IRC log of #zope3-dev for Thursday, 2005-12-15

« Wednesday, 2005-12-14 Index Friday, 2005-12-16 »
*** adamSummers has joined #zope3-dev00:21
*** zbir has quit IRC00:34
*** elro has quit IRC00:37
*** GaryPoster has quit IRC01:05
*** benji has quit IRC01:21
*** sashav has quit IRC01:26
*** sashav has joined #zope3-dev01:26
*** natea has quit IRC01:44
*** natea has joined #zope3-dev01:51
*** philiKON has joined #zope3-dev01:51
*** J1m has quit IRC02:10
*** MiUlEr has joined #zope3-dev02:38
*** philiKON has quit IRC02:41
*** GaryPoster has joined #zope3-dev02:41
*** alga has joined #zope3-dev02:48
*** tarek has quit IRC03:02
*** deo has quit IRC03:12
*** efge has quit IRC03:14
*** yota has quit IRC03:24
*** tiredbones has left #zope3-dev03:43
*** stub has joined #zope3-dev03:44
*** alga has quit IRC03:58
*** natea_ has joined #zope3-dev04:24
*** alienoid has quit IRC04:25
*** niemeyer has quit IRC04:29
*** natea has quit IRC04:34
*** natea_ has quit IRC04:58
*** zagy has quit IRC05:17
*** zagy has joined #zope3-dev05:19
*** GaryPoster has quit IRC05:28
*** GaryPoster has joined #zope3-dev05:30
*** GaryPoster has quit IRC05:40
*** natea has joined #zope3-dev05:50
*** WebMaven has quit IRC05:54
*** natea has quit IRC06:27
*** natea has joined #zope3-dev06:28
*** MiUlEr has quit IRC07:07
*** dobee has joined #zope3-dev09:01
*** rom|weg is now known as romanofski09:02
romanofskimoin :)09:02
*** MacYET has joined #zope3-dev09:05
*** sashav has joined #zope3-dev09:05
*** sashav has quit IRC09:14
*** natea has quit IRC09:20
*** natea_ has joined #zope3-dev09:20
*** dobee has quit IRC09:23
*** zagy has quit IRC09:44
*** dobee has joined #zope3-dev09:53
*** MJ has quit IRC10:14
*** zagy has joined #zope3-dev10:24
*** bradb has joined #zope3-dev10:24
*** sashav has joined #zope3-dev10:32
*** d2m has quit IRC10:35
*** MJ has joined #zope3-dev10:43
*** sashav has quit IRC10:46
*** sashav has joined #zope3-dev10:46
*** MrTopf has joined #zope3-dev10:49
*** ignas has joined #zope3-dev11:18
*** elro has joined #zope3-dev11:22
*** mgedmin has joined #zope3-dev11:24
*** odie has joined #zope3-dev11:47
*** adamSummers has quit IRC11:52
*** BjornT has joined #zope3-dev11:53
*** xenru|clone has joined #zope3-dev11:53
*** j-w has joined #zope3-dev11:56
*** xenru has quit IRC12:01
*** BjornT_ has quit IRC12:08
*** d2m has joined #zope3-dev12:09
*** MJ has quit IRC12:10
*** MJ has joined #zope3-dev12:17
mgedminsvn.zope.org is down again :(12:22
mgedminBerkeley DB error while checkpointing after Berkeley DB transaction for filesystem /svn/repos/main/db12:22
mgedminDB_INCOMPLETE: Cache flush was unable to complete12:22
*** povbot has joined #zope3-dev12:38
*** agroszer has joined #zope3-dev12:45
*** andrew_m has quit IRC12:52
*** niemeyer has joined #zope3-dev13:08
*** efge has joined #zope3-dev13:13
*** tonico has quit IRC13:21
*** elro has joined #zope3-dev13:22
*** faassen has joined #zope3-dev13:25
*** mgedmin has quit IRC13:33
*** J1m has joined #zope3-dev13:44
*** agroszer has quit IRC13:51
*** mgedmin has joined #zope3-dev13:54
*** MJ has quit IRC13:56
*** MJ has joined #zope3-dev13:57
*** tiredbones has joined #zope3-dev13:57
*** srichter has quit IRC14:11
*** BjornT has joined #zope3-dev14:11
*** philiKON has joined #zope3-dev14:12
MacYETchinakon14:13
philiKONdeutschlandyet14:14
*** mgedmin has quit IRC14:17
*** andrew_m has joined #zope3-dev14:21
*** agroszer has joined #zope3-dev14:23
*** tonico has joined #zope3-dev14:26
philiKONwhat do you guys think about http://www.zope.org/Collectors/Zope3-dev/51614:30
J1mInteresting14:33
J1mTis definately needs to be resolved.14:34
J1mNot entirely sure how. :)14:34
J1mI'm inclined to agree that ids should be strings.14:34
J1mBut then the user folder things should then not allow users to pick ids.14:35
J1mThis would be a good question for the list I think.14:35
philiKONJ1m, ok14:36
*** mkerrin has joined #zope3-dev14:40
*** J1m has quit IRC14:44
*** mgedmin has joined #zope3-dev14:52
philiKONTheuni, ping14:53
Theuniblubb15:01
