IRC log of #zope3-dev for Friday, 2005-02-25

J1mI still get the failure without running the deprecation tests00:02
srichtermmh, ...00:02
srichterok, I gotta go again; I will be back later00:03
J1mSomebody is doing classProvides on DeprecationProxy00:03
*** ChanServ sets mode: +o hazmat00:11
*** mgedmin has quit IRC00:14
*** sashav is now known as sashav|zZz00:22
*** niemeyer has quit IRC00:38
*** gintas has quit IRC00:43
srichterJ1m: did you figure it out?00:54
J1mno00:57
J1mI don't have time to chase it00:57
srichterI'll try to look at it tonight00:57
srichterI need to get my stuff done first as well00:57
J1mif Amos doesn't fix it soon, I'll probably reverse his commit.00:57
J1mI did send him a note00:57
srichteryou actually just need to comment out the test he added00:58
srichterhis cleansing of the IZAPI interface was good00:58
J1mThere *is* a problem with deprecated modules00:58
srichtermmh, then this might be a general proxy problem?00:59
srichteron a more positive note, I developed a user preferences system today, which uses interfaces for the configuration options01:00
srichterpreferences can then be accessed from anywhere having a location using ++preferences++prefgroup/prefname01:01
srichterthe system is for apidoc only right now, but nothing in their is apidoc-specific01:02
*** benji_york has quit IRC01:12
*** GaryPoster has quit IRC01:14
*** niemeyer has joined #zope3-dev01:16
*** srichter has quit IRC01:41
*** J1m has quit IRC01:59
*** srichter has joined #zope3-dev02:17
*** ChanServ sets mode: +o srichter02:17
*** tarek_ has quit IRC02:24
*** bskahan has quit IRC02:27
*** hazmat has quit IRC02:45
*** GaryPoster has joined #zope3-dev02:52
*** sashav|zZz has quit IRC02:55
*** jdz_ has quit IRC03:04
*** alga has quit IRC03:20
GaryPostersrichter ayt?03:20
srichterGaryPoster: yes03:27
GaryPoster:-) hey.  Wanted to talk about the timezone thing w/you briefly.03:27
GaryPosters'ok?03:27
srichteryep03:28
GaryPosterk.  Jim wants the utf timezone to go in zope.date (and therefore the adapter in zope.app.date I guess).  I wanted you to weigh in on the localizing approach in particular because I didn't think what I proposed would fit in with the i18n parsing stuff that still needs to be written03:29
GaryPostereh--that first sentence was because I wanted you to know that the stuff would not be going directly in i18n packages03:29
srichterI think zope.date would be overkill03:30
srichterI think we should play nice with Stuart and put it in pytz03:30
GaryPosterYeah, that's kinda what I thought.03:30
GaryPosterI haven't interacted with him.  If you think that will work that's great by me03:31
GaryPosterI'll contact him and make sure he's cool with it I guess.03:31
GaryPosterSo the localizing thing in regards to the i18n parsing:03:31
srichterI think he is really open to suggestions03:31
GaryPosterok, cool03:31
srichterafterall he checked in his code under ZPL03:31
GaryPostertrue.03:31
srichteractually, it is already done03:32
GaryPoster\me opens up code...03:32
srichtersee pytz/reference03:32
GaryPosteroh, ok03:32
GaryPosterno03:32
GaryPosterthat doesn't pickle. :-(03:32
srichterit has a very small footprint UTC timezone03:32
GaryPosterright, no I tried that.03:32
GaryPosterProbably trivial to make it pickle03:32
srichterwell, let's make it pickable03:32
GaryPosterYeah.  OK, fine by me.03:33
srichterwe pull off the same trick as for global adapter services03:33
GaryPoster??03:33
srichtersee zope.component.site towards the end of the file03:33
srichterthere you see how we reduce GlobalAdapterRegistry instances03:33
GaryPosterok will do.  next question:03:34
GaryPosterI'm looking at zope.i18n.format03:34
GaryPosterline 151-ish03:34
GaryPoster# TODO: Find timezones using locale information03:34
srichterright, right now we can only detect timezone names/abbreviations that were typed in in English03:35
