?01:43
srichterprojekt01: I have been busy all day today, so I did not work on it01:44
srichterprojekt01: but Helmut kept fixing the directives01:44
srichterprojekt01: I think it is pretty much working01:44
srichterI will be working on it tomorrow01:44
projekt01Ok, can you drop me a note if you have a stable state. Then I can implement the boston skin again.01:45
* projekt01 hopes to read all this mails so I can work again ;-)01:46
srichterhe he01:47
jennerMaybe someone could give me a hint about this photo issue -- I've got everything working (moved code from to photo[slide].py, fixed the tests and configure.zcmls), all tests run fine, but neither photo nor photoslide show up in the ZMI01:47
srichterprojekt01: so Uwe told me you called the NeckarSprint "The best of breed (Zope 3 Sprints)"01:47
srichterjenner: can you check in the fixes, so I can have a quick look?01:48
srichtereven though we did not get a change to look over your MVC ideas :-|01:48
jennersrichter: I have no svn write acces... I can make a patch against current trunk though01:49
srichteryou shoudl get access :-)01:49
projekt01srichter, I will implement a MVC prototype first, so we can discuss this heavy part based on a real example.01:50
srichterok, cool01:51
projekt01btw, the first components a "protocol router" and a "controller" are working already.01:52
srichter:-) ...whatever this means :-)01:52
projekt01I agree, but Java-servlet guys know what I mean ;-)01:53
jennersrichter: <- it's quite big 'cos is now photo.py01:53
srichterjenner: can you E-mail me the diff?01:54
srichterjenner: also, are you ok, if I look at it tomorrow?01:54
jennersrichter: np01:54
jennersrichter: srichter at
projekt01srichter, here's a "protocol router" description:
srichterjenner: yes01:56
srichterprojekt01: mmh, I would naturally say that our protocol router is the publisher01:57
projekt01more or less, but the protocol are not the requests01:58
projekt01a protocol is defined as more or less a specific mime type not a request01:59
srichterI see01:59
projekt01btw, this documentation is really a good base for sync our naming conventions ;-)02:01
jennerjavas (or better suns) naming conventions are sometimes a bit misleading, they tend to be misused as buzzwords :)02:05
srichterwe noticed this last week at the sprint as well02:06
jennersrichter: mail sent02:08
projekt01I agree, but better to have one definition then everybody uses it's own words for the same thing ;-)02:08
jennerbut _don't_ read it now :)=02:08
*** efge has joined #zope3-dev02:09
*** efge_ has joined #zope3-dev02:16
* jenner is off to bed02:20
*** jenner has quit IRC02:20
*** efge has quit IRC02:27
*** yota has quit IRC02:32
*** zbir has joined #zope3-dev02:36
*** efge_ has quit IRC02:40
*** philiKON has quit IRC02:42
*** tav_ has joined #zope3-dev02:43
*** tav has quit IRC02:50
*** stu1 has joined #zope3-dev03:25
*** projekt01 has quit IRC03:28
*** niemeyer has joined #zope3-dev03:46
*** niemeyer has quit IRC04:02
*** niemeyer has joined #zope3-dev04:07
*** stu1 is now known as stub04:08
*** niemeyer_ has joined #zope3-dev04:13
*** niemeyer has quit IRC04:14
*** niemeyer_ is now known as niemeyer04:14
*** efge has joined #zope3-dev04:23
*** niemeyer has quit IRC04:55
*** matt_ has quit IRC05:45
*** tav has joined #zope3-dev06:11
*** hazmat has joined #zope3-dev06:20
*** ChanServ sets mode: +o hazmat06:20
*** tav_ has quit IRC06:22
*** hazmat has quit IRC07:31
*** efge has quit IRC07:34
*** hazmat has joined #zope3-dev07:49
*** ChanServ sets mode: +o hazmat07:49
*** deo has quit IRC07:54
*** jhauser_ has joined #zope3-dev07:55
*** jhauser has quit IRC08:03
*** dobee has joined #zope3-dev08:21
*** sashav has quit IRC08:42
*** zagy_ has joined #zope3-dev08:49
*** zagy_ is now known as zagy08:49
*** hazmat has quit IRC09:39
*** MacYET has joined #zope3-dev09:40
dobeehi all, i got some odd behaviour with tal:translate ..09:42
dobee   <div>09:42
dobee    <div i18n:translate="" tal:content="python:u'Title'"  i18n:domain="zope"/>09:42
dobee    <div i18n:translate="" i18n:domain="zope">Title</div>09:43
dobee   </div>09:43
dobeei think both things should be translated, but they don't, only the second div,09:43
dobeeis this a bug or do i miss something, thx09:43
*** MJ has quit IRC09:47
*** hdima has joined #zope3-dev09:49
*** tekNico has joined #zope3-dev10:02
srichterdobee: please put up your question again, now that hdima is here; he has recently worked on the i18n namespace10:03
dobeehi all, i got some odd behaviour with tal:translate ..10:04
dobee  <div>10:04
dobee    <div i18n:translate="" tal:content="python:u'Title'"  i18n:domain="zope"/>10:04
dobee    <div i18n:translate="" i18n:domain="zope">Title</div>10:04
dobee  </div>10:04
dobeei think both things should be translated, but they don't, only the second div,10:04
dobeeis this a bug or do i miss something, thx10:04
dobeehdima: ayt?10:04
srichterhdima: I think it is a bug, ut I am not sure10:05
dobeehdima: i did another test because i thought it is implemented this way, to prevent the usage of composed message ids10:06
dobee  <div tal:attributes="title python:u'Title'"10:06
dobee         i18n:attributes="title"10:06
dobee         i18n:translate="" tal:content="python:u'Title'"  i18n:domain="zope"/>10:06
dobee    <div i18n:translate="" i18n:domain="zope">Title</div>10:06
dobeethe title attribute gets translated10:06
d2manyone knows if the zope3 buildout scripts  for windows are in SVN ? cannot find them10:06
dobeehdima: can you give me a hint where the i18n expressions are handled, so i have a look at the source10:07
srichterd2m: you mean the release stuff or the buildout package from Benji, or the stuff Tim uses to build the Windows installer?10:07
srichterdobee: the TAL code is very hard to read :-)10:08
srichterdobee: also, send a message to the list on this10:08
dobeehdima: ok10:08
d2msrichter: i suppose that thing tim_one is doing (he is posting the windows releases to
srichterd2m: yep10:09
d2msrichter: i already found how zope2.8 is creating the releases ( and stuff)10:10
srichterd2m: I am pretty sure, the Windows stuff is not online, but let's have a look at the MakeARelease wiki page, where Tim put some notes about the Windows build10:13
d2msrichter: ok, i posted to zope-dev ML yesterday about creating/changing the zope3 windows installer -- maybe i should repost to zope3-dev to get time_ones attention ?10:15
srichterd2m: see:
srichterd2m: he seems to use standard distutils stuff10:16
d2msrichter: thanks, will search for the packages10:18
*** sashav has joined #zope3-dev10:21
srichterdobee: ah, it's you Bernd :-)10:26
dobeeyes, dobe is reserved10:26
srichterI see, you could use dober :-)10:27
dobeesounds like a dog ;-)10:27
srichterI am almost done with the complex example; I think you were totally right when you pointed out that sorting is not part of the view but the controller10:27
*** sashav_ has joined #zope3-dev10:27
srichterfactoring out this functionality into a separate utility gave us a very simple implementation and clean separation10:28
srichterdobee: right, that's what I was hinting at ;-)10:28
dobeehave you already checked in some stuff?10:29
srichterI will in a moment10:30
srichterI have to finish writing the sort example10:30
