IRC log of #zope for Friday, 2010-10-22

« Thursday, 2010-10-21 Index Saturday, 2010-10-23 »
*** lucmult has quit IRC00:05
*** Subdino has quit IRC00:19
*** Hypergraphe has quit IRC00:36
*** lucmult has joined #zope00:42
*** smita has quit IRC00:44
*** Hypergraphe has joined #zope00:48
*** vipod has quit IRC01:02
*** gwik has joined #zope01:03
*** thetet has joined #zope01:03
*** gwik has quit IRC01:03
*** redir has quit IRC01:05
*** TresEquis has quit IRC01:06
*** menesis has quit IRC01:07
*** Hypergraphe has quit IRC01:09
*** lucmult has quit IRC01:13
*** Theuni1 has joined #zope01:16
*** dvschramm is now known as dvschramm|away01:17
*** lucmult has joined #zope01:22
*** nitrogenycs has left #zope01:30
*** MatthewWilkes has quit IRC01:39
*** ccomb has quit IRC01:40
*** J1m has quit IRC01:46
*** tiwula has quit IRC01:57
*** yuvilio has quit IRC02:14
*** giampaolo has quit IRC02:18
*** dayne has quit IRC02:25
*** Theuni1 has quit IRC02:56
*** redir has joined #zope03:01
*** alga has quit IRC03:05
*** redir has quit IRC03:07
*** davisagli is now known as davisagli|away03:18
*** daMaestro has quit IRC03:23
*** mcdonc has quit IRC03:33
*** thetet has quit IRC04:04
*** davisagli|away is now known as davisagli04:08
*** alvaro_o_ has quit IRC04:10
*** davisagli is now known as davisagli|away04:45
*** sm has quit IRC04:51
*** Gogo|tty has quit IRC04:51
*** Gogo|tty has joined #zope04:52
*** yuvilio has joined #zope04:56
*** sm has joined #zope04:56
*** acsr1 has joined #zope05:22
*** acsr has quit IRC05:25
*** yuvilio has quit IRC05:33
*** lucmult has joined #zope05:42
*** redir has joined #zope05:52
*** lucmult has quit IRC05:55
*** benji has quit IRC06:32
*** kursor has joined #zope06:56
CIA-38shane 1.4 * r117807 relstorage/ (CHANGES.txt relstorage/adapters/oracle.py setup.py):06:58
CIA-38Oracle: always connect in threaded mode. Without threaded mode,06:58
CIA-38clients of Oracle 11g sometimes segfault.06:58
CIA-38shane * r117808 relstorage/ (CHANGES.txt relstorage/adapters/oracle.py README.txt):06:58
CIA-38Oracle: always connect in threaded mode. Without threaded mode,06:58
CIA-38clients of Oracle 11g sometimes segfault.06:58
CIA-38Also added Oracle 11g to the list of supported databases. :-)06:58
CIA-38shane * r117809 /relstorage/tags/1.4.1: Tagged release 1.4.106:58
CIA-38shane * r117810 relstorage/relstorage/tests/README.txt: A couple more pg_hba.conf entries to make test setup easier.06:58
CIA-38shane * r117811 relstorage/relstorage/blobhelper.py: ZODB 3.8 compatibility06:58
CIA-38shane * r117812 relstorage/relstorage/blobhelper.py: Added note about copied code.06:58
*** davisagli|away is now known as davisagli07:02
*** davisagli is now known as davisagli|away07:04
*** TheJester has quit IRC07:40
*** davisagli|away is now known as davisagli07:40
*** davisagli is now known as davisagli|away07:54
*** febb has joined #zope07:54
*** davisagli|away is now known as davisagli07:55
*** kursor has quit IRC08:26
*** febb has quit IRC08:41
*** JaRoel|4D has quit IRC08:48
*** davisagli is now known as davisagli|away08:51
*** zagy has joined #zope08:52
CIA-38shane * r117813 relstorage/ (CHANGES.txt setup.py): Tagged 1.5.0a109:10
CIA-38shane * r117814 /relstorage/tags/1.5.0a1: Tagged 1.5.0a109:10
CIA-38shane * r117815 relstorage/setup.py: Back to dev09:10
CIA-38shane 1.5.0a1 * r117816 relstorage/setup.py: pedantic version number09:10
*** vipod has joined #zope09:16
*** smita has joined #zope09:23
*** jakke has joined #zope09:24
*** sashav has joined #zope09:26
*** d2m has joined #zope09:28
*** sm has quit IRC09:33
*** sm has joined #zope09:34
*** Theuni1 has joined #zope09:37
*** JaRoel|4D has joined #zope09:43
*** eperez has joined #zope10:03
*** menesis has joined #zope10:12
*** sylvain has joined #zope10:13
*** smita1 has joined #zope10:15
*** smita has quit IRC10:18
*** cna has joined #zope10:29
*** planetzopebot has quit IRC10:33
*** planetzopebot has joined #zope10:34
*** vipod has quit IRC10:47
*** vipod has joined #zope10:47
*** MJ has joined #zope10:49
*** sunew has joined #zope11:07
*** neo|4D has joined #zope11:30
*** alga has joined #zope11:32
*** nitrogenycs has joined #zope11:36
*** mitchell`off is now known as mitchell`11:47
