/srv/irclogs.ubuntu.com/2008/11/18/#launchpad.txt

=== leonel_ is now known as leonel
jmlHobbsee: is https://answers.launchpad.net/launchpad/+question/50248 being tracked in the bug? does the question need to stay open?00:16
Hobbseejml: Pass.  I don't do translation stuff at all00:18
jmlsorry, wrong URL00:19
jmlhttps://answers.launchpad.net/launchpad/+question/5089300:19
jmlHobbsee: ^^ that one00:19
* jml needs a global kill-ring00:19
Hobbseejml: i'm not sure.  I guess not, as long as something actually happens with the bug.00:23
jmlHobbsee: ok. I've closed the question. Re-open it if you reckon it'll help.00:34
Hobbseejml: ok, thanks00:34
=== kiko is now known as kiko-zzz
mwhudsonis Fabien Tassin here?01:33
Hobbseemwhudson: fta?01:35
mwhudsonah yes01:36
mwhudsonfta: hello01:36
jmlis there an equivalent of https://bugs.launchpad.net/bugs/<bug_id> for questions?01:38
ftamwhudson, hi01:40
mwhudsonfta: so about the google chrome import01:40
ftamwhudson, the import is broken, and now the project is was tied to has been removed too. I can reask for an import but i assume it will fail too for the same reason01:42
mwhudsonfta: i'm not sure that's a good assumption01:42
mwhudsonfta: in other words, i think it's worth a try01:42
ftathe svn url will be the same01:42
ftai don't see how it will be different from ~1 hour ago when the old project was still alive01:43
mwhudsonwon't the url be http://src.chromium.org/svn/trunk/src rather than http://src.chromium.org/svn/trunk/src/chrome ?01:43
ftayes, but one is a subset of the other, but should import just fine01:44
fta-but+both01:44
ftabtw, this one is broken too https://code.edge.launchpad.net/~vcs-imports/v8/trunk01:45
ftav8 is part of the chromium project too01:45
mwhudsonthe two-bit summary of the v8 failure is that google code sucks01:46
mwhudsoni guess i can try to use svnsync to get a local copy of the repository and import from that01:46
mwhudson<fta> yes, but one is a subset of the other, but should import just fine01:47
mwhudsonfta: when you've worked in the crufty business of vcs imports for as long as me, you stop making assumptions :)01:47
ftai have local import of both in bzr, it's just fine, i also made a git import to compare with bzr, fine too01:48
mwhudsoni bet you didn't do the imports with cscvs01:49
ftaindeed01:49
mwhudsonwhere as the vcs imports service on launchpad does01:53
mwhudsonfta: what is it you actually trying to achieve here?01:53
mwhudson+are01:53
=== vorian is now known as heHATEme
ftathose imports were not from me, but i'm using them to monitor the status of the project. I recently created my local imports because the ones on lp were broken, and also because i'm creating packages01:55
=== jml changed the topic of #launchpad to: https://launchpad.net/ | Channel logs: http://irclogs.ubuntu.com | Need help? Community help contact: - jml (afk)
=== heHATEme is now known as vorian
=== matsubara is now known as matsubara-afk
ftamwhudson, if it's too complicated, nm. i'll setup a real local sync when i have time.01:57
mwhudsonit's not necessarily complicated01:59
mwhudsonbasically, i would invite you to set up the imports you want, as you want them01:59
mwhudsonand ignore/have deleted the broken ones01:59
ftaok, will do. thanks for your time02:00
fta3am here. 'night all.02:00
mwhudsonnight02:02
CarlFKwho is respncible for "2) The version of the package you are using, via 'apt-cache policy packagename' or by checking in Synaptic."03:11
CarlFKit would be nice if 'packagename' was the name of the package I am reporting on03:11
CarlFKso that I can cut/paste that, and cut/paste the results03:11
HobbseeCarlFK: instead of the source package it belongs to?03:12
jmlCarlFK: where does that text appear?03:12
CarlFKwhats the difference?03:12
CarlFKhttps://bugs.edge.launchpad.net/ubuntu/+source/xawtv/+filebug#form-start03:14
CarlFKmay need to fill in a summary: https://bugs.edge.launchpad.net/ubuntu/+source/xawtv/+filebug "x" enter, "No, I'd like to report a new" bug" scroll down03:15
HobbseeCarlFK: some sources contain many binaries.  launchpad is organised by source packages03:15
CarlFKoh... I think I kinda get it.  "In what package did you find this bug?"  is on the same page.  so it would have to do ajaxy validation then replace packagename with something03:19
=== abentley1 is now known as abentley
=== barry is now known as barry-away
=== hakr is now known as h[a]kr
=== jml changed the topic of #launchpad to: https://launchpad.net/ | Channel logs: http://irclogs.ubuntu.com | Need help? Community help contact: - jml