dobeesrichter: stefan martin said that he'll work on the grid stuff today, and we gonna have a phone call to reimplement it the way we discussed it, (sort, viewlets for cells, directives)10:30
dobeesrichter: cool10:31
dobeesrichter: is the attribute injection still there? i mean viewlet interface attribute names from the tal namespace10:32
srichterit even moved to the content provider level10:33
dobeeyou want to keep it?10:33
srichteryes, it is a fundamental feature of the provider TALES namespace10:33
srichterit is very important for scripters10:33
srichteras a experienced develop you would probably use it less10:34
*** jenner has joined #zope3-dev10:35
dobeehm, in the grid stuff i'll try to have some other approach, so that when one overrides the grid template does not have to use the same variable names10:35
srichterI don't understand10:36
srichtercan you give an example?10:36
dobee moment10:36
dobeeok, i can't remember the exact syntax10:39
srichterthat's ok :-)10:39
dobeeyou remember i need the key of the row in the grid viewlets, in order to get unique ids,  then i could use an  interface with a rowkey field with that specific name, and therefore each template that uses this viewlet must have the rowkey variable  in the iteration10:41
*** MacYET has quit IRC10:41
*** sashav has quit IRC10:41
dobeethat seems a little bit odd to me10:41
*** MacYET has joined #zope3-dev10:41
*** sashav_ is now known as sashav10:41
dobeei wann do something like this in the viewlet class self.manager.getKey(self.context)10:42
srichterthis last line is a good API call, I think10:42
hdimasrichter: currently tal:content/tal:repeat should returns a message id (not a string) to be translated10:42
srichterhdima: mmh, interesting; I would think that any string returned from tal:content should be translated if the i18n:translate flag is set10:43
srichterhdima: what are the arguments for and against that behavior?10:44
srichterdobee: I think the easiest would be to have an adapter like this: key = getAdapter((context, manager), IKeyProvider)10:45
* hdima need to think10:45
srichterdobee: this way the adapter can decide how to create the key10:45
srichterdobee: in the simplest case it would simply be the name of the context object10:46
hdimasrichter: Some time ago there is has been a discussion about tal:content behavior on the mail list10:50
hdimaAnd we already have the bug #455
*** kickingvegas has joined #zope3-dev10:51
kickingvegasanybody home/10:51
*** kickingvegas has left #zope3-dev10:52
*** kickingvegas has joined #zope3-dev10:52
srichterhdima: do you think you can address this bug before the 3.2 release?10:53
hdimasrichter: yes10:54
hdimaI'll need to think more about the issue...10:55
*** ChanServ sets mode: +o tav10:57
kickingvegascan I ask a newbie question?11:01
kickingvegastrying to figure out how the Zope3 in 30 minutes example works ...11:02
srichtermmh, I'll try, though I don't know this example well11:02
jennerhi there11:02
kickingvegasor is there a better example to follow?11:03
d2mkickingvegas: there are 2 links on the bottom of that page11:04
kickingvegasI only see one to worldcookery.com11:05
srichterwell, there is also my book, which you can order :-)11:05
jennersrichter: have you had a chance to look at my mail? :)11:05
srichterjenner: no, I am still in the pre-working stage; I just got up, because I am jet lagging11:06
srichterdobee: the completed example is in11:06
jennersrichter: taking a nice cup of IRC to live up? :)11:06
srichternope, I just do that on the side11:07
srichterI am really trying to get some of the remaining sprint tasks done11:07
srichterdobee: do you remember what we wanted to make available in TAL templates connected to a viewlet?11:08
kickingvegasyes, I've looked at both texts; one thing that I have not been able to glean from either of them is how page templates interface with the code; in particular how a zpt page that is added through the ZMI access functions in an instantiated product object;11:08
kickingvegasI'm just not getting how namespaces work here.11:09
srichterkickingvegas: we are not supporting TTW development, including ZPT Pages; they are just there as demos11:09
srichterthat's the reason you (a) don't find that in either book, and (b) will not find any docs on that11:10
d2mkickingvegas: the 'Views' section in the 30minutes doc show how ZPTs are used now11:11
srichterright, and that's the method described in the books as well11:12
kickingvegasokay; then can I write a directory tree of zpt pages as a product?11:12
srichterkickingvegas: yes11:12
srichterkickingvegas: if the templates are not views of obejcts, you register them as resources11:12
kickingvegasand is the directory tree of zpt pages really the preferred way to develop for zope3?11:12
srichteryes, Python-based packages are really the preferred way now11:14
srichterwe are really addressing the Python developer in the Zope 3 right now11:14
srichterwe will try to support scripters as soon as we can11:14
*** stub has quit IRC11:15
kickingvegasbut with the registered directory tree concept, I still have to register each zpt page in the zcml file, right?11:16
dobeesrichter: sorry, 'been away11:17
dobeeyour q: do you remember what we wanted to make available in TAL templates connected to a viewlet?11:17
srichterkickingvegas: yep11:17
srichterkickingvegas: explicit is better than implicit11:17
srichterkickingvegas: also all the logic comes from view classes11:17
dobeesrichter: only 'view' which is actually the viewlet11:17
srichterdobee: I am waiting .... :-)11:17
srichterdobee: ok11:18
d2mkickingvegas: if you want to do a website with zope3 look at (theres a download of the whole site to install and run)11:18
dobeesrichter: the same for viewletmanager template, but as i remember the template of a vm is not api?11:19
srichterthe template is definitely not API11:20
d2mkickingvegas: the whole site needs 5 page templates for display11:20
srichterbtw, I am thinking about registering a "provider" directive in zope.contentprovider; any thoughts?11:20
dobeehm, yes a bit like the view directive, then maybe the template gets api ;-)11:22
dobeewe will use it in the grid stuff11:22
kickingvegasd2m:  okay - will look at the example code11:23
d2mkickingvegas: use (its the website as you see it)11:24
kickingvegasd2m: downloading it as I type -11:24
kickingvegasalso: still not clear how to walk up the name space to call other instantiated objects via zpt11:31
srichterkickingvegas: you are not supposed to11:31
srichterall logic, including walking up a path should be done in the Python view class11:32
kickingvegasokay then to view two or more objects then I need to create a new product that views them?11:34
srichternote that the term "product" has been deprecated; use "package" instead11:35
srichterkickingvegas: yep, you need the package to create the objects anyways11:35
*** projekt01 has joined #zope3-dev11:36
srichterprojekt01: do you think a genric "provider" or "contentprovider" directive would be useful?11:37
kickingvegasokay - thanks for all the input!11:39
*** kickingvegas has quit IRC11:40
jenneris there a way to set the log level to DEBUG?11:41
srichteryeah, but I forgot how ;-)11:42
srichterit should be soemwhere in zope.conf11:42
srichternot sure you will get any more info though11:42
