IRC log of #zope3-dev for Tuesday, 2005-02-15

*** Aiste has joined #zope3-dev00:01
*** oling has quit IRC00:02
*** oling has joined #zope3-dev00:02
*** Jim7J1AJH has quit IRC00:10
*** Jim7J1AJH has joined #zope3-dev00:13
*** SteveA has quit IRC00:23
*** SteveA has joined #zope3-dev00:26
srichterand there it goes!00:43
J1mYee ha!00:43
srichter802 files modified :-)00:43
J1mI guess I should be ready to revover the server.00:44
J1mI guess I should be ready to recover the server.00:44
srichtermaybe enough people are scared, so that they do not immediately update :-)00:44
srichterJ1m: the mail was too big for the list :-)00:45
srichter1.3 MB00:45
J1mI just approved it00:47
srichterok, thanks00:47
srichterI feel so much lighter now ... probably won't last long, since the complaints will trickle in soon...00:49
philiKONcongrats srichter00:57
benji_yorksrichter, I either commend or condem you.  Don't know which yet.  :)00:59
srichterbenji_york: I think the latter ;-)00:59
GaryPosterthanks, srichter!01:01
*** GaryPoster has quit IRC01:01
*** Aiste has quit IRC01:07
*** SteveA has quit IRC01:12
*** benji_york has quit IRC01:15
*** sashav has joined #zope3-dev01:18
J1msrichter, how would you like us to report problems?01:19
srichterJ1m: good question; I guess the collector or E-mail will be fine01:20
srichterof course, I would like people to try a little bit before sending a mail01:20
J1mshould we send email to you? The list?01:21
srichterthe list01:21
srichterI read it frequently enough01:22
srichterI think the biggest worry we will have is ZODB conversion01:22
srichterbecause most of the people using the trunk will be already at generation 101:22
srichterfor the trunk/development, it should be possible to run the evolution script, even though you are already at the target generation01:23
srichterJ1m: do you think you could provide some support for that? (I did not want to touch that code; it looked so clean ;-)01:23
*** SteveA has joined #zope3-dev01:24
J1msupport for what?01:24
srichterif you are let's say at generation 1, you need to be able to run the evolve1 (0 -> 1) script again01:25
J1mHm, I don't think that's a good idea01:26
J1mevolve scripts should be able to count on the generation they start with.01:27
srichterbut there is no telling what generation people are on01:28
J1mSure there is. It's recorded in the database.01:28
srichterfor example, I have ensured that people running X3.0 can convert01:29
srichterbut if a person running the trunk already ran that 0 -> 1 conversion before (for example when we changed from error logging service to utility), then s/he cannot run the evolution script I provided01:30
J1mIf we previously provided a conversion from 0->1, then we need to up the generation number and provide evolution to 2.01:31
srichterso during development you say we could go up 5-6 evolutions?01:32
srichterso X3.0 is at evolution 0 and X3.1 is at evolution 6, for example?01:33
srichterthen I'll note this is not the case now01:34
srichteralso, it makes writing evolution scripts harder01:35
srichterbut maybe this is the lesser evil01:35
J1mNo, it makes them easier, because they only do a small thing.01:35
srichterI restructured the evolve1 script quiet a bit to make it more streamlined.01:35
philiKONsrichter, i thinkt he evolve1 script should be left the way it was and new scripts should be introduced01:36
philiKONmakes it easier for people already running the trunk01:36
srichterso anyways, I think we can fix the situation by providing an empty evolve1 script and rename evolve1 to evolve201:38
J1mShould the getGlobalSiteManager have an Adapters attribute?01:38
J1mIt looks like the hanler method in zope.component.metaconfigure used to take a service name.01:39
J1mThat means "Adapters".01:39
J1mSo that's broken01:40
J1mI guess I'll try to fix it.01:40
srichterI was hoping that noone uses handler() directly but call the basic directives instead01:41
srichterok, time to go home01:41
srichtersee ya tomorrow01:41
