IRC log of #zope3-dev for Saturday, 2005-12-17

J1mI'd vote for you.00:00
benjijust ask GaryPoster, he agrees that if I were a mear ambasidor, I'd have already started WW300:00
ClaesBasIs there some kind of command like "mkzopeinstanceupdate"00:01
benjinot to mention put general spelling ability back 100 years00:01
GaryPosterHey--can I talk about the gross tzinfo thing here?00:01
benjiClaesBas, do you want to see changes made to your skeleton, or to the Zope 3 source (in subversion)00:01
benjisure, GaryPoster, why not00:01
GaryPosterSo the thing that (evil) Benji told me to look at is a really bad non-API use of pytz.00:02
GaryPosterIt's bad because it instantiates timezones that can be pickled, but cannot be unpickled, so they are lying in wait for the unwary.00:02
benjiGaryPoster, is right, all the evil twins have goatees, therefore I am an evil twin00:02
benjiyep, very evil00:02
benjiwhy, do you think, it does that?00:03
GaryPosterBecause it wasn't designed to be used the way it is being used here.  It isn't the right class for the job.00:03
benjiso it's a misuse more than a bug00:03
ClaesBasI try to learn this Zope thing and done Benjis examples which says that you should do a mkinstsance! I was thinking how to update the instance after a "svn update" of main source....00:03
GaryPosterWell, it's a misuse of pytz, which is a very nasty bug in zope.i18n00:04
GaryPosterSo I have three possible solutions00:04
benjiamongst the possible solutions are...00:04
GaryPoster1) I assert that these things should not be persistable, and I have a subclass of pytz.reference.FixedOffset that pukes when you try to persist it, perhaps with a helpful message like "Hey!  convert this time to UTC, buddy!"00:05
benjiI like that one, continue.00:05
GaryPoster2) I assert that these things may be persisted, even though you probably shouldn't, and let people hang themselves if they don't know any better, simply using  pytz.reference.FixedOffset without any warnings.  This will work, but it will make the ZODB bloat if people do it a lot, since there's no code for using a singleton for a given offset (so you'd have multiple 0 offset timezones, multiple 60 minute offset timezones, etc.).00:07
benjiand what is behind door number 3?00:08
GaryPoster3) I copy all, or the pertinent bits, of into zope.i18n, since it has a timezone implementation that does the right thing, and use it.  I can't use it in because it would create a nasty dependency.00:09
GaryPosterI prefer 1 too00:09
ClaesBasDoes a instance (mkzopeinstance) need any update after a "main source update" anytime?00:10
J1mWhat will 1) break?00:10
benjiyep, definately 1, cleaner and lets people know early when they've done something stupid00:10
J1mGaryPoster, what will 1) break?00:11
benjithe darn tests are failing and I don't know why00:11
GaryPosterwell, benji, you said you liked #1, not that you preferred it. ;-)  J1m, number 1 will change the current behavior, which is to break at *unpickling*, to break when pickling.  So, if it breaks anything, it was already broken, but in a more subtle and unpleasant way.00:11
benjino tracebacks; all it says is Total: 7023 tests, 0 failures, 3 errors00:11
benjithat's cause you weren't done yet, GaryPoster, didn't want to overcommit :)00:12
GaryPosterhee hee00:12
GaryPosterwhat do you think, J1m?00:12
J1mbenji, I'm aware of that.00:12
J1mIt's my fault.00:12
benjioh, good, I was afraid it was me :)00:12
J1mI'm trying to fix it.00:12
J1mI mean it was you.00:13
J1mBut I'll fix it for you.00:13
benjithanks, Jim, you're so nice to me00:13
J1mGaryPoster, surely, since existing code is causeing these tp be pickled, making pickling of them raise an error will break something.00:14
*** fcorrea has quit IRC00:14
benjiJ1m, it might break something, but in a here-we-just-saved-you-major-debugging-time way00:15