srichterfor example, MEZ (Mitteleuropaeische zeit) is not understood03:35
srichterso we need a way to check against the names in the locale as well03:36
GaryPosterAh, so this is not a "user has not entered a time zone, so I better use their local time zone".  That's what I thought (was afraid) it was.  OK, then, my stuff doesn't affect that03:36
GaryPosterOK, cool.  Anything else you wanna weigh in on?  Otherwise I'll let you go. :-)03:37
srichterI just want to go as much standard as possible03:37
srichteri.e. play nice with pytz, use zope.i18n.format whereever possible, and get rid of datetimeutils03:38
GaryPostercool, sounds good with me03:38
srichteroh, yeah03:39
GaryPosterOK, I'll get in touch with Stuart.  Thanks!  Oh and BTW enjoyed what I've read of the book so far--yay.  Great resource.  Jim said you were tired of it by the end, but it's *great* to have,03:39
srichterI would like Jim to make an edict that all dates/times are stored with UTC as timezone03:39
srichterlike we did for unicode in human readable strings03:40
GaryPosterYeah, I'm working on that with him.  This is kind of working up to that.  He sees that it is similar.  He doesn't love it but he's willing for it to be the standard.03:41
srichterok, cool03:41
GaryPostertalk to you later.  thanks again03:41
srichteryou are welcome; thanks for working on this03:42
srichterbtw, if you find bugs in zope.i18n.format, let me know03:42
GaryPosterok will do03:43
*** Awayblia has quit IRC04:17
*** niemeyer has quit IRC05:03
*** GaryPoster has quit IRC05:16
*** GaryPoster has joined #zope3-dev05:17
*** GaryPoster has quit IRC05:18
*** `anthony has quit IRC06:36
*** `anthony has joined #zope3-dev07:48
*** z|a is now known as zagy08:37
zagymoin08:37
Damasceneshit08:42
Damascenei did it again08:42
zagyhaa haa09:06
zagywhat's the time at your place?09:06
*** Theuni has joined #zope3-dev09:18
*** hdima has joined #zope3-dev09:24
*** `anthony has quit IRC10:00
*** Aiste|away has quit IRC10:26
*** Theuni is now known as T43un110:50
srichterhey, do you guys know the CSS, if I want to change the colot of an achor, but only if href is specified?11:12
T43un1a[attribute..] something.11:19
T43un1but that doesn't work in IE11:19
*** T43un1 is now known as Theuni11:19
srichtermmh, I might not care about IE :-); it is just a color; it will use default otherwise11:20
*** sashav|zZz has joined #zope3-dev11:28
*** sashav|zZz is now known as sashav11:29
Theunihttp://www.w3.org/TR/CSS21/selector.html#attribute-selectors11:30
*** zagy has left #zope3-dev11:41
srichteroh, this is cool11:42
Theunijupp11:42
* srichter needs to brush up his CSS skills11:43
*** zagy has joined #zope3-dev11:45
*** SteveA has joined #zope3-dev11:50
*** vlado has joined #zope3-dev11:51
*** Aiste|away has joined #zope3-dev12:05
*** Aiste|away is now known as Aiste12:34
*** hazmat has joined #zope3-dev12:41
*** tziade_ has joined #zope3-dev12:43
*** tziade_ is now known as tarek_12:43
*** admp has joined #zope3-dev12:56
*** philiKON has joined #zope3-dev12:56
*** GaryPoster has joined #zope3-dev13:03
GaryPostersrichter: I'm getting zope3/src/zope/app/apidoc/ifacemodule/configure.zcml", line 70.213:08
GaryPoster    ConfigurationError: ('Unknown directive', u'http://namespaces.zope.org/apidoc', u'preferencesGroup')13:09
GaryPosterBut that may be an oddity of my setup13:09
srichterGaryPoster: update your package-includes13:09
srichterGaryPoster: yes it is :-)13:09
GaryPosterI'll look into it later this morning but wanted to let you know--ah ok will do13:09
GaryPosterThanks :-)13:09
srichterI fixed the include files, so you guys have an easier time13:10
srichter(as requested by Fred)13:10
srichterGaryPoster: btw, I might have thought of a way to increase startup speed13:11
GaryPosterReally?  That would be very cool.  what?13:11
srichterkeep track of the actions a particular ZCML file cerates and store a pickle13:11
