/srv/irclogs.ubuntu.com/2009/10/05/#launchpad-dev.txt

lifelessjml: testtools is in debian testing now00:30
lifelessjml: and you might like lp:~lifeless/subunit/protocol00:31
=== statik` is now known as statik
mwhudsonoh oh for someone to have a pre-imp call with03:22
mwhudsonpreferably jml03:23
lifelessmwhudson: I cn do a quick one if you are desperate03:30
mwhudsonlifeless: it's ok, i think03:31
mwhudsonthanks though03:31
mwhudsonlifeless: i don't guess you know if there's a method already to find all official source package branches for a distroseries?03:31
mwhudsoneasy enough to add, and i've looked in the obvious places so...03:32
lifelessI don't know, no03:33
mwhudsonsoyuz is the mind killer04:04
=== abentley1 is now known as abentley
mwhudsonbug 21764404:50
mupBug #217644: ResultSet aggregates do not respect distinct option <Launchpad Foundations:Invalid> <Storm:Fix Released by jamesh> <https://launchpad.net/bugs/217644>04:50
jameshmwhudson: that bug should be fixed in the version of Storm you're using.04:51
mwhudsonjamesh: i can probably delete a bodge then, that's good04:51
jameshIt now applies the aggregate to a subselect, so distinct, offset and limit all work.04:53
=== Ursinha is now known as Ursinha-sprint
adeuringgood morning08:31
noodles775_hi adeuring08:31
adeuringhi noodles775_!08:31
al-maisanGood morning!08:42
mrevellMorning09:28
Fly-Man-morning09:36
=== Ursinha is now known as Ursinha-sprint
=== intellec` is now known as intellectronica
Ursinha-sprinthey bigjools, what version of kde are you running?10:22
bigjoolsUrsinha-sprint: 4.3 in Karmic10:22
Ursinha-sprinthm10:22
bigjools4.3.1 in fact10:22
Ursinha-sprintbigjools, I'm using 4.3 in jaunty, and almost no icons are showing in the system tray applet10:23
Ursinha-sprint4.3.1 as well10:23
Ursinha-sprintdo you have the same problem?10:23
bigjoolsUrsinha-sprint: are you using intel gfx?  (no I don't)(10:23
Ursinha-sprinthmm, don't think so, why?10:23
bigjoolsit's buggy as hell in jaunty10:23
Ursinha-sprintoh10:23
bigjoolsand I had some issues like that10:24
bigjoolsUrsinha-sprint: check #kubuntu for help I guess10:24
Ursinha-sprintthanks bigjools10:28
bigjoolsUrsinha-sprint: no prob - I do have other issues though, since 4.3 it hangs when logging out/shutdown etc :(10:28
Ursinha-sprintoh :(10:29
Ursinha-sprintbigjools, the only problem I have is with amarok vs. all the other stuff that needs sound10:29
Ursinha-sprintbesides this one with the tray10:30
bigjoolsUrsinha-sprint: oh yeah, I used to have that.  apt-get purge pulseaudio fixes that ;)10:31
Ursinha-sprintbigjools, lol10:31
Ursinha-sprinteverybody hates pulseaudio10:31
* wgrant thwacks bigjools with the "there is no problem unless you have filed a bug" bat10:31
Ursinha-sprintyes, it's a pita most of the time10:31
bigjoolsI can file bugs until I am blue in the face10:31
=== mrevell changed the topic of #launchpad-dev to: This is Launchpad Development Channel | Week 0 of 3.1.10 | PQM is OPEN | https://dev.launchpad.net/ | Get the code: https://dev.launchpad.net/Getting | On-call review in #launchpad-reviews | Use http://paste.ubuntu.com/ for pastes | This channel is logged: http://irclogs.ubuntu.com/
deryckMorning, all.11:02
bigjoolshowdy deryck11:04
allenapHi jml, I'm having problems with devscripts.autoland. It's stopping utilities/ec2 from working because it requires a couple of libs that are not installed, launchpadlib and lazr.uri.11:50
jmlallenap, wow11:51
jmlallenap, I thought they were developer dependencies?11:51
allenapjml: I don't think so... I'll look...11:51
jmlallenap, karmic or jaunty?11:51
allenapjml: Jaunty.11:52
mrevellallenap: If I pastebin you my short traceback from ec2, can you tell me if it's the same thing you're getting?11:54
allenapjml: No, they're not in any of the deps afaict. lazr.uri does not seem to be available for installation on my machine, without sorting out some PPAs.11:54
allenapmrevell: Sure.11:54
mrevellallenap: http://pastebin.ubuntu.com/286108/11:54
jmlallenap, lazr.uri is in download-cache11:55