GaryPosterMaybe so--don't you figure that code has already been fixed, though, since it would cause broken pickles?  Alternatively, if we do care about old00:15
* benji must be on some kind of talk-with-hyphens-kick00:15
GaryPosterto continue... :-)00:15
GaryPosterif we do care about old pickles, then we would have to change the original pytz class that is being used in an inappropriate way to support the inappropriate usage...00:16
J1mSo you assert that code is no longer trying to pickle these things, but might be tempted to in the future?00:16
J1mI'm not asking about old pickles.00:16
J1mI want to know if this change is going to break existing code.00:17
ClaesBasThe silence of my question says: The code created from "mkzopeinstance" never need to be upgraded after a "main source update"! Or?00:18
GaryPosterI assert that if code is trying to pickle these things, they were or are in for some nasty hurting.  Since there is an XXX, it seems others have already discovered the hurting, and presumably changed their code to adjust, since Zope is still broken.  If I try the change and run the tests and nothing breaks, will that be good enough?00:18
GaryPosterOK, cool00:19
GaryPosterClaesBas: yes, sometimes you want to make a new instance, or at least compare instances to see how the skeleton changed.  Once you apply any similar changes to the new instance that you had to the old, you should be able to use the old database, just copying it in.00:21
GaryPoster(or moving it in)00:21
ClaesBasBut there is no standard command or solution?00:21
GaryPosterNo, not to my knowledge.00:22
benjiClaesBas, we generally move the old instance out of the way, build a new one, then put the database(es) back00:23
ClaesBasSo everytime you do a new "mkzopeinstance" do a extra one to be able to diff against in the future....00:23
benjiClaesBas, I've often thought of adding a "diff" and "update" command to mkzopeinstance, but haven't had sufficient inclination00:24
* benji just notices that I repeated Gary, oops00:25
*** efrerich has quit IRC00:28
dman13ok, I implemented my fix for the xml attribute quoting in principals.zcml.  (I ran out of disk space trying to check out the 3.2 branch)00:29
dman13Along the way I noticed there is ./zopeskel and ./src/zope/app/server/zopeskel00:30
dman13the former appears to reference twisted;  mkzopeinstance appears to use only the latter, and the latter doesn't have a principals.zcml template00:30
dman13I'm not sure what I ought to do about that (if anything)00:30
ClaesBasWhen in future do you think documention/Books is going to be in level with source/concepts of Zope3 (I mean more stabilized)?00:33
*** gumpa has joined #zope3-dev00:33
benjiClaesBas, that day may never come, technology books are notoriously always out of date00:34
GaryPosterj1m, In Zope 2, when we wanted to disable pickling, we had both __getstate__ and __reduce__ raise errors.  Is it necessary (or good form) to have *both* raise an error?00:35
GaryPosterI guess I'm confused. :-/00:35
dman13is this really valid XML?   login='ad"min'00:35
J1mYou only need to have __reduce_ex__ raise an error.00:36
GaryPosterOK, cool, thanks00:36
benjiyep, dman1300:36
J1massuming that the object is not persistent. :)00:36
dman13benji: thanks.  I haven't seen that in practice before (and haven't read the spec in-depth)00:37
dman13benji: but saxutils.quoteattr() spit it out like it was :-)00:37
ClaesBasZope3 is never going to be a "big" plaform if you continue to race like you do today, with no chance for anybody else more then you "core-devs" to understand the product.... What is the "masterplan"?00:43
dman13the masterplan is to make z3 perfect, then retire with massive wealth00:43
ClaesBasBut the perfect product/concept is moving very fast...00:44
benjiClaesBas, there is definately a desire to "solidify" the core system, that is pretty much starting now00:44
