IRC log of #zope3-dev for Friday, 2006-09-22

floxharobed: about  Catalog, u can look the zope/app/catalog/browser/README.TXT00:00
WebMavenedgordon: it's the difference between an 'order' and a 'special order'.00:00
xenruharobed: from my point of view because they not first on this market00:00
floxit explains how, zope.index and are mixed together00:00
xenruand they come for more money on Java market00:00
edgordonyeah, i have them both. but i just think it is important for people who are only casually interested for them to be sitting on the shelf00:00
xenruharobed: zope3 is perfect product, but last years you need to belive in it, and only last year use00:02
harobedxenru, your are nexuo ingenior ?00:03
WebMavenedgordon: here is something sneaky: Order the book, look at it when it comes in, then don't buy it. It then goes on the shelf.00:04
edgordonoh, right00:04
philiKONWebMaven, no it doesn't00:04
philiKONat least not in germany00:04
philiKONhere it would go back to the publisher00:04
xenruharobed: I havn't any relations with Nuxeo or their developers00:05
WebMavenHere in the states, it depends on the chain, and what sort of distinction they make between an 'order' and a 'special order'.00:05
WebMavenA store has some discretion as to what they stock. if a title is one that a chain does stock, but (by chance or design) not in your local store, it will stay in the store.00:06
philiKONright, ok00:07
WebMavenusually, this has something to do with whether it is kept by the distributor in their warehouses.00:07
*** olarcheveque has joined #zope3-dev00:08
xenruAiste: hi00:08
xenruAiste: r u here?00:08
Aistexenru: yes, hya00:09
harobedwhy nuxeo remplace ZODB by JCR ?00:09
WebMavenharobed: their rationale is pretty clearly stated in their announcement.00:14
harobedok, it's because ZODB can't work with more 3T of data00:15
olarchevequeI found irc log about zope3 and AMF... where I can find some doc about it?00:15
WebMavenUmm, no.00:15
philiKONharobed, if i had 3T of data, i wouldn't use the zodb, but i would still use zope perhaps00:15
rockynuxeo fell under industry pressure imho00:16
WebMavenharobed: and that's not what they said.00:16
WebMavenolarcheveque: AMF?00:17
olarchevequeprotocol to link with flash00:18
philiKONrocky, something like that00:18
WebMavenrocky: Seems more like customer pressure.00:19
rockyin this case i believe they're one in the same00:19
WebMavenrocky: OK.00:20
WebMavenWhat's really odd is that the rationale they're using made more sense 2 years or so ago.00:21
WebMavenThese days, more integration efforts are centering on REST-like SOAs.00:21
WebMavenI know of one company here in the US that came under this very same kind of pressure from their customers a few years ago, and they made the switch.00:23
WebMavenBut their productivity suffered so much, they switched back to Zope.00:23
*** dobee has quit IRC00:24
WebMaventhey just told their customers that if they wanted the rapid turnaround and flexibility that they were providing, only Zope would cut it.00:24
WebMavenSo now they have a very good answer to the question "Why don't you swith to Java or .NET?"00:25
harobedwhat is it ?00:26
philiKONharobed, WebMaven just explained00:27
*** gumpa has quit IRC00:27
philiKONread it again :)00:27
*** gumpa has joined #zope3-dev00:28
harobedfor my case, I work in small web agency. Currently, our developpement are in PHP. Next, what is the best choise : Java or Zope 3 ?00:29
philiKONharobed, what do you expect us to say00:29
harobed:) Zope00:29
edgordonyeah, sort of the wrong place to ask that00:30
*** alecm has quit IRC00:31
*** J1m has quit IRC00:31
*** benji has quit IRC00:31
*** niemeyer has quit IRC00:31
*** srichter has quit IRC00:31
*** jinty has quit IRC00:31
*** WebMaven has quit IRC00:31
*** srichter has joined #zope3-dev00:31
*** alecm has joined #zope3-dev00:31
*** J1m has joined #zope3-dev00:31
*** benji has joined #zope3-dev00:31
*** niemeyer has joined #zope3-dev00:31
*** jinty has joined #zope3-dev00:31
*** WebMaven has joined #zope3-dev00:31
*** sets mode: +o srichter00:31
floxharobed: do zope, even if u do not trust the zodb00:31
harobedbut size of my data isn't big00:31
floxi hv some worries about slowness of ZODB....00:32
harobedflox, and ?00:32
floxwhen ZODB becomes large00:32
harobedwhat is your solution ?00:32
floxucan choose a relational DB, as well00:32
harobedswitch is easy ?00:33
floxPostgreSQL, or other open source or commercial packages00:33
xenruIf you have so mush data then you can use other storages00:33
floxharobed: sqlalchemy should help00:33
floxbut i did not try myself (yet)00:33
xenrulaunchpad ppl use it00:34
floxu heard about "Object-Relational Mapping"?00:34
floxi read a good article, some time ago, where someone explain his experience with Zope and SQLAlchemy00:35
floxsuccessful experience, of course00:35
floxharobed: See comment #16 on this blog post00:36
edgordonharobed, i was moving from zope 2, but my first z3 project used sqlos and I feel it went pretty well.00:38
harobedflox, yes, I know this post00:41
xenruedgordon: can u gimme some advice about sqlos? I don't know much about this project may be links or a little description?00:41
floxergordon: i guess that there's 2 leader ORM package for Zope00:42
floxeither ('sqlos' + SQLObject) or ('z3c.zalchemy' + SQLAlchemy)00:42
*** olarcheveque has quit IRC00:42
floxit lacks a feature/performance comparison....00:43
edgordonyeah, i can't remember what i read that scared me off from sqlalchemy00:43
edgordonthe SIP app is a pretty good starting point for a sqlos app00:43
edgordonreading through that really helped me figure stuff out00:44
