IRC log of #zope3-dev for Friday, 2009-11-20

*** sm__ has joined #zope3-dev00:00
*** sm has quit IRC00:00
*** sm__ is now known as sm00:01
*** sm_ has quit IRC00:07
*** romanofski has joined #zope3-dev00:12
*** ChrisW has joined #zope3-dev00:12
ChrisWHi All, anyone know how to run a specific DocFileSuite using zc.recipe.testrunner?00:13
ChrisWtrying to run just the protocol tests for ZEO...00:13
*** Theuni1 has quit IRC00:20
philiKONChrisW: iirc you can select tests using regexp00:25
ChrisWhey stranger :-)00:25
philiKONbin/test -t file.txt00:25
ChrisWwhat brings you here? :-)00:25
philiKONhey ChrisW00:25
philiKONwell, #zope3-dev still is on my connect-upon-login list for irc ;)00:25
ChrisWhow's the physics going?00:26
philiKONgoing well00:26
philiKONstill doing zope?00:26
*** J1m_ has quit IRC00:26
ChrisWyep, but less than before00:27
ChrisWstill use lots of bits of zope (buildout, test runner, component architecture, etc) but not the whole stack00:27
ChrisWalthough still supporting some Zope 2 apps, which is much nicer in Zope 2.12, what with buildout and all00:27
ChrisWyou sitll using python at all?00:27
ChrisW(btw: -t worked, thanks!)00:28
philiKONin my spare time yeah00:28
philiKONi.e. never00:28
ChrisWno zope though, I'm guessing? ;-)00:28
philiKONbeen doing lots of Mathematica and C++. now mostly C++00:29
ChrisWheh, ouch :-D00:29
philiKONMathematica is fun. functional programming FTW00:29
philiKONand well, when i say C++, it's really not much OOP00:30
ChrisWyeah, but C++ ;-)00:30
philiKONwe're using it for numerics00:30
ChrisWand C is even more :-/00:30
philiKONnumber crunching etc.00:30
ChrisWI may have to do some of that soon too00:30
ChrisW*dusts off hard hat*00:30
philiKONsure, some days i want to poke my eyes out00:30
philiKONjust to end it all00:30
philiKONbut usually it isn't C++'s fault00:31
philiKONit's my colleagues who've never heard of version control.00:31
philiKONor that you read code more than you write it00:31
*** benji has quit IRC00:31
ChrisWany unit tests? ;-)00:31
philiKONor that we're actually collaborating so code isn't just written for one person00:31
*** J1m_ has joined #zope3-dev00:32
ChrisWheh :-)00:34
ChrisWhey J1m_, just about to send an "interesting" mail to zodb-dev...00:36
J1m_How so?00:36
ChrisWthe version parameter to the history method00:36
ChrisWthe mail makes it all clear, 1 sec...00:36
ChrisWon its way00:38
*** cbcunc is now known as cbcunc|trizpug00:42
*** hath|away is now known as hathawsh00:44
ChrisWalso wondering about zc.recipe.rhrc vs zc.zodbrecipes:server vs zc.recipe.deployment: so, zc.zodbrecipes:server puts scripts in /etc/init.d, zc.recipe.rhrc does the same, what do I do about the name clash?00:47
*** sm has quit IRC00:52
*** run|vm|winxp has joined #zope3-dev01:01
ChrisWJ1m_: what's the ZEOStorage you're referring to?01:01
*** sm has joined #zope3-dev01:03
*** run|away has quit IRC01:06
*** lurkymclurklet-1 has joined #zope3-dev01:07
*** markusleist has quit IRC01:07
*** ChrisW has left #zope3-dev01:07
*** menesis has joined #zope3-dev01:08
*** markusleist has joined #zope3-dev01:09
*** markusleist has quit IRC01:11
*** markusleist has joined #zope3-dev01:13
*** J1m_ has quit IRC01:21
*** lurkymclurkleton has quit IRC01:23
*** gary_poster has quit IRC01:39
*** sm has joined #zope3-dev02:04
*** davisagli has quit IRC02:13
*** hazmat has quit IRC02:21
*** run|vm|winxp has quit IRC02:47
*** nathany has quit IRC02:47
*** davisagli has joined #zope3-dev02:56
*** run|vm|winxp has joined #zope3-dev03:03
*** drudi has joined #zope3-dev03:06
*** run|vm|winxp| has joined #zope3-dev03:08
*** run|vm|winxp has quit IRC03:08
*** run|vm|winxp has joined #zope3-dev03:09
*** run|vm|winxp| has quit IRC03:10
*** davisagli has quit IRC03:11
*** junkafarian has joined #zope3-dev03:18
*** hathawsh is now known as hath|away03:26
*** drudi has quit IRC03:35
*** davisagli has joined #zope3-dev03:39
*** dunny has joined #zope3-dev03:41
*** davisagli has quit IRC03:53
*** hath|away is now known as hathawsh04:05
*** fcorrea has joined #zope3-dev04:07
*** alga_ has quit IRC04:08