=== jamesh_ is now known as jamesh
=== jml changed the topic of #launchpad to: https://launchpad.net/ | Channel logs: http://irclogs.ubuntu.com | Need help? Community help contact: -
jmlthe truth is I never left you.06:31
=== abentley1 is now known as abentley
Ryan52Why is the "Importance" for bugs read only? how do I get the ability to change them?07:15
Ryan52(if this is the wrong place to ask this, then sorry, please point me to the correct place)07:16
persiaRyan52, You need to be a member of the right bug team for the project against which the bug is filed.07:16
Ryan52"Ubuntu Bugs" is the only team listed in the "Also Notified". Does that mean that's the team I need to be a part of?07:16
persiaOh, for bugs in Ubuntu, you need to be in ubuntu-bugcontrol07:17
persiaBut for bugs in #ubuntu, #ubuntu-bugs is a better place to ask.07:17
Ryan52ok, thanks07:18
Ow1hi.08:47
Ow1were the unlicences translations removed from Launchpad ?08:48
kiko-zzzOw1, not sure they were done yet, but they are scheduled for today08:49
Ow1do you know how should I format the URL to point to a translation08:49
Ow1and to be valid for all users08:50
Ow1I know there is something with +me08:50
kiko-zzzvisit the translation page for the string and use that URL?08:50
Ow1but tha url contains my username08:50
kiko-zzzOw1, does it? can you give me an answer?08:51
Ow1yes08:51
Ow1I can see it08:51
kiko-zzzsorry s/answer/example/08:51
Ow1a08:51
Ow1https://translations.edge.launchpad.net/~adiroiban/+licensing08:52
Ow1this is for my account08:52
kiko-zzzOw1, well.. that's not a URL for a translation. can you explain what you want to give people?08:52
Ow1I want to ask them to accet the BSD licence08:52
Ow1accept08:53
kiko-zzzOw1, gotcha! that's not the same as you asked before :)08:53
kiko-zzzOw1, translations.launchpad.net/+me/+licensing08:53
kiko-zzzsorry08:53
kiko-zzzhttps://translations.launchpad.net/people/+me/+licensing08:53
Ow1many thanks!08:54
* wgrant thinks that /+me should work.08:54
Ow1so for any further usage i should replace "~adiroiban" with "people/+me"08:54
Ow1it is working08:54
kiko-zzzyep08:55
Ow1Right now if someone has not accepted the BSD licence, his/her translations will be removed from Launchpad.08:55
kiko-zzzright08:55
Ow1What happens if he/she accept the licence tomorrow?08:55
kiko-zzzit will be too late08:56
Ow1is there a way to recover them?08:56
kiko-zzzdanilo-afk, ping08:56
wgrantHm.08:56
kiko-zzzI don't believe so08:56
wgrantI thought I saw an email saying they wouldn't be removed due to non-answering.08:56
* wgrant finds it.08:56
kiko-zzzreally?08:56
Ow1yes. mee to. I also saw that email from Danilo08:56
wgrantWhat about people who haven't answered the question yet.. is there08:57
wgrant> going to be a mail-out to those translators, or the administrators of08:57
wgrant> the projects concerned?08:57
wgrant"Those translations will be kept."08:57
wgrantFrom danilos himself.08:57
wgrantWhat an awful paste.08:57
* wgrant kicks Thunderbird.08:57
oojahWhere did this email get sent?08:57
Ow1ubuntu translators lists08:58
wgrantoojah: launchpad-users08:58
oojahRight, ok.08:58
wgrantAnd ubuntu-translators, it seems.08:58
oojahI must've missed it.08:58
wgrantRe: Important Launchpad news: removal of non-BSD licenced translations08:59
wgrantOn 2008/11/1508:59
kiko-zzzOw1: william's right and I stand corrected; there's a point that I hadn't noticed originally08:59
wgrantActually, probably 2008/11/14 unless you're in Australia or NZ...08:59
oojah:)08:59
wgrantI cannot understand how somebody behind Sun's NAT can drop out so frequently.09:01
persiawgrant, Awkward placement of WiFi transmitters?  When attending Sun-sponsored events, I've often found I had short drops.09:02
oojahI don't really follow l-u so it's a good job I noticed this conversation. I wonder what percentage of translators have/haven't responded yet.09:03
luisbg_I want to change the status of this spec https://blueprints.launchpad.net/ubuntu/+spec/ubuntu-studio to done, can I do it or does it have to be pete savage (the registerer)?11:19
cody-somervilleluisbg_, either a launchpad admin, ubuntu driver, or Pete will have to change it11:21
luisbg_cody-somerville, thanks! long time no see11:23
luisbg_any launchpad admin in the house? :P11:23
=== andreas__ is now known as ahasenack
* paolettopn e' away (ciao a tutti!)11:41
=== paolettopn is now known as paolettopn_zzz
=== paolettopn_zzz is now known as paolettopn
* paolettopn e' away (ciao a tutti!)11:42
=== paolettopn is now known as paolettopn_zzz
=== paolettopn_zzz is now known as paolettopn
