[12:01] (cprov/#launchpad) sabdfl:  /home/cprov/lp_dump-sffmproject.sql.gz
[12:02] <kiko> sabdfl, we're all afrai^Wexcited, too.
[12:03] (sabdfl/#launchpad) i just had someone asking me about it 
[12:03] (cprov/#launchpad) sabdfl: I'm requesting some extra fields like SF "Last Releases" or FM "branches" to Morgan 
[12:04] (cprov/#launchpad) sabdfl: then we can have more details to add product and its series/releases, as real release date, version, developers notes, etc
[12:07] (cprov/#launchpad) sabdfl: some projects as ACPI has almost perfect information, but others are lost cause the names involved doesn't make much sense
[12:07] <BradB> kiko: now that you're an OS X user, you can optimize tla/arch!
[12:09] <jblack> bradb: Heh. You've already got me looking at the problem tomorrow. 
[12:09] <jblack> Would you rather me look tonight? 
[12:09] <BradB> jblack: you can, if you want
[12:09] <BradB> there's some good resources on developer.apple.com
[12:10] <BradB> i'll set you up an account
[12:11] <kiko> I'm crying tears of blood as we speak
[12:12] <BradB> it does zero-filling...no idea of ext3 does that
[12:12] <BradB> i can't imagine that being the bottleneck though
[12:12] <BradB> s,of,if,
[12:13] <jblack> bradb: Want me to email you my ssh key? 
[12:17] <kiko> bash-2.05a$ grep -ri open_memstream *
[12:17] <kiko> python/tag.cc:   FILE *F = open_memstream (&bp, &size);
[12:17] <kiko> wtf is open_memstream.
[12:19] (carlos/#launchpad) kiko: I was playing with python-apt some weeks ago.... and dude, it sucks!!
[12:19] (carlos/#launchpad) :-P
[12:20] <kiko> that is SUCH an understatement
[12:20] <kiko> it sucks more than the niagara falls
[12:22] (carlos/#launchpad) kiko: it does not have any documentation + it relies too much on real file objects 
[12:30] (cprov/#launchpad) kiko: what can you expect from C++ code ? apt, dselect, etc sucks as much as you can imagine :)
[12:31] <BradB> jblack, kiko: i'm killing you temporarily now
[12:32] <jblack> Ok.
[12:32] <kiko> pkgSourceList::~pkgSourceList [in-charge] ()
[12:32] <jblack> I'm out. tell me when I can get back in
[12:32] <kiko> carlos, dude, what does that undefined symbol mean?
[12:34] (carlos/#launchpad) kiko: no idea, perhaps it's a function that creates a pseudo file in memory, like a StringIO, ask Matt, I think he know it better, I was playing with it, not looking at the source code :-D
[12:44] <kiko> some threading problem
[12:44] <kiko> friggin OS X
[12:46] (sabdfl/#launchpad) what's a good way to get a python script to time itself?
[12:47] (sabdfl/#launchpad) import time()
[12:47] (sabdfl/#launchpad) erm 
[12:47] (sabdfl/#launchpad) import time
[12:47] (sabdfl/#launchpad) time.xxx?
[12:47] <kiko> time.clock()
[12:47] <kiko>    Return the CPU time or real time since the start of the process or since
[12:47] <kiko>     the first call to clock(). 
[12:48] <kiko> so you can place one clock() at the end, or calls around the critical region you want to test.
[12:48] <kiko> assuming you want to time something.
[12:55] <kiko> BradB, sabdfl: so where do I file a bug on apt-pkg not running on OS X?
[12:55] <kiko> it hangs.
[12:57] (sabdfl/#launchpad) kiko: is apt-pkg a python package in warty?
[12:57] <kiko> let's find out.
[12:57] (sabdfl/#launchpad) there's a python-apt
[12:57] <kiko> yes, that's it.
[12:57] (sabdfl/#launchpad) if that's got the code you want, file it in thw ubuntu bugzilla using python-apt as a component and assign to mdz
[12:57] (sabdfl/#launchpad) i just spoke with him, he's very relaxed :-)
[12:58] <kiko> I should join #canonical, but xchat hates me.
[01:13] (dilys/#launchpad) Merge to rocketfuel@canonical.com/launchpad--devel--0: New layout on Sourcepackage Page and Nicole improves (patch-620)
[01:28] <kiko> BradB, can you please create an account on bugzilla.ubuntulinux.org?
[01:29] <kiko> argh.
[01:31] <BradB> i just created bradb@bbnet.ca
[01:32] <kiko> thanks.
[01:35] <kiko-fud> bug filed, fud reloadage!
[03:36] (dilys/#launchpad) Merge to rocketfuel@canonical.com/launchpad--devel--0: doap tests (patch-621)
[04:08] <jblack> So, anybody else interested in the OS X problems with arch ?
[04:08] <kiko-fud> me.
[04:08] <kiko-fud> I'm in fud-mode
[04:08] <kiko-fud> but I'm interested
[04:08] <jblack> I'm still diving into it, but it looks like the problem is in the revision libraries.
[04:09] <jblack> I don't know what, as of yet.
[04:09] <jblack> First off, OS X is slow. really slow. Timings tell me that its *1/6* the speed of a linux system.
[04:09] <jblack> But there's something really weird going on in revision libraries.
[04:18] <jblack> bradb: How deep into swap are you? 
[04:20] <BradB> jblack: top?
[04:22] <kiko-fud> free?
