[00:03] <ardk> Hi, I'm trying to install launchpad on a new server (9,04). When I run "make schema" all stop. I don't have py. launchpad/lp-branches/devel/bin/py: not found.
[00:03] <ardk> In the directory bin, I only have buildout. Thanks for any help
[00:05] <BBHoss> mars or sinzui: could it possibly be a memory issue? http://pasternak.superalloy.nl/pastes/2487
[00:05] <mars> ardk, could you please paste the output of running 'make' in ~/launchpad/lp-branches/devel ?
[00:06] <sinzui> BBHoss: maybe. I think so
[00:06] <ardk> utilities/shhh.py python2.4 bootstrap.py --ez_setup-source=ez_setup.py \
[00:06] <ardk> 		--download-base=download-cache/dist --eggs=eggs
[00:06] <ardk> utilities/shhh.py ./bin/buildout configuration:instance_name=development
[00:06] <ardk> utilities/shhh.py make -C sourcecode build PYTHON=python2.4 \
[00:06] <ardk> 	    PYTHON_VERSION=2.4 LPCONFIG=development
[00:06] <ardk> utilities/shhh.py LPCONFIG=development /home/ardk/launchpad/lp-branches/devel/bin/py -t buildmailman.py
[00:06] <ardk> make: *** [compile] Error 127
[00:06] <sinzui> BBHoss: launchpad requires about 2G to develop with.
[00:06] <mars> ardk, could you please use http://paste.ubuntu.com ?
[00:06] <BBHoss> sinzui: i just want to play around with it, can it not just use swap?
[00:07] <mars> ardk, it keeps the channel traffic down
[00:07] <sinzui> BBHoss: I do not now the details of downloading and memory with the python process. But BBHoss will barely run in 1G of memory
[00:07] <mars> BBHoss, I think you would have to ask a bzr guru if that is causing an error.  As for running a LP instance on 256MB of RAM
[00:09] <BBHoss> hmm yeah
[00:09] <BBHoss> i know it would be slow, just wanted to know what it would do
[00:09] <ardk> Sorry, http://paste.ubuntu.com/231802/plain/
[00:10] <sinzui> BBHoss: I did get Launchpad to run on 512 two years ago. I could run the test suite or all the services. I could browse and use bugs and answers.
[00:11] <mars> sinzui, could run the test suite?  Doesn't that need a gig or so?
[00:11] <sinzui> mars no. I thought I was an idiot and could not setup Launchpad
[00:11] <mars> ardk, rocketfuel-setup ran to completion, without errors?
[00:11] <sinzui> mars: Once I bought a new computer the next week
[00:12] <ardk> Yes, no error with rocketfuel, and no error with ./utilities/launchpad-database-setup $USER
[00:13] <sinzui> mars: and the one nice affect of the GIL is that on a two core machine, you can still work while the test suite is running
[00:13] <mars> ardk, try "$ bin/buildout" in launchpad/devel?
[00:15] <mars> sinzui, FWIW, I'll try to finish off my CHR tomorrow
[00:16] <sinzui> mars: I'll do the projects...I usually do about 30 on the weekend so the person on Monday does not cry
[00:16] <ardk> mars: an error whith buildout http://paste.ubuntu.com/231823/plain/
[00:16] <mars> sinzui, I'll see what I can do on the projects, too
[00:17] <mars> ardk, ok, try 'rocketfuel-get' then.  It looks like your devel/sourcecode directory may be broken, or possibly devel/download-cache
[00:17] <mars> I should have suggested rocketfuel-get first
[00:18] <mars> that is turning out to be the primary indicator of a working devel setup
[00:19] <ardk> mars: I'm trying rocketfuel-get thanks
[00:20] <ardk> mars: Exactly the same errors after rocketfuel-get
[00:21] <mars> ardk, could you please paste the full output?
[00:24] <ardk> mars: the output of rockefuel-get http://paste.ubuntu.com/231841/plain/
[00:26] <BBHoss> i dont see why bzr wouldn't be able to handle the small memory, wget seems to get it fine
[00:27] <mars> ardk, ok, so ~/launchpad/lp-sourcedeps/sourcecode has links, and ~/launchpad/lp-branches/devel/sourcecode has links in it?
[00:27] <mars> ardk, and ~/launchpad/lp-branches/devel/download-cache is full of .tar.gz files?
[00:28] <ardk> mars: yes , and in download-cache I have a subdir "dist" full of files .tar.gz
[00:28] <mars> ok
[00:28] <mars> that's as it should be
[00:28] <mars> ardk, and there is an 'eggs/' directory?
[00:31] <ardk> mars: Yes, list of eggs http://paste.ubuntu.com/231862/plain/
[00:32] <mars> ardk, ok, I have 118 .egg files in mine
[00:33] <mars> ardk, I have to leave now, maybe try 'python bootstrap.py', see what happens.  Otherwise, hang around here, or ask on the dev list: launchpad-dev@lists.launchpad.net
[00:34] <mars> good luck!