*** pcardune_ has joined #zope3-dev04:14
*** pcardune_ has quit IRC04:15
*** stub has joined #zope3-dev04:16
*** hathawsh is now known as hath|away04:19
*** hath|away is now known as hathawsh04:24
*** pcardune has quit IRC04:31
*** hathawsh is now known as hath|away04:34
*** hath|away is now known as hathawsh04:53
*** hathawsh is now known as hath|away05:09
*** aaronv has quit IRC05:23
*** hath|away is now known as hathawsh05:28
*** pcardune has joined #zope3-dev05:48
*** hathawsh is now known as hath|away06:01
*** matthal has quit IRC06:05
*** hath|away is now known as hathawsh06:23
*** fcorrea has quit IRC06:31
*** davisagli has joined #zope3-dev06:39
*** sm has quit IRC06:48
*** afd___ has joined #zope3-dev07:02
*** jpcw has joined #zope3-dev07:09
*** afd__ has joined #zope3-dev07:20
*** afd___ has quit IRC07:38
*** stub has quit IRC07:41
*** stub has joined #zope3-dev07:42
*** hathawsh is now known as hath|away07:47
*** run|vm|winxp has quit IRC07:52
*** run|vm|winxp has joined #zope3-dev07:53
*** Theuni1 has joined #zope3-dev07:57
*** hath|away is now known as hathawsh08:09
*** afd__ has quit IRC08:20
*** jfkw has quit IRC08:21
*** jukart has joined #zope3-dev08:25
*** pcardune has quit IRC08:44
*** matthal has joined #zope3-dev08:45
*** Theuni1 has quit IRC08:48
*** matthal has quit IRC08:49
*** runyaga has joined #zope3-dev08:50
*** Theuni1 has joined #zope3-dev08:52
*** run|vm|winxp has quit IRC09:02
*** matthal has joined #zope3-dev09:07
*** agroszer has joined #zope3-dev09:17
*** __mac__ has joined #zope3-dev09:19
*** davisagli has quit IRC09:22
*** davisagli has joined #zope3-dev09:23
*** sweh has joined #zope3-dev09:38
*** alecm has quit IRC09:48
*** hathawsh is now known as hath|away09:52
*** jpcw has quit IRC10:02
*** goschtl has joined #zope3-dev10:05
*** matthal has quit IRC10:09
*** alga has joined #zope3-dev10:15
*** jpcw has joined #zope3-dev10:18
*** markusleist has quit IRC10:18
*** hath|away is now known as hathawsh10:31
*** romanofski has quit IRC10:31
*** markusleist has joined #zope3-dev10:35
*** agroszer has quit IRC10:48
*** agroszer has joined #zope3-dev10:48
*** tarek has joined #zope3-dev10:56
*** hathawsh is now known as hath|away11:01
*** hath|away is now known as hathawsh11:06
*** reinout has joined #zope3-dev11:10
*** alga has quit IRC11:18
*** hathawsh is now known as hath|away11:21
*** hath|away is now known as hathawsh11:22
*** junkafarian has quit IRC11:22
*** jhauser has joined #zope3-dev11:42
*** hathawsh is now known as hath|away11:43
*** romanofski has joined #zope3-dev11:55
*** alga has joined #zope3-dev11:58
*** davisagli has left #zope3-dev11:59
*** hath|away is now known as hathawsh11:59
*** junkafarian has joined #zope3-dev12:03
*** junkafarian_ has joined #zope3-dev12:17
*** yvl has joined #zope3-dev12:18
*** tarek has quit IRC12:20
*** philiKON has quit IRC12:25
*** alga has quit IRC12:26
*** junkafarian has quit IRC12:31
*** junkafarian_ is now known as junkafarian12:31
*** philiKON has joined #zope3-dev12:43
*** markusleist has quit IRC13:06
*** jhauser_ has joined #zope3-dev13:36
*** jhauser has quit IRC13:41
*** jhauser_ is now known as jhauser13:41
*** hathawsh is now known as hath|away13:42
*** aaronv has joined #zope3-dev13:49
*** tarek has joined #zope3-dev13:50
*** reinout_ has joined #zope3-dev13:51
*** projekt01 has joined #zope3-dev14:06
*** reinout has quit IRC14:06
*** reinout_ is now known as reinout14:07
*** afd_____ has joined #zope3-dev14:13
*** fcorrea has joined #zope3-dev14:14
*** __mac__ has quit IRC14:14
*** __mac__ has joined #zope3-dev14:15
*** romanofski has quit IRC14:15
*** romanofski has joined #zope3-dev14:16
*** __mac__ has quit IRC14:16
*** dunny has quit IRC14:16
*** __mac__ has joined #zope3-dev14:16
*** afd____ has joined #zope3-dev14:25
*** afd___ has joined #zope3-dev14:26
*** afd_____ has quit IRC14:31
*** romanofski has quit IRC14:43
*** afd____ has quit IRC14:44
*** yotaff has joined #zope3-dev14:48
*** goschtl has quit IRC14:52
*** tarek has quit IRC14:53
*** ignas has quit IRC15:16
*** afd__ has joined #zope3-dev15:45