GaryPosterah--sort of Twisted-ish, sounds like13:12
srichterif a pickel for a ZCML file exists, use it13:12
srichterI dunno13:12
srichterthe only problem will be that we need to ensure that everything is pickable13:12
GaryPosterYeah, I think that's what their runner does13:12
srichterwhich is non-trivial and I think it is the reason noone has done it13:12
SteveAmarius and i talked about this at length13:13
SteveAthere are some serious problems with doing this, to do with what you actually get in a zcml action13:13
SteveAlike, generated classes / functions13:13
srichterSteveA: can you explain?13:13
GaryPosterright.  Also, some folks don't like that behavior in Twisted; I don't know why and don't have enough experience about this, but Shane, for instance, is not a big fan.  Might be worth seeing if anyone knows of abd drawbacks13:14
srichterright, but this can be fixed13:14
SteveAa lot of the problem is in just reading in all the python modules13:14
SteveAhave you profiled zope3 starting up?13:14
srichterwell, as long as you recognize when a file changes, you are good13:14
srichterit's like Python's PYC file stuff13:14
SteveAi think a big win could be in lazily loading modules13:14
SteveAi think twisted does something like this, holding an interface or class as a string until it is actually needed13:15
srichterSteveA: I remember Marius saying that the validation/conversion was the problem13:15
srichterSteveA: right, but I don't want to go down that road, since we want to use the plain schema facilities13:16
GaryPosterI've got to get off for now; I'll be back on later today.  bye13:16
srichterif I can make all of our tricky Python objects pickable, the problem should be solved13:16
*** GaryPoster has quit IRC13:16
srichterSteveA: We have to load all the objects anyways, since this sets up the CA13:17
SteveAsrichter: maybe extend actions to have a picklable there too, that represents its state13:17
SteveAsrichter: that's the point -- modify the CA to allow laziness13:17
srichterSteveA: ok, I think that's out of scope; at least for me13:18
srichterSteveA: what did you mean by your previous comment "extend actions to have a pickable"13:18
SteveAphone, brb13:18
srichtermmh, I could check how long it takes to read all classes in zope by running my class registry crawler over the dir13:20
SteveAi mean, when you make an Action, sometimes the Action is picklable, and sometimes it is not13:21
srichteryes, right13:21
SteveAmaybe explicitly add a "here is a picklable that represents me" attribute to an action, that is optional13:21
SteveAif you don't use it, then that zcml action isn't picklable13:21
srichterah, interesting13:22
SteveAwhen the Action is not itself picklable.  but if you do use it, then it is.13:22
srichterand if any action in a file is not pickable, we do not create a compiled ZCML file13:23
srichterthis way we could migrate slowly13:23
srichteron the other hand I could just check whether an action is pickable and if not, ignore it13:24
srichterI think I understand the entire pickling process well enough now that I could make pretty much all the tricky objects pickable13:24
*** sashav has quit IRC13:25
*** sashav has joined #zope3-dev13:26
*** hazmat has quit IRC13:26
SteveAsrichter: you'll be reading binary pickles soon ;-)13:26
srichter:-)13:26
srichterbut I think it would speed up things quiet a bit, since we could save converting values13:27
srichterSteveA: what about just storing the handler and its arguments?13:28
srichterall the conversion is already done at this time...13:28
srichterSteveA: I hope I will have a chance to chat with Marius about this today13:34
*** philiKON has quit IRC13:42
*** philiKON has joined #zope3-dev13:44
*** mgedmin has joined #zope3-dev14:04
mgedminsrichter, a while ago I tried the same approach you suggested (pickling zcml actions)14:27
mgedminI have a patch lying around somewhere if you're interested14:28