*** srichter has quit IRC01:42
*** mexiKON has joined #zope3-dev01:49
*** philiKON has quit IRC02:00
*** mexiKON is now known as philiKON02:01
*** hazmat_ has joined #zope3-dev02:08
*** hazmat has quit IRC02:24
*** d2m has quit IRC02:27
*** __gotcha has quit IRC02:33
*** __gotcha has joined #zope3-dev02:34
*** J1m has quit IRC02:36
*** srichter has joined #zope3-dev02:36
*** ChanServ sets mode: +o srichter02:38
*** SteveA has quit IRC02:52
*** SteveA has joined #zope3-dev02:55
*** hazmat_ has quit IRC03:08
*** xenwryly has quit IRC03:35
*** SteveA has quit IRC03:49
*** SteveA has joined #zope3-dev03:51
*** projekt01 has quit IRC03:55
*** philiKON has quit IRC03:56
*** __gotcha has quit IRC04:34
*** __gotcha_ has joined #zope3-dev04:34
*** oling_ has joined #zope3-dev05:03
*** oling has quit IRC05:11
*** SteveA has quit IRC05:18
*** SteveA has joined #zope3-dev05:21
*** __gotcha_ has quit IRC05:34
*** __gotcha has joined #zope3-dev05:34
*** SteveA has quit IRC06:13
*** SteveA has joined #zope3-dev06:22
*** __gotcha has quit IRC06:34
*** __gotcha has joined #zope3-dev06:34
*** bska|mobile has quit IRC06:45
*** SteveA has quit IRC07:41
*** SteveA has joined #zope3-dev07:44
*** Theuni has joined #zope3-dev07:59
*** Aiste has joined #zope3-dev08:44
*** sashav has quit IRC08:45
*** SteveA_ has joined #zope3-dev08:49
*** stub has joined #zope3-dev08:54
*** SteveA has quit IRC08:54
*** SteveA_ is now known as SteveA08:54
*** Aiste has quit IRC09:02
*** sashav has joined #zope3-dev09:33
SteveAI do so much want to change "zope.Public" to "public"09:38
*** oling__ has joined #zope3-dev09:43
*** oling_ has quit IRC09:51
*** d2m has joined #zope3-dev09:56
*** SteveA has quit IRC10:13
*** SteveA has joined #zope3-dev10:14
*** tziade_ has joined #Zope3-dev11:18
*** SteveA has quit IRC11:21
*** SteveA has joined #zope3-dev11:24
*** MalcolmC has joined #zope3-dev11:55
*** mgedmin has joined #zope3-dev11:55
*** mgedmin has quit IRC11:57
*** __gotcha has quit IRC12:01
*** oling___ has joined #zope3-dev12:02
*** oling__ has quit IRC12:12
*** Aiste has joined #zope3-dev12:20
*** SteveA has quit IRC12:22
*** SteveA has joined #zope3-dev12:39
*** J1m has joined #zope3-dev12:43
*** niemeyer has joined #zope3-dev12:56
*** mgedmin has joined #zope3-dev12:59
*** __gotcha has joined #zope3-dev13:00
srichterSteveA: permission ids must contain a dot13:11
srichtermgedmin: did you see my comments to your bug report?13:14
*** philiKON has joined #zope3-dev13:15
srichtermgedmin: I checked in some tests and fixes last night to make things work better; I also implemented almost all of the date/time formatters now13:15
SteveAsrichter: I know.13:16
SteveAzope.Public isn't really a permission.13:16
SteveAit is a token that means "don't invoke most of the security system"13:16
SteveAyou can rename any other permission in your system.13:16
SteveAit should be "public"13:16
srichterit should be easy to change13:17
SteveAthe "zope." in it is rather different than in "zope.ManageContent" for example.13:17
SteveAI have had to add to my system documentation an explanation for why we need a zope.Public permission when our other 6 permissions are launchpad.Whatever13:17
SteveAbecause it makes it hard to understand13:17
SteveAi also have to very carefully explain why it is special13:18
SteveAit *is* special, so I think it should have a special name.13:18
SteveAit could be aliased as zope.Public, as a deprecation measure.  But, I'd like it to be just "public".13:18
J1mI'm OK w that13:18
* SteveA and team is having a spec-writing fortnight, and these documentation/explanation issues are high on his mind13:19
SteveAcool, thanks jim.  do I need to write a proposal?13:19
* SteveA decides to write a proposal.13:20
SteveAmark's agreed for me to submit a talk or two on launchpad to EP200513:20
SteveAi think we've done some interesting things with the zope security system, and now that we (almost) have what we did documented internally, i can make that into a talk.13:21
SteveAand, the whole "here's a large zope3 application" thing will be interesting13:21
SteveAif I can only work out what to say about it13:21
rejjI'd be interested in seeing any examples of complex z3 apps that are not CMS based13:23
SteveAthis one doesn't even use the zodb13:24
SteveA(although it might later on for some things)13:24
SteveAi don't really know if you'd call it a CMS or not.  It certainly doesn't manage arbitrary content.13:24