*** ignas has joined #zope3-dev15:45
*** J1m has joined #zope3-dev15:51
*** tarek has joined #zope3-dev15:51
*** benji has joined #zope3-dev15:56
*** mgedmin has joined #zope3-dev16:00
*** lurkymclurkleton has joined #zope3-dev16:02
*** lurkymclurkleton has joined #zope3-dev16:02
*** afd___ has quit IRC16:03
*** gary_poster has joined #zope3-dev16:11
*** sm has joined #zope3-dev16:16
*** drudi has joined #zope3-dev16:43
*** tisto has joined #zope3-dev16:51
*** __mac__ has quit IRC16:54
*** jfkw has joined #zope3-dev17:10
*** drudi has quit IRC17:13
*** alga has joined #zope3-dev17:19
*** sweh has quit IRC17:26
*** alecm has joined #zope3-dev17:33
mgedminI want something for long-running requests17:36
mgedmina progress indication of some kind...17:36
mgedminoh, wow, a BrokenModified exception17:37
mgedminfirst time I see one17:37
*** aaronv has quit IRC17:38
*** runyaga has quit IRC17:41
*** pyqwer_ has joined #zope3-dev17:51
*** afd__ has quit IRC17:54
*** mcdonc has quit IRC17:54
*** projekt01 has quit IRC17:55
*** pyqwer has quit IRC17:57
*** zagy has quit IRC18:04
*** mcdonc has joined #zope3-dev18:15
ignasmgedmin, hmm, is long running request progress even compatible with wsgi way of doing things?18:15
*** mcdonc has quit IRC18:15
mgedminwho cares18:15
*** mcdonc has joined #zope3-dev18:15
ignaszope does ;) Zope is out of web server business ;) and that is a server level feature I think18:15
ignasat least the implementations I have heard of18:16
*** mcdonc has quit IRC18:16
mgedminwsgi supposedly lets me return response early18:16
ignaslike "hanging requests" + inter request communication18:16
mgedminso I  could print "in progress + a  bit of javascript for async polling of status updates", then go on working18:16
mgedminsharing my progress via the browser session indexed by some job id18:17
mgedmindoes _zope_ support "return response, then go on working"?18:17
mgedminI'm fine tying up one of my 4 request threads for this18:17
mgedminmy use case is sysadminy maintenance tasks where browser timeouts are unpleasant18:18
*** mcdonc has joined #zope3-dev18:19
*** drudi has joined #zope3-dev18:19
*** mcdonc has quit IRC18:26
*** reinout has quit IRC18:29
*** stub has quit IRC18:33
*** davisagli has joined #zope3-dev18:36
*** yvl has quit IRC18:36
*** hazmat has joined #zope3-dev18:41
*** ChanServ sets mode: +o hazmat18:41
*** davisagli has quit IRC18:41
gary_postermgedmin: zc still uses zc.async for stuff like that, fwiw and afaik18:46
gary_posterit's designed for just that sort of thing18:46
*** mcdonc has joined #zope3-dev18:48
*** nathany has joined #zope3-dev18:49
ignasmgedmin, now that I think about it, yes, gary_poster and you are correct, it is doable, pretty easily even18:53
ignasI was thinking about file upload with a progress bar usecase18:53
mgedminhm, zc.async sounds complicated18:53
ignaswhich is more complicated18:53
* mgedmin reading docs18:53
ignasthan spawning an async job, and polling for results using javascript18:54
*** alecm has quit IRC18:54
* gary_poster tried to not make zc.async complicated :-/ . Failed, at some level at least.18:54
ignasgary_poster, by the way - what happens to zc.async with wsgi servers that spawn more than one application?18:55
ignasI thought of using zc.async for "crontab" style stuff some time ago18:56
ignasbut was afraid that if ngnix or apache would spawn more than one wsgi application instance, I might start getting duplicate events, because one "thread" won't know about the other one18:56
ignasso used plain crontab18:56
ignasas I can be pretty sure there only is "one" instance of it running18:57
ignasor is zc.async more suited for "call this long running stuff, tell me when done" kinds of tasks18:57
ignasand not - let's have a thread next to my web application doing something18:58
mgedminafaiu you can have your zc.async dispatcher in a separate process18:59
* mgedmin skimmed the docs very quickly, glazing eyes over the very numerous doctest chunks18:59
ignashmm? so a separate application?19:00