dman13no, perfect is not moving,  we're just trying to figure out what it is00:44
dman13(that's one perspective, anyways)00:45
benjiplus, there is great attention given to backward compatibility (but that's a tangent to what you were asking about)00:45
ClaesBasThings like "apetizers" which not works disturbs me....00:46
dman13ok, check out r40837 (and issue 522) if you want to double-check what I just did00:47
ClaesBasI mean for example  jeffrey Shell's TODO application00:47
dman13should I merge that change to the trunk now or is someone going to merge all the 3.2 bugfixes at the end of today?00:50
ClaesBasI'm not going to disturb you anymore today, maybe I should wait some months to see if there is documention ( =chance to understand) then or find another platform follow....00:53
dman13ClaesBas: there is documentation for zope3, both online and in print.00:54
dman13ClaesBas: once you understand the principles of the platform and the main pieces, the rest is pretty natural00:54
dman13ClaesBas: at least, that has been my experience00:54
ClaesBasHow many hours how you invested in this?00:55
dman13ClaesBas: and you always have the source to fall back on to help understand something (I find it helpful, and the source is always more precise than the documentation)00:55
dman13I don't know.  Quite a few all total, but I created 2 applications based on it.00:56
ClaesBasThe problem I think is that it's a fast moving target....00:56
dman13it's somewhat of a problem.  Since the 3.0 release the API has remained stable, and the newer work mainly just adds more functionality.00:57
dman13I started working with Z3 in 2003.  One of the applications (which I am working on adding new functionality to now) uses a cvs snapshot from that timeframe00:58
dman13once I understood the concepts, following discussions on zope3-dev became easier because I had an idea what was going on even if I didn't know the internal details of that particular item00:59
dman13I think zope is a good platform and like working with it00:59
ClaesBasTry to follow Richters book and do the examples...00:59
dman13I don't have the print version, but I used the online version (not completely up-to-date) before the print one was released01:00
benjiClaesBas, you might want to ask srichter, I was under the impression that most or all of the examples should still work01:00
ClaesBasMy point is: You need to be somekind of core-dev or very experienced of Zope2 to be able produce anything with Zope301:01
ClaesBasI read on the list there is not going to be any updates of the books in the near future...01:02
benjiI don't think you need to be a core developer of Z3 to use it (but who am I to speak); and I think having extensive Z2 experience is actually a henderance01:03
*** gumpa has quit IRC01:03
ClaesBasI'm not trying to be rude to you,,,, I think Zope is very cool.....01:03
benjino offence taken, I welcome constructive criticism01:04
ClaesBasBTW I liked your (Benji) example Zope 3 in 30 minutes!01:05
benjiactually, that one is by Baiju, mine is here
jintyClaesBas: I am neither a core dev or experienced in Z2 and I use Z3 sucessfully01:07
dman13I agree with benji that zope2 experience can hamper learning zope3 because they are somewhat different paradigms01:07
dman13(however, learning z3 made the cmf make sense to me)01:07
ClaesBasSorry I meant your (promise) hello-thing!01:08
GaryPosterJ1m: the good news is, no functional tests fail when I make the non-persistable tzinfo change.  The bad news is, that just means that the code wasn't tested in any of the functional tests.  I had an error that should have made them puke.  It made some unit tests puke, but no functional tests.  I have to go home.  Do you have a quick suggestion/opinion/decree, or can I ask you to send me an email when you have a moment?01:09