* paolettopn e' away (ciao a tutti!)11:44
=== paolettopn is now known as paolettopn_zzz
Hobbseepaolettopn_zzz: any chance you can get rid of that script?11:46
=== doko_ is now known as doko
=== danilo-afk is now known as danilos
kaliladanilos: urgent - please check your email (Djihed)12:39
mrevellHey, does anyone here speak Turkish?12:48
=== mrevell is now known as mrevell-lunch
=== barry-away is now known as barry
=== mrevell-lunch is now known as mrevell
vvinethey, is it normal that I get "edit upstream link association" and "delete upstream link association" links on a project overview page14:26
vvineteven though I should not have any permissions on that project14:26
vvinet(I didn't try the links - dont want to )14:26
persiavvinet, It's normal for things not yet associated, in the hopes that people would only use them if they have the correct data.14:43
persiadelete upstream link association sounds odd.  I suspect it would take you to an error.  If not, hrm.14:44
vvinetI see it here: https://launchpad.net/ubuntu/+source/kdebase-workspace14:45
vvinetI have not clicked the delete button, as it is a POST form and could therefore have an effect14:48
vvinet(in theory)14:48
persiaBetter not to click it then.14:49
=== matsubara is now known as matsubara-lunch
=== bac` is now known as bac
=== sale_ is now known as sale
CoMp4c7i have an app which uses openid. I can login with all providers I have tested (myopenid, claimid, verisign) but not with launchpad, any ideas?15:44
=== kiko-zzz is now known as kiko
=== matsubara-lunch is now known as matsubara
RainCTkiko: can you help CoMp4c7? :)15:58
kikoCoMp4c7, hmmm. well, it's a problem for sinzui and flacoste. what's the problem you're having?15:58
CoMp4c7i am getting the errror:  Server denied check_authentication15:59
sinzuiCoMp4c7: This may relate to OAuth. My own scripts (launchpad API) require an Authorization token16:00
* sinzui thinks16:00
CoMp4c7i can login succesfully with the same app running in local16:01
sinzuiCoMp4c7: what do you mean by local? Your localhost? Your home directory?16:05
CoMp4c7sorry, localhost16:05
sinzuioh CoMp4c7 by openid, do you mean your Launchpad OpenID url?16:06
CoMp4c7i am trying to log in with http://login.launchpad.net16:07
* sinzui wonders if this is borked because Launchpad does not use OpenID for its own logins16:09
CoMp4c7sorry, i don't understand what are you saying16:12
flacosteCoMp4c7: are you a beta test users?16:12
CoMp4c7yes16:12
sinzuiSince this works on localhost, I wonder if this is a cookie issue. Launchpad knows you are login by the cookie in the request16:14
=== h[a]kr is now known as hakr
CoMp4c7if I remove all cookies in browser it should not happen?16:19
flacosteCoMp4c7: where are you trying to log in?16:22
CoMp4c7http://db.freevial.org16:24
flacosteCoMp4c7: when I enter https://login.launchpad.net in there, i get the 'Authenticate to http://db.freevial.org' screen on Launchpad16:29
flacostewhich allows me to click 'Sign In'16:29
flacosteok,16:29
flacosteit seems to fail for me also after that16:29
CoMp4c7do you have any ideas about what could be the problem?16:31
flacosteCoMp4c7: not really, i'd say a problem in Freevial openid implementation16:31
flacosteCoMp4c7: well16:32
flacostei think what is happening is that its trying to use check_immediate/check_authentication to validate the just provided credentials16:32
flacosteand we don't support that mode16:32
flacosteand the spec allows us to do this16:32
flacostecheck_immediate/check_authentication is for validating credentials without user intervention, a valid response is 'I cannot do this without user intervention' which it seems they decide to fail16:33
flacostebut this is a valid return code16:33
CoMp4c7and why do you think it is not a problem on localhost?16:34
flacostewhat do you mean localhost?16:34
RainCTflacoste: (He's the author. The source is available at http://tinyurl.com/freevial-db-edit, btw)16:35
CoMp4c7when app is running in localhost16:36
flacostehmm16:36
flacosteok16:36
flacosteactually, i just checked check_authentication spec16:36
flacosteand what i was just saying is incorrect16:36
flacostecheck_authentication is used by a consumer to valide the signature16:37
flacosteif we say it's invalid, it's because probably of a request parameter16:37
CoMp4c7but if the problem is with the request, it also had to fail when the app is runnig in localhost16:42
flacosteCoMp4c7: what is Auth/OpenID ?16:44
CoMp4c7flacoste: sorry, i don't understand what you are asking for16:46
flacosteCoMp4c7: what OpenID library are you using?%16:46
