IRC log of #zope for Tuesday, 2010-03-02

mgedminmeeting in 30 mins?16:32
*** seppo14 has quit IRC16:32
*** akm1 has joined #zope16:33
smreckon so16:34
*** kobold has joined #zope16:36
smgreat idea Theuni116:37
*** dayne has quit IRC16:38
*** sm_ has joined #zope16:39
*** baijum has joined #zope16:40
*** pepeu has joined #zope16:42
*** incubaton has joined #zope16:43
*** alvaro has joined #zope16:44
*** sidnei has quit IRC16:44
*** srichter has joined #zope16:46
Theuni1sm: thanks. let's see how this works out ;)16:48
*** vigith has quit IRC16:48
incubatonis anyone could help me with kss stuff ?16:50
*** aaronv has joined #zope16:51
betabugincubaton: you probably want to ask in #plone16:51
incubatonbetabug: ok :)16:51
*** aatiis has left #zope16:51
betabugoh, and probably you want to ask a question, not just ask if someone can help16:52
*** aatiis has joined #zope16:52
*** laurynas has joined #zope16:55
*** sidnei has joined #zope16:57
incubatonbetabug: i don't want to disturb this chat if no one can help, i fould help on #plone so you helped me, thx betabug16:58
betabugno problem, you're welcome!16:59
* Theuni1 takes a chair16:59
*** MatthewWilkes has joined #zope17:00
*** hannosch has joined #zope17:00
*** replaceafill has joined #zope17:00
* Theuni1 distributes coffee to everyone in the channel17:00
mgedminis the agenda posted somewhere on the web?17:00
mcdoncwhere's the beer17:00
* Theuni1 throws a crate towards mcdonc17:00
hannoschfirst time I'm in this channel - I'm the noob :)17:01
pyqwerHi all!17:01
*** MJ has joined #zope17:01
Theuni1So, everybody. Let's start with the experiment of having developer IRC meetings here.17:01
*** yvl has joined #zope17:01
* Theuni1 pokes hannosch in the eye17:01
Theuni1The agenda is in the archive of the mailinglist. See here:
betabugright here?17:01
*** allisterb has quit IRC17:01
* mgedmin wonders if the channel is logged17:01
Theuni1right here.17:01
betabugfun :-)17:01
Theuni1betabug: we thought we take it out in the public ;)17:01
betabuggood idea!17:01
*** povbot has joined #zope17:01
*** ignas has joined #zope17:02
Theuni1Does anybody have comments about the agenda?17:02
pyqwerSeems o.k. to me.17:02
pyqwerUnclear about Zope 3.5 <-> BlueBream, though.17:02
*** J1m_ has joined #zope17:02
Theuni1Ok, lets go.17:03
*** yusei has joined #zope17:03
Theuni1Issue: test runners/nightly builds17:03
Theuni1Alan Runyan asked to talk about this and I'm interested in it myself.17:03
*** allisterb has joined #zope17:03
Theuni1I think we have regular test runs for Zope 2, but everything else is in the dark.17:04
hannoschwe have a lot of them in various shapes as listed at
mgedminzope 3.4 kgs has nightly test runs, but since it doesn't change, those are a bit pointless17:04
baijumWhere is Zope 2 build running ?17:04
mgedminI'm willing to set up more buildbots, if somebody helps me figure out how to run the relevant test suites17:05
hannoschzope2 is on lurker's private server17:05
Theuni1hannosch: the last two seem broken17:05
hannoschthe zope2/cmf/plone stuff is all custom scripts and not buildbots17:05
baijumThere was one for BlueBream here:  (but server is down now)17:06
srichtermgedmin: great offer; I think you should link up with bajium17:06
Theuni1so we're definitely missing regular runs of the ztk packages and the whole compatibility enchilada, right?17:06
Theuni1also, the mails that go to the zope-dev mailinglist seem to only cover zope 217:06
hannoschthe health agency one basically runs all
srichterbaijum: Marius (mgedmin) has been running buildbots for a long time very reliably and it would be one less headache for you17:07
baijumA single Build master would be enough for all projects ?17:07
*** TresEquis has joined #zope17:07
baijumsrichter: ah. cool17:07
*** RUNYAGAs has joined #zope17:07
Theuni1hannosch: ah, they seem to have integrated builds17:07
mgedminjust share your master.cfg's and I'll see what I can do17:07
RUNYAGAsjust start?17:07
baijumRUNYAGAs: just few minutes..17:08
hannoschthe main thing this needs is someone coordinating it - we have enough servers17:08
*** jpcw has joined #zope17:08
Theuni1hannosch: right.17:08
* Theuni1 looks for someone to label "Da buildbot man"17:08
*** ccomb1 has joined #zope17:08
* srichter wants to appoint mgedmin, but waits for him to volunteer :-)17:09
RUNYAGAsmight have someone at enfold put me down for details17:09
*** ccomb has quit IRC17:09
* mgedmin is unsure what that position entails17:09
betabugfree drinks from me, when you come here on vacation! :-)17:09
Theuni1mgedmin: i think it needs someone to hunt after dead builders and keep an overview whether we have enough.17:09
srichterok, that's a valid question: what does the buildbot guy have to do?17:09
Theuni1and help people set up new ones17:09
baijumWe need 64 bit Windows & Linux slaves now ...17:09
TresEquiswho has the list of buildslaves available?17:09
* mgedmin has 64-bit Linux slaves17:09
* mgedmin has no Windows slaves17:09
RUNYAGAswe have windoz17:10
Theuni1TresEquis: that list is kinda unobvious because we have multiple buildbot installations17:10
baijumRUNYAGAs: 64 bit ?17:10
*** aaronv has quit IRC17:10
* mgedmin announces irc logs for this channel at
*** kevingill has joined #zope17:11
*** RUNYAGAs has quit IRC17:11
Theuni1mgedmin: i think the actual amount of work for the buildbot volunteer would be to help tame the various installations match our requirements. it looks like the actual buildbot installations we find individual helpers for.17:12
srichterTheuni1: ok, so the buildbot man should maintain the build master and nag people about build slaves17:12
Theuni1srichter: well, and maybe help people install separate masters17:12
*** RUNYAGAs has joined #zope17:12
baijumI would suggest to use to setup slaves17:12
srichterTheuni1: right17:12