*** kursor has joined #zope11:58
*** ccomb has joined #zope12:03
*** lucmult has joined #zope12:20
*** cna has quit IRC12:21
*** cna has joined #zope12:21
*** teix has joined #zope12:23
*** gwik has joined #zope12:24
*** evilbungle has joined #zope12:25
*** lucmult has quit IRC12:27
*** sashav has quit IRC12:31
*** milosn has quit IRC12:37
*** thetet has joined #zope12:47
*** alexpilz has joined #zope12:48
*** lucmult has joined #zope12:48
*** alexpilz has quit IRC12:50
*** alexpilz has joined #zope12:50
*** menesis has quit IRC13:02
*** hartym has joined #zope13:06
*** hartym has joined #zope13:06
*** fredvd has quit IRC13:07
*** milosn has joined #zope13:14
*** lucmult has quit IRC13:19
*** astoon has joined #zope13:21
*** astoon_ has joined #zope13:25
*** ccomb has quit IRC13:28
*** ccomb has joined #zope13:33
*** lucmult has joined #zope13:38
*** lucmult has quit IRC13:46
*** gwik has quit IRC13:53
*** gwik has joined #zope13:54
*** zagy has quit IRC13:56
*** zagy has joined #zope13:57
*** ccomb1 has joined #zope14:05
*** ccomb has quit IRC14:09
*** mr_jolly has joined #zope14:12
*** thetet has quit IRC14:15
*** mr_jolly has left #zope14:27
*** dayne has joined #zope14:28
*** giampaolo has joined #zope14:28
*** menesis has joined #zope14:31
CIA-38zagy * r117817 gocept.selenium/ (src/gocept/selenium/selenese.py CHANGES.txt): - API expandsion: ``getSelectOptions``14:34
*** lucmult has joined #zope14:38
*** lucmult has quit IRC14:45
*** yuvilio has joined #zope14:54
CIA-38zagy * r117818 gocept.selenium/ (CHANGES.txt setup.py): Preparing release 0.814:56
CIA-38zagy * r117819 /gocept.selenium/tags/0.8: Tagging 0.814:56
CIA-38zagy * r117820 gocept.selenium/ (CHANGES.txt setup.py): Back to development: 0.914:56
*** nitrogenycs has quit IRC14:56
*** lucmult has joined #zope14:59
*** nitrogenycs has joined #zope14:59
*** hever has joined #zope15:02
*** regebro has joined #zope15:06
*** lucmult has quit IRC15:15
*** benji has joined #zope15:50
*** neo|4D has quit IRC15:57
*** hever has quit IRC16:00
*** alexpilz1 has joined #zope16:07
*** alexpilz has quit IRC16:10
*** TheJester has joined #zope16:15
jakkeHi channel, another zope/grok question16:17
jakkeI have an adapter, defined in my main application16:17
jakkethis adapter is supposed to provide a custom error_view for http errors16:18
jakkebut this code is never called16:18
betabugzope/grok??16:18
d2mjakke: which error?16:19
betabugdidn't you talk about Zope 2 the last time?16:19
d2mheh16:19
betabugZope 2 might be called "stone age", but it's not grok16:19
jakkeit looks to me that the subscriber for the IObjectCreatedEvent is not the right subscriber to trigger the code16:19
d2mjakke: http error16:20
jakkeit's custom code; so I don't really know what's really zope2; what's zope five; and what's custom; sorry :s16:20
jakkeI want a nice error page for a 40416:20
betabugzope five != grok, either16:20
jakkemeanwhile I'm coming to the point where I know which code should be called16:20
d2mhow would an object creation throw a 404 ?16:20
*** cna has quit IRC16:21
betabuggood question16:21
jakkethat's what surprises me too16:21
jakkeI think the subscription (done by grok) is not completely what it should be16:21
d2mjakke: what are you trying to do?16:21
jakkeI have a traverser, trying to access an object in the zodb16:21
jakkebut if the object does not exist, I get a standard ZPublisher error page16:22
*** agroszer has joined #zope16:22
jakkethis should be a nice-in theme- error page16:22
jakkethere is code to do that; basically triggered by a NotFound error16:22
jakkeso i updated the code for the errors I want to be covered, that a NotFound exception is raised16:22
jakkebut the code that should be triggered; is bypassed, and ZPublisher comes into play right away16:23
jakkeso I'm missing the hook between the NotFound error and the custom code for the view16:23
*** alexpilz1 has quit IRC16:23
jakkebut I'm kind of lost with the grok registration and what covers what16:24
jakkeso I'm looking for a way to debug the registration and triggering path16:24
betabugI'd still say you should simply customize standard_error_message16:24
betabugor else set the status to a 404 and return whatever html you want16:25