mgedminbut it will probably be easier to start from the beginning14:28
srichtermgedmin: yeah, just as a starting point14:28
mgedminWhen I got rid of all lambdas and dynamically generated classes, I got the whole config to pickle14:28
mgedminbut it failed to unpickle afterwards ;)14:28
srichterawesome14:29
srichterok, I gotta go14:29
mgedminthe trick with __reduce__ and global services poses a chicken and egg problem14:29
srichterI will chat with you later some more14:29
mgedminok, see you14:29
*** srichter has quit IRC14:29
*** J1m has joined #zope3-dev14:33
*** BjornT has joined #zope3-dev14:38
*** mgedmin has quit IRC14:51
*** niemeyer has joined #zope3-dev14:55
*** srichter has joined #zope3-dev15:13
*** Aiste has quit IRC15:13
*** ChanServ sets mode: +o srichter15:13
*** `anthony has joined #zope3-dev15:19
*** philiKON has quit IRC15:23
*** Aiste has joined #zope3-dev15:32
*** GaryPoster has joined #zope3-dev16:03
*** philiKON has joined #zope3-dev16:08
*** mexiKON has joined #zope3-dev16:12
*** zagy_ has joined #zope3-dev16:14
*** philiKON has quit IRC16:20
*** zagy has quit IRC16:21
*** Aiste has quit IRC16:26
*** mexiKON has quit IRC16:30
*** Theuni has quit IRC16:31
*** mkerrin has joined #zope3-dev16:32
*** gintas has joined #zope3-dev16:33
*** mgedmin has joined #zope3-dev16:45
*** Aiste has joined #zope3-dev17:01
*** hdima has quit IRC17:03
*** jdz_ has joined #zope3-dev17:34
*** admp has quit IRC17:40
*** bskahan has joined #zope3-dev17:54
*** philiKON has joined #zope3-dev18:06
*** J1m is now known as J1m|away18:07
*** tarek_ has quit IRC18:07
*** admp has joined #zope3-dev18:10
*** philiKON has quit IRC18:11
*** vlado has quit IRC18:15
*** zagy_ has quit IRC18:17
*** alga has joined #zope3-dev18:21
*** Theuni has joined #zope3-dev18:35
*** mkerrin has quit IRC18:53
*** zagy has joined #zope3-dev18:54
srichtermgedmin: wanna chat about optimizing ZCML a bit more?18:59
mgedmina bit, yes19:00
mgedminI am very interested in short z3 startup time19:00
srichterI know19:01
srichter:-)19:01
mgedminhowever, when I did that experiment, I determined zcml parsing eats about 1/3 of the startup time19:01
mgedminso the gain is somewhat limited19:01
srichtermmh, what is the rest?19:01
srichterI know that networking has to do with it too19:01
srichterwhen I am at home on wireless it takes 4 secs to start up and here at school where I am on a cable it takes 10 secs19:01
Damascenehmm that shouldn't be the case though.19:02
Damasceneunless the os does weird binding19:02
SteveAi got it down to 2 secs when i removed most of the package includes and parts of zope.app19:02
srichterit has probably something to do with hostname lookup19:02
SteveAthe main issue is with the amount of python modules that need to be loaded19:02
mgedminsrichter, profile19:03
srichterSteveA: yeah, I know; I added about a second when I registered all the book chapters for apidoc; I need to check that it does lazy loading; this would speed up things again19:03
srichterwell, I guess about 2 secs are loading modules19:03
mgedminsrichter, http://mg.pov.lt/blog/profiling.html19:04
srichterthat's about what it takes to initialize the class registry19:04
SteveAfor all of zope?  i expect a lot more19:04
SteveAi think it may be valuable to have a zcml include that says "this is an additive only include"19:04
SteveAthat could be a route to optimisation19:04
srichteror frustration ;-)19:05
SteveAwe all know that optimisation is the route of all evil19:06
srichter:-)19:06
srichterwell, I guess a 30% speedup due to faster ZCML loading would be a good thing19:07
* mgedmin nods19:08
srichteryou know, another possible benefit from storing pickles is that we might be able to implement refresh19:09
srichtermgedmin: so did you try to store it all in one pickle?19:13
srichteror did you do it on a per file basis?19:13
mgedminsrichter, one pickle19:13