*** volvox has joined #zope3-dev00:46
xenruflox: do you know how they store objects in database?00:47
edgordonbye all00:47
*** edgordon has quit IRC00:47
xenruas table with many rows or as one pickled object?00:47
floxxenru: i did not try yet...00:47
floxi just read that information, and i prepare to use it, as soon as i hv time for that00:48
*** oferw has joined #zope3-dev00:48
xenruok, thx anyway00:48
xenruAFAIK, is on zope00:48
xenruis this true?00:49
xenruwhat you use as storage?00:49
oferwxenru: hi00:49
xenruoferw: hi00:49
oferwxenru: you looking for z3 examples?00:50
floxJ1m: not for media, i guess?00:50
J1mfor stories, not images00:50
floxon our project, we store SVG and JPEG in the ZODB....00:51
floxi guess it is bad design00:51
J1mit isn't00:51
J1mwe have a number of customers in which we store images.00:51
xenruno, I'm looking for possible solutions for our project here in moscow, it will be something like and youtube00:51
J1mreally big files are problematic.00:51
J1mbut images aren't that big.00:51
floxJ1m: true00:52
oferwxenru: you know what youtube is based on?00:52
floxJ1m: we have done a load test, last week...00:52
floxnot many peaople connected, only 40...00:52
floxbut some pages took very long time to load00:52
oferwflox: what is you site about?00:52
xenruoferw: no, because we need only to know how them process videos00:52
floxmore that 1 min (on LAN)00:52
J1mk, make your app faster.00:53
xenruflox: do use cache?00:53
floxoferw: this is an intranet website, with documentation00:53
floxand with something like 'zc..comment'00:53
floxto put comment on documents00:53
floxxenru: no cachecurrently00:54
xenruand check are your server is in debug mode00:55
floxdevmode, u mean?00:55
floxyes, it was in devmode, i think00:55
floxbut i think that devmode is only for APIDOC thing?00:55
floxi am wrong?00:56
xenrunot only00:56
xenrumay be J1m can explain you00:56
floxok, and i should disable logging, too, next time00:56
xenruswithch off, this give you speedup00:56
floxit was first load test. to have an idea how it behaves with  many people connected00:56
floxnext test will be more clever00:57
oferwflox: all the site is based on z3?00:57
floxxenru, u suggest an external cache like Squid? or to use zope3 things?00:57
xenruwe using Apache in all previous projects, but now I want to test squid00:58
J1mIf your pages are taking a minute, there is probably something wrong with the app.00:58
xenruApache taking about 95% of all requests00:58
J1mOr your hardware is very slow.00:58
J1mor both00:58
oferwJ1m: or your pages with too much information :-)00:58
floxhardware was good enough (2Go RAM, and dual CPU, at least)00:59
floxbut on Windows 2003 Server (maybe the problem)00:59
xenruhow many memory is free?00:59
floxaround 900 Mo free during the test, AFAIR01:00
floxZOPE took around 300-400 MB all the time01:00
floxslightly increasing during the 1 hour test01:00
xenruand you can add zodb cache and add zope threads01:00
*** volvox has quit IRC01:00
floxzope threads = ZEO01:00
floxJ1m: maybe the app01:01
floxwe do lots of thing with xml.minidom and with PrincipalAnnotations01:01
xenrumay be I missing something, but zope2 use threads, don't know how it is now with twisted01:02
*** benji has quit IRC01:02
projekt01xenru, did you see z3c.extfile01:02
floxlike store a "browsing history" per  user as PrincipalAnnotation01:02
xenruprojekt01: not yet, looking01:03
projekt01Jim, I heard about using ZODB without a history? Is htis possible and if yes how?01:03
*** MJ is now known as MJ|sleep01:04
projekt01xenru, it stores files in a separate filestorage and gives you the hash of the file back.01:04
J1mprojekt01: there have been storages that that didn't keep old revisions.01:04
J1mI don't know if there are any now.01:04
xenruprojekt01: many thanks, this is very useful information01:05
projekt01J1m, Ok, what is your recommended way to implement a counter for a heavy visited site?01:05
J1mI don't have one.01:06
*** robcampbell has joined #zope3-dev01:06
xenruprojekt01: we implementigh it as separate server with just C++ http server with a little functionality that only write a log files01:06
floxprojekt01: maybe u subscribe on an event (like EndRequestEvent) and  u increment some int attribute on the ISiteManager01:07
xenruand after that we analyse this log files01:07
projekt01xenru, I guess this is a good idea01:08
projekt01flox, we hav to avoide any kind of write transaction for this usecase.01:08
floxi see ;-)01:09
floxi hae DBConflicts, too01:09
projekt01Does anybody know if there is open source counter which can be integrated, perhaps python based01:10
projekt01with integrated I mean running as a standalone app and reading counts by a interval01:10
*** gumpa has left #zope3-dev01:11
* xenru hope we open one day our01:11
projekt01xenru, what do you think about a to read the apache log and count there some urls?01:12
xenruyou can use webalizer or something commercial, or google analitics01:12
projekt01hm, or write some own external app like you did.01:13
*** J1m has quit IRC01:13
xenruwe analyse many aspects of user behavior, even what links on page press user and ajax hits01:14
projekt01xenru, why did you not ue the Apache log files and write a own server? Any important reason?01:15
xenruincluding from what places come user, what search engines they use, what they do after they come and when return01:15
projekt01is the reason to have a custom format and more info?01:15
xenrudo you analyse how works google analitics or commercial services?01:16
projekt01I just need a counter whihc can count hits.01:17
xenruthen apache logs should be enough01:17