mrevellallenap: from what you've said, I assume it is11:55
allenapmrevell: Yes, I think that's related.11:55
mrevellallenap: cool, I'll hang onto your coat-tails...11:55
allenapjml: Yeah, but utilities/ec2 runs with the system python, not the buildout managed one.11:55
jmlallenap, oh right.11:55
jmlffs.11:55
jmlallenap, I guess the right thing to do is make the imports conditional then.11:56
allenapjml: Yeah, I had some grief with a similar thing last week.11:56
jmlallenap, I'm 90% sure that when we switch to Python 2.5 this problem will go away.11:56
allenapjml: Yeah. I can so that if you're busy.11:56
jmlallenap, thanks.11:56
jmlallenap, I've actually got today off, so I'd much prefer it if you do it :)11:57
allenapjml: Yes, we can run everything with the buildout bin/py.11:57
jmlallenap, I can review the patch if you'd like though.11:57
allenapjml: Sure, no worries. Thanks for being around on your day off :)11:57
jmlallenap, np.11:57
jmlallenap, you should install launchpadlib & lazr.uri btw11:57
jml'ec2 land' is sweet as.11:57
allenapjml: I'll try to make it so the command appears with a note that it won't work until those packages are installed.11:58
allenapjml: I look forward to trying it :)11:58
jmlgood idea11:58
maxbThe problem with recommending that people install the system package of lazr.uri is that it breaks launchpad dev12:00
allenapmaxb: Ah, that's troubling.12:01
maxbIt's the namespace package / setuptools / aaargh situation12:01
allenapmaxb: I think I'm going to have a painful day today.12:02
maxbuhoh :-/12:02
allenapBut first, lunch. Never hack setuptools on an empty stomach.12:03
maxbgood policy!12:03
wgrantI would generalise that statement.12:03
=== Ursinha-sprint is now known as Ursinha-afk
gmbwgrant: Indeed. s/hack setuptools/.*/12:12
* gmb -> lunch12:12
wgrantgmb: Not quite.12:12
gmbwgrant: Well, I suppose eating on an empty stomach is acceptable.12:12
mrevellallenap: When you're back from your lunch, and the you've resolved the pain in your day, could you please share the analgesic with me? :)12:14
=== mrevell is now known as mrevell-lunch
=== mrevell-lunch is now known as mrevell
allenapmrevell: I'm working on it now :)13:13
allenapmrevell: I'll see if I can get a quick fix for you.13:13
mrevellallenap: Hey, no rush beyond what you're doing already :)13:17
allenapmaxb: You said that installing lazr.uri causes launchpad dev to break. Can you elaborate on that?13:24
=== Ursinha-afk is now known as Ursinha-sprint
maxballenap: Hi13:52
allenapmaxb: Hello again.13:52
maxbSo the problem is that setuptools installs a *-nspkg.pth file which registers the toplevel lazr package with a __path__ pointing to the system directory13:53
maxbThis defeats attempts to add any other lazr.* via PYTHONPATH13:53
maxbAnd then you have a bit of a conundrum. I haven't figured out how to solve it, or if it is soluble13:54
allenapmaxb: Ah, that's grizzly. If it's installed via a deb then this won't be a problem, will it? Not that there's a lazr.uri deb easily available.13:54
maxbYes, it is a problem, and there's lazr.something debs available in karmic, hence me noticing the problem13:55
maxbpython-lazr-uri | 1.0-0ubuntu1 |        karmic | all13:56
ybenitezfhello14:28
allenapmaxb: Thanks.14:28
ybenitezfi from Cuba and need help with getting LP as explained in https://dev.launchpad.net/Getting, may can some of you helpme14:30
ybenitezf?14:30
maxbWhat do you need help with?14:32
ybenitezfmy internet connection is so poor that running rocketfuel-setup is not an option forme, i get the tar.gz from the page, but none wiki page explain what to do from there, also the tar.gz is a little outdate14:34
ybenitezfcan you give me links or something for getting LP running ?14:36
allenapjml: Can I call in that review please? https://code.edge.launchpad.net/~allenap/launchpad/packages-for-autoload-bug-443061/+merge/1286714:38
intellectronicaybenitezf: i think you can use both. download the tarball (though i think it's pretty old) and use rocketfuel-setup to do the rest (bzr will be smart and not try to download revisions it already has)14:42