jennerhm, where should transaction be imported from? DeprecationWarning: This will be removed in ZODB 3.6: use transaction.get() instead of get_transaction()11:50
*** J1m has joined #zope3-dev11:51
srichterjenner: grep in the code for it :-)11:51
srichterJ1m: good morning11:51
jennersrichter: I was afraid you'd suggest that :)11:51
srichterjenner: it is used at several places11:51
srichtershoudl be in the transaction package somewhere though11:51
dobeehdmi: just info: this is a workaround for the tal:content not translated issue11:53
dobee   <div i18n:translate="">11:53
dobee    <metal:block tal:replace="python:u'Title'"/>11:53
jennerlooks like it's as simple as "import transaction"  :)11:53
dobee   </div>11:53
J1msrichter, morning11:53
J1msrichter, there is no way principals.zcml should be used when testing11:53
* srichter notes that he is jet lagging :-)11:53
srichterJ1m: mmh11:54
srichter(btw, I am not sure why those tests worked before)11:54
srichterall it tries to do is load site.zcml11:54
srichteras a test; I guess we could resort to ftesting.zcml11:55
J1mI haven't had a chance to look into this yet, but it appears that we are now using site.zcml.11:55
J1mwe should be using ftesting.zcml.11:55
J1mThat's what it's there fore.11:55
J1mThat's what we did before.11:55
J1mI don't understand why we are now loading site.zcml.11:56
srichterI wonder how easy it will be to point at this, since the point of those tests is to simulate a startup11:56
srichterJ1m: note that this is not a problem of the ftests11:56
J1mI don't know what tests you are refering to.11:56
srichterThey are unit tests that try to bring up the server11:56
srichterbasically we are testing that we can generate a WSGI application11:56
J1mThere should be a *very* specific zcml file that only has enough to test bringing up the server,11:57
J1mYou can't predict what will be in site.zcml.11:57
srichterWell, for what I care it could be an empty ZCML file ;-)11:57
J1mThen use an empty zcml file.11:58
srichterit just has to load11:58
J1mBut by no means should you be using site.zcml.11:58
srichterok, I can use ftesting.zcml11:58
J1mYou should use something far more limited.11:59
*** VladDrac has joined #zope3-dev11:59
J1mYou should use something minimal.11:59
J1mNote that someday, we'll get rid of ftesting.zcml.11:59
srichteryeah, but it also has to exist11:59
srichterfine by me11:59
J1mwhat also has to exist?12:00
srichterthe site file I am specifying12:00
srichterlast test12:00
J1mThen stick it in, but make it as small as possible to achiev he goals of the test.12:00
J1m(I don't have time to look at that right now.)12:01
srichterJ1m: ok, fixed12:10
*** stub has joined #zope3-dev12:11
*** MJ has joined #zope3-dev12:12
*** mgedmin has joined #zope3-dev12:14
*** tziade has joined #zope3-dev12:16
srichtertziade: do you think you will have a chance of writing a small proposal today?12:17
tziadehello srichter12:17
tziadeabout the performance thing?12:17
srichtertziade: it would be really nice, if we would have a way to inspect the timing outcome, if desired12:17
tziadeok i'll do that12:17
srichterfor example, in a debugging session you might want access to the timing info12:17
tziadehow do you see this timing info: a sort of statitics table for the whole testing session, like hotshot does, in a file ?12:18
tziadeor some kind of global a in memory info12:19
tziades/a in/in12:19
jennersrichter: photo had no addMenuItem directive in its browser/configure.zcml, that's why it didn't show up in ZMI :)12:19
srichterprojekt01: ping12:19
srichtertziade: that would be cool :-)12:19
J1msrichter, I suggest waiting for a little email discussion re test timing.12:19
srichtertziade: it would be very nice to optionally display the timing info12:19
srichtermaybe we can make this helper function a test global12:20
srichterjenner: yipee, you figured it out; please sign up as a contributor and contribute back :-)12:20
srichterJ1m: I suspect that not many people are interested12:21
srichterprojekt01: ping, ping, ping :-)12:21
J1msrichter, I know we are, but we don't have much time to provide input this week.12:22
VladDracare there any tools that generate static api documentation out of interface definitions?12:24
srichterVladDrac: wget on API doc tool12:24
srichterVladDrac: it would be very cool, if someone could use pdftemplate (now supporting tiny RML) to generate PDF versions12:25
VladDractiny rml?12:25
srichterit is a subset of the comercial RML package from reportlab12:25
VladDracah ok12:25
srichterit allows you to write markup to generate PDF files12:25
srichtermgedmin: I think we should give pdftemplate a try in SchoolTool as well12:26
projekt01srichter, pong12:29
srichterprojekt01: good morning12:29
projekt01good morning, what's up12:29
srichterprojekt01: do you think a generic "contentprovider" or "provider" directive would be useful?12:30
srichterprojekt01: I realize that we could use the adapoter directive fairly easily, but we could make some additional assertions that might be useful to the end user12:30
projekt01yes, I guess normal user can use it.12:30
*** MacYET has quit IRC12:31
srichteralso, I removed the weight attribute in the viewlet directive, but allow any type of keyword arguments now12:31
srichterthat are set as attributes on the viewlet12:31
projekt01most the time adapters enough, but I guess it's hard to get the right registration settings for endusers.12:31
projekt01Ok, that good12:31
srichterthe viewlet manager can then decide what to do with the attributes12:32
srichterand using properties, the viewlet can easily convert them to whatever it needs12:33
srichteronce we have some experience, we might describe the attributes in a schema12:33
projekt01yes, this sounds good to me.12:36
projekt01srichter, perhaps I should add a PDF API doc package in our collective where will support a view like /@@apidoc.pdf as a example how powerful the pdftemplates are.12:36
srichterthat would be really cool12:37
projekt01srichter, is there a simply API to api doc or do I have to write each view as a pdf template?12:37
srichterI don't understand the question, can you rephrase, please?12:37
projekt01I think the API doc right now can't be output it's content as one large view.12:38
projekt01This would be useful for a pdf12:38
srichterah, I see what you mean12:38
srichterwell, the first step would be to get each single interface view as a PDF12:39
srichterlater I want batching, like: Give me a PDF with all interface documentation for package zope.XXX12:39
projekt01Yes, this is not a big deal since we can additional register a pdf view like every other page template view.12:40
projekt01I guess each pdf view would be the same with the ending ".pdf" instead of ".html"12:40
*** Alef has joined #zope3-dev12:40
srichterand you could even check in the views to the Zope 3 core and use a feature to check whether pdftemplate is installed12:42
srichteroh wow, Zope 3 is getting cooler by the minute :-)12:42
*** tav has quit IRC12:42
projekt01Ok, let me implement this next week.12:43
*** d2m has quit IRC12:43
*** MJ has quit IRC12:43
srichterprojekt01: btw, you really need an HTTP view of your repository :-)12:43