rejjI'm interested in complex applications that just happen to render themselves via html13:25
rejjrather than publishing "content"13:25
SteveAthen launchpad qualifies13:26
rejjthen I'd like to see whatever you've got ;) ... assuming you make slides available online, or whatever13:27
rejjsince the chances I'll be at ep2005 are approximately zero13:27
SteveAI guess if "zope.Public" becomes "public" then <allowed interface="..." /> directives in ZCML can be replaced totally with <required permission="public" interface="..." />.  One less thing to learn.13:45
J1mIt can be replaced with almost that now13:46
J1mI really don't see zope.Public as the big deal that you do.13:46
*** bradb has joined #zope3-dev13:55
*** SteveA has quit IRC14:05
mgedminsrichter, thanks14:07
mgedminI don't understand why 'DDD' is not allowed by the spec14:07
mgedminwhat if I want to generate fixed-length filenames like YYYY-DDD?14:07
mgedminok, I don't, actually, and probably that's an abuse of the _localisation_ mechanism anyway14:08
rejjbut not so unreasonable that it is safe to assume noone will ever want to do it14:08
mgedminoh, the collectors now have a nice domain name (
*** SteveA has joined #zope3-dev14:16
srichtermgedmin: as I said, I am willing to change that; it is a trivial fix anyways14:21
srichternow, the problem is that you probably also want to parse this format, right?14:21
mgedminnot yet14:22
srichterso I'll add support for "DD" and "DDD" for formatting14:23
mgedminthe IDatetimeFormat docstring gave me the impression that the formatter is generic14:25
mgedminand knows about numbers and texts14:25
mgedminand you can specify any number of any characters14:25
mgedmin(well, there were some special cases like YY)14:25
*** J1m has quit IRC14:27
srichterthat's not the impression I got from reading the text, but we could certainly support it14:28
srichterbtw, I just E-mailed Stuart about pytz inclusion14:28
*** __gotcha has quit IRC14:35
*** srichter has quit IRC14:42
*** zagy has quit IRC14:51
*** zagy has joined #zope3-dev14:52
*** Awayblia is now known as Workblia14:56
*** projekt01 has joined #zope3-dev15:27
*** SteveA_ has joined #zope3-dev15:35
*** SteveA has quit IRC15:37
*** SteveA_ is now known as SteveA15:37
*** srichter has joined #zope3-dev15:45
*** ChanServ sets mode: +o srichter15:46
srichterdoes anyone here know Stuart Bishop's nickname?15:48
srichtermgedmin: ok, I agree with you observation about the formatting now after rereading the documentation15:54
srichterSteveA: oh that's right; thanks15:55
srichterstub: are you there?15:55
SteveAprobably busy for 45 minutes15:57
srichtergot him15:57
srichtermgedmin: mmh, rewriting my code to support n occurrences of a letter is a bit of work; luckily the pattern parser does the right thing already16:02
srichterso it is the formatting and parsing code that needs to be adjusted16:02
*** regebro has joined #zope3-dev16:10
*** gintas has joined #zope3-dev16:13
srichtermgedmin: good news! the latest locale files contain all timezone names in their native language, so that we will be able to support true language-based timezone parsing and formatting16:31
srichteralso, it seems that the data was even more extended16:31
srichteryou have now three choices for month and day names: abbreviated, narrow (i.e. one letter) and complete name16:32
*** J1m has joined #zope3-dev16:32
*** oling___ has quit IRC16:33
*** benji_york has joined #zope3-dev16:33
srichterJ1m: in the next days we need to update the locale data files; there are some awesome updates, like fully localized timezone names16:33
J1mfirst we need to recover from the merge.16:35
J1mExisting databases are hosed for me.16:35
J1mI can start Zope, but I get lots of breakage.16:36
srichterwhere, for example?16:36
*** alga has joined #zope3-dev16:36
J1mI need to recover from shallower problems forst.16:37
J1mI need to recover from shallower problems first.16:37
J1mI'll probably have to get back to you later.16:37
J1mFor now, I need to use the tag.16:37
J1mYour text in zope/deprecation/README.txt doesn't match reality.16:38
srichterhow so?16:40
*** apauley has joined #zope3-dev16:42
J1mThe first argument to the `deprecated()` function is a list of names that16:43