maxbybenitezf: You should read https://dev.launchpad.net/Getting and the source of rocketfuel-setup14:43
intellectronicaybenitezf: anyway, you don't have any way of doing LP development without getting the complete branch, regardless of whether you do it with bzr or by getting a tarball. the first download will be quite bug, but after that you'll only need to update a few revisions every time, which is not so terrible even with a slow connection14:44
ybenitezfi have tried that, but rocketfuel-setup (from tarball) complains about bzr version even when i have instaled version 2.0.x.x, i have been reading https://dev.launchpad.net/Getting all the weeked14:46
maxbYou'll need to get a newer rocketfuel-setup.14:47
ybenitezfyep14:47
maxbhttps://dev.launchpad.net/Getting tells you how14:47
ybenitezfbut when executed it fails downloading14:47
maxbYou're not giving us enough detail about your problems to enable us to hep14:48
maxb*help14:48
ybenitezfsorry14:51
ybenitezfmine school use a proxy server, i have not direct internet connection, all that i need is an updated tar.gz like the one in https://dev.launchpad.net/Getting14:51
ybenitezfcan you help me  with that14:51
ybenitezf?14:51
maxbNo, because that's not actually what you need.14:52
maxbThe size of the download should not actually be that different whether it's by tarball or bzr.14:52
ybenitezfahh14:52
ybenitezfthere is a problem with that, with the tar.gz i can use a download manager, running rocketfuel-setup can fail and i have to start all over again14:53
ybenitezfi'm a poor connection man, ;-)14:54
maxbDo you have the tar.gz already?14:54
ybenitezfyes the old one14:54
maxbThen just run "bzr pull" in it, and let it download all the updates14:57
ybenitezfok14:57
ybenitezfi will try again14:57
ybenitezfsee you later maxb, thanks for your help14:58
maxbI think we should remove herb's tarball (and possibly replace it)15:02
maxbFor one, it's of db-devel not devel. For another, it's layed out as a standalone branch, not a shared repository15:03
maxbAnd if we're optimizing for download size, there's no need for it to contain a working tree either15:03
leonardrbigjools, can i ask you a question about ppas?15:08
bigjoolsleonardr: go for it15:09
leonardrbigjools: is it possible to delete one of your ppas?15:09
leonardri'm working on https://answers.launchpad.net/soyuz/+question/84542 as part of chr15:09
bigjoolsleonardr: no, the best we can do is to disable them15:09
bigjoolsleonardr: assign to LOSAs15:09
leonardrok, thanks15:09
bigjoolsnp15:09
ybenitezfmaxb, i have untar the tar.gz and run "bzr pull" from ~/launchpad/lp-branches/db-devel but fails with an error: bzr: ERROR: Not a branch: "/home/archives/rocketfuel/wiki/db-devel/"15:19
maxboh15:19
ybenitezfjust for asking, anyone can facilitate launchpad.tag.gz update15:20
maxbleonardr: There's a slight edge case. If a PPA has had no uploads ever, it can be deleted15:21
=== stub1 is now known as stub
maxbleonardr: https://answers.edge.launchpad.net/soyuz/+question/79720 <-- question where I asked for one of my PPAs to be deleted and cprov explained things15:23
leonardrmaxb, thanks15:23
=== jtatum` is now known as jtatum
maxbybenitezf: bzr pull --remember lp:~launchpad-pqm/launchpad/devel15:25
maxbAlso, rename the db-devel directory to devel15:25
ybenitezfmaxb: ok15:26
ybenitezfmaxb: run that from devel it self or from parent dir ?15:29
maxbFrom inside devel15:29
ybenitezfmaxb: I'm at it, seems to be taking time15:30
maxbHrm, we *really* need to remove or update that tarball. A modern bzr version packs the repository down from 219MB to 150MB15:32
maxbHmm. he left. Anyway, it downloaded 25484KB for me15:36
Ursinha-sprinthi barry :)15:55
barryUrsinha-sprint: hi!15:55
Ursinha-sprintbarry, you are running karmic, right?15:55
barryUrsinha-sprint: nothin' but! :)15:55
Ursinha-sprintbarry, brave man!15:55
Ursinha-sprintI'm trying to have one instance of lp running on karmic but no success15:56
barryUrsinha-sprint: or... crazy! :)15:56
Ursinha-sprintit complains the lack of something.loom15:56