*** tav has joined #zope3-dev12:45
projekt01Ok, I have some time to fix all parts in Tiks next week after I replaced the missing pagelet with viewlets ;-)12:45
projekt01Then we are ready to announce some great libraries like pdftemplate or JSON tree viewlets where everybody can use out of the box for custom projects.12:47
*** tziade has quit IRC12:47
srichternow I just have to convince you to use zpkgtools12:48
projekt01no way12:48
srichterat least for the standalone packages12:48
srichterI just suggested to Marius to use podf template and it will make our packaging a million times easier, if zpkgtools is supported12:49
*** faassen has joined #zope3-dev12:49
srichterok, time to boot up12:49
J1mwhat are podf templates?12:50
projekt01since zpkgtools is not a package builder and we can't build a installer where is able to install zope3 to D:/myfolder, there is now way to use it for windows!12:50
J1mprojekt01, what do you mean zpkgtools is not a package builder?12:51
*** ajung has joined #zope3-dev12:51
J1mThat's exactly what it is.12:51
projekt01Jim, pdftemplate is a Reportlab integration for write views via TAL where generate PDF files.12:51
J1mso podf was a typo of pdf.12:52
projekt01J1m, we can't export a package structure where we can use for different install builder.12:52
J1mI don't know what that means.12:52
J1mwhat specifically would you like to do that you can't do with zpkgtools?12:53
projekt01zpkg doesn't export a package structure where is useable for other install builder like WISE or BitRock.12:53
J1mprojekt01, zpkg generates a distutils package.12:54
projekt01zpkg can't install a zope3 instance to D:\myfolder on windows12:54
J1mdistutils can generate packages in various binary formats.12:54
*** d2m has joined #zope3-dev12:55
J1mright, distutils only installs "libraries".12:55
J1mdistutils needs to be taught to install applications.12:55
J1mThis is a distutils issue, not a zpkg issue.12:56
*** tarek has joined #zope3-dev12:56
projekt01Ok, this means there is no way to build a "*.exe" where is able to install Zope3 to D:\myfolder?12:57
jennerhm, If I'd like to have a custom zmi_views, how can I add the default contents tab? like <page ... attribute="Contents"... />?12:57
J1mprojekt01, you should ask fred.12:57
J1mThere *may* be a way to do this, but it runs against the graon of distutils.12:57
J1mWe need to change the way disutils works.12:58
J1mWe need to change the way distutils works.12:58
SteveA"runs against the groan of distutils" ;-)12:59
J1mprojekt01, I'll note that you are the first person who has complained about this, so we have considered this a low priority.13:00
srichterJ1m: we had a very long, loud and intense discussion during the sprint about this :-) so I guess it is a very loud complaint :-)13:01
J1mI didn't hear it. You should have been louder. :)13:02
projekt01I really think there should be a way to use BitRock for build cross platform installers. I really think this is the next generation cross platform install builder.13:04
projekt01Additional to install packages via a package manager this is the best way to install applications on windows, linux, Solaris, Aix etc.13:07
* mgedmin has a question about the new immutable MessageIDs13:12
mgedminIf I had code like this previously:13:12
mgedmin        msg = _('${instructors} -- ${courses}')13:12
mgedmin        msg.mapping = {'instructors': instructors, 'courses': courses}13:12
mgedminhow am I supposed to convert it to the new immutable message IDs?13:12
* projekt01 immutable sounds bad in this case13:13
* mgedmin looks at the wiki13:14
* mgedmin reads
* mgedmin wishes for shorter wiki URLs13:14
d2mmgedmin: can be replaced by
J1mmgedmin, you convert it to a new message id with a different mapping.13:14
mgedmind2m, yes, but the server redirects it to the full name, and I cannot copy&paste the short url from the browser's location bar13:15
* mgedmin discovers the new syntax13:15
d2mmgedmin: true13:15
mgedmin        msg = _('${instructors} -- ${courses}') % {'instructors': instructors, 'courses': courses}13:15
mgedminbut that's what the wiki page says13:15
mgedminwas it changed during implementation?13:16
J1mok, maybe I'm wrong13:16
* mgedmin will run unit tests and see if the % syntax works13:17
*** mkerrin has joined #zope3-dev13:20
srichtermgedmin: this syntax is at least not documented13:20
srichtermgedmin: if it works, could you add a test in messages.txt?13:21
mgedminfun fun fun13:21
ajung    sun sun sun13:21
mgedminwhere is the documentation?  zope/i18nmessageid/messages.txt?13:21
* srichter notes that the syntax mgedmin is looking for looks nice13:22
mgedminbut it does not seem to work13:22
mgedmincan I just do13:23
J1mBut what does not seem to work?13:23
mgedmin        msg = _('${instructors} -- ${courses}', mapping={'instructors': instructors, 'courses': courses})13:23
mgedminJ1m, the msgid % mapping syntax13:23
mgedminzope.i18n.translate(msgid % mapping) gives me the msgid back13:23
mgedminwith no substitutions13:23
J1mRight, I wasn't aware of that. :)13:23
mgedmininterleaved irc conversations are fun13:24
* mgedmin goes to read the real documentation13:24
srichterspecifying the mapping only in the constructor does not always seem useful; I agree with Marius that the above syntax should work13:24
srichteroh nevermind, scratch the above comment13:25
projekt01ajung, btw, jim washington is very happy with the new integration for integrating the JSON server13:26
projekt01ajung, Jim wrote in mail to me: Tell Andreas and Tarek that they are brilliant!13:26
srichterprojekt01: good to hear13:27
VladDracI've seen the commitmessages13:27
* VladDrac needs that definately for his amfserver (very similar to xmlrpc/json)13:27
VladDraccool work13:27
mgedminok, the _('...', mapping={...}) syntax works13:27
mgedminI assume the i18n message extractor can cope with it?13:28
srichtermgedmin: oh, good point; we have to try that13:28
faassenmgedmin: we actually use that syntax in Zope 2 with Silva. :)13:29
srichtermgedmin: at least send a mail to the list as a reminder to check it out, if you do not have time to try it13:29
mgedminI can try it now13:30
mgedminif I figure out how to13:30
ajung sr: can we merge the publisher stuff into the trunk?13:31
srichterajung: already done13:33
ajungups, cool13:33
* ajung goes to the daily project-bashing-meeting13:33
VladDracthe publisher stuff will be part of 3.2?13:33
srichteryep, it's in the trunk, so it will be in 3.213:34
srichterI am planning on a technology preview this weekend13:35
srichterI mainly want to give philiKON and mkerrin a chance to fix some outstanding issues13:35
srichtermkerrin: how is the FTP/SFTP stuff coming?13:35
mgedminI wanted to add a comment on
mgedminsaying that the proposal was implemented in the trunk, but without the msgid % mapping syntax13:37
mgedminhowever I cannot modify that page when I've logged in13:38
d2mmgedmin: you need another permission to edit, i change it13:39
d2mmgedmin: please try loin and edit again13:39
*** srichter has quit IRC13:41
mgedmind2m, I got a comment box, entered my comment painstakingly, submitted the form and got a13:44
mgedminInsufficient Privileges13:44
mgedminYou do not have sufficient privileges to view this page.13:44
J1mDamn wiki spam. Damn spam.13:45
mgedminwoohoo, svn ci doc/CHANGES.txt gives me a "Connection closed unexpectedly"13:45