mgedminI was doing a feasibility study19:13
srichterok19:14
mgedminI piled a lot of ugly hacks on top of each other etc19:14
srichterok19:14
mgedminI gave up at this problem:19:14
mgedminconfiguration actions that define e.g. adapters19:14
mgedminreferred to the global adapter service19:15
mgedminpickling global services uses this __reduce__ hack19:15
mgedminto store only the service name in the pickle19:15
mgedminunfortunatelly, the same pickling was used for the action that actually created and registered the global service19:15
mgedminso you cannot load the pickle, because the action that registers the global service tries to access the service by name19:16
srichterI see; we might not have this problem anymore19:16
mgedmingreat19:17
srichterwe are still doing the same trick, but there is no adapter registry creating going on anymore19:17
*** zagy has quit IRC19:28
*** MalcolmC has quit IRC19:28
*** Theuni has quit IRC19:30
*** tvon|desk has quit IRC19:30
*** philiKON has joined #zope3-dev19:51
*** tvon has quit IRC20:10
*** gintas has quit IRC20:27
*** J1m|away has quit IRC20:41
*** Aiste has quit IRC20:45
*** WebMaven_ has joined #zope3-dev20:51
srichteranyone here with docbook expereince?20:56
srichterwould it be possible (without much pain) to use docbook for API documentation?20:57
WebMaven_srichter: no *real* experince, but I played with it some a few years ago.20:57
srichterWebMaven_: is it easy to use? Or will I end up in frustration trying to define my own elements?20:58
WebMaven_can you qualify 'use' and 'pain' in this context?20:58
srichterwell, I am not ready to learn all the details about docbook; I just want to create the docbook files using ZPT and then compile it20:59
srichterand get a PDF20:59
srichterand all that with as little effort as possible20:59
WebMaven_Should be easy, then.20:59
srichterI would use TeX, but ZPT is not non-SGML friendly20:59
WebMaven_you will almost certainly *not* need to define your own elements.21:00
srichterare the elements in docbook flexible?21:00
srichterok21:00
WebMaven_generally yes.21:00
WebMaven_warning: there are a lot of them.21:00
srichterthat's good21:00
WebMaven_but they are generally very simple and obvious to use.21:00
srichterWebMaven_: do you have a latest Zope 3 running?21:00
WebMaven_Not at the moment.21:01
WebMaven_(not on a box I can access form here anyway)21:01
WebMaven_do you need me to set up a sandbox for something?21:01
srichterthen use: http://130.64.83.73:8080/++apidoc++21:02
srichtergandalf21:02
srichter12321:02
srichtergo to the ZCML Reference21:02
srichterdo you think we could generate decent docbook code for this?21:02
srichterin particular the attribute details21:03
srichterthis is probably the most tricky stuff21:03
WebMaven_I don't think I'm getting a response from that server...21:04
srichteryou should21:05
srichterare you sure the URL is correct?21:05
srichterI know it is public; I show people stuff all the time21:05
WebMaven_Dang. It's the stupid corporate firewall here, blocking port 8080.21:06
WebMaven_I keep forgetting about that.21:06
srichterWebMaven_: try again on port 8021:09
WebMaven_works, looking now.21:13
srichterI can see it :-)21:14
WebMaven_Yes, I think DocBook can be made to do this fairly easily.21:14
srichtercool21:15
srichterany idea where I should start reading about it?21:15
WebMaven_You can stick with generic elements to start (section, paragraph, itemizedlist, heading, etc) and later try transitioning to programming specific elements, which might not be a perfect fit.21:16
WebMaven_http://www.docbook.org/21:16
srichterok, thanks21:17
srichterwhat are the tools to generate PDFs or validate the XML21:18
srichterthe docbook book stinks; no complete examples; I hope people do not think the same about my book :-\21:20
srichterok, found some examples ;-)21:22
WebMaven_some examples are there (chapters, articles, etc.)21:22