philiKONTheuni, are you going to merge r40777 to the trunk as well?15:03
philiKON"Fixed bug 486: Made function getSiteManagerDefault compatible with sub-sites."15:03
TheuniphiliKON:  i wanted to, than i saw that the changelog doesn't provide a space to put the log messag ein15:05
philiKONso?15:05
philiKONsrichter will take of changes.txt15:05
Theunii see, so i can just throw it in without taking care of changelog? great.15:06
philiKONyup15:07
Theunii'll check it in then15:22
philiKONgreat15:24
*** srichter has joined #zope3-dev15:35
*** jinty has joined #zope3-dev15:38
*** zbir has joined #zope3-dev15:53
*** ChanServ sets mode: +o srichter15:53
*** _anguenot has joined #zope3-dev16:19
*** efge has quit IRC16:37
*** efge has joined #zope3-dev16:37
*** andres_ has joined #zope3-dev16:42
andrew_mdoes anyone know what the id of the test user mgr/mgrpw is when running a functional test?16:43
*** ruda_porto has joined #zope3-dev16:43
mgedminandres, I do16:43
mgedminit's mgr:mgrpw16:44
mgedminthere's also globalmgr:globalmgrpw, IIRC16:44
mgedminthe first one gets permissions through a local security grant on the root folder16:44
andrew_moh.. i've seen that, so that's the ID i can use for getRolesForPrincipal() for example?16:44
mgedminthis lets you catch mistakes such as missing __parent__ references in your content objects16:44
mgedminthe second has a global security grant16:44
mgedminandres, that was the username and password16:45
mgedminthe principal ID was something different16:45
*** odie has left #zope3-dev16:45
mgedminlook at some ftesting zcml file to find the functional test user's principal ID16:45
mgedminI think it is zope.manager, but I am not sure16:45
*** romanofski has quit IRC16:46
andrew_mahh, ok. i've tried that already. must be something else wrong then. thanks anyway16:46
*** GaryPoster has joined #zope3-dev16:48
philiKONhey GaryPoster16:53
GaryPosterhey philiKON :-)16:54
*** andres has quit IRC17:07
*** xenru|clone has quit IRC17:10
*** xenru has joined #zope3-dev17:10
*** andres has joined #zope3-dev17:27
*** andres_ has quit IRC17:43
*** ignas has quit IRC17:56
*** yota has joined #zope3-dev17:57
*** J1m has joined #zope3-dev18:04
J1msrichter, ping18:04
philiKONJ1m, is r40787 going into 3.2 branch? (zeo fixes with twisted)18:05
* MacYET hands a bell18:05
J1mphiliKON, of course18:07
philiKONok, good18:07
MacYET3.2b2 will released next week?18:07
J1mWe have deeper twisted problems btw18:07
philiKONJ1m, i saw it18:07
philiKONi was just curious18:08
J1mMacYET, srichter and I were thinking this weekend, if possible.18:08
MacYETp18:08
MacYETok18:08
andrew_mphiliKON: any idea if the locking package makes it into the Zope 3.2 release?18:08
MacYETthat was the original plan18:08
J1mMacYET, that doesn't mean anothe rz2 beta has to happen this weekend.18:08
philiKONandrew_m, if it's not in the beta1, it's not going to be in the final either18:08
MacYETj1m. i'll wait 4 u :918:09
J1mandrew_m, if it wasn't in beta 1, it won't be in the final18:09
J1min stereo!18:09
philiKON;)18:09
andrew_myes.. ok, thanks18:10
philiKONandrew_m, but you can vote for it to be in 3.318:10
*** MJ has quit IRC18:10
GaryPosterandrew_m: fwiw, we'll be putting an alternate locking package (zope.locking) as a separate project RSN18:10
andrew_mnah.. think i will extract it and bundle it with our product somehow18:10
philiKONJ1m, btw, i'm +10 on time-based deprecation. i first didn't see the need, but now i've come to realize that it would make our lives wrt Five a lot easier18:10
andrew_mGaryPoster: ok, thanks for letting me know!18:11
MacYETandrew_m: try to corrupt philiKON or jim :)18:11
andrew_mMacYET: haha.. do you know if they like dutch cheese?18:11
J1mTheuni, SteveA and others considering attending a PyCon Zope sprint, I've decoded to go.18:11
J1mdecided18:12
philiKONJ1m, i might be coming, too18:12
J1mawesome!18:12
J1mI hope you can18:12
philiKONi would be working on local five-based components in cmf, though18:12
philiKONno direct z3 stuff18:12
SteveAJ1m: would this be before or after?18:12