*** alga has joined #zope17:13
srichterBTW, the Python guys also went with one Buildbot czar now, since it did not work out with a group17:13
Theuni1right. he still has others supporting him, right?17:13
mgedminso what would the testing be about?  ZTK?  Zope 2?  Bluebream?  All of the above?17:13
srichterTheuni1: yes, but he does exactely what you are proposing17:13
Theuni1mgedmin: I think we need to care for the whole "zope-dev"17:14
Theuni1srichter: ah17:14
hannoschmgedmin: I think we should coordinate testing on Zope projects, that includes Grok and Zope217:14
*** thetet has quit IRC17:14
Theuni1right. one of the points of the dev meeting here is to get us all to talk more to each other and help us do stuff.17:14
hannoschwe can aggregate all the details into one mail to zope-dev per day17:14
TresEquis"cool kids" at PyCon seem to want to run Hudson17:14
baijumPython many many masters:
*** ccomb1 is now known as ccomb17:15
Theuni1TresEquis: yeah, i looked at hudson and i liked it a lot. for some reason i'm avoiding a tool discussion ;)17:15
* mgedmin has the unfortunate tendency to assume many responsibilities and then crash under the load17:15
Theuni1but you're right. hudson *is* awesome.17:15
baijumOne master for each project ?17:15
Theuni1if *i* would take over the nightly build responsibility, i'd look into hudson ;)17:15
Theuni1mgedmin: how about this: we postpone the details until next week, and you ponder it until then?17:16 etc.17:16
ccombI will try to set up a new buildbot tonight or tomorrow, until the previous one can be restored17:16
ccombfor the ZTK and bluebream17:16
mgedminTheuni1, fine, I'm always in favour of postponement ;)17:16
srichterTheuni1: and collect over the next week the projects and branches we want to support in the buildbot17:16
Theuni1srichter: right17:17
Theuni1next topic17:17
Theuni1ZTK open issues17:17
Theuni1That's kinda fuzzy. :)17:17
Theuni1We have some open ends since Martijn dropped out of the steering group.17:17
baijummost of them are still part of Zope 3:
* mgedmin wants every ZTK and zope.* package to have its own bug tracker on launchpad17:17
Theuni1And I can't even really remember all of them.17:17
Theuni1Ah, wait.17:17
Theuni1I'm not talking about bugs.17:18
Theuni1I'm talking about general organizational issues.17:18
mgedminah, okay17:18
hannoschwe have general government / process, definition of goals, roadmap, release manager and more17:18
Theuni1Like: we started the ZTK effort, but supposedly nobody uses it yet and we're not sure what the roadmap actually ist.17:18
* Theuni1 scratches his head on where to start17:18
TresEquisis Grok close to using it yet?17:19
Theuni1any grok folks around?17:19
TresEquisZ2 could go back, now that tempers have cooled, if we agree on goals17:19
sidneiwe are using the ztk on launchpad and landscape (or about to)17:19
ignaswhat about Schooltool?17:19
Theuni1sidnei: cool17:19
Theuni1sidnei: how did the zope.*/* split work out for you?17:19
* mgedmin didn't think to join #grok and invite the locals17:19
ignasyvl, are you using ZTK?17:19
baijumAFAIK, Grok use a particular svn revision for 1.1 release17:20
hannoschI think Grok is pretty close. there's some harder / testing things that'll take them more time17:20
J1m_what "it" are you referring to? ZTK or some ZTK development process?17:20
sidneiTheuni1, no major issues, we are using very little of, mostly things that are pulled as dependencies of other things.17:20
TresEquisbaijum: how is BB using ztk / zopeapp?17:20
yvlignas, kind of.  Some revision of it.17:20
Theuni1J1m_: which "it" it?17:20
*** sm is now known as sm-afk17:21
baijumTresEquis: BB also use a particular revision for 1.0 release, cann't wait for ZTK 1.0 ...17:21
J1m_what "it" are y'all taking about using?17:21
* mgedmin thinks "it" is the ZTK itself17:21
hannoschJ1m_: the ztk.cfg file in subversion17:21
baijumBB 1.0 and Grok 1.1 use same revision of ZTK17:21
TresEquiswe need to get "released" versions of ztk and zopeapp out ASAP17:21
sidneiall the patches we had pending have been submitted and new releases made, thanks to TresEquis for granting me rights.17:22
pyqwer+1 to release17:22
J1m_We're using the ZTK packages and I think we're using the .cfg file.17:22
TresEquisCan we just call the r#### used by Grok 1.1 and BB 1.0 "ZTK 1.0"?17:23
baijumAnother thing is creating Windows binaries for ZTK package17:23
* mgedmin dreams of a build daemon...17:23
mcdoncdoes it make sense to bring up the idea of giving a name and a separate identity the "bicycle toolkit" (z.component, z.configuration, z.interface)17:23
Theuni1TresEquis: should we agree on goals before a release? ;)17:23
baijumGrok is waiting for Windows binaries for 1.1 final release17:24
mgedmingoal #1: make a release17:24
TresEquisI don't think we disagree, after the split of zopeapp.cfg17:24
mgedminthere, simple!17:24
Theuni1mcdonc: it makes sense, but i'll put that on the backburner ;)17:24
agroszermgedmin, a well configured buildbot as build daemon?17:24
RUNYAGAsbinaries of each package for win, correct? not ztk as one big binary, ?17:24
mgedminagroszer, tell me how to run all of ZTK's tests and you'll get a buildbot in a day or two17:24
Theuni1TresEquis: how do we establish that we agree? :)17:24
baijumRUNYAGAs: individual package binaries17:24
Theuni1and how are the goals defined? :)17:24
* mgedmin doesn't do windows, which might disqualify him from the buildbot czar position17:25
hannoschcan we appoint the Zope Foundation board to find a release manager for the ZTK?17:25
Theuni1hannosch: the foundation has a task for figuring out MacYET2's successor for Z217:25
TresEquisZF board doesn't really drive development process17:26
Theuni1so we could also think about the same issue for the ZTK then17:26