J1mmgedmin, what is your id?13:45
mgedminsomebody up there doesn't want me to contribute anything to Zope today? ;-)13:45
mgedminmgedmin, I guess13:45
d2mmgedmin: you got the right privileges (i think at least that has not changed) -- maybe you need to clear a cookie ( logout-login)13:45
mgedminfun fun fun13:45
J1mGood old BDB ...13:46
d2mJ1m: 'powermember' is it for the wiki - no ?13:46
* J1m goes to unf&^%
mgedmineven more fun13:46
d2mok, must just be caching then13:46
mgedminI just successfully submitted my comment13:46
mgedminand got back the same page13:46
mgedminwithout my comment13:46
mgedminand the browser's Back button is disabled13:47
* mgedmin is not happy to have to retype the comment again13:47
mgedminnever mind, the comment is there, only mutilated13:47
mgedminmy extract of an interactive '>>> foo' session disappeared somewhere13:47
* mgedmin adds a :: in the preceding paragraph13:48
mgedminit's there in the sources, but invisible on the rendered page13:48
*** philiKON has joined #zope3-dev13:48
J1msvn is unscrewed again, for the moment.13:49
* mgedmin gives up, as he has to be in a lecture in 10 minutes13:49
mgedminsvn commits work again, thanks J1m13:50
mgedminif somebody has time to take a look at the markup of
mgedminand figure out why my comment is not rendered in the output13:50
mgedminthat would be very nice13:50
* mgedmin has to run now13:50
*** mgedmin has quit IRC13:50
*** J1m has quit IRC13:51
*** kaczordek has joined #zope3-dev13:51
philiKONhi y'all13:51
zagyphiliKON: hi13:56
*** tziade has joined #zope3-dev13:59
jennerdo I need anything special to register a utility except providing a <utility/> config directive in configure.zcml?14:00
philiKONthat does it14:05
philiKONi think it's quite enough :)14:05
*** mgedmin has joined #zope3-dev14:06
*** mgedmin is now known as mgedmin|uni14:06
philiKONhi mgedmin|uni14:06
mgedmin|unihi, philiKON14:06
philiKONmgedmin|uni, i'll take care of the wiki page14:07
mgedmin|uniwhat sort of markup is used on the wiki?  structuredtext?14:07
philiKONi need to sync it with the doctest anyways14:07
mgedmin|uniso that's why it doesn't work like restructured text! :-)14:07
jennerhm... ok14:07
jennerzope.component.getService is deprecated, right?14:08
philiKONservices are deprecated14:08
philiKONas a whole14:08
jennerstrange... a subclass of TextLine defines a property called "allowed_values", but I can't see any references to it, neither in the package code nor in zope.schemas or somewhere else...14:18
*** seb_ has joined #zope3-dev14:20
*** seb_ is now known as yotaff14:20
yotaffhi :)14:20
philiKONjenner, which subclass?14:25
jennerphiliKON: ResizeUtilityName (I'm trying to revive the photo package)14:26
VladDracjenner: which photopackage?14:28
VladDracthe one in svn?14:28
VladDracah ok14:28
* VladDrac already once revived it, I guess it could use some additional work yes :)14:29
VladDrachas been on my todolist for quite a while14:29
* VladDrac 's just loaded with work14:29
jennerso... any hints? :)14:33
jennerafaik the purpose of subclassing the TextLine is to return all available resizing utilities (which were added using <utility .../>)14:34
philiKONyou would really want this to be a tuility vocabulary14:39
philiKONno TextLine but a Choice14:39
philiKON  <vocabulary14:39
philiKON      name="Permissions"14:39
philiKON      factory=""14:39
philiKON      interface="" />14:39
philiKONand then:14:39
philiKON  permission = Choice(..., vocabulary="Permissions")14:40
philiKONnow you just have to adapt that to resize utilities14:40
jennercool, thanks14:41
jennerworks like a charm :)14:45
jennerexcept that now I get an exception if I try to create a photo:14:47
jennerUnpickleableError: Cannot pickle <class ''> objects14:47
*** andres has quit IRC14:53
*** zbir has quit IRC14:55
*** kaczordek has quit IRC14:56
*** ignas has joined #zope3-dev15:08
*** andres has joined #zope3-dev15:12
andresFunny. I got zope to using 780mb ram... Just by uploading a file...15:12
andresok, the file was 100mb big...15:13
*** SteveA_ has joined #zope3-dev15:15
jennerweird... apparently it's the vocabulary -- if I create the Choice field using values=[<utility names>] then all works just fine, if I skip values and provide a vocabulary=<name of vocabulary in configure.zcml> then I get this UnpickleableError15:21
jennerah, <vocabulary ... nameOnly="True" /> fixed that15:26
jennerwhatever nameOnly may be...15:27
*** MrTopf has joined #zope3-dev15:30
*** ajung has left #zope3-dev15:39
jennerVladDrac: I've got a working version (without PhotoSlideFolder since it seems to be kind of useless now) --
*** SteveA_ has quit IRC15:47
*** einheit has joined #zope3-dev15:51
*** einheit is now known as SteveA_15:51
*** zbir has joined #zope3-dev15:55
philiKONjenner, that means that only the name of the utility will be returned, not the utility itself (which is probably not picklable)15:57
jennerphiliKON: just found it in the source, but thanks anyway :)15:58
*** horizon5 has joined #zope3-dev16:12
*** efge has joined #zope3-dev16:19
*** niemeyer has joined #zope3-dev16:24
*** alga has joined #zope3-dev16:27
* runyaga reads faassen's scaling zope down and almost goes nuts16:27
jennerrunyaga: why?16:29
runyagabecause I have been whining about it for years16:30
*** srichter has joined #zope3-dev16:33
*** ChanServ sets mode: +o srichter16:33
*** bradb has joined #zope3-dev16:38
efgerunyaga: you mean about the fact that scaling down is a must, or that scaling down is not important for real life apps ?16:39
runyagascaling down is not important for you and I16:39
runyagascaling down is a must16:39
runyagabanging rocks (scaling zope down) is not interesting (for you and I) but necessary16:40
SteveA_runyaga: is faassen saying what i've been saying?16:47
SteveA_or something different?16:47
runyagawhat are you saying?16:47
runyagahe isnt talking about removing zodb16:47
SteveA_i'm saying make the zmi and zodb separate things from the zope3 'core' stuff16:47
runyagaeven more simple ;-)16:47
runyagajust zope down scale down to the simple human16:48
runyagamaybe ubuntu can make a z3 distro, Ubuntu Zope - Zope for Humans16:48
SteveA_aha, i see16:48
SteveA_i agree with that too16:48
zbirrunyaga: have you looked at Jim's bobo side project?16:49
runyagai dont understand why we cant inline python into zpt (that is tangentale to the blog entry but relevent)16:49
runyagazbir, not really - i'm familiar with it16:49
runyagabut people want Zope16:49
runyagathey want to scale into Zope 316:49
runyagathey want to grow into Zope 316:49
runyagapeople's comfort level goes up when they use a system that makes them feel like they are standing on the shoulders of giants and they trust the giants wont stumble16:50
efgeinlining python is a good thing to get basic developers to join, but it's a recipe for disaster in the medium term16:50
runyagathe good outweighs the bad16:51
zbirfor the rock bangers, perhaps. The world of web development isn't a zero-sum game16:51
efgeI don't want to basic developers to people who can't think further than "doh I want the same as PHP" rock bangers16:51
efge... want to cater to ...16:51
dobeeseems that i've found a bug16:52
runyagaif they work in a framework they can grow.. they start that way.. then they learn how to refactor16:52