WebMaven_googling for "docbook to pdf" revealed: http://www.linuxfocus.org/English/May2000/article152.shtml21:22
srichterhow is the compiling done? (to anything)21:23
srichterI am pretty sure you need to have some style info somewhere21:23
admpquite often docbook->foss->latex->pdf21:23
srichterfoss?21:24
admpFO=formatting objects21:25
WebMaven_OK, it looks like there is a db2pdf utility (also named docbook2pdf)21:26
mgedminhttp://www.xml.com/pub/a/2005/01/19/print.html is interesting (they suggest CSS instead of XSLT + XSL), but I don't know if the book mentioned in the article used docbook or something else21:26
WebMaven_srichter: these utilities are part of the docbook-utils package in debian: http://packages.debian.org/stable/text/docbook-utils21:27
srichterI think there is an equally-named package for MDK21:27
srichterok, so I'll try attack the problem and see how far I can get21:32
*** BjornT has quit IRC21:32
WebMaven_srichter: again, I suggets you start with generic elements like section, para, title, etc, and only then start trying to use elements like ooclass21:37
srichterwhat is ooclass?21:38
srichteryep, I will do that21:38
srichterWebMaven_: did you work out a topic for PyCon?21:39
WebMaven_Do you mean, was one of my talk proposals accepted?21:39
srichterno, I mean the Zope 3 sprint21:40
WebMaven_nothings's been settled, there are a few ideas in the wiki.21:40
WebMaven_(I think the page needs to be refactored a bit, it's getting messy)21:41
*** Aiste has joined #zope3-dev21:41
WebMaven_Do you have a suggestion?21:41
srichterwhat's the URL?21:41
srichterwell, I checked in this user pref stuff today, but it needs a reality check21:41
WebMaven_http://www.python.org/moin/Zope3AppSprint21:41
srichterand some more experienced people in the problem domain21:41
srichterif there are Z2 hackers with CMS experience coming, you could make it very cool21:42
WebMaven_user prefs for...?21:43
srichteruser preferences21:43
srichteryeah, look at API doc :-)21:43
srichterit basically uses schemas to describe the preferences21:44
srichterright now you can only have preference groups, which simple schemas21:44
srichterbut one could also develop a containerish preference group that contains other groups21:45
srichterand so on21:45
srichteranother topic would be to actually implement the docbook generator we were just talking about21:45
WebMaven_Hmm. Looks interesting: http://130.64.83.73/++apidoc++/Book/preferences/show.html21:45
srichterI think several items in the list online are too big for a simple sprint21:46
WebMaven_Me too.21:46
WebMaven_We're only going to pick one.21:46
srichterright, that's the docu21:46
srichterof course :-)21:46
WebMaven_For the sprint, my strategy will be to piuck something that we think is too small, and make sure it has a lot of polish when we're done.21:47
srichteryes, that's a good way of doing it21:47
WebMaven_Are you interested in participating?21:48
srichterunfortunately, I am in Germany21:49
WebMaven_Oh, that's right, you're not coming to PyCon.21:50
WebMaven_:-(21:50
srichterwell, its fortunate and half of the week I will be in Austria skiing with my dad and brother (yeah 8-)21:50
philiKONsrichter, where will you be?21:51
WebMaven_Bah. A big cold blue room with white carpeting.21:51
srichterand speed :-)21:51
WebMaven_srichter: what sort of simple killer-app would you suggest we work on?21:53
srichterthinking...21:54
WebMaven_(I remember when Squishdot was driving Zope installs)21:54
srichterthat one would be probably feasible21:55
srichterbut you need some nice UI to go with that21:55
*** BjornT has joined #zope3-dev21:55
srichteryou can reuse a lot of existing code, like for discussions21:55
WebMaven_Sure.21:55
WebMaven_I don't know if it would be exciting today, though.21:56
WebMaven_(PHPBB and the like have more-or-less taken over)21:56