Theuni1TresEquis: it doesn't really, but we already have something like that on the agenda17:26
TresEquisit might help with finding (most likely *paying*) release managers17:26
TresEquisa small stipend17:26
*** MyCatVerbs has joined #zope17:26
TresEquisbut the dev community has to drive process17:27
hannoschTresEquis: it's currently the only recognized "authority" and the community doesn't seem to self-organize17:27
mcdoncmy fear is that "the ztk" is actually several toolkits.. and while it's a reasonable goal to finish the ztk packaging and release as it is currently defined, i think we'd get a lot of mileage out of identifying those toolkits and decentralizing their maintenance and release process a bit17:27
* mgedmin notices the approaching 30 minute limit for the meeting17:27
RUNYAGAswhat else re ztk? could someone write up how to run trunk ztk tests? seems actionable17:27
Theuni1mgedmin: yes. thanks.17:27
mcdoncas a long term thing i mean17:27
TresEquisZ 2.13 *will* use ZTK if it gets released17:27
mgedminactually we are covering the topics pretty well17:27
TresEquisjust to answer that question17:28
Theuni1RUNYAGAs: should be possible, because there's something pre-configured out in the wild already. i'll do that.17:28
mgedminall that remains is to declare that Zope 3 is now called BlueBream 1.017:28
algacan we have a vote?17:28
TresEquismigration path17:28
algaabout the name17:28
mgedminalga, isn't it a bit late?17:28
mcdoncwe vote by volunteering i think17:28
TresEquisyup:  dynocracy ;)17:29
RUNYAGAsi agree17:29
TresEquiswho does the work, gets to vote ;)17:29
RUNYAGAshe who does wins17:29
Theuni1every commit gives you one vote17:29
RUNYAGAsthat it ?17:29
Theuni1ok, timelit.17:29
TresEquisthanks for doing this, Theuni117:29
Theuni1I'll go through the log and write up a summary of today.17:29
pyqwerTheuni1: Thanks for organizing this!17:30
Theuni1Hope you enjoyed that and we'll continue on the mailinglist and talk again next week.17:30
baijumTheuni1: thanks !17:30
Theuni1no worries. thanks for participating :)17:30
algaTheuni1: thanks!17:30
sidneiif there's one thing i can add to a virtual wishlist is to get gary's buildout branches reviewed and merged *wink*17:30
hannoschhey, we didn't get into a fight :)17:30
ignasso the mighty club in the hands of Grok is a Trout!? :D17:30
mgedminso, people who want buildbotting of their projects (ztk, bluebream, grok, zope2), email and tell me the steps to check out, build and run the test suite17:30
*** nilo has joined #zope17:30
mgedmin+1 for gary's buildout branches17:30
betabughmm, while you guys are you all here, I wanted to know if you are aware that the current 2.12 install procedure is a bit of a block for newbies?17:31
betabugs/are you/are/17:31
* J1m_ knocked off 2 of gary's branches on Sunday17:31
baijumbetabug: I tried to create a doc here:
*** J1m_ has left #zope17:31
algais irc'ing forbidden when driving?17:32
*** akm1 has quit IRC17:32
mgedminalga, if not, it should be17:33
*** MacYET2 has left #zope17:33
betabugbaijum: the instructions in the release docs somewhere gave various options: easy_install, virtualenv, buildout17:33
*** yusei has left #zope17:33
betabugthe buildout given was b0rken, people referred me to plone's buildout config instead17:33
baijumbetabug: In fact there are many options now..17:33
* baijum need to go .. sorry17:34
betabugwhich is not necessarily a good idea17:34
*** Takayuki has left #zope17:34
betabugbaijum: ok, thanx anyway!17:34
betabugwe were able to tell people here in #zope: "look at xy in your abc directory" - that's no longer possible17:34
mgedminspeaking of zope 2, I had lots of pain upgrading random people's older zopes to 2.thelatestwhatever (11? 12?) and then having their sites crash with Unicode errors left and right17:34
*** baijum has quit IRC17:35
betabugmgedmin: that was the change of ZPT in 2.1017:35
*** mr_jolly has joined #zope17:35
betabugmost likely17:35
mgedminactually zpt pages worked mostly fine, but those people I helped were using DTML documents everywhere17:35
betabugit needs some work to fix it in people's code17:35
betabughmmm, no experience with that then17:35
mgedminI probably shouldn't have brought up this topic17:35
betabugthe install procedure is a much bigger problem IMHO as one of the #zope answerbots17:36
mgedminbut it might've been a reason for those people's decision to build the next version of their site in drupal or django :/17:36
betabugit also lead to some distros dropping zope out of their package system17:37
*** mr_jolly has left #zope17:37
*** laurynas has quit IRC17:38
betabugwell, looks like there's no interest to discuss this17:38
*** phimic has quit IRC17:39
ignasnot sure what's there to discuss, vierualenv + buildout is a kludge, grok is doing it better, someone needs to fix it, but no one knows how ;)17:39
betabughow is grok doing it?17:40
ignasthey at least are packaging one big tarball for all the eggs17:40
ignaswhich makes installation go smoother17:40
ignasalso - IIRC Zope2 is doing all kinds of fake egg stuff17:40
TresEquisnot in 2.1017:40
*** RUNYAGAs has left #zope17:41
TresEquissorry, 2.1217:41
*** alga has quit IRC17:41
TresEquisyou need buildout to develop Zope2, but not to install it17:41
betabugto tell you the truth, I believe that if people had an option to get a tarball that worked like 2.11 as a choice to the eggified install procedure, most (by a large margin) would get the tarball17:41
mgedminI wonder how distros would look at one big tarball full of eggs17:41
mgedminnot well, I bet17:41
TresEquisit installs fine into a virtualenv17:41
ignasmgedmin, not well, it would conflict with all the eggs17:41
TresEquisthere is no reason not to ship an "egg basket" tarball and use it as a local index, for folks who don't want to go to PyPI for every egg17:42