dobeedef ChoiceDisplayWidget(field, request):16:52
dobee    return zapi.getMultiAdapter((field, field.vocabulary, request),16:52
dobee                                IDisplayWidget)16:52
dobeebut the field.vocabulary attribute is always None if the field has a vocabularyName attribute16:53
dobeeso if you do zapi.getMultiAdapter((field, self.request),IDisplayWidget)16:53
dobeeyou get a ComponentLookupError16:54
dobeesrichter: ayt?16:54
dobeehave you read my last message16:55
srichterI don't know much about the those display widgets16:56
*** faassen has quit IRC16:56
srichtermmh, if field vocabulkary is None, the field seems not to be bound16:56
dobeei thinnk it is a bug, because the field.vocabulary attribute gets not set, but it is specified in the interface of it16:56
srichteris the edit case working?16:57
srichterif so, have a look there16:57
projekt01runyaga, efge, migration will become a real pain if you use inline code in custom templates! We really have to teach people to separat code from the presentation layer !!!16:57
dobeethe proplem is that if you define a field with a vocabulary name the vocabulary init is deffered,16:58
dobeethe vocabulary is only used in the _validate method, but then not stored in self.vocabulary16:58
srichtermmh, that does not seem right16:59
srichterwhen the field is bound, the vocabulary attribute should be set16:59
dobeemy speech16:59
dobeeshall i fix it17:00
*** efge_ has joined #zope3-dev17:00
*** zbir_ has joined #zope3-dev17:01
srichterdobee: yep, of course :-)17:03
srichterphiliKON: let me know once you tracked down the deprecation bug17:03
*** J1m has joined #zope3-dev17:05
philiKONsrichter, yes17:06
philiKONsrichter, didn't find time to work on it tonight :(17:06
srichterI will give you and mkerrin some time to fix that sort of thing and then I am going to attempt to create an alpha or technical preview17:07
srichterthis will also give a me a chance to see how we can package twsited17:07
philiKONoh, already17:07
srichteryeah, that was the plan17:07
srichteras soon as we have the twisted interation, we always wanted to release something so people could tell us early on if things are broken17:08
runyagaprojekt01, people dont want to learn17:08
philiKONsrichter, note that this only occurrs when running functional tests17:08
philiKONand it's just a deprecationwarning17:09
philiKONthat occurs because apidoc imports the oldstyle messageid class17:09
srichterI know; I still would like to get rid of it, if possible17:09
projekt01runyaga, zope3 isn't a scripter tool, it's a real application server. But it doesn't mean there can't be a scripter level. Perhaps only the core developer will add it because they don't need it anymore.17:10
* runyaga smiles17:10
runyagathats right.. zope 3 isnt a scripters tool17:10
*** faassen has joined #zope3-dev17:11
projekt01I personaly would never propose to write code in the presentation layer since we have such great posibilities.17:11
philiKONsrichter, of course17:11
*** efge has quit IRC17:12
philiKONi'll try asap17:12
J1msrichter, are you available for a phone call?17:13
projekt01runyaga, perhaps python scripts are fine, but such code has nothing to do in templates anymore.17:13
srichterphiliKON: if it's done by the weekend, I will be very happy17:13
*** zbir has quit IRC17:13
srichterJ1m: yes17:13
faassenconcerning scaling down and scaling up.17:14
faassenZope 3 needs to scale down to simpler apps.17:14
faassenbut at least we get to scale down, instead of having to scale up PHP.17:14
faassenwe're in a much more comfortable position with that. let's not forget that. :)17:14
srichterI totally agree about scaling down and I have plenty of ideas, just no time :-(17:15
runyagaso is j2ee *wink* but it comes out of the box with scaling down17:16
srichterI think the whole TTW story has still a huge, huge potential17:16
runyagai agree.. paul said at goldegg sprint lots of people said w/o TTW they wouldnt use zope17:16
srichterit would actually not take much to write a scripter version of Zope 317:17
faassenwell, lots of people said at least they wanted to reload code without restarts.17:17
philiKONbut there are also a lot of people that say without a 10 minute tutorial, they won't GET INTO zope 317:17
faassenI was surprised to hear so many poeple say that..can it reload code without restarts?17:17
faassenwell, we need a brief tutorial.17:17
faassenwe need a better website.17:17
srichterthe problem is that so many people use MACs17:17
runyagafaassen, benji's brief tutorial is good starting point17:17
faassenI think that will be a good step on the way. we'll iteratively scale down Zope 3, and it's not like steps haven't been taken over the years to do that either.17:17
srichterI just realized last week how slow they are17:17
philiKONsrichter, why is that a problem?17:18
srichterif I take Zope 3 without API doc, I can start it up in less than 3 secs17:18
runyagafilesystem access?17:18
srichterbecause a lot of people had start up time at around 10 secs17:18
srichterand ST comes up in less than 217:18
projekt01...we need some time...17:18
srichteralso, I tend to develop much more with unit doctests, so refreshing is really a non-issue17:20
runyagaunit tests are 6 months down the road from someone first getting their toes wet17:20
runyagaatleast the scripter17:21
srichterthat's really too bad for them17:21
J1m(I wonder is srichter sees what I type when I dialog him.17:21
runyagaJ1m, you need you register with nickserv17:21
srichterfor the scripter I would only support TTW code17:21
srichterin which case you do not need the restart trick17:21
srichterJ1m: no, I don't17:21
srichterJ1m: are you registered?17:21
runyagaJ1m, freenode has been nice enough to block all /msg that come from people whom are not registered w/ nickserv17:22
srichterprivate messaging only works with registered users due to spam, I read17:22
J1mhm, I thought I was ...17:22
*** J1m has quit IRC17:22
*** J1m has joined #zope3-dev17:23
tziadesrichter, yup17:24
srichterwe got it worked out :-)17:24
tziadethis sux, freenode should warn you17:25
andrestziade, they do17:26
philiKONthey do17:26
philiKONat least when you /msg17:26
andresbut only as a normal server message17:26
philiKONno idea what dialog is17:26
andresnot as a private message.17:26
philiKONJ1m, this is probably why the other day your messages didn't get through to me17:26
tziadehm, when i open a private dialog it wont17:26
philiKONand you thought i was ignoring you :)17:26
andrestziade, look where all the messages from the server are printed17:27
tziadeoh yeah, this freenode window i never look at ;)17:28
andrestziade, exactly thats the problem.17:29
andresbtw, i have the following problem: I have List() schema, but when i try to get a widget for it, i get component lookup errors17:31
andresThis is exactly like its done in the tests. What is blocking my mind?17:34
*** ignas has quit IRC17:34
*** jinty has joined #zope3-dev17:35
* projekt01 is looking for a method to read form fields from the request and extract the name/value pairs17:39
mgedmin|uniprojekt01, getWidgetsData?17:39
*** tarek has quit IRC17:40
projekt01I have to read the field/value from a simply HTML form where was inserted in a wysiwyg editor17:41