Ursinha-sprintlol15:56
barryUrsinha-sprint: hmm.  i haven't seen that.  do you have the full traceback or error?  note that this karmic was not an upgrade, it was a fresh install15:57
Ursinha-sprintbarry, hmm, I don't have it here now, only in my eeepc15:57
Ursinha-sprintthat was a fresh install with alpha 5 and I've been updating that since then15:57
barryUrsinha-sprint: dang.  yes, that's basically what i did too.  it definitely works (though i haven't tried building anything today yet)15:58
henningeabentley: ping16:01
abentleyhenninge: pong16:01
henningeabentley: Hi, there still seems to be a problem with the job system.16:02
abentleyhenninge: What is the problem?16:02
henningeabentley: Even with thumper's latest patch, we still have jobs stuck in "running".16:03
henningeabentley: https://pastebin.canonical.com/22916/16:03
henningeabentley: I can relate the rosetta-branches cases (job_type=3) to OOPSes in the script.16:04
abentleyhenninge: Is the database reaper tearing down the DB connection in those cases?16:04
henningeabentley: I don't know. How would I be able to tell?16:05
abentleyhenninge: It would show in the oops and the log.16:05
abentleyhenninge: Can you point me at an example oops?16:07
=== deryck is now known as deryck[lunch]
=== matsubara is now known as matsubara-lunch
=== Ursinha is now known as Ursinha-afk
=== sinzui1 is now known as sinzui
=== deryck[lunch] is now known as deryck
=== beuno is now known as beuno-lunch
=== matsubara-lunch is now known as matsubara
mrevellnight all!18:22
=== beuno-lunch is now known as beuno
rockstarabentley, hi18:37
abentleyrockstar: hi18:46
rockstarabentley, chat?18:48
abentleysure18:48
sinzuibarry: ping19:03
barrysinzui: pong19:03
sinzuibarry: I am hacking on announcement headings and titles. Look at https://edge.launchpad.net/launchpad-project/+announcement/345019:04
sinzuibarry: ^ Do I need to add a Heirarcy to set the traversed objects for the page title? There are no breadcrumbs19:05
barrysinzui: if there are no breadcrumbs, then i don't think you need a hierarchy just for the page title.19:05
barrysinzui: but in this case why is there not breadcrumbs?19:06
sinzuibarry: I do not think a breadcrumb is need (yet) because I want to see19:06
sinzuiLaunchpad Registry >> Annoucement: I did something cool19:06
sinzuibarry: That is what I am asking you. I do not know why there ar no breadcrumbs in any context19:06
barrysinzui: right, that's what i think i'd like to see too19:06
barrysinzui: ah19:06
barrysinzui: so the default hierarchy probably should work for you.  i wonder why it's not19:07
barrysinzui: isn't launchpad-project the root context?19:07
sinzuiIt is19:08
sinzuihmm19:08
sinzuibarry: is canonical url used in hierarchy? I ask because series and milestone do not have a hierarchy, but their canonical URL define the parent object19:09
barrysinzui: not directly.  Hierarchy looks for an IBreadcrumb for each object that has been traversed in the request.  it's built from those19:10
barrysinzui: perhaps put a break point in Hierarchy lib/canonical/launchpad/browser/launchpad.py and see if def items() gets hit?19:11
sinzuiThus we have no traversed objects...because launchpad-suite is not in being shown19:11
barrysinzui: it could be a problem with the template, or perhaps there's a custom +hierarchy adapter?19:11
sinzuiahh19:11
* sinzui looks19:11
sinzuino19:12
* barry is not sure what "no" means19:13
sinzuiI do not see anything suppressing the breadcrumbs19:14
sinzuiThere is no Hierarchy or Breadcrumb adapter.19:14
barrysinzui: you said above there are no traversed objects.  do you mean that request.traversed_objects is empty when you hit Hierarchy.objects() ?19:15
sinzuibarry: I assume there are not because I cannot explain why there are no breadcrumbs19:15
barrysinzui: best thing to do is put a breakpoint in Hierarchy.items() and see what happens19:16
sinzuibarry: I started looking at this when I saw the team membership pages were also missing breadcrumbs. That bug mentions both Hierarchy and Breadcrumb, but if neither of these are the case, we have a much harder problem to fix these.19:16