TresEquisbut it takes effort to maintain17:43
ignasand being a release manager of such a beast17:43
ignasis one of the nastiest jobs ever17:43
betabugthe old style also had everything in a structure that related to what you import17:43
TresEquisbetabug: we aren't going back to that17:43
betabugwhich is great when hunting through the source code17:43
hannoschTresEquis: that's basically what the Plone installers do these days. all of them including the windows one use buildout under the hood - but ship everything in one big tarball / installer17:43
mgedminbetabug, if you use buildout, try the omelette recipe17:43
ignasbetabug, use omelette17:43
betabugTresEquis: why? what's the reason? (just asking to understand)17:43
mgedminor ctags17:44
TresEquiswe *are* going to keep using eggs17:44
moo---betabug: omelette symlinks everything to one folder space which looks like python module structure17:44
TresEquisthe monolithic tree is impossible to maintain17:44
betabugwell, the "period" doesn't help understand things17:44
moo---betabug: makes development easier17:44
kiorkyomelette is good but not perfect, i prefer ctags.17:44
ignasmakes integration with the rest of the python world easier17:44
*** Arfrever has joined #zope17:45
betabugdoes it make integration with the rest of the unix world easier?17:45
mgedminI wish ctags could find packages and modules, not just classes and functions17:45
hannoschbetabug: unix isn't the only world17:45
ignasbetabug, yes, in fact - 20 eggs are better than a tarball that contains code used by "other" packages17:45
betabugbut windows is only a lame copy anyway ;-)17:45
hannoschusing eggs means we can leverage whatever people built to make Python distributions work well under whatever OS package managers17:46
ignasbetabug, when noone used zope packages, except Zope users you could have a tarball17:46
betabugthat certainly is a point17:46
betabugand the omelette thing is good information17:47
mgedminsetuptools is the best thing that happened to the Python world during the last 10 years17:47
betabugwhat do people do to be able to install without a net connection?17:47
mgedminbut we're all still collectively struggling to find the best way of doing things17:47
TresEquiswe don't17:47
ignasbetabug, now you have zope.component and all the stuff, that is used by quite some packages, and as soon as you say "hey i'll package it for myself in my huge tarball" you are making it hard17:47
TresEquismgedmin: agreed, even if the python-dev folks are too l4m3 to know it17:47
ignasmgedmin, the least bad way ;)17:47
mgedminbetabug, there are tools that assemble a set of all the eggs needed for later offline installations17:47
mgedminI don't remember the names :/17:48
betabugmgedmin: ok, that would be something worth looking at17:48
mgedminzc.sourcerelease perhaps?17:48
TresEquisbetabug: we run "private" indexes for individual projects, to avoid depending directly on PyPI17:48
ignasbetabug, how do you install your deb/rpm packages that have 100 of dependencies offline?17:48
*** afd_ has joined #zope17:48
TresEquisbut we don't try to install without any net at all17:48
betabugignas: I don't use deb/rpm :-)17:48
TresEquis("we" being the teams I work with)17:48
mcdoncit is possible...
mcdoncalthough with 70 dependencies, not very practical17:49
betabugI'm sometimes on expensive connections (3G) or totally off the net for a while17:49
ignasbetabug, use buildout cache17:49
hannosch"we" used zc.sourcerelease a couple of times when deploying inside organizations with insane firewall setups. it does its job17:49
moo---betabug: also this pip tool has something called "pybundles"17:49
srichterbetabug: you can have a local PyPI mirror; or download the packages into a dir and use find-links17:50
ignasbetabug, buldout can cache eggs and share egg directories for installed eggs17:50
betabugvery good information coming here :-)17:50
moo---betabug: for example, delivarance can be downloaded as such 80 mb bundle17:50
ignasbetabug, download-cache = /home/ignas/.buildout/cache17:50
TresEquis'compoze' can be used to fetch source dists and build a local index17:50
ignasbetabug, in ~/.buildout/default.cfg17:50
betabugignas: great!!!!17:50
ignasbetabug, and you won't have to download the same egg twice, ever, (unless you run out of disk space, and a tarball gets "cut" which buildout does not know about)17:51
betabugall that stuff should be documented in one place :-)17:51
mgedmina few years ago everybody was doing their own web framework; now they're all doing their own packaging tools too!17:51
ignasbetabug, that overloads the newbie17:51
ignasbetabug, and it is in pypi zc.buildout page17:51
ignasbetabug, you mentioned you get confused with virtualenv + easy_install + buildout17:51
betabugignas: sure, but install instructions in 3-4 variations, of which 1 is only working through complaining on irc...17:51
ignasschooltool is using17:52
mgedminignas, pypi zc.buildout page is scary!17:52
ignas1 way that always works17:52
mgedmin is maybe a bit friendlier17:52
ignaseven if it is suboptimal for some people17:52
betabugthe thing is that "zope 2" has to have clear, consice install instructions, baijum definitely is right to put things into form again17:53
* ignas found it takes way less explaining when you only have 1 official way of installation, even if it has twice as many steps as it could have taken for *some* people17:53
*** sysspoof has quit IRC17:53
betabugthe python aces will recognize the infos they need and speed ahead anyway17:53
mcdoncwhat is the problem with doing easy_install zope2 now?  does that not work?17:54
betabugso, "documentation" is certainly one answer to my gripes17:54
ignasso I start with - virtualenv -> official buildout.cfg -> bootstrap -> buildout -> run!17:54