jakkethat is edited; but it has a grok.subscribe decorator; don't know why16:25
d2mjakke: 'it' is a five.grok view?16:27
jakkeyes; from five import grok; and class ErrorView(grok.View); where I set the context(myApp); and grok.require('zope2.View')16:29
*** alexpilz has joined #zope16:30
d2mjakke: how about grok.context(INotFound) ?16:30
giampaolommm maybe I'm crazy, but... is Zope server concurrent? I've just tried to call a view with a "sleep(100)" call in it and the server does not serve any other client as long as time.sleep() returns16:30
betabuggiampaolo: in a default install there are 4 threads16:30
giampaolobetabug: mmm what does that mean exactly?16:31
betabugthat 4 requests can be answered concurently16:31
*** eperez has quit IRC16:31
betabugso if 4 requests hit a sleep(100)", the 5th that comes along will hang16:31
Theuni1betabug: or at least started processing16:31
jakked2m: nope; don't get into the custom code16:32
betabugyeah, well, I'm probably symplifying16:32
Theuni1:)16:32
giampaolowell, I'm only making one request (well, two, considering the other browser window which attempts to connect)16:32
Theuni1just to make sure no wrong understanding comes up. that's one of the most misunderstood things about threads I meet16:32
d2mjakke: create your own error handling classes and register them for the interface you need16:32
*** astoon has quit IRC16:33
Theuni1people tend to think their system will automatically answer more requests when they crank up the threads to tens or even more16:33
jakkeI get from my traverser : __getitem__ into ZPublisher.BaseRequest - publishTraverse; and so further up16:33
d2mjakke: here is an example in grok context: http://svn.zope.org/grokapps/gbepastebin/src/gbepastebin/error.py?view=auto16:33
betabugTheuni1: yeah, the db connections are limited too16:33
Theuni1well16:33
Theuni1not with a hard limit anymore16:33
giampaolooh gosh... my Zope can't handle more than 1 request at time... this is not encouraging at all16:33
Theuni1giampaolo: how many CPUs do you have?16:33
jakked2m; that's what I think is strange: there are classes to handle it; they are suppose to be registered; but the code doesn't get called16:33
betabugthere are also people who reduce that to 1 thread and run more instances on one ZEO server16:34
* Theuni1 does.16:34
jakkesomewhere in the chain something is wrong, and I haven't found the way to debug it16:34
giampaoloTheuni1: two, but does that make any difference since Zope uses threads?16:34
Theuni1I do keep few instances with 10 threads for serving small static stuff16:34
Theuni1giampaolo: Python can only run on 1 CPU effectively.16:34
Theuni1And if 1 CPU does multiple things, well guess what, those things will each run slower.16:34
giampaoloTheuni1: I know, but here it seems that ONE single long-running requests blocks everything16:35
giampaoloit's unacceptable16:35
jakked2m: thanks; might be the wrong NotFound I'm using; I'll move on the the one from zope.publisher!16:35
betabugbut if you have >1 instances, they can get a CPU each :-)16:35
Theuni1giampaolo: right. that's the trade-off16:35
betabuggiampaolo: hmmm, I still don't think this is happening in a normal install16:35
giampaolowhat trade-off? if Zope can only handle 1 request per time than why is it using threads in the first place? =)16:35
d2mjakke: here is a more recent example in five.grok https://svn.infrae.com/infrae.wsgi/trunk/src/infrae/wsgi/errors.py16:36
betabugI have some pretty long running processes and other requests get done while that runs16:36
*** morphex has joined #zope16:36
bigkevmcdgiampaolo: have a look at Launchpad.net, it's running Zope, and is quite capable of handling more than one request at a time...16:36
betabuggiampaolo: check how your zope is set up, in zope 2 this is in INSTANCE_HOME/etc/zope.conf16:36
giampaolobigkevmcd: my point is that I don't get what should I expect from zope if I put a "time.sleep(10)" in my view16:37
bigkevmcdgiampaolo: why would you do that?16:37
betabuggiampaolo: my point is that I don't see that behaviour, so something is wrong16:37
bigkevmcdgiampaolo: what are you attempting to do?16:37