philiKONSteveA, sprints are after18:12
J1myup18:12
SteveAokay18:12
J1mI haven't made detailed plans yet, but I expect to attend the conference.18:13
J1mIn part to allow time for pre-sprint brainstorming and planning.18:13
srichterJ1m: pong18:14
J1msrichter, we have a big problem with twisted.18:17
J1mBig big problem18:17
J1mBig big big problem. :)18:18
* J1m waits for srichter to say "Wha?"18:18
*** agroszer has quit IRC18:18
MacYETwha18:19
J1mNo. srichter  has to say "wha". Those are the rules. :)18:19
J1mSigh18:20
J1mOK18:20
J1mWe are now calling application code *before* getting all of the input.18:20
J1mThis means we can tie up application threads waiting for input.18:21
*** dobee has quit IRC18:21
J1mI hope srichter will show me around the thread-management part of the twisted integration.18:21
srichterWhhaaaaaaaaaaaa?18:22
J1mOK18:22
J1mActually, that's:18:22
J1mWaaaaaaaaaaaa!18:22
srichterWaaaaaaaaaaa!18:22
J1malright18:22
srichterhe he18:22
MacYETkindergarten :)18:23
J1mwhere is the thread-dispatching logic?18:23
srichterin the wsgi layer of twisted.web218:23
J1mI was afraid you were going to say that.18:23
TheuniJ1m:  i booked my flights today and i'll be around until march 418:23
J1mTheuni, cool18:24
srichterit basically creates a deferred that is called in a thread18:24
srichterI am sure mkerrin is more knowledgable18:24
J1mOh, there's mkerrin18:24
srichtermkerrin: ping! :-)18:25
Theunibtw: i started sketching out the smtp server integration for the twisted smtp server to go into zope ...18:25
srichtercool18:25
J1mI probably need to bring this up on the web-sig.  I think the wsgi spec is flawed in this regard,18:26
*** agroszer has joined #zope3-dev18:27
J1msrichter, I'm rather worried about making twisted the default for 3.2.18:27
J1mThis stuff isn't adequately baked.18:27
srichterok18:29
srichterI guess extending the generation scripts to choose twisted or ZServer would be good18:29
srichterif we turn off twisted, then noone will look at it again and figure out the problems18:30
srichter(personally I have zero time till Christmas to do anything)18:30
*** j-w has quit IRC18:30
J1mwell, certainly people who care about twisted should be willing to use it for their projects.18:31
J1mGaryPoster, says he's willing to develope with it.18:31
GaryPosterYup.18:31
J1mBut, some flaws like this will only be evident in production.18:31
GaryPosterYup.18:31
J1mMy suggestion is to add an option to mkzopeinstance to choose which server to use.18:32
J1mI can think o a straightforward but long-term problematic way to make this work now.18:33
*** sashav has quit IRC18:33
philiKONJ1m, +1 to mkzopeinstance fix for now18:39
J1msrichter, do you know how the thread limit is set?18:43
*** romanofski has joined #zope3-dev18:46
srichterJ1m: in zope.conf18:47
srichterthere is a call either server does that sets it in the code18:47
mkerrinsrichter, Jim: were you looking for me?18:48
J1mmkerrin, yes18:48
J1mHere's the problem:18:48
J1mWe are now calling application code *before* getting all of the input.18:49
J1mThis means we can tie up application threads waiting for input.18:49
J1mThis is bad18:49
*** andres_ has joined #zope3-dev18:50
J1mmkerrin, are you familiar with the thread-dispatching strategy that twisted wsgi support is using?18:50
*** MJ has joined #zope3-dev18:51
*** stub has quit IRC18:52
mkerrinJ1m: I know it a far bit18:52
J1mcool18:52
J1mDo you understand the problem I'm raising?18:53
mkerrinJ1m: yes and no18:54
J1mapplication threads are expensive resources.18:54
J1mThis is why twisted would rather not use them. :)18:54
J1mAnd why we try to limit their number.18:55
J1mBecause we limit their number, they are also expensive resources.18:55
J1mIt is very important to limit the amount of time a request spends using one.18:55
J1mWaiting around for input is not a good use of the application's time.18:56
J1mIt's what Twisted's select loop is good at.18:56
J1mWe need to rearrange things so that the appplication code doesn't start until all of the input has been gathered.18:57