CoMp4c7flacoste: PHP OpenID library by JanRain16:48
flacosteCoMp4c7: version?16:50
CoMp4c7flacoste: 2.1.216:51
flacosteCoMp4c7: ok, i'm trying to see if I can find a record of that error on my side17:01
flacosteCoMp4c7: from the top of my head, I'd suspect difference some of the parameters that change from one request to the next one17:02
flacosteon the live site, but not on localhost17:02
flacostebut this is really hand-wavy17:02
flacosteCoMp4c7: in particular, make sure that getReturnto() returns an identical value from within the two requests17:05
flacosteCoMp4c7: when mode='try' and when mode='continue'17:05
flacosteCoMp4c7: other things to consider, is /tmp configuration in both instances?17:07
flacosteCoMp4c7: and is apache configured similarly in both instances17:07
flacostei'm thinking about configuration which makes that the store isn't really used17:08
flacosteand so means that it falls back to the check_authentication dumb mode17:08
flacostewhich is an error if the request started by a valid association (saved to the store, but lost when the reply comes back)17:09
flacostethus going down the check_authentication route (which fails, because a shared secret was used initially)17:09
CoMp4c7flacoste: but if store is not working fine, it would fail with other providers17:09
flacosteCoMp4c7: maybe17:10
flacostedepends on the implementation of the other providers17:11
flacostewhich protocol they are using17:11
flacoste1.x or 2.x17:11
CoMp4c7about apache, I don't think both servers are configured similary, db.freevial.org is in a shared hosting17:11
CoMp4c7but I checked the problem is not a mod_encoding problem, as many people say in Google17:12
CoMp4c7flacoste: i have manually typed returnTo to be sure it doesn't change and i get the same error17:18
flacosteCoMp4c7: ok, i'll have to look at my logs, but first i need some lunch, will you be around later?17:18
CoMp4c7i will be here next 3 hours i think17:19
CoMp4c7thank you very much for your help!17:19
=== hakr is now known as h[a]kr
=== h[a]kr is now known as hakr
=== paolettopn_zzz is now known as paolettopn
* paolettopn e' away (ciao a tutti!)18:05
=== paolettopn is now known as paolettopn_zzz
=== paolettopn_zzz is now known as paolettopn
* paolettopn e' away (ciao a tutti!)18:07
=== paolettopn is now known as paolettopn_zzz
=== paolettopn_zzz is now known as paolettopn
=== merriam_ is now known as merriam
ahasenackso, my use case of PPA is to build packages for me. I only use PPA because I need these packages for older distros (dapper, for example), otherwise I would be building them on my machine only (hardy).18:37
ahasenackso I tend to upload packages, build, tweak, upload again, etc18:37
ahasenackthe problem is that PPA checks the versions, and I can't just delete a package and upload it again, PPA wants me to bump the version all the tame18:38
ahasenacktime18:38
ahasenackso isn't PPA for me?18:38
persiaahasenack, Probably not.  You might be interested in a local installation of pbuilder or sbuild.18:38
cprov-lunchahasenack: why the version suffix bother you ?18:39
ahasenackcprov-lunch: I use my ppa for build tests, and once I'm satisfied I use the landscape's ppa "for real"18:39
beunoahasenack, that's the nature of apt18:39
ahasenackso, basically I only want a build machine for the older distros18:39
ahasenackI don't mean to use my PPA to publish these builds18:39
cprov-lunchahasenack: it very hard for use to guess that.18:40
ahasenackuse?18:40
ahasenackcprov-lunch: if you mean "users", well, that's it, I don't want users for my ppa18:41
cprov-lunchahasenack: you can workaround that easily by working on $real_version~ppaN18:41
cprov-lunchahasenack: I meant 'us', but nevermind18:41
ahasenackcprov-lunch: unless the problem I was debugging and testing involved the versioning itself... :(18:43
ahasenackcprov-lunch: which is what it does, so I can't mangle the version-release part because I actually need to test an upgrade path18:43
persiabeuno, Hrm?  How is it apt?  apt doesn't mind if you upgrade, downgrade, crossgrade, etc. (although things might break).  I thought it was for repo sanity.18:44
cprov-lunchahasenack: apart from being more difficult and confusing the version tilde suffix should allow you to test upgrade paths as well18:45
cprov-lunchahasenack: but I understand it might become inconvenient in certain cases.18:45
ahasenackcprov-lunch: can I have more than one ppa attached to the same account?18:45
cprov-lunchahasenack: no18:46
ahasenackI read somewhere about private ppa18:46
ahasenackok18:46