[00:34] <ardk> Thanks
[00:39] <mars> ardk, oops, typo, make sure that is 'python2.4 bootstrap.py'
[00:39] <mars> otherwise you'll have a mess on your hands, untangling the 2.5 files from the 2.4 ones.
[00:39] <ardk> mars: Yes I've corriged, and I have a problem with lazr.smtptest 1.1
[00:58] <maxb> Launchpad seems to contain both a monolithic zope tree and various zope modules - can anyone explain how that works?
[00:58] <maxb> I'm trying to investigate zope.proxy misbehaviour, but it's not clear which zope.proxy is actually in use
[01:22] <wgrant> maxb: It seems to be the one in the monolithic Zope.
[01:22] <wgrant> Which makes little sense.
[01:22] <wgrant> In fact it looks like all of the eggs are shadowed.
[01:28] <sidnei> wgrant: are you referring to that zope.proxy issue? i think i missed the context of that
[01:29] <wgrant> sidnei: Ah, yes, you were away while he said it. maxb is unsure as to which zope.proxy is in use in Launchpad.
[01:30] <wgrant> There is a 3.5 egg, and what I presume to be a 3.4 from the monolithic tarball.
[01:30] <wgrant> But the monolithic zope 3.4 is in lib/, which is always earlier in sys.path AFAICT.
[01:31] <sidnei> wgrant: that's correct. we should use the 3.5 egg as soon as gary merges his branch
[01:31] <wgrant> sidnei: So those eggs are just sitting there doing nothing at the moment?
[01:32] <sidnei> wgrant: yes, at the moment. i also suggested patching the zope 3.4 from lib with the fix from the 3.5 egg
[01:32] <sidnei> in the meantime that is
[01:33] <sidnei> wgrant: i believe this is the branch gary is working on: https://code.edge.launchpad.net/~gary/launchpad/zbuildout
[01:33] <sidnei> wgrant: it's only pending some test failures afaict to be merged
[01:33]  * wgrant mutters something about private branches.
[01:37] <sidnei> wgrant: oh, drats. let me see if i can toggle that bit
[01:39] <wgrant> sidnei: You probably can't.
[01:39] <sidnei> i wonder if bzr does remote branching
[01:39] <wgrant> It's unlikely a non-admin can now that the security policy is relaxed.
[01:42]  * sidnei watches as something happens with https://code.edge.launchpad.net/~sidnei/launchpad/zbuildout
[01:45] <wgrant> sidnei: Aha, thanks.
[01:46] <sidnei> still pushing, but seems like it might be done soon
[02:36] <jml> sinzui, actually, I added the rule to link lp:12345 to a bug as well. there ought to be no conflict due to [0-9] and [a-z] having no intersection.
[02:39]  * wgrant discovers why /var/tmp/fatsam is called that.
[08:20] <maxb> Well, the interesting thing is that even applying the zope.proxy fix from upstream, I still seem to experience the bug
[08:24] <wgrant> maxb: Is the bug causing any issues?
[08:25] <maxb> It's causing MemoryErrors all over the place for me
[08:25] <wgrant> Oh.
[08:25] <wgrant> Of course.
[08:25] <wgrant> Stupid Python.
[08:25] <wgrant> That's why Storm doesn't define it in the first place.
[08:25] <maxb> Well, really I can't blame it for having difficulty instantiating a 4 billion element list
[08:26] <wgrant> Perhaps.
[08:27] <maxb> The annoying thing is that if I manually instantiate a zope.proxy.ProxyBase(object()), it seems happy to return correct results
[08:28] <wgrant> From 'make harness'?
[08:29] <maxb> I've not come across that one, I'm just working with "make run", or a standalone script in the style of utilities/make-lp-user
[08:29] <wgrant> Ah.
[08:29] <wgrant> make harness will get you an interactive Python with Zope configured and lots of useful objects.
[08:30] <wgrant> I believe there's also a make iharness, which is probably even more useful.
[08:31] <maxb> The other somewhat perplexing thing is that I'm now managing to get a zope.security.interfaces.ForbiddenAttribute: ('__len__', <storm.store.ResultSet object at 0x611e090>)
[08:31] <wgrant> That's good.
[08:31] <wgrant> That's what the proxy is there for.
[08:31] <maxb> when the only change from code that didn't do that was to delete some operations frm my minimal test script
[08:33] <wgrant> Hmm, actually.
[08:34] <maxb> Get a storm resultset, print len(rs) -> 2**32-1, transaction.commit() results in ForbiddenAttribute.
[08:34] <maxb> However, add "for i in rs: pass" before the commit, and no ForbiddenAttribute happens
[08:35] <wgrant> Special...
[08:35] <maxb> hehe, yes, "special"
[08:37]  * wgrant looks with bemusement at build__datecreated__key
[08:37] <wgrant> That makes no sense.
[08:38] <cody-somerville> wgrant, Are you pleased to finally have access to lp source code? :)
[08:39] <wgrant> cody-somerville: I am!