Theuni1giampaolo: no, you can get unblocked with more threads, however, if you have X threads and X requests hang during processing then it will be blocked16:38
giampaoloI have an admin iterface in my site which updates the db so when I press "update db" it does a lot of long running stuff. It might take 5-10 minutes to complete16:38
bigkevmcdgiampaolo: if it's some sort of primitive job system, then you can enqueue a job, and have the browser poll for the result of the job16:38
giampaoloor even more16:38
regebroYeah, some batching or polling would be better there.16:38
betabugor put that into a 2nd instance that only the admin has access to16:39
regebroOr somebody will push the button ten times, and all your threads will be blocked and they will call you and ask why the server crashed. :)16:39
betabugmuch less work for custom programming with ZEO + a 2nd instance16:39
giampaoloI can fix this issue by running this task in a separate process or thread manually started by me, but I was just surprised that a simple "time.sleep()" and ONE request freezes everything16:40
betabuggiampaolo: no it does not16:40
giampaoloreally, it seems this stuff is NOT concurrent16:40
d2mgiampaolo: http is not concurrent16:40
*** astoon has joined #zope16:40
giampaolod2m: I meant zope server16:41
betabugwell, as I said, I've got some pretty long running processes and Zope uses those threads concurrently16:41
d2mgiampaolo: which is a http server16:41
betabugyou can even get to the control panel and see those other requests16:41
giampaolod2m: I've heard of concurrent http server =)16:41
d2mthats what is called 'async'16:41
giampaoloservers*16:41
betabuggiampaolo: maybe you have only one db connection and *that* is blocking? what db is it that you're using there?16:42
giampaolobetabug: no, it's not the db which is blocking16:42
giampaoloI'm sure of that16:42
betabughow do you know?16:42
giampaolobecause if I put a simple sleep() in my view everything gets frozen16:42
giampaoloso no db is implied16:42
giampaolod2m: concurrent != async16:43
d2mnow, why would you want a sleep in a view? it is rendered to html and returned all at once16:43
jakked2m: I changed the subscription to NotFound; but the error page doesn't even know the kind of error anymore16:43
jakkeso registration is back to the zope application16:44
d2mjakke: INotFound16:44
jakkedoesn't help16:44
d2mfrom zope.publisher.interfaces import INotFound16:44
giampaolod2m: sleep() is just a replacement for any long-running task you can imagine. Replace sleep() with db.long_running_task() if you want16:44
jakkeI use zExceptions NotFound; and that is recognised as a NotFound error in the default error page16:45
betabugwell, it still doesn't reflect my experience, I'd guess the instance is set to 1 thread only16:45
jakkeit's "just" not picked up by the adapter16:45
*** goschtl has joined #zope16:45
giampaolobetabug: if you run 1 thread only then there's no reason for Zope to use threads at all. It could have sticked with an async approach (select(), poll(), anything) resulting in much better performances16:46
betabugsigh16:49
betabugI don't run 1 thread only16:49
betabugI'm suspecting that your instance is running on only 1 thread16:49
giampaolooh sorry16:49
betabugbecause (as I said *multiple* times already) I don't see the described behaviour here16:50
betabugand I've been sticking sleep() into test code a long time before16:50
betabugplus I've really long running processes and they run in parallel with other, normal requests, no problem16:51
giampaoloit seems there's nothing in my zope.conf which suggests I'm running on 1 thread16:51
betabugwhich zope version is it?16:52
giampaoloI'm using grok 1.1, not sure what version of Zope it uses, hang on16:53
betabugaha16:53
d2manother grok refugee16:53
betabugis there a channel for grok?16:54
betabug(just wondering)16:54
d2myes, #grok16:54
jakkelol16:54
jakkemaybe I should move there ;-)16:54
giampaololol you think it's grok?16:54
d2mmaybe, not been there in a while16:54
betabugwell, searching the web for [grok number of threads] seems to point to some wsgi setup things16:54
d2mgiampaolo: grow is wsgi based, which wsgi server do you use?16:54
d2ms/grow/grok16:55
giampaolommm zope.app.wsgi?16:55