projekt01ok, I guess I give it a try and if it's working I will add a z3c.couter package01:19
xenrubecuse we giving session and "persistent" cookies for visitors and analyse them for a long period01:19
*** harobed has quit IRC01:19
projekt01Ok, I see, we jut need a counter for the new release of
xenruI think you should use apache logs01:21
xenrunot on zope side01:22
xenruapache should cache pages before requests come to zope01:22
oferwprojekt01: this is all z3?01:23
projekt01xenru, I agree01:23
projekt01oferw, what do you mean?01:24
xenruoferw: looks like a plone01:24
projekt01porting a plone app to the next state ;-)01:24
oferwprojekt01: if this site a based on z301:25
projekt01this is plone site, the new version will be based on z3 with a z3 XML-RPC transcoding server01:26
oferwprojekt01: looking good in general01:26
projekt01based on lovely.remotetask which is located on at:01:27
projekt01oferw, thanks but that was not my work;-)01:28
*** philiKON has quit IRC01:32
oferwprojekt01: :)01:32
*** rocky is now known as rocky|Zzz01:35
*** oferw has quit IRC01:54
*** regebro has joined #zope3-dev01:58
*** regebro has quit IRC02:01
*** natea|away has quit IRC02:01
*** projekt01 has left #zope3-dev02:13
*** natea has joined #zope3-dev02:21
*** yota has quit IRC02:55
*** robcampbell has quit IRC03:02
*** jinty has quit IRC03:09
*** replicant has joined #zope3-dev03:10
*** natea has quit IRC04:00
*** stub has joined #zope3-dev04:33
*** strichter has joined #zope3-dev05:27
*** srichter has quit IRC05:27
*** niemeyer has quit IRC05:35
*** replicant has quit IRC05:41
*** romanofs1i has joined #zope3-dev05:41
*** romanofski has quit IRC05:57
*** tiredbones has quit IRC06:18
*** Londo_ has quit IRC06:18
*** baijum has joined #zope3-dev06:26
*** Londo_ has joined #zope3-dev06:35
*** tiredbones has joined #zope3-dev06:35
*** natea has joined #zope3-dev07:23
*** dobee has joined #zope3-dev08:01
*** natea has quit IRC08:05
*** eins has joined #zope3-dev08:09
*** stu1 has joined #zope3-dev08:11
*** tiredbones has quit IRC08:23
*** Londo_ has quit IRC08:23
*** alecm has quit IRC08:23
*** WebMaven has quit IRC08:23
*** tiredbones has joined #zope3-dev08:23
*** Londo_ has joined #zope3-dev08:23
*** alecm has joined #zope3-dev08:23
*** WebMaven has joined #zope3-dev08:23
*** natea has joined #zope3-dev08:24
*** stub has quit IRC08:29
*** MJ|sleep is now known as MJ08:48
*** natea has quit IRC08:52
*** stu1 is now known as stub08:55
*** tiredbones has quit IRC08:56
*** alecm has quit IRC08:56
*** WebMaven has quit IRC08:56
*** Londo_ has quit IRC08:56
*** tiredbones has joined #zope3-dev08:57
*** Londo_ has joined #zope3-dev08:57
*** alecm has joined #zope3-dev08:57
*** WebMaven has joined #zope3-dev08:57
*** flox has quit IRC09:00
*** febb is now known as febb_09:09
febb_hello all09:09
*** febb_ is now known as febb09:09
*** wrobel has joined #zope3-dev09:17
*** wrobel has left #zope3-dev09:18
*** wrobel has joined #zope3-dev09:27
*** hdima has joined #zope3-dev09:27
*** Aiste has quit IRC09:45
*** nouri has quit IRC09:45
*** SteveA has quit IRC09:45
*** Aiste has joined #zope3-dev09:45
*** nouri has joined #zope3-dev09:45
*** SteveA has joined #zope3-dev09:45
*** yota has joined #zope3-dev09:46
*** stub has quit IRC09:49
*** jinty has joined #zope3-dev09:50
*** dlk has joined #zope3-dev09:57
*** MJ has quit IRC10:02
*** mgedmin has joined #zope3-dev10:02
*** kobold has joined #zope3-dev10:11
*** wosc has joined #zope3-dev10:12
*** philiKON has joined #zope3-dev10:15
*** hdima has quit IRC10:30
*** Aiste has quit IRC10:36
*** MJ has joined #zope3-dev10:48
*** flox has joined #zope3-dev10:48
*** strichter is now known as srichter10:52
*** ChanServ sets mode: +o srichter10:53
*** Aiste has joined #zope3-dev10:55
*** harobed has joined #zope3-dev11:01
*** romanofs1i is now known as romanofski11:05
*** ktwilight has quit IRC11:15
*** timte has joined #zope3-dev11:20
*** rocky|Zzz has quit IRC11:24
*** ktwilight has joined #zope3-dev11:38
*** faassen has joined #zope3-dev11:39
*** ktwilight has quit IRC11:40
*** ktwilight has joined #zope3-dev11:46
*** batlogg has quit IRC12:08
*** batlogg has joined #zope3-dev12:11
*** projekt01 has joined #zope3-dev12:26
*** dunny has joined #zope3-dev12:32
*** regebro has joined #zope3-dev12:39
*** strichter has joined #zope3-dev12:48
*** srichter has quit IRC12:48
*** dunny has quit IRC13:04
*** strichter is now known as srichter13:13
*** ChanServ sets mode: +o srichter13:13
*** mkerrin has joined #zope3-dev13:26
*** rock1 has joined #zope3-dev13:44
*** rock1 is now known as rocky13:46
*** wosc has left #zope3-dev13:46
*** hdima has joined #zope3-dev13:56
*** J1m has joined #zope3-dev14:06
*** J1m has quit IRC14:11
*** J1m has joined #zope3-dev14:12
*** J1m has joined #zope3-dev14:13
nouriLooks like installation of zope.component from requires zope.testing, but it's not in the required list.14:20
philiKONnouri, only a testing requirement, right?14:21
nouriI import zope.component and get ImportError14:21
*** alga has joined #zope3-dev14:21
philiKONthat's certainly a bug14:22 imports zope.testing14:22
philiKONsuperficially, it's a packaging bug14:22
philiKONnouri, hmm, sounds like you have the 3.2 eggs?14:22
nouriyes, are there any other eggs?14:22
philiKONhmm, no14:22
philiKONbut 3.3 fixes this :)14:22
* philiKON tries to think of a way to get nouri build 3.3.0 eggs :14:23
* nouri tries to think of something else :p14:23
nouriSeriously, I might look into this at some point, but not today nor tomorrow14:23
philiKONgogogo :)14:24
nouriIn the meantime we could fix the zope.component egg, no?14:24