barrysinzui: bug number?19:18
sinzuibarry: bug 429663, you filed it19:18
mupBug #429663: ~team/+invitation/team does not have breadcrumbs <post-3-ui-cleanup> <Launchpad Registry:Triaged> <https://launchpad.net/bugs/429663>19:18
barryah yes, looks like you have the same problem.  this is a more general problem for which we do not yet have a fix19:20
sinzuihmm19:21
sinzuiWell Hierarchy is seeing the Project, Announcement, and View in the correct order19:21
sinzuiWhen I solve this, I think I will take the team bug too19:22
barrysinzui: hopefully there will be a generic fix for this19:24
sinzuibarry: I see I misunderstood the direction the breadcrumbs are made19:24
sinzuibarry: project has one, announcement does not. Since there is only one, we are not showing any19:24
barrysinzui: right19:25
* barry -> otp19:25
* rockstar lunches19:27
dobeyhrmm19:31
dobeywhat does "Nominated for Trunk by <$user> (approve/decline)" mean on a bug?19:31
sinzuibarry: Answer is simple, though the rule is not clear. We need to register TitleBreadcrumb for every object that uses .title.19:33
sinzuidobey: It means you are asking a series driver/release manager to commit to fixing a bug during the development of that series19:35
dobeysinzui: so in other words, it makes absolutely no sense at all, since "trunk" should always be in development :)19:35
sinzuidobey: Yep! If you use milestones to target bugs, the bug will wrong be associated to a series. I never use this feature. I would like to remove it19:36
sinzuis/wrong/wrongly/19:37
dobeysinzui: yeah, i've never even seen it before until today, when someone who hasn't ever even commented on the bug, decided to nominate it for trunk19:37
barrybac: sorry about that, ready to continue?20:14
barrysinzui: can we do a generic fallback for TitleBreadcrumb?20:15
sinzuino. I think the fallback is to DisplayNameBreadcrumb. I saw thast everything that has a .title also has TitleBreadcrumb in zcml20:16
barrysinzui: i'm just worried that if we have a lot of views using title, that's a lot of extra zcml hacking too20:17
sinzuititle is rare compared to displayname.20:18
barrysinzui: right20:18
sinzuiTitle is arguably bogus because in most cases we construct it20:18
barrysinzui: agreed.  sounds like a plan20:19
sinzuifor announcement, headline is actually title20:19
sinzuihmm20:19
sinzuioh20:19
sinzuibarry: maybe announcement is wrong because it does not have a displayname. just a headline presented as a title20:20
barrysinzui: what if you added a @property for displayname that returned the title?  or is that too ugly of a hack? :)20:21
sinzuiI am doing that now :)20:21
bacbarry: yes20:21
barrysinzui: excellent :)20:21
barrybac: calling20:21
sinzuiEdwinGrubbs: ping20:34
sinzuiEdwinGrubbs: unping20:39
rockstarmwhudson, when you're around, ping.20:59
mwhudsonrockstar: hello20:59
rockstarmwhudson, morning.21:00
rockstarmwhudson, so I'm working on getting the branch upgrade job stuff working with the upgrade-out-of-the-way concept.21:00
mwhudsonrockstar: oh right21:01
rockstarmwhudson, abentley says that we'll be doing the upgrades on crowberry, which is fine.  However, I'm not sure how where to sprout the branch that needs to be updated.21:01
rockstarmwhudson, I'm wondering if we have a helper function to figure that out yet.21:02
mwhudsonrockstar: tempfile.mkdtemp() ?21:02
abentleyrockstar: I'd be inclined to mkdtemp.21:02
mwhudsonrockstar: or am i misunderstanding?21:02
rockstarmwhudson, abentley, I thought we saw something where it was leaking file references in a reviewers meeting recently.21:03
abentleymwhudson, rockstar: I suppose a confounding factor is that we'd like to do it on the same partition.21:03
mwhudsonabentley: yeah, that's true21:03
abentleyrockstar: Not mkdtemp.21:03
rockstarabentley, after looking at the docs, I must have been confusing temp dirs and temp files.21:04
rockstarabentley, does it really need to be done on the same partition?21:04