[08:39] <maxb> aha, the same 64bitness bug exists in zope.security.proxy as well as zope.proxy
[08:40] <cody-somerville> wgrant, I was told that I'm the only person outside of the launchpad team to contribute code to launchpad in a couple of years. Hopefully that'll change real quick now! :)
[08:41] <wgrant> cody-somerville: I've already two approved branches that will land once devel reopens on Monday.
[08:41] <cody-somerville> wgrant, awesomeness
[08:41] <cody-somerville> wgrant, I'm sure you'll help pave the way for all sorts of folks to be able to contribute effectively
[08:41]  * wgrant hates gina's config system.
[08:42] <cody-somerville> zconfig or... ?
[08:42] <wgrant> lazr.config. But it's not a good fit.
[08:42] <cody-somerville> Yea
[08:43] <cody-somerville> I examined lazr.config for one of my internal projects and I decided to go with a different solution
[08:43] <wgrant> lazr.config is good. Just not for this.
[08:47] <maxb> Hmm, no vcs-import for zope.security
[08:48] <wgrant> It would have been part of the main Zope 3 import until recently.
[08:58] <maxb> OK, woo, now I actually have a functioning webapp running on karmic :-)
[08:58] <wgrant> Which Python?
[08:58] <maxb> 2.5
[08:58] <wgrant> Aha.
[08:58] <wgrant> Using all the default deps except zope.security?
[08:59] <maxb> I just grabbed and applied patches of the relevant revisions from zope.security and zope.proxy to the sourcecode/zope tree
[09:00] <wgrant> Is this the first time you've had the webapp running?
[09:01] <maxb> No, but before now it tended to OOPS a lot
[09:01] <wgrant> I count that as a yes.
[09:01] <maxb> Even an OOPSing webapp was proof I was on the right track :-)
[09:02] <wgrant> It's nice running a local Launchpad - it's so fast!
[09:03] <maxb> On my good laptop, yes
[09:03] <maxb> Not so much on my netbook!
[09:04] <wgrant> Probably not.
[09:21] <wgrant> maxb: Found anything yet that doesn't like 2.5?
[09:42]  * wgrant needs an SSD
[10:03] <maxb> I suppose I should now figure out how to run the testsuite
[10:05] <wgrant> maxb: make check
[10:05] <wgrant> Making sure you have good cooling.
[10:05] <wgrant> And many, many hours.
[10:06] <wgrant> I should run the full test suite here at some point - I've so far only run the bugs tests in full, which took 50 minutes.
[10:08] <maxb> hmm. possibly a laptop sitting on a bed is not the best environment for long duration high CPU work
[10:15] <wgrant> A lot of it's high-disk, as the database has to be replaced after each test.
[10:15] <wgrant> I'm not sure if I was reading things properly, but the output of the lp.bugs tests suggested that 30 of the 50 minutes were spent doing DB setup.
[14:14] <maxb> Does anyone know when the dev wiki is going to be opened for editing? I'd like to scribble some notes on my experiences of getting LP to work with Python 2.5
[14:45] <cprov> good morning, folks
[14:48] <wgrant> Morning cprov.
[16:20] <sidnei> hey wgrant, pushing that branch finished at some point during the night: https://code.edge.launchpad.net/~sidnei/launchpad/zbuildout
[18:59] <maxb> sidnei: am I simply not driving bazaar properly, or is there litte interesting in that branch yet?
[19:19] <sidnei> maxb: what do you mean?
[19:23] <maxb> Unless I'm failinf to use bzr properly, it seems to contain a group of changes not related to the buildsystem, and a one-line fixup
[19:24] <maxb> bzr log -n0 -r ancestor:../devel..
[19:28] <sidnei> maxb: that command doesn't show much indeed. use diff instead of log, you will see that there's a lot of changes
[19:29] <sidnei> maxb: it looks quite unpolished though, there's quite some code that has been commented out, presumably because it doesn't pass the tests yet
[19:29] <sidnei> maxb: i think using a patched sourcecode/zope3 is a better approach until gary makes more progress
[21:29] <timut> hi all
[21:30] <timut> i'm installing launchpad on hardy 8.04, anybody have a list of dependency or lib packages prerequisite before install?
[21:56] <maxb> are you following https://dev.launchpad.net/Getting ?
[21:57] <maxb> dependencies are dealt with as part of rocketfuel-setup
[22:03] <timut> maxb: yups, but i have problem with bzr and python-pyrex
[22:05] <timut> Setting up python-pyrex (0.9.6.4-1ubuntu1) ...
[22:05] <timut> pycentral: pycentral pkginstall: not overwriting local files
[22:05] <timut> pycentral pkginstall: not overwriting local files
[22:05] <timut> dpkg: error processing python-pyrex (--configure):
[22:05] <timut>  subprocess post-installation script returned error exit status 1
[22:05] <timut> Errors were encountered while processing:
[22:05] <timut>  python-pyrex
[22:06] <timut> E: Sub-process /usr/bin/dpkg returned an error code (1)
[22:17] <maxb> timut: I would suggest temporarily moving aside anything including pyrex in its name from the /usr/lib/python2.4/site-packages/ directory and retrying the package setup with dpkg --configure -a