* nouri has no idea where to look14:24
philiKONyou probably want the 3.2.0 tag?14:25
philiKONbetter ask tres how to proceed14:26
J1mfaassen: did you figure out the buildout problem you were having yesterday?14:29
*** niemeyer has joined #zope3-dev14:39
*** ktwilight has quit IRC14:40
*** ktwilight has joined #zope3-dev14:42
*** eins has quit IRC14:43
*** povbot has joined #zope3-dev15:16
J1mI'll try it...15:18
J1mworked great for me.15:22
J1mHere's my buildout.cfg:15:22
*** benji has joined #zope3-dev15:22
nouriIs there an easy way to use the testrunner once zope.testing is installed?15:23
*** ignas has joined #zope3-dev15:24
J1mI see what the problem is.15:24
J1mThe documentation for custom is wrong.15:24
J1mHere's what worked for me:15:25
J1mparts = pl15:26
J1mrecipe = zc.recipe.egg:custom15:26
J1megg = python-ldap15:26
J1mfind-links = /home/jim/s/python15:26
J1mno aaargh15:26
*** nathany has joined #zope3-dev15:26
J1mparts = pl15:26
J1mrecipe = zc.recipe.egg:custom15:26
J1meggs = python-ldap15:26
J1mfind-links = /home/jim/s/python15:26
J1mwell, I still don't see why this failed for you.15:26
faassenyes, thats' weird.15:27
J1mThis worked for me too:15:27
J1mparts = python-ldap15:27
J1mrecipe = zc.recipe.egg:custom15:27
J1mfind-links = /home/jim/s/python15:27
J1m(I hate this irc client.)15:28
faassenI have some additional sutff, like include-dirs and such.15:28
J1manyway, the option to specify the eggs to get is eggs, not egg.15:28
faassenhm, in your second recipy.15:28
J1mIt defaults to the part name.15:28
faassenthere's no 'egg =15:28
faassenand that's python-ldap in my case.15:28
faassenrecipe = zc.recipe.egg:custom15:29
faassenegg = python-ldap == 2.2.015:29
faassenfind-links = .15:29
faassenis what I got.15:29
J1mRight, which is why I don't know why it isn't working for you.15:29
faassenI'll change the part name, see what happens.15:29
faassenor first, I'll dump the 'egg' line.15:29
J1mOne issue with buildout is that it doesn't report unrecognized options.15:30
J1mThat's because of a missguided desire to suport ConfigParser's DEFAULT feature.15:30
faassenokay, removing the egg line didn't work.15:30
faassenrecipe = zc.recipe.egg:custom15:31
faassenfind-links = .15:31
faassenis what I'm trying now.15:31
J1mI think I'm gonna start compliaining about unrecognized options.15:31
faassenyeah, sounds good. :)15:31
faassenstill no go..15:31
J1mThis will make DEFAULT unusable, but that's not much of a loss.15:31
*** stub has joined #zope3-dev15:31
J1mwhat do you get from ls .15:31
J1min your buildout directory.15:32
faassenhm, so you have 'eggs = '15:32
faassenand I have egg =?15:32
J1mThe option is really named eggs.15:33
J1mBut it defaults to the part name.15:33
faassenso that shouldn't be an issue.15:33
J1megg is ignored.15:33
faassenokay, renamed it to 'eggs', still no luck.15:33
J1mmay I see a directory listing please. :)15:33
faassenyes, it's rather full though, I'm trying this in an isolated buildout now.15:34
faassenanyway, the thing in there is: python-ldap-2.2.0.tar.gz15:34
faassenargh, in an isolated buildout it does work. that is, it crashes.15:34
J1mit crashed because it can't find the include files?15:35
faassenyeah, exactly. that's fine.15:36
faassenat least I assume it's that.15:36
faassenso now the question is, why didn't that work in my more extensive buidout.15:36
faassendirectory structure coming up.15:36
* baijum is experimenting with MakingARelease15:36
baijumJ1m, I cann't run this, ../../zpkgtools/bin/zpkg -r Zope-3.3.0 -C Zope.cfg Zope (I think it's because the tag is '3.3.0' instead of 'Zope-3.3.0')15:36
J1mbaijum: why are you doing this?15:37
baijumjust experienting :)15:37
J1mI made the last zpkg-based releases yesterday, I hope.15:37
*** mgedmin has quit IRC15:37
J1mwell, I don't have time to help you with this.15:38
*** ktwilight has joined #zope3-dev15:38
faassenanyway, tihs is really weird.15:38
J1mIt is rather pointless.15:38
faassenit works in an isolated buildout, but in my full buildout it just fails.15:38
J1mI'm still waiting for that pastebin page15:38
J1mthere it is15:38
faassenJ1m: yeah, sorry, I should find a better paste thing.15:38
faassenJ1m: anyway, concerning the last zpkg release, does that mean 3.3.0 final is out?15:39
faassenJ1m: I only see a c1 on zope.org15:39
J1mYes, waiting for philiKON to finish the process.15:39
faassenah, cool. :)15:39
J1mI uploaded the tar ball and exe yesterday.15:39
faassendoesn't seem to be appearing  yet on's zope 3 releases page yet, I guess that's philiKON's job.15:40
* baijum saw this and give it a try15:40
J1mphiliKON: philiKON15:40
J1mphiliKON: philiKON philiKON philiKON 's job15:41
baijummay be the doc should be changed ?15:41
*** philiKON has quit IRC15:41
faassenbaijum: I think we still need to dowork on..15:41
faassenJ1m: heh. he left. :)15:41
faassenbaijum: we still need to work on how it will work in the future, finalize that, before that page can be updated.15:41
faassenbaijum: since I'm going to do maintenance releases I might bug J1m about it at some point. :)15:42
J1mfaassen: may I see your full buildout.cfg in the larger buildout?15:42
faassenJ1m: yes.15:42
faassenit's slooooow.15:43
J1mI noticed15:44
faassenI just ran it all afresh in a clean buildout, same problem.15:50
faassenreally weird.15:50
faassenhm, it doesn't appear to be the find-links in the buildout section that's interfering. I'm just trying various things out now.15:52
J1mYeah, I'm at a loss.15:52
J1mYou could try snipping things out until it works.15:52
faassentrying that.15:53
faassenoh, concerning setting up my own test layer. I think I can see how to do that..but is there any way to get to the INSTANCE_HOME during the test running?15:54