GaryPosterOK, gotta run, I'm late. :-/  I'll check in when I can.01:11
*** jinty has quit IRC01:23
ClaesBasI promised you earlier not to disturb you anymore today.... Keep up the good work (with maybe more documention and examples) in the future!01:24
*** ClaesBas has left #zope3-dev01:24
*** GaryPoster has quit IRC01:31
*** yota has quit IRC01:52
*** J1m has quit IRC01:56
*** romanofski has joined #zope3-dev03:02
*** stub has joined #zope3-dev03:26
*** GaryPoster has joined #zope3-dev03:27
*** romanofski is now known as rom|zzZZ03:35
*** GaryPoster has quit IRC04:03
*** GaryPoster has joined #zope3-dev05:03
*** stub has quit IRC05:21
*** hazmat has quit IRC05:46
*** GaryPoster has quit IRC06:17
*** GaryPoster has joined #zope3-dev06:21
*** GaryPoster has quit IRC06:21
*** philiKON has joined #zope3-dev06:39
*** philiKON has quit IRC06:50
*** philiKON has joined #zope3-dev06:51
*** deo has quit IRC07:01
*** dman13 has quit IRC09:47
*** yota has joined #zope3-dev10:09
*** mgedmin has joined #zope3-dev10:19
*** tahara has joined #zope3-dev11:04
*** benji has quit IRC11:25
taharaHi.I have a question about manner of using 'removeSecurityProxy'11:46
philiKONtahara, go ahead and ask11:47
taharanow I try to fix the gadflyda test page.11:47
taharabecause I use gadflyda's test, then occured forbidden attribute error.11:48
philiKONforbidden attributes are an indicator of missing security declarations11:49
taharaGadflyAdapterCursor does not have any security declarations.11:50
taharabut I think if a local utility component(gadflyda) is not security proxied,11:51
taharathen does not occur error. and11:52
taharaif I got gadflyda instance using getUtility, it is not proxied, but got instance using traverse, it is proxied.11:54
taharaso I don't know which is a correct way to implement local utility...11:55
taharaI'm sorry to my bad english...11:55
philiKONnot sure what that has to do with implementing a local utility11:55
philiKONretrieving a local utility would most probably happen via getUtility11:55
philiKONbut i guess the regular traversal should also work11:56
philiKONincluding cursors retrieved from the utility11:56
taharabut when I write a browser package,11:56
taharaself.context is proxied component,11:57
philiKONright, right11:57
taharathen should I use removeSecurityProxy for context?11:57
philiKONpreferably not11:57
philiKONproivde security declarations11:58
philiKON<class class="....GadflyCursor"> <allow interface="...IZopeCursor" /> </class> I think11:58
taharaok I see. so gadflyda needs more configuration.11:58
taharaI'll post this report to collector after dinner:-)12:02
*** BjornT has joined #zope3-dev12:23
*** povbot has joined #zope3-dev14:40
*** drzoltron_ has joined #zope3-dev15:06
*** tahara has joined #zope3-dev15:22
*** mgedmin has quit IRC15:30
*** SteveA has joined #zope3-dev15:31
*** drzoltron_ has quit IRC15:39
*** Aiste has joined #zope3-dev15:49
*** J1m has joined #zope3-dev16:01
taharadoes anyone commit this bugfix?
taharagadfly may not important feature, but every beginner get start with this.16:03
taharaso I do mind this error.16:03
*** tiredbones has quit IRC16:06
J1mtahara, do you want to commit this bugfix?16:06
J1mare you a contributor?16:07
taharaNo I'm not.16:07
taharaI want to be, but not yet.16:07
J1mwould you like to be?16:07
taharayes, ofcourse!16:07
J1mso submit a contributor agreement and I'll set you up.16:08
taharaI use zope in four years in japan. but I'm not good in English:-(16:08
taharaoh Thanks!!16:08
J1mThat's OK, I don't know Japanese at all.16:08
J1mBut they still let me contribute.16:09
taharaok I'll send agreement.16:09
taharathank you so much!16:09
J1mSeriously, communication is more important than coding.  English is the common language for Zope development.  I very much appreciate the efforts of people work hard to communicate in English to contribute.16:11