srichteroh, you know what you could attack?21:56
srichtercode generation :-)21:56
WebMaven_??21:57
srichterIt should be very easy to create new content types and make them usable21:57
srichteryou basically do a UML class model and then import it into Zope 321:57
srichtergenerate interfaces and classes on the fly21:58
WebMaven_Well, is that a Zope3 App Sprint, or is it more properly a Zope3 Sprint (which there is one)?21:58
srichteroh, there is a separate Zope 3 sprint?21:58
srichterI see21:58
WebMaven_http://www.python.org/moin/Zope3Sprint21:58
srichteryou know, noone has written a blogging app for Zope 3 yet, with RSS and all21:59
srichtermaybe this would be a good one21:59
srichterblogging is certainly a hot topic21:59
WebMaven_Yes, that is on my personal short-list.21:59
WebMaven_It should be simple to start, and can use a lot of polishing work.22:00
srichteryep, that's certainly such a tool22:01
WebMaven_(spell-check, autolinking, trackback, pingback, FOAF, OPML, RSS, ATOM,22:01
srichterof course, if you want to work with existing code bases, you could work on ZWiki and BugTracker; both need some serious polishing22:01
WebMaven_quoting, categories, nested categories, category feeds, skins,22:02
srichter:-)22:02
WebMaven_We could do that too.22:03
srichtera lot of the functionality you would need for the blogger is actually already in the bug tracker22:04
srichterlike threaded comments22:04
mgedminsrichter, consider adding <label> elements in ++preferences++...edit.html22:04
srichterthe bug tracker would also be good to work on, since the Zope 3 community could directly benefit from it22:05
srichtermgedmin: I removed the label, because I found the hint stuff tedious at best22:06
srichterbut I can make it a simple label again22:06
philiKONit's not about the hints but AAA compliance22:06
srichtermgedmin: of course, feel free to work on it as well22:06
philiKONjust to mention one thing where the markup counts, not the way it's displayed22:07
philiKONanyway, gotta go22:07
*** philiKON has quit IRC22:07
mgedminsrichter, I wasn't talking about hints22:07
srichterI know22:07
mgedmin<label> lets you click on the big text to toggle te setting instead of targeting a small checkbox22:07
srichterooh, that's cool22:08
srichterI did not know that22:08
srichterhow will it know which checkbox to use?22:08
mgedmin<label for="elementid">label text</label>22:08
mgedmin<input id="elementid" ... />22:09
srichterahh, great22:09
srichterI will do that22:09
mgedminshortcut: <label>label text <input ... /> maybe more text </label>22:09
srichterok22:09
mgedminnice work on the apidoc, by the way22:10
mgedminI wish I had time to explore it now22:10
mgedmingotta work, though22:10
srichterthanks22:10
srichtermgedmin: did you see the interface details screen?22:10
srichterI really love the show/hide section feature (I have no clue whether it will be useful, but it is cute ;-)22:11
mgedminfunky javascript stuff?22:11
mgedminyes, I like it22:11
srichterthat's what the settings are for22:11
mgedminI recently added similair javascript magic to my planet: http://mg.pov.lt/blog/planet-mg.html22:12
srichterthe good thing is that it works in many browsers22:13
srichterit is very simple DOM access22:14
* mgedmin nods22:14
WebMaven_srichter: you could add live searching to the interface finder... that would be cool.22:14
mgedminI have found quirksmode.org very useful22:14
mgedminsrichter, do you write unit tests for the javascript code?22:15
srichterWebMaven_: it does that already; or what else do you mean?22:15
*** tvon has joined #zope3-dev22:15
srichtermgedmin: nope22:15
srichterI tried the JS unittest package once, but it is just painful22:16
WebMaven_http://objectrealms.net/news/livesearch22:16
srichterwhat does live search do?22:16
WebMaven_try typing in the search box.22:16
mgedminJsUnit is painful22:17
srichtermgedmin: yep22:17