cprov-lunchahasenack: I don't know exactly what to suggest, usually people go for PPA because they want history and a coherent repository, if you only care about building in a pristine area you can use pbuilder18:47
cprov-lunchahasenack: play as much as you want locally and upload the 'final' version to PPA18:47
=== hakr is now known as h[a]kr
ahasenackcprov-lunch: can I use pbuilder to build for older distros? Like dapper? Just install a dapper chroot, etc?18:49
cprov-lunchahasenack: bear in mind that there is always a safe way around version (either using ~ or + as suffix), unless you use epoch :(18:49
ahasenackcprov-lunch: the package is already using ~18:49
ahasenackso I can't mangle it twice18:49
cprov-lunchahasenack: yes, you can have multiple chroots with pbuilder18:49
ahasenackeven the landscape ppa, we don't use its publishing capabilities in landscape, we have another repository elsewhere that is the official one18:50
ahasenackwe basically use ppa as a build mechanism, that's all18:50
* ahasenack searches for space on his laptop for 3+ more distros18:50
cprov-lunchahasenack: you have the history on PPA18:51
ahasenackcprov-lunch: we have it in bzr branches18:51
ahasenackcprov-lunch: we don't grab the previous source build to build a new version, our debian/* structure is in bzr18:52
cprov-lunchahasenack: not the binary as it was built originally, I suppose18:52
=== jason__ is now known as DBO
ahasenackcprov-lunch: no, those we archive elsewhere too18:52
cprov-lunchahasenack: it's just a little frustrate, on both sides, we could publish stuff with not checks what so ever, but if you look to the ubuntu history all 300k sources have coherent versions. Why is that hard to get on a project basis ?18:56
=== beuno_ is now known as beuno
ahasenackcprov-lunch: I understand your side, believe me18:56
ahasenackcprov-lunch: my biggest problem is with builds for older distros, which I don't have at hand to test and "practice"18:57
ahasenackcprov-lunch: so PPA is basically unforgiving regarding any mistakes18:57
cprov-lunchahasenack: note that I'm not blaming you, it's a general problem, bzr gets this too.18:57
ahasenackcprov-lunch: even if the mistakes were never "released"18:57
ahasenackcprov-lunch: again, it's because I use it as a build bot and have no intention of publishing the packages until they are fine18:57
cprov-lunchahasenack: it's very fascist in this aspect18:58
ahasenackcprov-lunch: hmm, am I wrong in thinking that even failed builds get blacklisted? I vaguely remember something about that18:58
cprov-lunchyes, even FTBFS versions get blacklisted.18:59
persiaahasenack, You can use ~~ to further mangle ~, but you *really* want a local build service for your use case.18:59
cprov-lunchpersia: really ? ehe18:59
ahasenackcprov-lunch: so, if people use ppa to build+publish, wouldn't they probably have a local build service anyway to test first?19:00
ahasenackcprov-lunch: if failed builds get blacklisted, that would waste a version/release19:00
ahasenackcprov-lunch: anyway, sorry, it's another topic19:00
cprov-lunch`dpkg --compare-versions '1.0~1' 'lt' '1.0~~1' || echo 'it is'`19:00
persiacprov-lunch, '~' isn't magic.  It's just a character.  It happens to sort beneath null.19:00
persiaSo ~~ sorts before ~19:01
ahasenackthanks for the tips all, I will try to get some linodes as build bots for my use case19:01
cprov-lunchpersia: I didn't assume it to be magic, just restrict to a single use in debian versions. But you are right, it's smaller and pass the validation regexp.19:02
persiaahasenack, You might be interested in deb-o-matic, which is a framework to have multiple pbuilders running against multiple targets on a remote (controlled) host for build-testing.19:02
persiacprov-lunch, That's just convention.19:02
persia(and in Debian, if you need more than one, you've already shot yourself in the foot)19:03
cprov-lunchgreat, let's applaud !19:04
persiacprov-lunch, This brings back the debate as to use cases for PPAs :)  Has there been any discussion on the topic in the past couple months?19:05
NCommanderpersia, oooh, that sounds cool, where can I find this deb-o-matic?19:05
cprov-lunchpersia: no, just this project-based ppa thing hitting us harder and harder.19:06
persiacprov-lunch, Ran into that one with the Japanese team today as well.  I suspect that's probably a more important priority than sorting individuals.19:08
cprov-lunchpersia: but do you understand why is it hard for teams to follow a safe versioning path ?19:09
persiaYes.19:09