gary_posterignas: it's based on a shared ZODB via ZEO, so it has those standard characteristics (conflict resolution and so forth), and a shared scribbling space.  It's primary use case is dynamic "call this long running stuff, tell me when done" as you said.  I've always thought it would be pretty trivial to build from that into a crontab-like solution.The old Zope 2 predecessor had something like that.  It is superior to crontabs because19:00
gary_postersure that the next run never overlaps with the current one.  You could hack together a poor mans version with just callbacks.19:00
ignasoh, using zodb for syncing19:01
ignasmakes sense then19:01
gary_postermgedmin: you can have the worker threads in the same process.  It shows how to do that.  That's how I used it, though I generally made those processes not part of the load balancing for the main app.19:02
gary_posterthat meant for flexible deployment19:02
gary_posterif you discover you need more or less apps for the long-runnning stuff19:03
mgedmin"flexible" is often the opposite of "easy" ;)19:03
gary_postertoo true19:03
mgedminhaving to edit my zdaemon.conf and allocate new tcp ports doesn't look like something to take on lightly19:04
mgedminfor something that's a glorified debugging view rather than a core feature19:04
gary_postermm, that's if you want to have the monitor port feature, right?19:04
gary_poster...<shrug> sure.  It cam in handy.  We didn't see an easier way for similar functionality19:05
ignasmgedmin, can you commit a zodb transaction early in the process?19:06
mgedmingary_poster, "if"? cool, I assumed it was mandatory19:06
ignasmgedmin, if you want eye candy, you can probably spawn the actual job in an "iframe"19:06
gary_postermgedmin: heh, flexible, remember? ;-)19:06
* mgedmin books a spare afternoon to play with this19:06
ignasmgedmin, and commit "progress" into some variable in zodb19:07
ignasmgedmin, and poll for it using javascript19:07
ignasmgedmin, maybe you could do it using tools "available" to you19:07
mgedminignas, idea19:07
*** drudi has quit IRC19:08
mgedminbut ISTR firefox gave up after a few minutes and terminated the http connection19:08
mgedminwhich I think aborted the view on the server side19:08
mgedminmaybe not19:08
ignashmm, other way of doing it19:08
gary_posterthat's why he was suggesting the JS poll19:08
ignasis "streaming"19:08
ignasi think Zope can stream content directly19:08
ignasso you can return an iterable/stream19:08
ignasthat tracks progress and performs the task19:09
ignaswhile returning some html/javascript19:09
ignasto show progress in a nice way19:09
ignasI mean - people do "large file downloads" somehow in zope don't they?19:09
gary_posterI think for mgedmin's use case a variation of your first approach would be easier19:09
gary_posterscribble progress on some ZODB object19:10
gary_posterhave a simple view that returns value of progress19:10
gary_posterhave JS that polls that19:10
mgedminthat is the easy part19:10
ignasthe hard part is how to make the iframe not get "disconnected"19:11
mgedminforking the original request into "return html with js" and "do background processing with occasional zodb commits" is the difficult bit19:11
ignasmgedmin, forking?19:11
gary_posterwhat I just described is what you can do with zc.async.19:11
ignasmgedmin, why fork?19:11
gary_posterThe forking lets him return the original request back19:12
gary_posterwhile he just works elsewhere19:12
mgedminmaybe I should try "request.response.write(); request.response.close(); do_background_job_now_in_the_same_thread()"19:12
mgedmindoes request.response have write() and close()?19:12
*** pyqwer_ has quit IRC19:12
ignasmy question is - why "return the original request back"19:13
ignasa button that submits a form in an iframe19:13
ignasthat's how we do multiple file uploads for example19:13
* gary_poster forgets. He thinks that the publisher wants to get adapt the output to something or other and handle it itself19:13
ignasso the page you are in stays as it was, while long running thing19:13
ignasis chugging along in some <iframe style="display:none" />19:14