J1mshould be declared deprecated.16:43
J1mThat's what it says16:44
J1mThat's not what the code does16:44
srichterfor modules it is what it does16:44
srichterfor classes it is something else?16:44
J1mLook at the README.txt16:44
srichterdid I foget to document the usage in classes?16:45
J1mLook at the code shown16:45
J1mLook at the code in in that package16:45
J1mThe first argument is a string *not* a list or any other sequence.16:45
srichteroh, ok, I will update the README.txt16:45
srichterI forgot about this shortcut16:45
apauleyhi all16:46
apauleynewbie question: I'm playing with the buddydemo, typing the testcases in my interactive interpreter16:47
apauleyHow can I make my buddy objects persistent, so that when I create an object in the interpreter it also shows in zmi?16:48
philiKONapauley, oh, that's a little more complicated16:49
philiKONyou'd have to start your ZODB instance from the interpreter instance16:49
philiKONand get a hold of the root folder16:49
philiKONany special reason you want to do this?16:49
apauleyI'm familiarizing myzelf with zope3 in general, but actually I'm working towards writing a QT gui for one of the demo apps16:50
apauleySo I'm trying to figure out how I'd get hold of my objects and manipulate them with an interface other than a browser16:51
srichterapauley: the way to do it is via ZEO connections16:52
philiKONwell, wether a buddy object is in a folder or not doesn't matter wrt manipulating it16:52
philiKONapauley, if your QT doesnt' really need the ZODB, i wouldn't worry about this at all16:52
apauleyZEO connections? any docs for that?16:53
srichteryou might want to contact Uwe Oestermeier, who did a wxPython GUI already16:53
apauleyThanks, that might help me16:53
philiKONapauley, ZEO connections for debugging is frequently used with z2.7; a lot of the docs you find on that should be applicable to z3 as well since the ZODB impl is the same16:54
philiKONUwe actually uses ZEO to connect his wxPython GUI client to a Zope 3 ZODB16:54
apauleyphiliKON: I don't actually need ZODB for my purposes, I'm more interested in using mysql for storage16:58
philiKONthat's perfectly fine. i assume you just want to use zope3's component architecture which is quite possible using outside of zope16:59
philiKONpeople have already successfully used it together with SQLObject for example (an RDB to object mapper)16:59
apauleyYes, exactly17:02
philiKONyeah, then you probably don't want to go through all the ZODB hurdles if you're not going to use it anyway17:03
srichterI think some Python-based RPC protocol or even XML-RPC should suffice17:04
philiKONfor what?17:04
srichterdata communication17:04
philiKONah ok... if that's even necessary17:04
srichterwell, as far as I understnad him, he still wants an app server17:05
philiKONso far only storage has been the issue17:05
philiKONas far as i understand him he wants to build a QT app with the component architecture17:05
srichterwell, then things are much easier17:05
apauleyOK, so as far as I understand it would be easier to write a gui app that uses xmlrpc?17:08
philiKONapauley, well, do you need to communicate with a zope3 server?17:09
philiKON(we really don't know what you need to build)17:09
apauleyI need to build a point of sale system17:10
apauleyI'd like to use zope to implement things like mvc17:10
*** alga has quit IRC17:10
philiKONi mean, what kind of comopnents does this system have? fat client? app server? etc.17:11
apauleyOh, ok17:11
apauleyEach pos station should be able to function independently of the backend, and of each other17:12
apauleyIn case the network drops, for instance17:12
apauleyso each pos station would be a pc with a local view (eg QT) and a local storage component17:13
philiKONso, no server communication then?17:14
srichterright, this sounds exactely like Uwe's requirements17:14
srichterphiliKON: well, eventually he would want to sync with the server17:15
srichterand you need to do conflict resolution17:15
apauleyThe POS station submits messages to a local queue17:15
apauleyAnd the local queue delivers the messages to the backend17:15
philiKONsrichter, well, the stations actually don't change the base data, dunno17:15