faassenI want to point it to ftesting.zcml in the instance as I have absolutely no changes.15:54
J1mNo, Zope 3 doesn't really have such a concept.15:54
faassenso how do I get it to load up the package includes and such?15:54
J1mGood question. :)15:55
faassenI mean, all I need is an ftesting environment equivalent to the normal one, and etc/ftesting.zcml provides this.15:55
J1mI'm really annoyed at myself for not getting the tests working (or at least running) in the sharing example.15:55
faassenanyway, for now I can hack around it by running the tests starting in the instance part.15:56
faassenbut that's obviously sub-ideal. :)15:56
J1mYeah, unfortunately, the Zope 3 functional layer makes a lot of assumptions about what cwd looks like.15:56
J1mwhich is rather lame.15:56
faassenand rewriting that layer is trickier than it seems.15:57
J1mWe really need to get away from having important data outside of Python packages.15:57
faassenafter all, one still wants to find ftesting.zcml15:57
faassenJ1m: tricky for package-includes :)15:57
J1malthough this is a somewhat different issue.15:57
*** volvox has joined #zope3-dev15:58
*** edgordon has joined #zope3-dev16:00
J1mI'll have to think about the testing issue.16:03
faassenJ1m: okay, I think I reproduced the issue.16:04
faassenlike this:16:04
faassenparts = zope3 pl16:04
faassenfind-links =
faassenrecipe = zc.recipe.zope3checkout16:04
faassenurl = svn://
faassenrecipe = zc.recipe.egg:custom16:04
faasseneggs = python-ldap == 2.2.016:04
faassenfind-links = .16:04
faassenI think the zope3 recipe somehow makes the pl recipe fail to find anything.16:04
J1mlet me see if I can reproduce that.16:05
J1mI can't reproduce that.16:09
faassenhm, interesting.16:10
faassenlet me try again.16:10
J1mThe only thing I can think of to try now is to log into your site and try to debug it. :(16:10
J1myour machine16:10
faassenhm, unfortunately I can't easily open it up. let me try this again.16:10
J1mOr you could try sending me a tar ball of your buildout.16:11
J1mwhat version of Python are you using?16:11
faassenpython 2.416:11
faassenhand-compiled. :)16:11
J1mme too16:11
faassenyes, it's really the zope 3 recipe for me.16:12
faassenif I change this:16:13
faassenparts = zope3 pl16:13
faassento this:16:13
faassenparts = pl zope316:13
*** dobee has quit IRC16:13
faassensuddenly it'll work.16:13
faassenyou had it the other way around? :)16:13
J1mchange find-links = . to16:13
*** dobee has joined #zope3-dev16:14
J1mfind-links = full_path_to_current_directory16:14
faassenokay. will try again, though yesterday I worked iwth a full path a lot. that is, a file:/// path.16:14
faassennow I'll try it with a non-file path16:14
faassenyeah, it may be that the zope 3 recipe is doing os.chdir in ther.e16:14
faassenand that somehow the find story depends on that.16:14
faassenbut it's not the case.16:15
J1mI bet the zope3 recipe is doing a chdir and the buildout isn't reseting cwd.16:15
faassenwell, yea,h but it's not the case. :)16:15
*** alecm has quit IRC16:15
faassenI mean, I'm pretty sure. yesterday I tried it with file:/// urls, http:// urls and still no luck.16:15
*** Aiste_ is now known as Aiste16:16
J1mpretty sure doesn't count much at this point. :)16:16
faassenI'm confirming as we speak.16:16
faassenI mean, I made it a HTTP URL now.16:16
faassenit works if I put the pl section first16:16
faassenokay, interesting.16:17
J1mNow I can reproduce the failure using a relative path,16:18
faasseninteresting it also fails with file:///16:18
faassenit doesn't fail with http://localhost, but I'm 100% sure I had it not working for me when I tried that before, so I'm trying it in the cmplex buildout again.16:19
faassenyes, it still fails in my complicated buildout, even using http://localhost16:19
faassenwhich *works* in the simple buildout.16:19
faassenand a full file:/// url fails.16:21
J1mYeah, that fails for me too.16:21
*** ofer has quit IRC16:21
faassencan one offer paths to find-links (i..e non-urls?)16:22
faassenyea, I know you can.16:22
faassenas we used ., right.16:22
faassenI mean: .16:22
faassenokay, a full path works.16:22
faassenbut does it for my complex buildout? :)16:22
rockyok, i forbid anyone to ever do monkey patches besides me... that should relieve a lot of my headache :)16:23
faassenrocky: okay, sounds good. then we shoot you.16:23
faassenrocky: problem solved. :)16:23
faassenokay, in my complicated buildout a full path works.16:24
J1mIf I put python-ldap on it works.16:24
J1mI think I know what's going on16:24
faassenyeah, that's equivalent to my http://localhost?16:24
J1mMaybe not.16:24
faassenhey, weird, http://localhost works now.16:25
J1mBut you said easy_install worked from your localhost, which is odd.16:25
faassenthat's crazy. I know it didn't work before.16:25
faassenyeah, it works for sure in my test script.16:25
faassenbut I know it failed before. I mean, I was trying lots of things and I tried this one. :)16:25
J1mI'll note that setuptools is picky about the content-type returned from a server.16:25
faassenwell, this is a straight Apache directory serve thing.16:26
faassenyeah, now it consistently seems to find it.16:26
faassenbizarre. I know that didn't work before. :(16:26
J1mIt won's scan a page that doesn't have a text/html content type/16:26
faassenmy http://localhost is an apache directory listing, that's HTML.16:26
J1mbut it sounds like your localhost thing is working.16:26
J1mthat is, specifying localhost in find-links is working.16:27
faassenyeah, but it didn't work before. :(16:27
J1mSo the bugs:16:27