srichterWebMaven_: nothing happens, but thats probably because I use Konqui22:17
WebMaven_Oh.22:17
mgedminI wrote a small test runner for my planet: http://mg.b4net.lt/planets/mg/test.html22:17
mgedminso that I could rerun all tests just by reloading the page22:17
srichteris your panet a Zope 3 app?22:18
srichterWebMaven_: is that your library (live search)?22:18
mgedminno, it is planetplanet.org -- standalone python code + mark pilgrim's feed parser + htmlgen (yuck)22:18
srichterI see22:19
WebMaven_for each character you type, it does a 'live search' with an XMLHTTPRequest to get related searches. you can use the down arrow key to pick a result and hit enter, or you can continue typing your search query.22:19
srichteryeah, I started FF22:19
WebMaven_Google does it here: http://www.google.com/webhp?complete=122:20
srichterwell, if someone is willing to check in this code under ZPL in the Zope 3 repository, then I could add this functionality22:20
srichterright, I know about google doing this22:20
WebMaven_The ObjectRealms code is likely Plone specific.22:21
srichterwell, we basically would need the JS22:21
WebMaven_But the JS is relatively simple.22:21
srichterWebMaven_: that would be a good util to develop at the sprint :-)22:22
WebMaven_The only difficult part (I think) is getting good 'completions' to populate the search box with.22:22
srichterright22:22
WebMaven_You probably need a utility to keep track of popular searches, and order them by the number of results.22:23
srichteryep22:24
srichterI probably would do it in the form of code completion22:24
WebMaven_So, does Z3 have a catalog/indexing facility yet?22:25
srichterlike I start typing zope.in and it would suggest zope.interface22:25
srichterWebMaven_: yes22:25
WebMaven_yeah, that would be very cool.22:25
srichterbut you cannot make use of that in apidoc, because nothing is persistetn22:25
srichterbtw, that would be another great sprint topic22:25
WebMaven_we're just full of those today, aren't we?22:26
srichterJim wants a way to precompile all of the apidoc info and store it in a AODB22:26
WebMaven_;-)22:26
srichterincluding indexes et all22:26
srichterthis would make the apidoc runnable, even if not all of Zope 3 can start up22:26
WebMaven_Well, for code completion, you don't really need a full index.22:26
srichterright, that's the reason I suggested it :-)22:27
srichterin fact, the search should be very fast, once you fetched all the names and chaced them22:27
mgedminwhat's AODB?22:27
srichterZODB I meant22:28
WebMaven_Aspect Oriented Database. ;-)22:28
*** jdz__ has joined #zope3-dev22:34
*** jdz_ has quit IRC22:40
*** admp has quit IRC23:06
*** tvon has quit IRC23:07
*** tvon has joined #zope3-dev23:07
mgedminhow do I create a TestRequest that returns a string I want when I call str(request.URL)23:08
*** niemeyer has quit IRC23:08
mgedmin?23:08
mgedminapparently I need to do request = TestRequest(environ={'PATH_INFO': '/path/path/path'}) and then call request.traverse()23:13
mgedminick, I need loads of stubbing to get that work23:15
mgedminrequest.getURL = lambda *a: '/path/I/want' # hacky, but at least it works23:16
*** sashav has quit IRC23:30
algaProblems23:31
*** sashav has joined #zope3-dev23:31
algasuppose we have a permission schoolbell.create23:31
algathat lets users create objects23:31
algaif we try that, we get23:31
algaUnauthorized: (<zope.app.dublincore.annotatableadapter.ZDCAnnotatableAdapter object at 0x41ab230c>, 'modified', 'zope.app.dublincore.change')23:32
algain File "/home/alga/src/schooltool/Zope3/src/zope/app/dublincore/timeannotators.py", line 26, in ModifiedAnnotator23:32
algaI suppose the IZopeDublinCore adapter has to be unwrapped there, right?23:33
algaafter all, it is an event handler23:33
*** benji_york has joined #zope3-dev23:36
*** bskahan has quit IRC23:50
*** jdz__ has quit IRC23:53

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