*** SteveA has quit IRC17:15
philiKONthat local queue thing could even be totally independend of the app17:16
philiKON(e.g. a cron job)17:16
apauleyNo, the POS stations pretty much just read the data17:16
apauleyIt gets updated by the backend using the same queue mechanism17:16
philiKONwell, i'd take one step at a time, implement a QT app with zope3's component architecture and worry about data sync later17:17
philiKONsince the data sync happens on mysql level anyway, there's really  not much zope can do for you here17:17
*** SteveA has joined #zope3-dev17:18
apauleyYes, I'm also not too worried about the queue right now17:18
philiKONapauley, k, shouldn't be too hard then. if you've got any questions on how to use the comopnent arch. outside zope, this is the place17:18
apauleyI'd like to demo the basic concept of writing a component-based gui-app with a relational db to my management on monday17:20
apauleyTheoretically I would be able to replace a browser view with a gui component, right?17:21
apauleyWithout interfering with business logic17:21
philiKONapauley, except that most of the stuff that has browser views is for usage in zope3's app server so i'm not sure if you can actually reuse that much of it17:22
apauleyAnd I'd also be able to suddenly change my storage to mysql, without the rest of the system worrying too much17:22
philiKONor, if you decide to change to wxPython, you'd "only" have to rewrite the gui components17:22
apauleyI'm not really completely stuck on QT17:24
philiKONSteveA, what do you guys use for the schooltool GUI apps... wxpython?17:25
apauleyIf Uwe's wxPython stuff does what need to be done, I'd be happy to use that17:25
SteveAphiliKON: it was.  you need to ask alga or mgedmin or gintas17:26
srichterI read some comparisons and the real good Python people use Qt, but wxPython is much, much wider spread because it is easier to understand17:26
SteveApersonally, i've decided not to use wxwindows for a while.17:26
philiKONSteveA, how come?17:26
SteveApygtk is nice17:26
philiKONbut you're limited to X1117:27
SteveAwith wxpython, when you hit a wall, it has spikes in it17:27
philiKONor you need to ship gtk for windows17:27
mgedminpygtk! pygtk!17:27
SteveAor ship an ubuntu cd with your app ;-)17:27
srichterphiliKON: the WingIDE guys do this very successfully17:27
mgedminthere is a windows gtk+ port17:27
philiKONmgedmin, i know17:28
srichterin fact, John is helping to speed up pyGTK for windows17:28
philiKONbut on OSX, e.g., i want my interface to look like other mac apps17:28
philiKONwxpython can do that for me17:28
srichtermgedmin: so you use pygtk?17:28
mgedminin schooltool we use wxPython17:28
mgedminthat's why I'm advocating pyGtk now :)17:28
mgedminwxPython is nice17:28
mgedminjust not really finished17:28
mgedminand not really cross-platform17:29
SteveAand has those hairy wall spikes17:29
philiKONSteveA, not sure what you mean with that :)17:29
srichterwell, I am convinced now that if I would do the project I would use pyQT :-)17:29
philiKONSteveA, mgedmin, would you advocate one over the other for use with the CA?17:29
mgedmin(e.g. to get right-button menus in list boxes you have to register three event handlers -- two are only used on MSW, while the third one is only used on GTK)17:29
philiKONmgedmin, that sucks17:30
SteveAphiliKON: what window manager should the linux kernel use?17:30
mgedminseriously, though -- if you want cross-platformness, then wxPython is better17:30
philiKONSteveA, well, it's not *exactly* like that, is it?17:31
mgedminyou will have to work harder, perhaps, but your apps will have the native look and feel17:31
SteveAshtoom has a variety of UIs17:31
philiKONSteveA, i mean, it might be that one is more suitable than the other for separation of concerns17:31
mgedmin(is the Gtk2 port of wxPython stable yet?)17:31
SteveAjust as `anthony17:31
SteveAjust ask `anthony17:31
philiKONSteveA, k, thanks17:34
*** sashav has quit IRC17:42
srichtermgedmin: as soon as stub imports pytz I have a checkin ready that will support formating of timezones17:46