faassenI'm really confused, it must've been I somehow did something wrong and I do it right now, though.16:27
*** philiKON has joined #zope3-dev16:27
J1mfile urls don't seem to work.16:27
nouriAnyone have a good idea on what's happening here:16:27
philiKONso, who was looking for me eearlier16:27
philiKONever so annoyingly16:27
nouriOn component.getUtility16:27
nouriAttributeError: 'GlobalAdapterRegistry' object has no attribute '_null'16:28
J1mbuildout doesn't reset cwd after a recipe is run, which causes relative paths not to work.16:28
faassenphiliKON: that must've been J1m :)16:28
* nouri still trying out the zope.component 3.2 egg16:28
faassenright, that definitely appears to be a bug.16:28
J1m2 bugs16:28
philiKONnouri, you want the 3.3 egg. i know you do. make eggs. make eggs!16:28
faassenyeah, the second bug is a bit more surprising. :)16:28
J1measy_install has a fair bit of code in it that should be down in lower layers.16:28
faassenas I have no idea what's up there.16:28
philiKONJ1m, you were looking for me?16:28
faassenJ1m: so what's the second bug?16:29
J1mIt may be that there is some special handling of file urls that I'm not reproducing.16:29
faassenI guess.16:29
J1mphiliKON: where's the release? :)16:29
philiKONi told you i will probably not have time before saturday16:29
J1mfaassen: bug 1: file urls aren't handled correctly16:29
philiKONsomeone else (like the 3.3 release manager *cough* faassen *cough*) can beat me to it16:30
philiKONbut i have a book deadline monday16:30
philiKONwhich i can't miss16:30
faassenanyway, I'm still flabberghasted about the http://localhost issue, but I'll write that down to Martijn being confused terribly so he wasn't thinking rationally anymore.16:30
faassenphiliKON: I'm not the 3.3 release manager. I'm the bugfix release manager.16:30
J1mbug2: buildout doesn't reset cwd.16:30
faassenphiliKON: of 3.3 :)16:30
philiKONfaassen, uh huh16:30
faassenphiliKON: seriously, I never volunteered for Zope
philiKONpoint is, i don't have time before tomorrow, which i think i communicated16:30
philiKONfaassen, yeah, i know16:30
faassenphiliKON: I am sure it can wait for tomorrow.16:31
philiKONi think so too16:31
J1mphiliKON: I thought you were going to do that.  It's OK if you don't have time, but I wish I knew.16:31
faassenJ1m: isn't tomorrow okay?16:31
J1mok w me16:31
philiKONJ1m, again, i'm pretty sure i told you that i wouldn't have time before saturday16:31
faassenanyway, that's all resolved, and we tracked down nefariousness in buildout. :)16:31
J1mI thought you were going to do it last night.16:31
faassennow with a freshened heart and mind, I will proceed on my quest.16:32
J1mIt can wait till saturday imo16:32
philiKONJ1m, i came home last night and DSL was out16:32
J1mfaassen: I still need to figure out the testing issue.16:32
faassenJ1m: yeah, that's a problem.16:32
philiKONJ1m, called them, they only fixed the dsl link today16:32
faassenJ1m: okay, concerning zc.recipe.cmmi16:33
faassenJ1m: somehow my python-ldap part kicks in and fails before the cmmi recipe is being run.16:34
faassenJ1m: even though the openldap part comes before the python-ldap part in the parts listing.16:34
rockyso am i right in assuming you guys are building a buildout recipe that will setup an openldap server? or?16:34
faassenrocky: just a client for now.16:34
rockybecause i'm soooo going to need this for our bigger clients running Plone ;)16:34
faassenrocky: but just to compensate for that, I do have a buildout piece that sets up openoffice servers and all kinds of stuff.16:35
rockyoh awesome16:35
faassenrocky: now if you write a few buildout things that sets up zope 2, you'd make me happy.16:35
rockyfaassen: have you seen the topp.buildout work?16:35
rockyi thought it did that16:35
faassenrocky: yeah, sorry, I did see that. dont' know how well it works, didn't try it yet.16:35
J1mfaassen: is there an openldap subdirectiry in the parts directory?16:36
faassenJ1m: no.16:37
faassenJ1m: it just never reaches that part, it just divesi nto the failing python-ldap bit.16:37
J1mThat's odd16:37
faassenI do have it working in a simple setup,  at least it dives into configure now.16:39
faassenI'll experiment a bit more and let you know what I find out.16:39
J1mI've never seen the parts run out of order before.16:39
J1mok, I'm gonna get some breakfast.16:40
faassencould it be the custom egg recipe kicks in too soon somehow?16:40
faassenI mean, I don't know, might have a download phase, etc.16:40
J1mI can't think of a reason why it would.16:40
J1min a minute, I'm gonna ask you to try the localhost thing again.16:41
faassenthe localhost thing now consistently works for me, though.16:41
J1mtry now16:41
philiKONJ1m, you have a typo in the windows release. it's called 3.*2*.016:41
philiKONis it the right release or just a wrong name?16:42
philiKONit has   Zope-3.2.0.win32-py2.4.exe16:42
J1mah crap16:42
philiKONis that a build of the 3.2.0 release or 3.3.0 w/ wrong name?16:43
J1mProbably made the wrong release.16:43
faassenJ1m: anyway, trying the localhost thing again, did you upload a new version of a recipe or something?16:43
J1mwell, it was generated so I probably typos early on and made the wrong release. :(16:43
faassenhm, now it can't find the localhost anymore.16:43
J1mThat's because you weren't getting it from localhost.16:43
faassendo yhou think that find-links = ..16:43
philiKONJ1m, :( good thing that i caught it today, though16:43
faassenJ1m: well, when I type http://localhost I expect localhost to be accessed. :)16:44
faassenJ1m: weird though.16:44
faassenJ1m: I mean, why doesn't localhost work? I get a HTML page.16:44