betabugmcdonc: uhmm, for starters... my system didn't have easy_install17:54
ignasbecause that works even with ubuntu zope.component installed17:54
ignasworks even when you don't have root17:54
betabugand the install instructions sounded like they are supposed to be there17:54
mcdoncmm.. that definitely sounds like a personal problem17:54
*** pthulin has quit IRC17:54
ignasworks even if you don't know what virtualenv is17:54
betabugnope, they're not in a standard python install17:54
betabugthey == easy_install17:54
TresEquismcdonc: heh, I was just about to say that17:54
mgedmineveryone agrees that easy_install into the global site-packages directory is Evil17:55
ignasmcdonc, all problems are personal, question is - how often do people have these personal problems17:55
TresEquiswhich is why virtualenv17:55
mcdoncwell, for other projects we just document the shit out of installing easy_install17:55
betabugdjango, bfg tell you where to get easy_install17:55
TresEquisdownload it, run it to create the place to install Zope, then run that place's 'easy_install'17:55
mcdonczope2 doesnt have such a culture and never did17:55
*** runyaga has joined #zope17:56
betabug"not mentioning requirements" isn't anything one could call culture17:56
betabugold zope2 install instructions where complete17:56
mcdoncwell the release manager quit17:56
betabugsorry, didn't know that17:56
betabuglooks like bad timing too, as the release packaging changed at the same time17:57
hannoschI didn't know there's new users to Zope2 at all - I wonder why anyone would want to go with it today ;)17:57
mcdoncwell the thing is that this packaging stuff has been in use literally now for many many years17:57
mcdoncthis is not an excuse for not documenting it "in context"17:57
betabughannosch: "new users" is not necessarily someone starting with a new framework, a sysadmin taking over a server is in there too17:57
*** sashav has quit IRC17:57
mcdoncbut at some level its so overdocumented in multiple places (because it's not in the stdlib) that it's a little silly17:58
betabughannosch: but in fact, yes, there are new zope 2 users some times, zope 2 is a complete, trusted, working system17:58
ignasbetabug, it is one of the reasons why zope3 got kicked out of some/most distros when they moigrated to eggs17:58
hannoschbetabug: well a sysadmin has almost no chance of updating to a new Zope version as it has a development cost each time17:58
ignasbetabug, it took years to make it installable again, and they did not make "migration from old instances to new instances" thing at all17:58
betabugzope 2 is not as much a moving target, to some application builders that's a good thing17:59
ignasfrom what I understand your gripe is not just about new users, it is about "how do I migrate from zope 2.10 to zope2.12 in one easy step!?"17:59
betabuge.g. most zope upgrades in the 2.7 - 2.11 range are pretty easy17:59
ignasor is it zope 2.11 to zope 2.12, can't recall which one messed things up ;)18:00
srichterbetabug: I agree, the Zope community often underestimates the value of API stability18:00
betabugwell, for me the "new users" part is in a large part on "how will I answer questions here in #zope?"18:00
mcdonctbh, i think its much ado about nothing.. i just easy_install'ed zope2, did bin/mkzopeinstance, bin/zopectl fg and it just worked18:00
TresEquissrichter: they underestimate the value of *documenting changes*18:00
ignassrichter, api stability is easy to fix most of the time, installation + packaging procedure stability is way worse IMHO18:00
mcdoncso as long as you a) know how to install easy_install and b) know how to invoke it and c) optionally know how to install virtualenv, it's a total nobrainer18:01
betabugsure, document changes, and change only what's really, really necessary in APIs18:01
betabugmcdonc: sure, and a bit of documentation will help too18:01
*** runyaga_ has joined #zope18:01
*** runyaga_ has quit IRC18:02
* mgedmin figured out the 2.9 -> 2.12 upgrade with a bit of pain and blood along the way18:02
*** runyaga has quit IRC18:03
*** bobo_b has joined #zope18:03
betabugI must admit I haven't moved yet18:03
*** aaronv has joined #zope18:03
betabugtoo lazy for my own stuff, too busy with other projects for my company stuff18:03
betabugbut I'll get there :-)18:03
ignasmgedmin, how many times did you do it with Zope3 before?18:04
mgedminthat was more painful18:05
mgedminbut it's not comparable: custom product built on top of z3 versus a z2 instance with objects in the ZODB18:05
*** baijum has joined #zope18:05
ignasmy point is - you did a more difficult version first, and then the "easier thing" was "bloody painful"18:06
mcdoncyou're right that the zope install documentation is total fucking nonsense18:06
mgedminit's true, I already knew how to use buildout/virtualenv/easy_install18:06
* ignas can't even imagine how not being able to disassemble reassemble whole stack in the middle of the night with your left hand feels like18:07
betabugwell, thank you all!18:08
betabugI'll look through my logs to sort it all out for me18:08
ignasand i am not even expecting empathy about install procedure from people who use "write your own buildout recipe, it's easy" in their replies18:09
betabugany zope-devs passing through .gr on vacation or so, I'll offer free drinks when you pass by Athens!18:09
ignasbetabug, look at schooltool installation procedures18:09
ignasnot sure you will like them18:09
ignasbut still it's one more battle tested example18:09
mcdonci love how people treat docs like they treat plone18:09
mcdoncjust add more18:09
mcdonc(hannosch excepted of course ;-) )18:09
betabugI don't want more, I want 1, not 3-4 install options :-)18:10
mgedminevery system administration problem can be solved by adding more documentation?18:10
mcdoncexcuse the paste18:11