*** _projekt01 has joined #zope3-dev17:48
*** projekt01 has quit IRC17:48
*** rejj has quit IRC17:56
*** Jim7J1AJH has quit IRC18:10
*** Jim7J1AJH has joined #zope3-dev18:13
*** SteveA_ has joined #zope3-dev18:37
*** SteveA has quit IRC18:41
*** mgedmin has quit IRC18:56
*** sashav has joined #zope3-dev19:01
Workbliawhere could i find info about registering a local authentication utility ?19:04
*** Aiste has quit IRC19:05
*** gintas has quit IRC19:18
*** tziade_ has quit IRC19:22
*** apauley has quit IRC19:23
*** regebro has quit IRC19:23
J1msrichter, ayt?19:27
*** tvon has joined #zope3-dev19:32
*** tvon has quit IRC19:43
*** alga has joined #zope3-dev19:51
srichterJ1m: here now19:59
J1msrichter, is there a way to deprecate an entire module easily?20:00
srichterHere is what I did for (see
srichterdepmodule = DeprecationProxy(module)20:01
J1mgot it20:01
srichterdepmodule.deprecate(module.__dict__.keys(), msg)20:01
srichterI guess we could put such a helper function into zope.deprecation20:01
srichtersomething along the lines: deprecateModule(module, deprecated_module_path, msg)20:02
*** bradb has quit IRC20:03
*** GaryPoster has joined #zope3-dev20:04
J1mI actually want to deprecate a module location, not the module itself.20:04
srichterthat's exactely what is going to happen20:05
srichterso, maybe deprecateModulePath(path, module, msg)20:05
srichterwe could probably even autogenerate the message20:06
*** zagy has quit IRC20:09
*** SteveA_ has quit IRC20:25
*** bradb has joined #zope3-dev20:27
*** SteveA_ has joined #zope3-dev20:27
*** Theuni has quit IRC20:35
*** mgedmin has joined #zope3-dev20:36
*** MalcolmC has quit IRC20:38
*** `anthony has quit IRC21:09
*** Aiste has joined #zope3-dev21:13
*** `anthony has joined #zope3-dev21:20
*** niemeyer has quit IRC21:47
*** SteveA_ has quit IRC21:51
*** SteveA_ has joined #zope3-dev21:54
*** gintas has joined #zope3-dev22:12
* mgedmin tries to get the hang of zcml overrides22:17
mgedminI have a <browser: view ...> in one configure.zcml22:18
mgedminand another <browser:view ...> with the same name & interface in another.zcml22:18
mgedminI do <includeOverrides file="another.zcml" />22:18
mgedminyet I get a config conflict error22:19
*** bradb has quit IRC22:19
*** stub has left #zope3-dev22:20
benji_yorkIs the zcml small enough to paste here?22:20
*** SteveA_ has quit IRC22:21
mgedmindo you mean just the browser:view directive, or what?22:22
mgedminI have a package-include file22:22
mgedminthat <include package="foo" />22:23
mgedminand <include package="bar" />22:23
mgedminin foo/configure.zcml I have <include package=".browser" />22:23
mgedminin foo/browser/configure.zcml I have a <browser:page ... />22:23
mgedminin bar/configure.zcml I have <includeOverrides file="overrides.zcml" />22:23
mgedminin bar/overrides.zcml I have basically the same <browser:page ... />22:24
mgedminonly the class="" attribute differs22:24
mgedminI get two conflicts: one for an adapter that defines a menu item22:24
mgedminanother for the view itself22:24
mgedminI would like to know if this scheme is supposed to work22:24
mgedmini.e., can I use <includeOverrides> deep within a package's configure.zcml22:25
srichterthat I am not sure about22:26
srichterbut is foo called before bar?22:27
srichtermmh, strange; J1m wrote this feature....22:28
mgedminalthough in the error message itself22:28
mgedminmy overrides.zcml is mentioned above the other configure.zcml file22:28
mgedminis that a clue?22:28
benji_yorkHmm, I've delt with these before, but don't have a clear set of instructions to resolve them.  :(22:30
* mgedmin nods22:31
benji_yorkIs the includeOverrides after the include?22:31
mgedminbar/configure.zcml contains one <subscriber> and one <includeOverrides>22:31
mgedmin<include package="bar" /> is the very last line in Zope3/package-includes/myapp-configure.zcml22:32
mgedmingot it working22:32
*** jhauser_ has joined #zope3-dev22:32
benji_yorkWhat'd you do?22:32
mgedminI moved <includeOverrides> from bar/configure.zcml directly to Zope3/package-includes/myapp-configure.zcml22:32
mgedmin'bar' is basically a package that modifies certain things in 'foo'22:33
mgedminit is not usable without 'foo'22:33
mgedminI wanted to be able to add a single package-include file for bar and have it alter the behaviour of foo22:33
mgedmintherefore I put that <includeOverrides> into bar's configure.zcml22:34