abentleyrockstar: It doesn't have to be done on the same partition, but if it's not, it's more expensive to rename into place because the rename is actually copy & delete.21:05
rockstarabentley, yeah, I guess you're right.21:06
abentleyrockstar: And that also leaves a larger window for race conditions.21:06
abentleyrockstar: If you create it in .bzr/new, nothing bad will happen.21:06
=== jamalta is now known as jamalta-afk
EdwinGrubbsbeuno: Here is an example of a custom scrollbar that the yui slider could be styled to look like: http://wave.google.com/help/wave/images/ss2.gif21:25
beunoEdwinGrubbs, that may work better21:26
=== Ursinha-afk is now known as Ursinha
Ursinharockstar, sure :)21:33
rockstarUrsinha, hm, it looks like I can branch that url.21:34
Ursinharockstar, stupid doubt: if I change something in the lp branch, by pushing code to it, will it be mirrored to the original place?21:43
rockstarUrsinha, no, in fact, I don't think you can push to a mirrored branch.21:45
Ursinharockstar, well, I'm not able to branch his branch, get:21:45
Ursinhabzr: ERROR: Server sent an unexpected error: ('error', "An attempt to access a url outside the server jail was made: 'chroot-182925242704:///'.")21:45
rockstarmwhudson, ^^21:46
mwhudsonrockstar: ah :)21:46
rockstarUrsinha, are you branching from lp?21:46
mwhudson<mwhudson> rockstar, Ursinha: i think that error means a badly configured bzr+http server21:46
rockstarmwhudson, :)21:47
Ursinha:)21:47
mwhudsonrockstar: bzr info nosmart+http://bzr.malept.com/bazaar-overlay/ works21:47
Ursinharockstar, nope, I'm trying to branch from the original branch21:47
Ursinhaof course :)21:47
rockstarUrsinha, ah, okay.  So not-our-problem problem.  Those are my favorite kind.21:47
Ursinharockstar, bad boy21:47
rockstarUrsinha, so my initial assumption was correct, and it's a problem with the source.21:48
Ursinhabut what mwhudson, being able to do a bzr info shows what?21:48
Ursinha*what mwhudson said21:48
mwhudsonUrsinha: it's a problem with the smart server on their end21:48
Ursinhamwhudson, is that a way to workaround it, besides asking the user to fix his server?21:50
mwhudsonUrsinha: well i guess bzr could fall back to the dumb protocol, but basically no21:50
Ursinhahm21:51
Ursinhamaybe we need a better error message in the branch page?21:51
mwhudsonit's not formatted very nicely i guess21:51
Ursinhathis message makes everyone think that the problem will be fixed for itself21:51
Ursinhaor that nothing needs to be done21:52
mwhudsonwell, unless you're the one running the server, there isn't much that can be done21:52
Ursinhamwhudson, I mean, the message says that21:52
Ursinhamaybe we should show the error just like it's done with the imports?21:53
mwhudsonthat would be better, yes21:53
Ursinhamwhudson, well, I'll file a bug then21:53
rsalvetibecause there's no error saying that the main server is off or with problems21:53
rsalvetiit's just "updating" hehe :-)21:53
Ursinharsalveti, I think it should be better if displayed like when we have import errors21:53
rsalvetiUrsinha: cool :-)21:53
Ursinhait has a message saying that it failed and the log21:54
Ursinhas21:54
rsalvetiUrsinha: yep, because the current message let me think that this was a problem with launchpad itself21:54
Ursinharsalveti, maybe you should contact the owner and ask him to fix that21:54
Ursinhaor host it in lp itself21:54
rsalvetiI just noticed that it's using a mirror now that rockstar showed the external link21:54
Ursinhabecause it's bzr, no need to do that :)21:54
rsalvetiUrsinha: yep, to fix the branch, yes21:55
Ursinhain fact there's already a bug, it seems21:57
Ursinhabug 37622921:57
mupBug #376229: broken mirrored branches display "  Launchpad is processing new changes to this branch and will be available in a few minutes. Reload to see the changes." message <branch-puller> <ui> <Launchpad Bazaar Integration:Triaged> <https://launchpad.net/bugs/376229>21:57
mwhudsonUrsinha: oh, i had't scrolled down that far21:58
mwhudson:)21:58
Ursinha:)21:59
Ursinhathanks mwhudson and rockstar21:59
Ursinhamuch appreciated :)21:59
=== Ursinha is now known as Ursinha-afk