gary_posterignas, right, but when you are waiting for a response, Apache/Firefox/various other things let go after a while, giving up.19:14
gary_posterhe could make the iframe and ignore it I suppose, if the error message is not too obvious19:15
ignasyeah, but if you would "print out a couple of bytes directy to the response" once in a while19:15
gary_postermeh :-)19:15
ignasas if you were slowly serving some large file19:15
gary_postermaybe so19:15
* gary_poster would be going to zc.async, but then, he knows it.19:16
ignaswell - i am proposing a solution I know how I would implement in my pylons application ;)19:16
gary_posterheh, fair enough :-)19:17
ignasmy background is - allowing 500 mb file uploads/downloads without using lots of RAM19:18
gary_posterhm, interesting.  downloads don't seem so bad in Zope land, but uploads of that size might be interesting.19:20
gary_poster(that is, as you said, there is some streaming spelling or other, involving file handles IIRC)19:20
ignasduno, switched from Plone to pylons ;)19:21
ignasbut can recall that you can stream stuff for downloads19:21
ignasdid not fix uploads in our Plone version though19:21
*** tarek has quit IRC19:21
*** tarek has joined #zope3-dev19:22
*** goschtl has joined #zope3-dev19:23
* mgedmin now imagines somebody accidentally switching from Plone to Pylons due to a spelling error19:29
*** davisagli has joined #zope3-dev19:31
*** davisagli has quit IRC19:31
*** davisagli has joined #zope3-dev19:32
*** sm_ has joined #zope3-dev19:41
ignaseasy_install Pylone19:42
*** aaronv has joined #zope3-dev19:45
*** sm has quit IRC19:49
*** sm_ is now known as sm19:49
*** alecm has joined #zope3-dev19:53
*** goschtl has quit IRC20:00
*** hath|away is now known as hathawsh20:06
*** pcardune has joined #zope3-dev20:07
*** jukart has quit IRC20:13
*** ignas has quit IRC20:14
*** pcardune has quit IRC20:15
*** philiKON has quit IRC20:20
*** goschtl has joined #zope3-dev20:31
*** junkafarian has quit IRC20:40
mgedminfriday LATE evening, I must sit here and debug why zc.buildout ignores unzip = true20:49
mgedminbecause the staging server won't start up20:49
mgedminbecause it insists on writing into /home/mgedmin/.python-eggs, because simplejson is a zipped egg20:50
mgedminI hate the world20:50
davisaglimgedmin: set your PYTHON_EGG_CACHE environment variable to make it write somewhere else20:52
mgedminso it'll blow up next time I or the other sysadmin ssh's and does a sudo /path/to/buildout/bin/instance restart?20:53
davisaglimgedmin: also, unzip = true only has an effect at the time an egg is fetched, so if you already have simplejson it won't help20:53
*** agroszer has quit IRC20:53
mgedminrm eggs/simplejson*; bin/buildout; ls eggs/simplejson* -> doesn't unzip20:53
*** goschtl has quit IRC20:54
davisagliI run zope via the script generated by plone.recipe.zope2instance, which lets me configure it to set env vars20:54
davisaglie.g. environment-vars = PYTHON_EGG_CACHE ${buildout:directory}/var/.python-eggs20:55
davisagli(you still have to create that directory manually)20:55
mgedmincreating a ~/.pydistutils.cfg with [easy_install]  zip_ok = 0 works20:56
davisaglihuh, interesting20:57
mgedminsad, more likely21:00
mgedminlooks like
*** mgedmin has quit IRC21:05
*** runyaga has joined #zope3-dev21:12
*** hathawsh is now known as hath|away21:14
*** hath|away is now known as hathawsh21:16
*** jpcw has quit IRC21:19
*** pcardune has joined #zope3-dev21:22
*** agroszer has joined #zope3-dev21:27
*** kaeru has quit IRC21:34
*** tisto has quit IRC21:35
*** kaeru has joined #zope3-dev21:46
*** hathawsh is now known as hath|away21:51
*** hath|away is now known as hathawsh22:03
*** yotaff has quit IRC22:17
*** sm_ has joined #zope3-dev22:28
*** agroszer has quit IRC22:29
*** sm has quit IRC22:39
*** sm_ is now known as sm22:39
*** dunny has joined #zope3-dev23:15
*** mcdonc has quit IRC23:29
*** Theuni1 has quit IRC23:35
*** tarek has quit IRC23:35
*** aaronv has quit IRC23:45
*** tarek has joined #zope3-dev23:48
*** Theuni1 has joined #zope3-dev23:53

Generated by 2.15.1 by Marius Gedminas - find it at!