mcdoncInstalling Zope using zc.buildout18:11
mcdoncUnless you are `developing zope`__, you most likely want to be18:11
mcdonccreating a :ref:`buildout-based Zope instance <buildout-instances>`18:11
mcdoncrather that installing using buildout as described in this section.18:11
TresEquisnon-dev installs shouldn't use buildout18:12
TresEquisunless you are installing something *else* which needs buildout18:12
hannoschmcdonc: that's ChrisW writing docs :)18:12
mgedminwhat are the downsides?18:12
mcdonci'll try to fix it18:12
TresEquisbuildout is overkill, and it makes a *weird* environment18:12
ignasis it so bad18:13
TresEquisvirtualenv + easy_install is much better18:13
mgedminas to the first, it's true18:13
mgedminbut given we're talking about Zope, I'm not sure what "weird" means ;)18:13
TresEquiseggs, parts?  Really?18:13
*** allisterb_ has joined #zope18:14
TresEquiswith virtualenv, the eggs go in lib/python2.x, which is at least "normal"18:14
ignasTresEquis, precisely my point I am for 1 option even if it is a bit more complicated, than for 5 options going in the order of simple -> difficult, works sometimes -> works always18:14
TresEquisfor Python folks18:14
hannoschmcdonc: that doc is pretty weird. it first tells you a lot of stuff you shouldn't do and then cuts short on the thing you should actually do18:14
*** allisterb has quit IRC18:15
TresEquisactually, the first one (installing with virtualenv) is the easy way18:15
binbrainINSTALL page is just organized oddly, put the virtualenv + easy_install at the top, anything below this point is for actual Zope developers18:15
TresEquisbut it doesn't need the damn'ed 'bin/activate'18:15
*** webmaven has joined #zope18:16
hannoschright. but that cold at least have a link to something that tells you how to get virtualenv18:16
TresEquisI saw at least two people at PyCon sprints get hoked by activate18:16
mgedminsuggestion: keep the main INSTALL document short, move things like buildout into an appendix for advanced users who insist they need buildout18:16
ignassudo apt-get install virtualenv18:16
TresEquiswe shouldn't document 'easy_install' at all outside of virtualenv18:17
TresEquisand buildout should be on a "developers" page18:17
ignashow do you "update" an easy_installed Zope?18:17
*** KageSenshi has joined #zope18:17
hannoscheasy_install -U ?18:17
TresEquisignas: huh?18:18
TresEquishow do you update anything else?18:18
ignasdon't know18:18
ignasnot using easy_install much18:18
TresEquisyou install a new version, and change configs to point to it18:18
TresEquisbuildout is completely wrong for a sysadmin who is not also a developer18:19
ignaswhat configs? i install zope 2.12 that is using zope.component 3.4.15, someone updates zope.component and puts a version 3.4.16 in the index18:19
ignaswhat command do I run to get zope.component "updated"? just  bin/easy_install -i<Zope version> Zope2 ?18:19
TresEquisbin/easy_install -i ... -U zope.component18:20
TresEquisHowever, you *don't* just randomly update stuff installed with Zope2, normally18:20
ignasno, i don't know that zope.component was updated, i know that 6 months passed and there might be some new updates in zope2 dependencies18:20
TresEquisunless you *know* it fixes a bug whch matters to you18:20
ignasdoesn't there being an index for Zope2 "keep it stable" ?18:21
*** milele has quit IRC18:21
hannoschthere's an index per release18:21
*** felixhummel_ has joined #zope18:22
*** incubaton has quit IRC18:25
*** incubaton has joined #zope18:26
koshhi guys18:26
*** felixhummel has quit IRC18:27
*** afd_ has quit IRC18:27
mgedminTresEquis, problem is, how do I know if there are bugs I want fixed?18:30
mgedminit's one thing if I find the bug myself18:30
hannoschmgedmin: that's what release notes are for18:31
mgedminit's another thing if somebody discovers a security hole in zope.somenameIneverheardof that happens to be used by Z218:31
ignasmeh, buildout is way less work, bin/buildout -n  -> it gets the newest versions.cfg and installs all the new stable secure versions of packages18:32
mgedmindebian is even better: apt-get upgrade18:33
koshI certainly hope we get debian packages again18:33
mgedminboth depend on other people doing the hard work for you18:33
koshthat was far easier to maintain18:33
mgedminschooltool is pushing a lot of the ztk into debian/ubuntu, afaiu18:33
mcdoncfor my money, i'd rather have control than letting someone else decide what "new stable" means... the last time there was a critical remote vulnerability in any zope package was in like 2006 due to docutils filesystem inclusion, afaik18:37
*** felixhummel has joined #zope18:37
*** replaceafill has quit IRC18:37
*** goschtl has quit IRC18:37
ignasand i'd like to allow people chose18:37
ignasapt-get install upgrade exists, but a lot of admins do not use it...18:38
koshwell debian never switched an instance to a another version so if you where running 2.9 the most you would get is updated to 2.918:40
koshit would not switch stuff to 2.1018:40
mcdoncone answer is to use a different framework that doesnt have 60 dependencies... another is to track the ztk18:40
koshI never had problems with it and it saved a lot of time18:40
mgedminthis is getting a bit maudlin18:41
mgedminwhat's a good way to version-control zope 2's TTW objects?18:41
*** felixhummel_ has quit IRC18:41
koshwell I think that debian will get packages again18:41
* mgedmin always found the TTW story rather compelling -- talk about rapid feedback!18:41
koshthe idea of dropping zope entirely is just not a very good option, I have looked at the other webframeworks and things like security in them seems to be an afterthought at most18:42
*** allisterb has joined #zope18:42
koshI love the TTW part of zope 218:42
*** dbfrombrc has quit IRC18:42
* mgedmin used ZCVSFolder a long time ago18:42