gary_postermwhudson: well, good news, though it makes me a bit nervous :-P .  My bin/py script (a faux interpreter that can do most of the important interpreter tricks), generated from my zc.buildout branch, runs ``time ./bin/py -c ''``, um, about twice as fast as python24 on my system.  So, assuming it does everything we need, that may help a lot.22:21
mwhudsongary_poster: ah, cool22:21
mwhudsongary_poster: twice as fast as python2.4 -c pass or bin/py from trunk?22:23
gary_postermwhudson: how important is this?  If I'm right that this works fine, and this puts codehosting back to the previous level of performance is this CP-worthy, for instance?22:23
gary_postertwice as fast as python 2.422:23
maxbsite.py sure does do an awful lot of weird stuff :-)22:23
gary_posterheh, true22:23
mwhudsongary_poster: i'm slightly inclined against the cherrypick22:24
mwhudsonnoone has actually complained, we've just looked a bit bad :)22:24
gary_postermwhudson: ok.  that means it's more easy going for me. :-)22:24
gary_postermwhudson: but I'll get this in and ping you in case you want to run some tests22:24
mwhudsongary_poster: yeah, very interested to see this land22:25
gary_postercool22:25
gary_posterbarry: hey, just so you know, I'm in the process of changing bin/py from a shell script with PYTHONPATH hack to a Python script (with its own fun ;-) ).  So if you have not landed the mailman buildman change, then we should coordinate, in case my change messes up your plans.22:42
gary_poster...this reminds me that I may need to change _pythonpath.py ...22:42
gary_posterbut anyway, feel free to talk22:42
barrygary_poster: it should be landed now.  sounds interesting though :)  why do you think you'll need to change _pythonpath.py?22:43
gary_posterbarry: ok, cool.  If make builds and tests pass then I'll be happy. :-)  I'll maybe want to change _pythonpath to make it more careful, the way I had to change the buildout-generated scripts, to handle namespace packages and so on.22:48
gary_posterHandling dist-installation-style namespace packages and normal-installation packages in the same namespace require some excessively careful dances. :-/22:48
gary_posterrequires22:48
barry:/22:48
gary_posterbarry: I was reading about Py 2.6.3 fun.  :-(22:48
barrygary_poster: you probably want to run bin/test --layer=MailmanLayer even though the tests are not entirely stable22:48
gary_posterand setuptools22:49
barrygary_poster: it's what makes open source so rewarding :)22:49
gary_posterbarry: ah!  ok will do thank you22:49
gary_posterbarry: lol :-) :-/22:49
sinzuiThis is new. The test suite just opened up the help manual for  gnome terminal when the test failed.22:56
* sinzui hopes this is not a karmic freature22:56
barrysinzui: ouch.  which test?22:56
sinzuixx-announcements22:56
barrysinzui: of course it doesn't fail for me, but i've never seen that (and i've seen lots of test failures on karmic :)22:58
sinzuibarry: My change to remove an ancient page_title broke more than I anticipated. I suspect a view was shared by several templates.22:59
gary_postermwhudson: oh, bah.  :-/  I had clunked my change.  I was testing an earlier version that didn't do all the necessary dances.  With the completely new mechanism, doing the proper dance as I currently understand the steps, we're back to one-order-of-magnitude-ish slower. So, my change will fix some problems, but it is not faster.23:09
gary_posterOTOH, it also is not (even) slower, so that's at least something, from my perspective.  If the codehosting could use a significantly smaller subset of packages than the whole thing Launchpad needs, we could experiment with a callable just for that purpose...but a pre-forking process would be the better solution long-term anyway I suspect.23:09
* gary_poster runs to dinner23:09
gary_postergood night all23:09
mwhudsongary_poster: ah well23:10
mwhudsongary_poster: bye for now23:10
gary_posterbye23:10
barrysinzui: :(23:11
=== matsubara_ is now known as matsubara-afk

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!