J1mIn particular, we don't want to enter the app thread.18:57
mkerrinJ1m: Right - so what we need is our own implementation of twisted.web2.wsgi.WSIGResource so that we can do our own threading18:57
J1mOK18:58
J1mCan you briefly explain the strategy that the existing thing uses?18:58
J1mIt looks like it starts a new thread for each request, but I'm puzzled, because I can't see anything to limit the number of threads.18:59
J1mDoes the reactor limit the number of running threads>18:59
J1m?18:59
*** MacYET has left #zope3-dev18:59
mkerrinJ1m: yes - you limit the number of threads in the reactor18:59
mkerrinJ1m: zope.conf should be doing this now18:59
J1mzope.conf is where users type this.19:00
J1mI hadn't figured out what the Python behind that was doing.19:00
J1mI wonder if this shouldn't be handled at the wsgi layer.19:00
mkerrinJ1m: this is the threading issue you are talking about now19:01
J1mI haven't thought this through, but it sort of seems as though the thread-dispatching policy should be implementable as a wsgi middlewear thing.19:01
J1muh, I think so.19:01
J1mwhat other issue might it be?19:02
* J1m reviews wsgi pep19:02
mkerrinSorry - getting a bit caught up in different things here but yes it seems the threading should be done by the application and not in twisted.web2.wsgi19:03
J1mI'm trying to figure out of that is possible.19:04
J1mThe wsgi spec seems unclear on this point, but it seems to expect an application to call start_response before returning.19:06
J1mThe twisted wsgi implementation seems to rely on this.19:07
*** andres has quit IRC19:08
J1mIf this is true, I don't see how threading can be handled in the application.19:08
* J1m reads some more19:08
*** agroszer has quit IRC19:10
TheuniJ1m:  i think we'll deploy twisted as well19:11
*** Aiste has quit IRC19:12
J1mActually, the spec seems to allow start_response to be called after a return.19:16
J1mIt's not clear how far after the return.19:16
J1mmkerrin, are you the author of the twisted wsgi support?19:17
* J1m breaks for lunch19:17
mkerrinNo - srichter should be able to answer that19:18
srichterJames Knight is the author of Web219:19
srichter(which includes the web2 WSGI support)19:20
srichterI forgot his IRC nick though19:20
*** jinty has quit IRC19:20
*** dman13 has joined #zope3-dev19:23
*** Theuni has quit IRC19:24
*** MrTopf has quit IRC19:26
*** Aiste has joined #zope3-dev19:40
J1mI think I'll bring this up on the web-sig and twisted mailing lists.19:48
J1mI don't think I can commit time to fix this for 3.2.19:48
J1mI think I'll focus on making it easy to choose either zserver or twisted.19:48
*** elro has quit IRC20:00
tiredbonesIn my quest to learn zope 3 I pulled in many tutorial and made separate instances of each tutorial. Everything was working find until I just did an svn update.I was at the top of the zope tree when I ran the svn update. Now when I try to runzope, the cPersistence is not imported.I change my PYTHONPATH several times.still no go.DO I have to delete my instances before I updating?20:01
mgedmintiredbones, you may have to run 'make'20:01
mgedminsometimes 'svn up' renames your src/persistent to src/persistent.OLD and checks out a fresh copy20:02
mgedminafter that you need to recompile your extension modules20:02
tiredbonesmgedmin, What directory should I be in to run make? src?20:03
mgedmintiredbones, Zope320:04
tiredbonesthanks!20:04
*** alga has joined #zope3-dev20:20
andrew_mi need to assign a role to a principal like it can be done in etc/principals.zcml, but from python. what object should i use as context? the root folder?20:21
andrew_mah.. i think i was looking for the wrong thing, as usual :/20:26
*** mkerrin has quit IRC20:36
tiredbonesUnder one instance of zope I have no problem running "runzope". However, under another instance I'm getting a dictionary changes size during iteration on file configure.zcml. Can someone give me clue as to what is going on?20:41
*** andres_home has joined #zope3-dev20:42
*** sashav has joined #zope3-dev20:46
*** WebMaven has joined #zope3-dev20:49
tiredbonesWhat is the site's name one can paste some code in?20:49