*** dbfrombrc has joined #zope18:42
mcdonc"this is getting a bit maudlin" heh18:42
*** pyqwer has quit IRC18:43
*** JaRoel|4D has quit IRC18:43
*** allisterb_ has quit IRC18:44
mgedminI needed to refactor a website once18:44
mgedminit was produced by a designer using some kind of a design tool, exported to a bunch of very redundant HTMLs and imported into Z2 as ZPTs18:44
Theuni1which reminds me18:44
mgedminIIRC I extracted the data over webdav, edited it with vim on the filesystem18:44
koshah that situation I have not deal with before18:44
mgedminand ended up writing a small Python script to produce a .zexp from the files in the filesystem18:44
*** pepeu has quit IRC18:45
ignasyeah, I want to know how to do it too, but my usecase is 100Kloc of TTW code18:45
mgedminenjoying your new maintenance contract, are you?18:46
ignasits as if Object Oriented Reengineering Patterns was written about that thing ;)18:46
mcdonc100K lines of ttw code... perhaps just kill yourself... or at least fake your own death18:47
Theuni1i think that code will do that for him perfectly fine. wo reason to perform suicide. ;)18:47
ignasyou should see the quality, they use sql queries to perform date operations18:47
ignaslike "add 1 day to this date"18:47
*** mcdonc has quit IRC18:48
sp0cksbeardnot sure if that option is still supported, but FSDump used to be one way18:48
sp0cksbeardthere used to be a complementary product to map filesystem objects into ZMI18:48
*** digitalmortician has quit IRC18:48
sp0cksbeardversion controlling the dumped objects and then mapping the working copy back into ZMI18:49
Theuni1ignas: i have to do something similar18:49
Theuni1i'm looking into using genericsetup for that18:49
*** mcdonc has joined #zope18:50
koshin general though a giant horrible mess of code is a problem in any system18:50
*** incubaton has left #zope18:50
*** replaceafill has joined #zope18:50
kosha nicer way to manage a mess is a good idea but I don't find it to be an important major point18:50
ignasTheuni1, thanks, will have to read about it18:50
*** runyaga has joined #zope18:50
ignaskosh, TTW makes it more difficult really18:50
ignaskosh, i can't use most of my tools :/18:51
*** emrojo has quit IRC18:51
koshignas: but it made it easier to write the system in the first place18:51
*** replaceafill has quit IRC18:51
*** emrojo has joined #zope18:52
ignasyou really think so? code duplication is rampant, refactoring was never performed, because it is pretty inconvenient to do TTW, there are no tests, and no docs...18:52
*** daMaestro has joined #zope18:52
*** alga has joined #zope18:52
koshsounds like a lot of filesystem codebases also18:53
ignasI know, I know - you can write bad code in any language18:53
*** davisagli has joined #zope18:53
Theuni1that's wrong18:55
Theuni1a good fortran programmer knows how to write fortran code in any language!18:55
*** astoon has left #zope18:59
ignas Theuni1: did you know that PyPI is written in php written in python?18:59
*** JaRoel|4D has joined #zope19:02
*** giampaolo has quit IRC19:02
*** davisagli has quit IRC19:03
*** davisagli has joined #zope19:04
*** JaRoel|4D has quit IRC19:07
*** yvl has left #zope19:11
mcdonci dont suppose anyone knows whether the source of comes from zope2 checkout's docs/INSTALL.rst or not do they?19:18
hannoschmcdonc: it comes from a SVN tag19:18
mcdoncif i changed docs/INSTALL.rst eventually when a new release was made, it would get rendered tho?19:19
hannoschmcdonc: it reflects the last release, but otherwise changes in SVN will show up on that page at some point19:19
mcdoncok good19:19
mcdonci'm going to break it into 2 sections19:20
mcdoncone for windows people and one for unix19:20
hannoschmcdonc: the real Sphinx stuff comes from
hannosch"svn pg svn:externals" tells you what is used there19:21
*** tiwula has joined #zope19:24
davisagliTheuni1: btw, thanks for organizing the meeting, and for sending out notes for those of us who failed at getting up at 7 a.m. ;)19:24
Theuni1davisagli: you're welcome. i hope it helps. :)19:25
davisagliTheuni1: same time next week?19:25
hannoschit's been quite civilized and well-behaved, I was surprised :)19:26
Theuni1hannosch: yeah :)19:26
*** emrojo has quit IRC19:26
Theuni1.oO(sad thing you're surprised though *G)19:26
*** agroszer has quit IRC19:26
*** emrojo has joined #zope19:26
hannoschwell, judging our mailing list communication style, I was surprised to see some communication culture - not everyone speaking at the same time trying to get his point across19:27
Theuni1wait for next week *g19:27
TresEquisbetabug: I just reorganized the install docs a bit:
*** kobold has left #zope19:29
TresEquismcdonc: you too :)19:29
TresEquisI didn't do anything particular for Windows folks, though19:31
TresEquisjust moved the zc.buildout bits out into a separate file19:31
TresEquisand split the bits which didn't care how you had installed it out, too19:31
hannoschif you have Python, pywintypes and virtualenv installed, it should be quite the same19:32
mcdoncTresEquis: maybe just merge that to the trunk?19:35
*** aaronv has quit IRC19:38
*** g_n_o_s_i_s has joined #zope19:39
*** jbglenn has joined #zope19:42
*** nilo has quit IRC19:45
*** sm-afk is now known as sm19:46
* sm catches up.. awesome19:46
*** allisterb has quit IRC19:46
*** allisterb_ has joined #zope19:46
TresEquismcdonc: I'm gonna merge to 2.12 branch and the trunk19:47
*** jbglenn has quit IRC19:47
hannoschTresEquis: +1 and don't listen to ChrisW ;)19:48