d2mif you run on wsgiref -- that is single threaded16:55
* betabug wonders why it says "please mention zope version when asking questions" in the /topic16:55
d2mhehe16:56
*** hever has joined #zope16:56
betabugbut noooo, it must be zope's fault, zope is only handling a single request and for the last 10-15 years nobody noticed16:56
d2mbetabug: anyway, other zope related channels (except #repoze) are pretty low noise16:57
betabugsure, the noise is here :-D16:57
d2m:)16:57
betabugthere's a party in #repoze???16:57
d2mare the day16:57
d2mall the day16:57
betabugI was considering doing a fun project with repoze, just for the heck of it16:58
*** kursor has quit IRC16:58
*** Theuni1 has quit IRC17:01
*** alexpilz has quit IRC17:01
*** pepeu has joined #zope17:02
*** pepeu has left #zope17:02
*** sm has joined #zope17:03
*** benji has quit IRC17:08
*** benji has joined #zope17:15
*** benji has quit IRC17:19
*** benji has joined #zope17:20
*** J1m has joined #zope17:20
*** sunew has quit IRC17:23
*** d2m has left #zope17:23
*** d2m has joined #zope17:24
*** hever has quit IRC17:26
*** nitrogenycs has quit IRC17:27
*** f10w has quit IRC17:28
*** f10w has joined #zope17:30
*** dayne has quit IRC17:39
*** sm_ has joined #zope17:41
*** sm has quit IRC17:44
*** sm_ is now known as sm17:44
*** astoon has quit IRC17:50
*** MJ has quit IRC17:53
*** Theuni1 has joined #zope17:57
*** zagy has quit IRC18:02
*** dayne has joined #zope18:02
*** jakke has left #zope18:05
*** runyaga has joined #zope18:06
*** evilbungle_ has joined #zope18:11
*** evilbungle has quit IRC18:11
*** davisagli|away is now known as davisagli18:12
*** davisagli is now known as davisagli|away18:14
*** evilbungle_ has quit IRC18:15
*** nitrogenycs has joined #zope18:18
*** rnix has quit IRC18:18
CIA-38charlie_x 2.2 * r117821 Products.CMFDefault/Products/CMFDefault/browser/ (folder.py templates/folder_contents.pt): Backported changes from 2.3 regarding redirection after completion and CSS class for striping rows.18:20
CIA-38charlie_x 2.2 * r117821 Products.CMFDefault/Products/CMFDefault/browser/ (folder.py templates/folder_contents.pt): Backported changes from 2.3 regarding redirection after completion and CSS class for striping rows.18:33
CIA-38charlie_x 2.2 * r117822 Products.CMFDefault/Products/CMFDefault/browser/ursa.py: Backported correction of mime-type setting from 2.318:33
*** Theuni1 has quit IRC18:37
*** davisagli|away is now known as davisagli18:38
*** kursor has joined #zope18:47
*** goschtl has quit IRC18:53
*** yvl has quit IRC18:54
*** alvaro_o has joined #zope18:54
*** tiwula has joined #zope19:04
*** agroszer has quit IRC19:06
*** ccomb1 has quit IRC19:07
*** TresEquis has joined #zope19:22
*** mitchell` is now known as mitchell`off19:29
*** menesis has quit IRC19:36
*** benji is now known as benji-lunch19:46
*** runyaga has quit IRC19:47
*** sylvain has quit IRC19:49
*** kursor has quit IRC19:49
*** gwik_ has joined #zope19:53
*** daMaestro has joined #zope19:54
*** daMaestro has joined #zope19:54
*** gwik has quit IRC19:56
*** sunew has joined #zope19:57
*** yakoi has joined #zope19:59
*** yakoi has left #zope19:59
*** kleist has joined #zope20:01
*** cbcunc has quit IRC20:04
*** alexpilz has joined #zope20:04
*** agroszer has joined #zope20:14
*** gwik has joined #zope20:29
*** zagy has joined #zope20:33
*** benji-lunch is now known as benji20:43
*** sunew has quit IRC20:44
*** hever has joined #zope20:49
*** kursor has joined #zope21:01
*** kleist has quit IRC21:04
*** menesis has joined #zope21:17
*** redir has quit IRC21:19
*** hever has quit IRC21:20
*** sunew has joined #zope21:26
*** agroszer has quit IRC21:27
*** sunew has quit IRC21:41
*** zagy has quit IRC21:50
*** ccomb has joined #zope21:52
*** menesis has quit IRC21:57
*** zagy has joined #zope22:11
*** zagy has quit IRC22:16
*** chaoflow has quit IRC23:06
*** chaoflow has joined #zope23:06
*** teix has left #zope23:15
*** JaRoel|4D has quit IRC23:17
*** lynucs has joined #zope23:29
*** lynucs is now known as jham23:29
*** Subdino has joined #zope23:45
*** nitrogenycs has quit IRC23:48
*** jham has quit IRC23:57
« Thursday, 2010-10-21 Index Saturday, 2010-10-23 »

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