WebMavenpastebin?20:51
*** faassen has quit IRC20:51
WebMavenrafb.net/paste/20:51
tiredbonesThat's it. thanks20:52
*** sashav has quit IRC20:52
tiredbonesIf someone has some time could you look at this, ha ha - time you say. http://pastebin.com/465576. thanks20:57
*** bradb has quit IRC20:59
WebMaventiredbones: what's this for?21:00
tiredbonesWebMaven, I just did a svn up and I was checking to see if my instances where still working.21:01
tiredbonesthis seem to be something with the system.21:02
*** sashav has joined #zope3-dev21:06
mgedmintiredbones, is that all?21:08
mgedminon second thought, it would be21:08
mgedminzcml helpfully hides the full internal traceback21:08
tiredbonesmgedmin, what did you mean by your last statement?21:10
mgedminnever mind21:13
GaryPosteroh, the sarcasm. :-)21:14
*** BjornT has quit IRC21:15
J1mmgedmin, I wasn't aware that zcml hides any tracebacks.21:17
J1mIt usually shows 2 tracebacks.21:17
J1mThe full python traceback, to the best of it's abilities, and the zcml traceback.21:18
mgedmintiredbones, can you paste the full traceback (all of them)?21:18
tiredbonessure21:19
tiredbonesmgedmin, http://pastebin.com/46560421:23
mgedminfun fun21:25
mgedminbug inside the adapter registry21:25
* mgedmin is not touching that with a big stick21:25
mgedmintiredbones, I suggest you report this to the collector21:25
*** dobee has joined #zope3-dev21:26
J1mThat is a very odd error.21:26
mgedminadapter registry gurus should be able to figure out what is going on from the full traceback21:26
mgedminhopefully...21:26
J1mIf you look at the code, the only way to explain that is if there is a thread somewhere accessing the same data.21:26
J1mWe reassign a value in the dict.21:27
tiredbonesmgedmin, what is the collector and point me something to read on how to do that.21:27
J1mThere is no way that alone can cause this error.21:27
J1mtiredbones, is this during startup?21:28
J1mOr in a test?21:28
tiredbonesFYI, this instance is benji tutorial.21:28
tiredbonesthis is the startup.21:29
J1mSorry, that can't happen. :)21:29
tiredbonesWell, I first tried this instance and got the error, so then I trired the cookery instance and that was ok.21:30
mgedmintiredbones, the collector is http://www.zope.org/Collectors/Zope321:38
tiredbonesthanks, I'll see what I can do.21:39
J1mIf you submit an issue there, you need to give enough information for someone to reproduce the problem.21:39
tiredbonesI get 404 with that http.21:40
GaryPosterhttp://www.zope.org/Collectors/Zope3-dev21:41
tiredbonesHow can one give enough information about a problem one does not understand?21:41
J1mYou can give the steps you took to get the symptoms.21:44
tiredbonesokay:)21:44
* mgedmin can't do anything right today :(21:47
tiredbonesI have to take my son to his ski club and I'll report this to the collector when I come back, I did register thou.21:48
*** fermigier has joined #zope3-dev21:57
*** philiKON has quit IRC22:07
*** _anguenot has quit IRC22:11
*** mgedmin has quit IRC22:17
*** tiredbones has quit IRC22:19
*** BjornT has joined #zope3-dev22:19
*** dobee has quit IRC22:20
*** elro has joined #zope3-dev22:21
*** LaurenceRowe has joined #zope3-dev22:27
*** tiredbones has joined #zope3-dev22:30
*** elro has quit IRC22:39
tiredbonesI'm going to file a report with the collector, is it permissible to paste code there?22:41
*** alga has quit IRC22:53
SteveAbetter to attach it22:53
tiredbonesSteveA, I'm confuse. I was going to fill out their form, how do I attach. Any-Who, I rebooted my system and it works now.23:03
*** efge has quit IRC23:04
*** jinty has joined #zope3-dev23:17
*** LaurenceRowe has quit IRC23:23
*** zbir has quit IRC23:30
*** projekt01 has joined #zope3-dev23:42
projekt01srichter, ayt?23:43
*** romanofski is now known as rom|zzZ23:44
*** elro has joined #zope3-dev23:53
*** WebMaven has quit IRC23:58
« Wednesday, 2005-12-14 Index Friday, 2005-12-16 »

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