*** eperez has quit IRC19:50
*** emrojo has quit IRC19:55
*** r0ver has joined #zope19:55
*** MJ has quit IRC19:59
*** emrojo has joined #zope20:00
*** dunny has joined #zope20:00
*** hever has quit IRC20:01
*** sysspoof has joined #zope20:04
*** dunny has quit IRC20:04
*** Theuni1 has quit IRC20:05
*** deux_ has quit IRC20:05
*** nilo_ has joined #zope20:10
*** dunny has joined #zope20:11
*** jinty has left #zope20:16
*** Kabz|4D has quit IRC20:17
*** pthulin has joined #zope20:17
*** sylvain has quit IRC20:21
*** digitalmortician has joined #zope20:21
*** TomBlockley has quit IRC20:24
*** tiwula has quit IRC20:27
*** Kabz|4D has joined #zope20:29
*** ccomb has quit IRC20:35
*** ccomb has joined #zope20:35
*** TresEquis has quit IRC20:36
*** tisto has quit IRC20:37
*** g_n_o_s_i_s has quit IRC20:37
*** tiwula has joined #zope20:39
mgedminwhat's pywintypes?20:43
hannoscheh, pywin32 or whatever Mark Hammonds thing is called20:43
*** ccomb has quit IRC20:53
*** RaFromBRC has joined #zope20:55
*** MyCatVerbs has quit IRC20:56
*** MyCatVerbs has joined #zope20:57
*** jpcw has quit IRC20:59
*** baijum has quit IRC21:00
*** webmaven has quit IRC21:04
*** MyCatVerbs has quit IRC21:04
*** Kabz|4D has quit IRC21:06
TheJestervirtualenv is possibly the worst product name ever21:09
*** emrojo has quit IRC21:10
* mgedmin has seen worse21:11
TheJesterGiven it's not virtual in any way shape or form21:12
TheJesterjail, sandbox, something similar would have been far more accurate21:13
mgedminoh well too late now21:14
TheJesterchroot isn't very accurate either d8)21:14
mgedminat least it's googlable21:14
TheJesterAnd I expect anyone who starts telling new Zope users to use build-out to come here and support them when it doesn't work (which it never seems to)21:15
*** Arfrever has quit IRC21:15
mgedmingood point21:15
* TheJester expects to see ZRE being released any time now21:16
* TheJester laughs insanely21:17
* mgedmin doesn't even know what ZRE is21:17
TheJesterZope Runtime Environment d8)21:17
TheJesterala JRE d8)21:17
*** webmaven has joined #zope21:17
moo---TheJester: then maybe buildout should be fixed to report human-readable errors21:19
mgedminmoo---, I'm all for that!21:20
mgedminunfortunately it builds on top of setuptools, and there are maybe three people in the world who understand how setuptools works21:20
mgedminand one of them is busy creating a fork/rewrite21:20
*** kiorky_ has joined #zope21:23
moo---mgedmin: well, I call that progress21:23
*** kiorky_ has quit IRC21:24
mgedminit is21:24
TheJesterI'm sure it's great if you have exactly the same environment and don't want to deviate in any single way21:24
mgedminwhat is? buildout21:25
mgedminbuildout is great if you want to develop multiple packages that rely on the same dependencies but maybe different versions of them21:25
TheJesterOf course that's kind of difficult between operating systems and distros d8)21:25
mgedminbecause buildout can share the eggs and only install each one once21:25
mgedminotherwise you'd have either a multitude of virtualenvs, with copies of every egg in every env21:26
mgedminor you'd try to share your virtualenvs and feel the pain of incompatible versions21:26
TheJesterIt can share the eggs, but, telling it you already "have that" globally apparently requires some other nasty fake-egg juju21:26
mgedminwhat? no fake-egg juju21:26
mgedmina line in ~/.buildout/buildout.cfg21:26
TheJesterYOu need to stay here andd tell that to everyone who has problems with it d8)21:26
*** matthewwilkes_ has joined #zope21:26
mgedminbuildout should have its own irc channel21:27
*** MatthewWilkes has quit IRC21:27
TheJesterYes and the 3 ppl that understand it can stay there21:27
TheJesterthe user count is 1, is it you? d8)21:27
TheJesteror is it some pooor bugger waiting for help? d8)21:28
mgedminoh, does the channel exist?21:28
TheJesterIt seems to21:28
TheJesterhaha it has chanserv in it21:29
*** runyaga_ has joined #zope21:30
TheJesterThe joy of zope-in-a-tarball was that it was really simple for people to get it up and running and play with it21:30
TheJesterUntil it fell 3 versions behind on python d8)21:30
*** matthewwilkes_ is now known as MatthewWilkes21:33
*** cbcunc has quit IRC21:33
koshTheJester: I definitely agree that the tarball was far easier to get installed and working21:34
mcdoncTresEquis just changed the ZOpe install thing to feature easy_install as the one true way to get zope installed (with a pointer over to a separate doc for buildout)21:35
*** r0ver has quit IRC21:38
TheJesterSo it only works if you get the LATEST zope and install it before any of the sub-packages get updated21:40
*** r0ver has joined #zope21:40
mcdoncno idea what you mean21:40
TheJesterLucky you21:40
*** spamsch has joined #zope21:43
*** cbcunc has joined #zope21:46
*** jbglenn has joined #zope21:48
*** teix has quit IRC21:53
*** dunny has quit IRC22:04
*** dunny has joined #zope22:05
*** mcdonc has quit IRC22:06
*** teix has joined #zope22:07
*** benji has quit IRC22:12
*** benji__ has joined #zope22:13
*** benji has joined #zope22:16
*** webmaven has quit IRC22:18
*** teix has quit IRC22:23
*** avn has quit IRC22:27
*** aaronv has joined #zope22:28
*** avn has joined #zope22:28
*** alga has quit IRC22:32
*** webmaven has joined #zope22:34
*** TresEquis has joined #zope22:36
*** febb has quit IRC22:37
*** allisterb has joined #zope22:37
*** allisterb_ has quit IRC22:39
*** cbcunc has quit IRC22:50
*** nilo_ has quit IRC23:00
*** daMaestro has quit IRC23:06
*** sashav has joined #zope23:07
*** allisterb_ has joined #zope23:10
*** benji has quit IRC23:10
*** benji has joined #zope23:11
*** benji has quit IRC23:12
*** mcdonc has joined #zope23:12
*** benji has joined #zope23:12
*** kevingill has quit IRC23:12
*** allisterb has quit IRC23:12
*** cbcunc has joined #zope23:14
*** daMaestro has joined #zope23:14
*** mgedmin has quit IRC23:33
*** Kabz|4D has joined #zope23:36
*** jwhisnant has quit IRC23:48
*** Kabz|4D has quit IRC23:49
*** sysspoof has quit IRC23:50

Generated by 2.15.1 by Marius Gedminas - find it at!