*** hdima has quit IRC17:41
*** tziade is now known as tarek17:41
projekt01I have to use such "forms" since we dont' have persistent schema support ;-)17:42
mgedmin|uniprojekt01, FormEncode?17:43
* mgedmin|uni has no idea17:43
andresprojekt01, arent they in the request obj?17:44
projekt01Yes they are, I'm looking for a iterator where extracts the form fields and put it together to a string where I can use with pythons email.message_from_string17:45
projekt01I think I have to write it by myself17:45
projekt01something like: requestToMessage(request): return message, where supports file upload and will render a multi part message17:48
* projekt01 is going to look at nuxeo's mail stuff17:49
*** bradb has quit IRC17:50
tarekprojekt01, the webmail ?17:51
*** bradb has joined #zope3-dev17:52
tarekyou can have a look at the getRawMessage that handles the adding of files in a mail17:52
projekt01tarek, cool, can you point me to the svn url?17:53
*** mgedmin|uni has quit IRC17:53
tarekyup, i am looking at it now17:53
tarektake a look at17:54
tarekit create a "file part"17:54
tarekgetRawMessage renders a multipart message17:55
tarekby mixing a regularl message to attached files17:55
tarekthere's several ways to attach files, if you want more info we can continue on #cps17:56
tarekas this is out of z3 subject17:56
*** SteveA_ has quit IRC17:56
projekt01tarek, Ok, I see, thanks a lot, I deep into the code first17:57
dobeesrichter: do you plan to merge the contentprovider branch into trunk?18:06
*** stub has quit IRC18:07
*** bradb_ has joined #zope3-dev18:09
VladDraczope2.4 payback time18:10
* projekt01 can't also wait for a stable contentprovider version ;-)18:12
VladDracoh well18:12
VladDracit wasn't too hard18:13
VladDrac(debian sarge doesn't have psycopg and it can be a bitch to compile sometimes)18:13
VladDracah, unfortunately sqlobject 0.6.x has disappeared18:15
runyagai think its packaged in sip18:16
runyagacrap its not..18:17
VladDracyeah I found a local copy18:18
VladDracbut now I'm afraind someone committed some faulty code to the formlib repository18:18
VladDraclet's see18:18
runyagawould you send it to me? or email ian so they can hold older versions18:18
VladDracoh wait it's my own code :) (strange the error doesn't occur in dev)18:19
VladDracrunyaga: you mean ian doesn't have old releases himself?18:19
VladDracI have an original 0.6.1 tar.gz here18:19
VladDracrunyaga: does SIP use sqlos?18:20
runyagasip is where sqlos came from18:20
*** bradb__ has joined #zope3-dev18:20
*** bradb has quit IRC18:20
VladDracrunyaga thought so18:21
runyagayour using sqlos w/ formlib?18:22
runyagathats a great idea18:22
VladDracrunyaga I wanted to18:22
runyagai was going to do it this weekend18:22
VladDracbut I didn't manage to get it to work properly with 3.118:22
runyagato help out the rock banging cause18:22
VladDrac(actually it seemed it didn't even work with 3.0)18:22
VladDracso I implemented my own, similar but less transparent system18:22
VladDracmy system is actually pretty cool, I think combined with sqlos would be awesome18:23
VladDracsomething sqlos doesn't have (I think) is an actual implementation of object relations18:23
VladDraccurrently my system implements foreinkeys through vocabularies providing (objectid, label) items18:24
*** regebro has quit IRC18:28
*** bradb has joined #zope3-dev18:28
*** bradb_ has quit IRC18:34
*** ignas has joined #zope3-dev18:37
*** BjornT has quit IRC18:39
*** bradb__ has quit IRC18:42
VladDracrunyaga: anyway, there has been some recent sqlos work, make sure you svn up :)18:45
VladDracnot sure if stuff works properly with 3.1 now though18:45
*** dobee has quit IRC18:49
*** regebro has joined #zope3-dev18:51
andresVladDrac, it does18:52
VladDracandres ah cool ok18:52
andresat least 5 min ago ;-) At the moment im not that sure *g*18:52
* VladDrac should try it again then18:52
VladDractoo late to integrate it into my app now18:53
VladDracperhaps I can backport some of my relational code18:53
*** niemeyer has quit IRC18:54
*** projekt01 is now known as _projekt0118:56
*** _projekt01 is now known as projekt0118:56
*** projekt01 has left #zope3-dev18:56
*** horizon5 is now known as matt_18:57
*** projekt01 has joined #zope3-dev18:59
projekt01srichter, did you get my message?19:02
*** MrTopf has quit IRC19:02
*** dobee has joined #zope3-dev19:09
*** zbir_ is now known as zbir19:19
*** mgedmin has joined #zope3-dev19:19
*** sashav has quit IRC19:26
srichterprojekt01: yeah, sorry I was away19:41
*** BjornT has joined #zope3-dev19:41
*** runyaga has quit IRC19:48
*** tarek is now known as tarek219:57
*** MJ has joined #zope3-dev20:01
*** yota has joined #zope3-dev20:03
*** J1m has quit IRC20:03
*** MacYET has joined #zope3-dev20:10
*** s1 has joined #zope3-dev20:14
*** sm has quit IRC20:15
*** s1 is now known as sm20:15
*** MacYET has quit IRC20:16
*** ignas has quit IRC20:23
*** loreto has joined #zope3-dev20:24
*** regebro has quit IRC20:30
*** tarek2 is now known as tarek20:34
*** efge has joined #zope3-dev20:42
*** MJ has quit IRC20:47
*** jenner has quit IRC20:48
*** jenner has joined #zope3-dev20:51
*** philiKON has quit IRC20:54
*** efge_ has quit IRC20:55
*** MacYET has joined #zope3-dev21:00
*** tarek has quit IRC21:03
jenneris there a ready-to-go interface for serving files from FS? I mean content types with metadata in ZODB and real content on the filesystem ( like ExternalStorage for AT)21:09
srichternot in Zope 321:11
sm isn't showing the quoted code at the end; no workaround that I can find, short of upgrading the software21:11
d2msm, you could remove/change the comment email header21:13
MacYETups, i though i am on #plone .)21:13
jennersrichter: where would I start digging? zope.server.interfaces?21:14
*** tekNico has left #zope3-dev21:14
jennerI'd like to have some sort of a filestream_iterator like in z2 (ZPublisher.Iterators)21:16
*** SiggyF has joined #zope3-dev21:17
SiggyFI am reading through the zope3 book. I've got some questions. Can I ask them here or should I ask in #zope?21:18
MacYET#zope is for zope 2 lamers21:18
srichterjenner: zope.publisher.http21:18
srichterSiggyF: shoot21:18
SiggyFDo I have to restart the zope server everytime I change something?21:19
MacYETexcept for templates, yes21:21
MacYETor use z3reload21:21
*** faassen has quit IRC21:23
SiggyFOk thanks. I also wondered what a logical place is for my sources? The book suggests zope3/src.21:24
srichteryep, good place, if you use the SVN checkout21:25
srichterclearly INSTANCE/lib/python is the logical polace for distributions21:26
SiggyFAnd if I work with more developers, should I check a whole instance into svn or cvs?21:27
srichterno, just your package21:28
SiggyFOk. Thanks srichter and MacYET21:29
*** J1m has joined #zope3-dev21:29
jennerHm, it seems like subclassing and overwriting _setData and _getData is a good point to start at21:34
MacYETcan't you write your own schema field implementation?21:38
jennerMacYET: I'm not sure, but I think AT schemas and z3 schemas are a bit different, e.g. there's no way to define a storage type...21:41
MacYETright, but you can write your own backend :)21:42
MacYETlook at the FieldProperty implementation21:42