taharaI see.16:12
taharaI have beed experienced my communication problem in last pycon:-(16:13
J1mWe appreciate the effort.16:14
*** d2m has joined #zope3-dev16:15
philiKONtahara, it's never too late to learn a language :). i'm learning chinese right now, for example...16:19
J1mI wish I had time to learn a 2nd language, even an easy one like Spanish.  Someday.16:20
philiKONJ1m, never learned one in school?16:20
J1mNah. I was a goof off until fairly late in my school carreer.16:22
philiKONso there's still hope for me too ;)16:22
taharain japan, almost people have learned in english ten years, but can't speak, sigh.16:23
taharabut I try my best:)16:24
J1mYou can at least read and write it.16:24
philiKONand we can understand what you write :)16:25
*** tiredbones has joined #zope3-dev16:29
*** ChanServ sets mode: +o srichter16:31
*** natea has joined #zope3-dev16:41
*** ignas has joined #zope3-dev16:43
*** tarek has joined #zope3-dev16:48
*** tahara has quit IRC16:50
*** xenru has joined #zope3-dev17:10
*** BjornT_ has joined #zope3-dev17:20
*** MJ has joined #zope3-dev17:23
*** drzoltron_ has joined #zope3-dev17:27
*** BjornT has quit IRC17:34
*** efrerich has joined #zope3-dev17:45
*** BjornT_ has quit IRC18:08
*** BjornT has joined #zope3-dev18:08
*** dman13 has joined #zope3-dev18:15
*** drzoltron_ has quit IRC18:24
*** VladDrac has joined #zope3-dev18:37
*** sashav has joined #zope3-dev18:40
*** roym has joined #zope3-dev18:56
*** jinty has joined #zope3-dev18:57
*** jinty has quit IRC19:02
*** lifeless has quit IRC19:10
*** lifeless has joined #zope3-dev19:10
*** J1m has quit IRC19:14
*** zbir has quit IRC19:56
*** BjornT has quit IRC19:56
*** BjornT has joined #zope3-dev19:58
*** zbir has joined #zope3-dev19:58
*** MrTopf has joined #zope3-dev20:04
efrerichCan somebody say when the next 3.2 release is planned? I want inform the i18n-translators20:13
*** BjornT has quit IRC20:23
*** BjornT has joined #zope3-dev20:27
*** natea is now known as natea|away20:34
philiKONefrerich, my guess is first week of january20:37
philiKONbut srichter knows more20:37
*** mkerrin has joined #zope3-dev21:16
*** zbir has quit IRC21:41
*** tarek has quit IRC21:52
*** tarek has joined #zope3-dev21:53
*** newpers has joined #zope3-dev22:03
*** ChrisW has joined #zope3-dev22:10
ChrisWany btree wizzards here?22:11
*** efrerich_ has joined #zope3-dev22:15
newpersi think it's clear that there needs to be more zope3 demos out there, but does anyone agree with me that there needs to be more zope3 demos that don't use the zmi?22:27
*** ChrisW has left #zope3-dev22:28
philiKONnewpers, seen ?22:29
philiKONnewpers, plus, what zmi? ;)22:29
newpersthat list is extremely small22:30
newpersmaybe i'm confused on the target audience of zope322:30
philiKONmake it one entry bigger :)22:30
*** zbir has joined #zope3-dev22:31
philiKONseriously, 3 starter tutorial ain't that bad22:31
*** philiKON has quit IRC22:31
*** philiKON has joined #zope3-dev22:31
philiKONoops. sorry. hit command + q there22:32
newpersi guess i'm failing to remember that zope3 is still new22:32
*** efrerich has quit IRC22:33
philiKONbut it's not *that* new22:34
philiKONyou haven't said yet what it is that you want that the existing docs don't provide22:34
newpersit's not the docs that's the probloem22:35
newpersi just don't understand how #rubyonrails has over 350 people in there right now and #zope and #zope3-dev don't come close22:36
philiKONyou read my blog?22:37
newpersyeah, it was great22:37
philiKONthen you know why :)22:37
*** BjornT_ has joined #zope3-dev22:41
*** roym has left #zope3-dev22:50
*** BjornT has quit IRC22:53
*** MrTopf has quit IRC22:55
*** mkerrin has quit IRC22:55
*** jhauser has joined #zope3-dev23:28

Generated by 2.15.1 by Marius Gedminas - find it at!