cprov-lunchpersia: is it a social problem cause by the lack of "queue"  (review) ?19:09
persiaI wrote it up in a bug once.  I'll go try to find it.19:09
cprov-lunchs/cause/caused19:09
persiaNo.19:10
persiaIt's the use of PPAs for distribution vs. the use of PPAs for test-building/review.19:11
persiaI personally believe PPAs are currently positioned somewhere between the two, and not quite optimal for either, but I want to find my rant again.19:11
cprov-lunchpersia: right, I see your point19:13
persiabug #263301 is different, but my comment there sums up the two use cases.19:14
ubottuLaunchpad bug 263301 in soyuz "Can't re-upload a package with a different src tarball after deletion in PPA" [Undecided,New] https://launchpad.net/bugs/26330119:14
persiaThere's a couple other related bugs, but I don't see them right now.19:14
cprov-lunchpersia: let me check.19:15
persiaI have seen bug reports from users wishing each of the two use cases, so it's not clear-cut how to fix it.19:15
cprov-lunchpersia: right, your comment on that bug tends to agree with the current behaviour, towards orig.tar.gz coherency with ubuntu19:19
cprov-lunchpersia: that I think is consensus, right ?19:19
persiaI don't think there is consensus.19:19
cprov-lunchpersia: the version blacklisting is the part that causes disagreement, IMHO19:20
persiaWell, there's people who complain about the version blacklisting, and then there's people who complain about e.g. package signing.19:20
ahasenackany particular reason why a failed build also gets blacklisted? It's not like it was ever downloaded and installed: it didn't even build19:20
persiaI don't think it's best fixed on a per-bug level.  I think it needs a philosophical decision as to which use case to support, or to continue to straddle this boundary.19:21
persiaahasenack, The source package was published.19:21
ahasenackpersia: does it have to be published in this case?19:21
cprov-lunchpersia: we agree on this, it's way wider the fixing each isolated bug.19:22
persiaahasenack, Yes, because it needs to be published to be passed to the buildds.  It's part of the architecture of Soyuz.19:22
ahasenackok19:22
cprov-lunchpersia: oops, not literally, the source doesn't have to be in the repository to be built.19:22
persiaRight.  Last time we discussed this (Maybe August or early September), you thought it might come up in some of the LP meetings scheduled.  If it hasn't, that's fine.  I mostly just wanted to check :)19:23
persiaahasenack, Trust cprov.  He actually knows this stuff.  I'm just a user.19:23
ahasenackpersia: :)19:23
=== fta_ is now known as fta
cprov-lunchpersia: it was discussed many times since PPA are out, no decision was made, because it seems to run fine for the vast majority of weekend-users19:24
cprov-lunchpersia: we have only few problematic project-based PPA which suffer from this problem.19:25
persiacprov-lunch, Right.  It's "good enough" for most things, but annoying for anyone who wants to use it heavily.19:25
cprov-lunchpersia: yes, we can say that.19:25
persiacprov-lunch, Personally, I don't think that's necessarily bad, as long as it's by design.19:25
cprov-lunchtacking the problem w/o disrupting the rest of the users is mandatory at this point19:26
persiaIt's like getting a mid-range digital camera.  Not as light and portable as a pocket camera.  Not as featurefull as an SLR.  Still, the most popular range.19:26
persiaOn the other hand, if that is the decision, it would be good to determine that, and intentionally wontfix the bugs for users who want a review system or want a real publishing system for their own distro.19:27
cprov-lunchpersia: good example, but kind of depreciate what we want to be the best/easiest/fastest pkg management system out there, right ?19:27
=== h[a]kr is now known as hakr
persiacprov-lunch, Well, it's a matter of picking segment.  If you want to be good for test-build/review stuff, you need to be flexible enough that you're unsafe for repository work.  If you want to offer repos that people can trust, you're unsuitable for test-build/review.  Currently, it's somewhere in-between.19:29
persiaFor the same reason you can't put a dSLR in your pocket easily.19:30
cprov-lunch:)19:30
cprov-lunchpersia: well, currently PPA have very consistent repositories & history (versions are blacklisted and the repo is never ovewritten)19:30
persiaOne option we discussed previously was letting users select signed/strict vs. unsigned/loose for each PPA, but it was suggested that this option would be confusing, and that there would be no way to safely change it later.19:30