jennerMacYET: I'm trying to store photos on FS using the photo package code, photo basically subclasses BTreeFolder and stores images as objects, so I though I'll just store my own implementation of Image in the BTree (or actually a subclass of my own File implementation)21:45
*** zbir_ has joined #zope3-dev21:45
*** MJ has joined #zope3-dev21:46
smpage fixed, thanks d2m21:50
smnot sure where to document that, except here: "start stx docs in column 0 or risk messed up display of literal blocks"21:51
*** anguenot has joined #zope3-dev21:54
*** Alef has quit IRC21:54
*** zbir has quit IRC21:57
mgedminhow about "do not use stx, or risk messed up display" ;)21:58
andresI have slight problem: In the creation of an object i dont have a __parent__ but i need a reference to another object. Is there some possibility to get the root folder somehow?22:12
andreselse i have to change external software which i would not like.22:13
smmgedmin: that's another way to put it, though rst loves to sprinkle error messages in your doc :)22:14
srichterandres: yes, you can always have a utility return the root folder, since you have a context in the interaction22:15
mgedminrst does that, yes :(22:16
smunfortunately rst seems not an option here22:16
* sm would like a best-of-stx-and-rst22:16
smalso now that you mention it, there should be a way to make rst hide all those errors22:17
andressrichter, do i?22:17
srichteryou can also always get the current site using getSite(), which will provide you with a location22:18
mgedmin(unless you're not in the middle of processing a request, and getSite() returns None, of course)22:20
srichterof course22:23
andressrichter, hm. Im sure you are right, but i dont see where my failure is then. I have a class Projekt, which is sqlos based. In the __init__ of sqlos (or really sqlobject) from which my class inherits a method of my class gets called, because its the set property of an attribute of mine.22:23
andresAt this point __parent__ is undefined, which is the root of my problem.22:24
andresOr better __parent__ is None.22:24
srichterright, so use getSite or the context of your view22:25
andressorry, for beeing a bit stupid, but i dont see where i get the context...22:26
andresin this case.22:27
srichterI dunno much about sqlos, so I am not sure how to hook it up22:28
jennerhmm, how do I get the id of an object?22:30
*** efge_ has joined #zope3-dev22:32
andrew_mcan someone tell me how to turn off logging in zope (to speed up things for a demo)?22:33
srichterandrew_m: see zope.conf22:34
srichterandrew_m: deactivate not needed packages, like api doc and the like22:34
dobeejenner: which id?22:35
andrew_msrichter: i see.. thanks22:36
srichterjenner: zapi.getName(obj)22:36
jennersrichter: thnx22:37
dobeesrichter: when do you plan to merge roger-contentprovider into trunk22:38
*** tiredbones has quit IRC22:38
srichterdobee: I am not sure22:38
srichterdobee: in the next days I assume22:38
srichterdobee: feel free to develop against the branch though22:38
dobeewell this answered my question ...22:39
dobeewe plan to checkin some of the grid stuff this week22:39
mgedminwhat's grid?22:41
mgedmindoes this have anything to do with customizable container views?22:41
*** zbir has joined #zope3-dev22:41
srichterit is the completion of the flexible contents view from the sprint22:41
mgedminsounds cool22:42
*** Alef has joined #zope3-dev22:47
*** efge has quit IRC22:49
srichtermgedmin: yeah, we really need to start exploring the possibilities of viewlets for ST22:51
projekt01mgedmin,and pdftemplates ;-)22:52
mgedminfixing z3 forms would be nice22:52
*** zbir_ has quit IRC22:52
* mgedmin ought to take a look at zope.formlib again22:53
projekt01mgedmin, you mean the package?22:53
mgedminthere are some use cases where the existing form fields are cumbersome22:54
mgedmincross-field validation comes into mind22:54
projekt01mgedmin, you mean the invariants?22:54
*** MacYET has quit IRC22:55
*** tziade has joined #zope3-dev22:55
projekt01mgedmin, I think there is a existing refactoring from benji_york or others going on? right?22:56
projekt01or is this only a widgets refactoring?22:56
srichterbut zc.formlib is supposed to address some of the other issues, like the one Marius mentioned22:57
projekt01invariants, the schema base validation? right?22:57
projekt01I think we also can support this in the EditView22:58
mgedminprojekt01, I want to do some ad-hoc validation23:03
*** bradb has quit IRC23:03
mgedminlike data = getWidgetValues(schema, request, etc)23:03
mgedminif data['foo'] > data['bar']:23:03
mgedmin   report an error23:03
mgedminand by "report an error" I want to cause the widget of a field to display a custom error message23:04
mgedmine.g. "this field must be set to a number > 59 since you have checket the frobnitz checkbox above"23:04
mgedminwhen the "report an error" part is "show a message above the form" (and not "next to a widget"), I can do that easily23:04
mgedminbut stuffing a msg into a widget was impossible without setting private variables last time I tried it23:05
*** alga has quit IRC23:07
VladDracif I have a dictionary and I want to make it traversable (to be used with tal, i.e. dict/key), what would be the correct way?23:17
*** tziade is now known as tarek23:20
*** mgedmin has quit IRC23:21
*** tarek is now known as tarek_23:26
*** tarek_ is now known as tarek23:27
VladDracoh well, I'll just use dict['key']23:27
VladDracwhat does it mean if an adapter is 'for="*"'?23:31
jhauser_it adaptes to all interfaces23:32
jhauser_means you can call it on all objects23:32
VladDracI figured that, but that's not what seems to be happening23:32
VladDrac    for="*"23:33
VladDrac    factory=""23:33
VladDrac    provides="" />23:33
VladDracdoesn't that mean that any object is adaptable to ITraversable through DefaultTraversable?23:33
*** tarek has quit IRC23:33
VladDrachmm wait it does seem to be working, somewhat23:34
jhauser_at least you can call ITraverable on every object, but it does not gurantee, that it then is traversable23:35
jennerhmpf, my first naive attempt to create a FS storage failed --  Error type:  Error object: ('getSize', <photo.fsfile.FSImage object at 0xb461e9ac>)23:35
VladDracjenner: make it require the appropriate interface with appropriate permission?23:36
VladDraci.e. <require interface="...ISize" permission="zope.View" />?23:36
*** tziade has joined #zope3-dev23:37
VladDracok got my problem solved I think23:38
*** tziade has quit IRC23:40
*** tarek has joined #zope3-dev23:41
jennerVladDrac: doesn't work (not sure I do it right thoght)23:48
VladDracwell you need to fill in the ... of course23:49
jenner<require permission="zope.View" interface=".interfaces.IFSImage" />23:49
VladDracand IFSImage defines getSize() ?23:50
jenneryes, it implements IImage, so it has to23:51
VladDracIImage provides getImageSize() - is that different?23:52
jennerIImage inherits from IFile which provides getSize23:54
jennerin my case I have an IFSFile which implements IFileContent and IFSFile (the latter is a marker interface for IFile)23:56
VladDracwhat if you <require permission="zope.View" interface="" /> ?23:56
jennerfor <content class=".photo.Photo">?23:57
jennernope, raises a ConfigurationConflictError23:58
jennerhm, that's a bit messy, I think I have to check my interfaces again23:59

Generated by 2.15.1 by Marius Gedminas - find it at!