J1mYou were getting it from download.zope.org16:45
J1mI just removed it from there.16:45
J1mfaassen: what content type down your localhost give16:45
J1mI don't know.16:45
faassenI mean, it's text/html, it's HTML, it's a link, when I click on it in my browser it downloads.16:45
J1mI don't think it is a question of being localhostm as the tests all use a local server.16:46
faassenanyway, that thing is ..16:46
faassenJ1m: anyway, at least I wasn't insane. I could've sworn it didn't work with localhost before.16:46
J1mSo I don't know what it doesn't like about your localhost.16:48
J1mI don't understand why buildout wouldn't work and easy_install would, since they use the same code.16:48
J1mAnd I have no way to reproduce or debug it.16:49
J1mAnd I'm hungry. :)16:49
*** baijum has quit IRC16:53
faasseninteresting. I'll worry about ti late,r my localhost was just a last-ditch attempt.16:57
faassenI think it did work with http://localhost for easy_install though16:57
faassenI think I tried that. I'll try again.16:57
faassenyes, this works:16:57
faasseneasy_install -v --find-links=http://localhost python-ldap16:58
faassenReading http://localhost16:58
faassenFound link: http://localhost/python-ldap-2.2.0.tgz16:58
faassenBest match: python-ldap 2.2.016:58
faassenDownloading http://localhost/python-ldap-2.2.0.tgz16:58
faassenso it's definitely something in buildout that's screwing this up from not-working.16:58
faassenor possibly a slightly more recent setuptools16:58
faassenas this workingenv I made earlier this week.16:58
faassentesting that hypothesis.16:58
faassenhm, network connectivity issues here.16:58
faassenJ1m: okay. :)16:59
faassenJ1m: anyway, I can confirm easy_install works with my localhost and buildout doesn't.16:59
J1mDoes a trailing slash make any difference?16:59
faassenJ1m: hm, yes, it appears that might indeed work: http://localhost/ in buildout seems to make it work.17:01
faassenunless you placed that python-ldap back on :)17:02
faassenJ1m: anyawy, go eat. :)17:02
faassenJ1m: I think I have enough working pieces/workarounds to continue for a while. :)17:02
faassenJ1m: and you got tons of issues to ponder. :)17:03
faassenaway from keyboard for a while.17:03
*** dlk has left #zope3-dev17:05
*** hdima has quit IRC17:08
*** stub has quit IRC17:11
*** SmokeyD has quit IRC17:25
*** SmokeyD has joined #zope3-dev17:26
*** natea has joined #zope3-dev17:29
*** volvox_ has joined #zope3-dev17:30
*** volvox has quit IRC17:30
*** flox has quit IRC17:38
*** batlogg has quit IRC18:03
*** batlogg has joined #zope3-dev18:05
*** zagy has quit IRC18:05
*** alga_ is now known as alga18:06
*** stub has joined #zope3-dev18:06
*** Aiste has quit IRC18:13
*** batlogg has quit IRC18:19
*** volvox_ has quit IRC18:25
*** MJ has quit IRC18:33
*** philiKON has quit IRC18:47
*** ktwilight has quit IRC18:51
*** ktwilight has joined #zope3-dev18:51
*** xenru|clone has joined #zope3-dev18:53
*** ignas has quit IRC18:57
*** bradb has joined #zope3-dev19:01
*** stub has quit IRC19:01
*** kobold has left #zope3-dev19:07
*** mgedmin has joined #zope3-dev19:08
*** mgedmin has quit IRC19:08
*** xenru has quit IRC19:08
*** xenru has joined #zope3-dev19:08
*** mgedmin has joined #zope3-dev19:09
*** RaFromBRC has joined #zope3-dev19:09
*** batlogg has joined #zope3-dev19:15
*** xenru|clone has quit IRC19:15
*** dokai has quit IRC19:15
*** philiKON has joined #zope3-dev19:30
*** vbachs has joined #zope3-dev19:34
*** RaFromBRC has quit IRC19:34
*** alga has quit IRC19:35
*** mgedmin has quit IRC19:36
*** MJ has joined #zope3-dev19:48
*** zagy has joined #zope3-dev19:55
*** harobed has quit IRC19:58
*** nathany has quit IRC20:08
faassenJ1m: when extending another buildout, what happens with parts?20:11
faassenJ1m: like, I have a base.cfg which lists parts = a b20:11
J1myou don't extend buildouts. You extend configuration files.20:11
faassenJ1m: and then I have my buildout.cfg which extends..20:11
*** gumpa has joined #zope3-dev20:11
faassenJ1m: yeah, okay, yes, so when I'm extending config files..20:12
faassenthe base.cfg20:12
faassenif I define new parts there..20:12
faassenor new develop eggs, say.20:12
faassenI just override them completely?20:12
faassenI need to relist them?20:12
J1moptions are overridden in total.20:12
faassenit'd be nice to be able to reuse [instance] sections mostly.20:12
faassenand just add a few eggs to it.20:12
faassenand also to be able to reuse a develop section, and just add a few more development eggs, say.20:13
faassenbut that usecase isn't supported, right?20:13
J1mnot without creating extra options.20:13
J1mBut I could see the usefulness of letting an extending confguration refer to the bits it's extending.20:14
faassenanyway, I was just wondering whether I missed something, but it's the way I thought.20:15
J1mPerhaps you can think of a nice variable substitution syntax that means "get section x, option y, from the configuration that I am extending.20:16
faassenhm, does one need to be able to say a lot?20:16
faassenyou might just be able to get away with something like ..20:17
faassenhm, depends on whether you want to rename sectinos, I guess.20:17
faassenif your part names are the same.20:17
faassenthen it'd be enough to say: @20:17
faassenlike, zcml: @20:17
faassenor zcml: @ extrastuff20:17
faassenwell, on a separate line, that.20:17
faassenas you'd refer to the original section automatically.20:18
J1mI have no idea what you are refering to.20:18
faassenbut perhaps that's not flexible enough.20:18
faassenwhat I'm saying if if I have a fooconfig.cfg20:18
faassenwhich extends barconfig.cfg20:18