flacosteCoMp4c7: how would i recognize your localhost instance using referer in my log19:31
flacosteCoMp4c7: also i find it very strange, i don't seem to see any connections for the check_authentication call19:31
cprov-lunchpersia: once we release per-PPA repo signatures it will be state-of-art for the latter use-case19:31
persiaFor hosting a derivative distro?19:31
CoMp4c7flacoste: http://localhost/~arnau/freevial19:31
flacosteCoMp4c7: if I grep for "http://db.freevial.org" referral on login.launchpad.net,  I only see requests for the initial +openid page19:32
cprov-lunchpersia: yes, why not ?19:32
flacosteCoMp4c7: did you made any requests today from it?19:32
CoMp4c7yes19:32
persiacprov-lunch, I thought there were a few other bugs around that, but if that's the plan, it's a sensible direction.19:33
CoMp4c7i discovered that there are the same problem with claimid.com19:33
CoMp4c7i think it is a problem of the server19:33
persiacprov-lunch, Fixing the remaining little bits will probably even more annoy the test-build users, but they ought be using local build farms anyway (it's annoying when three people upload OOo at once, for example).19:34
flacosteCoMp4c7: ok, I cannot see anything interesting in the logs, unfortunately19:34
persiacprov-lunch, Also, it will probably end up being slightly more "complex" due to the key management systems, etc., although with good docs this shouldn't matter too much.19:34
cprov-lunchpersia: one key per PPA, makes the whole dance simpler.19:35
persiaAnything else is madness.19:36
flacosteCoMp4c7: ok, this is very weird19:36
cprov-lunchpersia: extra signatures can be made on top of the signing key to increase trust.19:36
persiacprov-lunch, Is the plan to use the user's email, or a constructed email for signing?19:36
cprov-lunchpersia: if the user really cares about sane distribution of his packages19:36
persiaSo if I do a keyserver search for "Emmet Hikory", would I find my PPA key?19:36
CoMp4c7flacoste: do you have any idea about where could be the problem?19:37
flacosteCoMp4c7: i just tried again and the URL of the page returned by launchpad - that displays the error - do contain a succesful auth reply19:37
flacosteCoMp4c7:19:37
flacostehttp://db.freevial.org/auth.php?mode=continue&janrain_nonce=2008-11-18T19%3A35%3A42ZbAvTGr&openid.assoc_handle=%7BHMAC-SHA1%7D%7B49231912%7D%7BTpfGdQ%3D%3D%7D&openid.claimed_id=https%3A%2F%2Flogin.launchpad.net%2F%2Bid%2FcDYAwTe&openid.identity=https%3A%2F%2Flogin.launchpad.net%2F%2Bid%2FcDYAwTe&openid.mode=id_res&openid.ns=http%3A%2F%2Fspecs.openid.net%2Fauth%2F2.0&openid.ns.sreg=http%3A%2F%2Fopenid.net%2Fextensions%2Fsreg%2F1.1&openid.19:37
flacosteop_endpoint=https%3A%2F%2Flogin.launchpad.net%2F%2Bopenid&openid.response_nonce=2008-11-18T19%3A35%3A46Z7Hj4Ht&openid.return_to=http%3A%2F%2Fdb.freevial.org%3A80%2Fauth.php%3Fmode%3Dcontinue%26janrain_nonce%3D2008-11-18T19%253A35%253A42ZbAvTGr&openid.sig=bC03e6utz%2FsL0gNQJyJdKvNDTro%3D&openid.signed=assoc_handle%2Cclaimed_id%2Cidentity%2Cmode%2Cns%2Cns.sreg%2Cop_endpoint%2Cresponse_nonce%2Creturn_to%2Csigned%2Csreg.nickname&openid.sreg.19:37
cprov-lunchpersia: key email & comment will be deliberately blank19:37
flacostenickname=flacoste19:37
persiacprov-lunch, And name?19:37
flacosteCoMp4c7: so it's in the processing of that URL that the error message is generated19:38
cprov-lunchpersia: name will be the ppa title + 'signing key'19:38
persiaSo something like "Emmet Hikory's PPA signing key" ?19:38
flacosteCoMp4c7: the error message talks about check_authentication failing, but I know for sure that no such request has been made to Launchpad19:38
cprov-lunchpersia: yup19:38
flacosteCoMp4c7: since only request is ever made (check_authentication implies that your server would send another request to Launchpad to validate the signature provided in the above URL)19:39
flacosteonly one request, i mean19:39
CoMp4c7flacoste: it could be a encoding problem, but I double check that...19:39
flacosteCoMp4c7: what encoding problem?19:39
CoMp4c7flacoste: when apache sends parameters to php, i read about a problem with mod_encoding, for example19:41
persiacprov-lunch, In that case, it sounds like you've made the decision, which is good.  I'd recommend going through and wontfixing bugs like 263301 with a note that PPAs are designed for distribution, and while they can be used to test-build or for review, it is the user's responsiblity to take appropriate care with versioning, package names, etc.19:42
flacosteCoMp4c7: could be a session problem19:42