[04:22] <kiko-fud> Processes:  69 total, 5 running, 64 sleeping... 180 threads            22:22:28
[04:22] <kiko-fud> Load Avg:  6.75, 3.72, 3.00     CPU usage:  70.7% user, 29.3% sys, 0.0% idle
[04:22] <kiko-fud> SharedLibs: num =  128, resident = 21.1M code, 1.36M data, 5.25M LinkEdit
[04:22] <kiko-fud> MemRegions: num = 5644, resident = 90.4M + 5.59M private, 71.6M shared
[04:22] <kiko-fud> PhysMem:  54.7M wired,  131M active, 63.2M inactive,  249M used, 6.79M free
[04:22] <kiko-fud> VM: 2.41G + 76.8M   570775(22) pageins, 555211(434) pageouts         
[04:22] <kiko-fud> loadavg 7? on a desktop?!
[04:23] <BradB> yikes
[04:23] <jblack> Heh. this is a low end system.
[04:24] <jblack> top takes 10% of the cpu
[04:25] <kiko-fud> low-end?!
[04:26] <jblack>   858 top         10.1%  0:23.02   1    15    18   348K   288K   444K  13.6M
[03:17] <sabdfl> hi SteveA_
[03:17] <sabdfl> tests seem nicely organised
[04:05] <SteveA_> hi
[04:06] <SteveA_> I'm about to go to a cafe (where it is warmer than at home) and make authenticated pagetests work.
[04:12] <sabdfl> SteveA_: this actually works!
[04:12] <sabdfl> you just need to be a bit careful
[04:13] <sabdfl> check the pagetests i checked in last night, many of them are authenticated
[04:13] <BradB> SteveA_: sabdfl had simple auth-using page tests working, as did I (except that Connection: closed problem, which apparently sabdfl didn't experience.) If, however, the specific example I gave in the bug report doesn't fail for you guys, I'd be a worried.
[04:13] <sabdfl> oh, ok, haven't read that bug report
[04:13] <sabdfl> will keep going till I hit it
[04:13] <sabdfl> BradB: morning :-)
[04:13] <BradB> hi :)
[04:13] <sabdfl> get ssh up and running?
[04:14] <BradB> yeah
[04:18] <SteveA_> I'd like to make the page tests do auth more obivously, so that you can read the test and see what's going on
[04:18] <SteveA_> right now, we have:   ... Authorization: Basic Zm9vLmJhckBjYW5vbmljYWwuY29tOnRlc3Q=
[04:19] <SteveA_> I want to replace that with simply: Launchpad-authorized-user: mark@canonical.com
[04:19] <BradB> does basic auth have an expiry time?
[04:19] <SteveA_> no
[04:19] <SteveA_> you'll still use normal auth in the browser
[04:19] <SteveA_> but, the tests will come out clearer
[04:20] <sabdfl> BradB: btw i needed to tweak one of your tests
[04:20] <sabdfl> the malone homepage one
[04:20] <BradB> changing data?
[04:20] <sabdfl> it was pretty much looking for a dump of the whole page
[04:20] <sabdfl> i tweaked it to do this:
[04:21] <sabdfl>   >>> print http(r"""
[04:21] <sabdfl>   ... GET /malone HTTP/1.1
[04:21] <sabdfl>   ... Cookie: wstyle=
[04:21] <sabdfl>   ... """)
[04:21] <sabdfl>   HTTP/1.1 200 Ok
[04:21] <sabdfl>   ...
[04:21] <sabdfl>   Content-Type: text/html;charset=utf-8
[04:21] <sabdfl>   ...
[04:21] <sabdfl>       <h1>Welcome to Malone!</h1>
[04:21] <sabdfl>   ...
[04:21] <sabdfl> so much, much shorter
[04:21] <sabdfl> basically just tests you got a 200 Ok and a h1 header that looks right
[04:21] <sabdfl> this way, we can mess around with the page content without it affecting the tests
[04:22] <sabdfl> i've tried to do all my tests this way, it will reduce the number of false failures
[04:22] <BradB> it's a fine line to walk :)
[04:22] <sabdfl> well, we can create tests that look for more specific features of a page
[04:22] <sabdfl> but right now we mainly want to ascertain whether or not a change breaks tests altogether
[04:23] <sabdfl> breaks pages, rather
[04:23] <sabdfl> so later we can test that, for example, given a virtual host x you get something on the page like x
[04:23] <sabdfl> and given a virtual host y you get something different
[04:23] <sabdfl> but right now we don't have that distinction
[04:24] <sabdfl> so just knowing the page rendered 200 OK is a huge help
[04:24] <BradB> i guess we'll find out over the next little while if that's enough coverage. :)
[04:26] <BradB> SteveA_: do you have a few minutes now to verify whether the bug report for auth'd page tests fails for you? i may not be around later to clarify if it somehow manages to not fail for you.
[04:30] <SteveA_> ok
[04:31] <SteveA_> which bug number?
[04:32] <BradB> 2114
[04:36] <SteveA_> BradB: I get at assertion error at step 9
[04:37] <BradB> ouch, that's another bug i reported
[04:37] <SteveA_>   File "/stuff/code/Launchpad/launchpad/sourcecode/zope/src/zope/app/tests/dochttp.py", line 97, in dochttp
[04:37] <SteveA_>     assert (request and response) or not (request or response)
[04:37] <SteveA_> AssertionError
[04:37] <BradB> 2115
[04:38] <SteveA_> yep
[04:38] <SteveA_> I can certainly look into bug 2115
[04:38] <SteveA_> until i've fixed that, I can't look into 2114
[04:38] <BradB> ok
[04:39] <SteveA_> thanks -- the bug report of 2114 is very clear and easy to follow
[04:40] <BradB> cool
[04:41] (dilys/#launchpad) Merge to rocketfuel@canonical.com/launchpad--devel--0: great WordWord renaming (patch-622)
[04:42] <sabdfl> dilys: you rock
[08:21] !dmwaters:*! Hi all! I need to do some quick rehubbing, this won't take long.