faassenand I want an extra egg in my instance in fooconfig.cfg20:18
faassenI have an [instance] section in there20:18
faassenwith a zcml option20:18
faassenand if I want tos ay, reuse everything from th barconfig.cfg for the zcml bits.20:18
faassenif the section names are identical, there's no other information necessary.20:19
faassenif not, you might want to do:20:19
J1mI'd want to leverage the ${} syntax.20:19
faasseneggs = @barinstance20:19
faassen                     extraegg20:19
faassenwhat I'm getting at is that there doesn't appear to be any reason to specify the option explicitly.20:20
faassenif you're going to reuse information from an option, the option name is implicit.20:20
J1mexplicit is better than implicit.20:20
faassenexcept when ti's not.20:20
faassenhere it's just redundant typing and failure prone.20:20
faassenso it's not better.20:20
faassenbecause it makes absolutely no sense to say:20:20
J1mI also wouldn't want to refer to a base configuration file name.20:20
J1mYou are extending a base configuration that comes from bany sources.20:21
faassenzcml = ${barinstance, eggs}20:21
J1mmany sources20:21
faassenI mean, it's always going to be zcml.20:21
J1mI don't follow you.20:21
faassenyou were saying, get section X, option y from the configuration that I am extending.20:22
faassenI'm sayin that option y is redundant and allowing people to specify it is failure prone at best.20:22
J1mPerhaps we allow it to be blank.20:24
faassenyeah, that at least follows the DRY principle. :)20:24
faassendon't repeat yourself.20:25
J1mI'll think about this.20:26
J1mI like the idea of being able to refer to the base configuration.20:26
J1mI'd like to extend the current syntax.20:27
*** RaFromBRC has joined #zope3-dev20:27
faassenyeah, that makes sense.20:27
*** rocky has quit IRC20:32
*** projekt01 has left #zope3-dev20:32
faassenJ1m: oh, by the way, that python-ldap is too horrible to work with easy install or buildout. :(20:32
faassenJ1m: it has a horrible hack in there to find paths, it appears, and such.20:33
J1mfind what sorts of paths?20:33
*** srichter has quit IRC20:33
faassenexec_startdir = os.path.dirname(os.path.abspath(sys.argv[0]))20:34
faassenpackage_init_file_name = reduce(os.path.join,[exec_startdir,'Lib','ldap',''])20:34
faassenI think that's fixable, they only want the version number or something.20:34
*** batlogg has quit IRC20:35
faassenbut then they use configparser to read a setup.cfg to load up various library paths.20:35
*** dobee has quit IRC20:35
faassenwhich are coded to usr/local/openLDAP in there20:36
faassenit doesn't appear those can be overridden. do paths set in always take precedence above those set by buildout?20:36
faassenas it's certainly completely ignoring whatever I tell it. :)20:36
J1mwell, buildout just runs and passes options to build_ext.20:37
J1mI'm guessing that it will ignore options you pass to build_ext.20:38
faassenand build_ext doesn't allow overrides?20:38
faassenif they're already set, you mean?20:38
J1mIt does, that's why it provides options.20:38
faassenI wonder why it ignores them then.20:38
J1mI don't know.20:38
J1mMore distutils foo than I have.20:39
faassenanyway, I'll try to look into that later.20:39
J1mHave you tried running build_ext yourself?20:39
faassenfor now I just want to go ahead with my ldap work instead.20:39
J1mUnfortunately, sometimes you just have to give up.20:39
J1mLike with pywin32.20:39
faassenoh, what's that? :)20:39
J1mThe win32 extensions.20:40
J1mYou need that to be able to use the sftp stuff on wiin32.20:40
J1mBut there's no working egg for it.20:40
J1mCause win32 wants to do weird things with .pth files and the windows registry.20:40
faassen(I just tried build_ext manually with --include-dirs)20:41
faassenit just ignores that information completely.20:41
J1mYup, sounds like an evil setup.py20:41
J1mso you would have to build a custom recipe, or install it by hand.20:41
faassenweird though. it just uses distutils.core.setup20:41
faassenoh, wait.20:42
faassenI see where it's failing, I suspect.20:42
faassenit uses ext_modules = [20:42
J1mMaybe if you pass those options to setup, it ignores command-line options. I dunno.20:42
faassenand *there* it passes the include_dirs and such.20:42
faassenand apparently those aren't overridden.20:42
faassenlike these are not direct arguments to setup.py20:43
faasseninstead they're passed along to Extension()20:43
faassenwhich is hooked into ext_modules = [Extension(include_dirs=something)]20:43
J1mPretty inflexible.20:44
faassenit's ane xample of why Python code can be bad for configuration. :)20:44
J1mI think it's more of an example of why explicit is better than implicit.20:46
J1mThis setup is trying to do too much.20:46
faassenperhaps it's an example of both. :)20:47
faassenmaybe it has to do this Extension bit to build a whole lot of c extensions, I don't know.20:48
faassenJ1m: out of here.. have a nice weekend!20:55
*** faassen has quit IRC20:55
*** RaFromBRC has quit IRC21:05
*** vbachs has left #zope3-dev21:10
*** regebro has quit IRC21:15
*** gumpa has quit IRC21:21
*** gumpa has joined #zope3-dev21:21
*** philiKON has quit IRC21:39
*** philiKON has joined #zope3-dev21:40
*** philiKON has quit IRC21:41
*** mkerrin has quit IRC21:59
*** jinty has quit IRC22:16
*** natea is now known as natea|busy22:42
*** Aiste has joined #zope3-dev22:51
*** Aiste has quit IRC22:51
*** sivang has quit IRC22:53
*** reco has quit IRC23:02
*** whit has quit IRC23:17
*** whit has joined #zope3-dev23:17
*** dunny has joined #zope3-dev23:50
*** TresEquis has joined #zope3-dev23:55

Generated by 2.15.1 by Marius Gedminas - find it at!