benji_yorkThat sounds reasonable.  I'm looking to see if we do something similar...22:34
mgedminI will try to make a small testcase to reproduce this, if I can scrounge up the time22:35
mgedminI just wanted to know if I was doing something that wasn't supposed to work22:35
benji_yorkWe do bascially the same thing.  I'd like to know why it isn't working in your situation.22:36
*** rejj has joined #zope3-dev22:47
*** jhauser has quit IRC22:47
benji_yorkrejj, made any headway on 2.4?23:14
mgedminhm, doctest.REPORT_ONLY_FIRST_FAILURE suddenly stopped working23:16
mgedminnow I see all diffs in my functional doctests23:16
mgedminI wonder why23:16
* mgedmin looking at
mgedminnope, not that23:18
mgedminFunctionalDocSuite used to check whether 'optionflags' was present in kw23:21
mgedminand did not touch it if it was there23:21
mgedminnow it unconditionally overrides it23:21
mgedminwhy was it changed?23:21
mgedminthat line was last changed by fdrake in rev 2778423:26
J1mIN an old checkout of mine, it overrides23:27
J1mWhy don't you just pass -1 to
mgedminbut in zope/app/tests/, rev 20922, it checks23:28
mgedminI do, it doesn't work either23:28
*** hazmat has joined #zope3-dev23:30
mgedminI mean, -fpv1 also shows multiple doctest failures in a single functional doc test file23:30
J1mWeren't you threatening to change
J1mAccording to:23:31
J1mnever mind23:32
mgedmin(I did change, but -1 worked after the change)23:32
mgedminI would like to compare the latest version of (before it disappeared) with the current
mgedminhow do I find out the latest available rev of a file that was removed23:33
*** ChanServ sets mode: +o hazmat23:33
*** gintas has quit IRC23:33
J1mI dunno.  I'm finding viewcvs not very helpful wrt this23:33
mgedminI did a binary search23:36
mgedminis the last revision that does not give me a 404 page23:37
mgedminlast change appears to be in rev 2908723:37
srichteryeah, I noticed this problem (with -1 not working for functional tests) as well23:37
srichterJim always claimed it worked for him ;-)23:37
mgedminit has a if 'optionflags' not in kw: check23:37
mgedminso, where does the current zope/testing/ come from?23:37
mgedminit was branched from trunk after rev 2848623:38
mgedminzope/testing/ was later modified in revs 28779, 28792, 28984, 29022, 2908723:39
mgedminit is likely that these changes were lost by srichter's merge23:40
WorkbliaWidgets have a required attribute that is not used by the widget code (it is set in constructor from the field the widget is for) yet on all occasions required testing is done - widgets look at the field directly23:40
* mgedmin feels vaguely sorry about ignoring srichter's plea to not change anything in some packages23:41
Workbliais it a feature - widget.required is deprecated and should be read only23:41
* mgedmin hasn't been reading keeping up with zope3-dev lately23:41
srichtermgedmin: J1m: that's aweful; why did the diff not pick it up? I did not even modify this code23:41
Workbliaor a bug - widget.required should be used instead of self.context.required in widget code23:41
mgedminsrichter, maybe because was moved from to
srichteroh, right23:42
J1msrichter, perhaps because the file moved23:42
srichtersvn was not clever enough23:42
srichterit should be able to follow a file23:43
*** Aiste has quit IRC23:43
mgedminI think svn follows moves, but not copies23:44
mgedminit just saw that a file was forked into two branches, then removed in one branch23:45
srichterI see23:47
mgedminsrichter, try svn diff svn:// svn://
mgedmin(what an ugly command line)23:48
srichterok, I see23:50
mgedminthe only regression is that final chunk23:50
mgedminas far as I can tell23:50
mgedminto see the diff but ignore changes in whitespace, do svn diff --diff-cmd=diff -x '-u -Bbw' svn:// svn://
srichterI agree only the option flas are different23:51
mgedminyippee, my functional tests pass again23:52
mgedmin(I relied on NORMALIZE_WHITESPACE, which was also accidentally lost)23:53
*** Aiste has joined #zope3-dev23:53
mgedminsrichter, do you want me to commit the fix for
mgedminI have it in my sandbox23:54
mgedmin(which is read-only, but then it's just a matter of svn switch --relocate)23:55

Generated by 2.15.1 by Marius Gedminas - find it at!