flacosteCoMp4c7: http://bazaar.launchpad.net/~freevial/freevial/php-db-edit/annotate/74?file_id=consumer.php-20080707143651-baxjp2psoha1dy4y-10 line 42519:43
cprov-lunchpersia: thanks for the encouragement, let me find my "hit-me" helmet and go on with that.19:44
lamalexhey guys, can anyone tell me how to make a member an administrator of a team/project?19:44
CoMp4c7flacoste: what I haven't tested yet if the store is set up correctly19:44
CoMp4c7and could be related with this19:44
persiacprov-lunch, No problem.  As I said before, it's not important to me which was chosen, but I thought the fence-straddling while leaving all the bugs open was a bad course to follow.19:45
beunolamalex, go to the team's page, go to "Show all members", and edit him to being an admin19:45
cprov-lunchpersia: thinking for a sometime, while fixing other problems, might have helped.19:46
persiacprov-lunch, Understood :)  It's great that you've gotten to the point where this becomes a sensible decision.19:47
lamalexbeuno: edit them how?19:47
lamalexi dont see the option anywhere19:47
flacosteCoMp4c7: in the file I pointed you at, look at 96419:48
lamalexah nevermind19:48
flacosteCoMp4c7: that's where it's failing, for some reasons it seems that it doesn't know about the association19:48
flacosteCoMp4c7: do you have a database available? you could change store implementation19:49
beunolamalex, do you have permissions to do so?19:49
lamalexyah, i found it, thanks19:49
flacosteCoMp4c7: there is a MySQL/PostgreSQL store implementation provided19:49
cprov-lunchpersia: I can only thank you for the help in make is 'sensible'.19:49
flacosteCoMp4c7: or SQLiteStore19:50
persiacprov-lunch, Feel free to ask me for a rant anytime.  If I don't have one stored up, I'll generate one.19:50
flacosteCoMp4c7: do you know if your server is a single host or a server farm?19:50
CoMp4c7flacoste: app uses a mysql database19:50
flacosteCoMp4c7: I suggest you switch to the MySQLStore implementation, that should be safer19:50
flacostejust in case, /tmp isn't shared across the different servers19:51
CoMp4c7flacoste: and faster19:51
CoMp4c7i will try that19:51
flacosteCoMp4c7: let me know if this solves your problem19:51
CoMp4c7php also saves Sessions to tmp19:51
persiaNCommander, https://launchpad.net/debomatic19:54
persiaahasenack, Actually, you might find that link useful as well.19:55
ahasenackpersia: hehe, was just checking it out, you mentioned deb-o-matic just a little time ago19:55
ahasenackpersia: have you used it?19:56
persiaahasenack, I've used it, but not hosted it.  No idea about the administration.  As a user, you just upload stuff, and it gets built, which is handy.19:57
ahasenackok19:57
cprov-lunchpersia: ehe, you are a priceless resource to us. We should send you LP t-shirts ;)19:59
persiacprov-lunch, I'm priceless when I recommend competing build-farm software?  I should start promoting trac here: I bet I'd get a tie pin :)20:01
kikocprov-lunch, and we can; just talk to marianna and she'll arrange. that is as soon as he stops promoting trac!!20:01
* persia stops promoting trac20:02
cprov-lunchpersia: see, you were doing well until you mentioned trac :)20:02
=== thumper_laptop is now known as thumper
cprov-lunchpersia: btw, suggesting pbuilder to users is the right thing to do, ubuntu developers do use it extensively before uploading sources, it's ideal for learning. I don't see any problem with that.20:04
beunocprov-lunch, I've been thinking if there's anything we can do on the UI that will make it more clearly that even after deleting a package, that "namespace" is used.20:04
persiaactually, many of us use sbuild, for closer similarity to the buildds.20:05
beunomaybe saying very clearly that the user has to upload a version above X?20:05
persiaIf the buildd chroots were easily available for download, we'd be based off those.20:05
cprov-lunchbeuno: deleted/superseded packages are listed with the appropriate filter. Are you thinking in something else ?20:07
beunocprov-lunch, do you have a quick link to where I can see how that looks now?20:07
cprov-lunchbeuno: source and binaries (if any) are available from librarian, all the metadata is listed, etc.20:07
cprov-lunchbeuno: go to a heavy PPA, let me find one.20:08
beunocprov-lunch, I'm thinking more along the lines of visually displaying to the owner of the PPA above *what* version they can upload from20:09
beunoso you don't have to dig into history, it's something that's right there on the main page20:09
cprov-lunchbeuno: uhm, ISWYM, interesting20:10
cprov-lunchbeuno: try https://edge.launchpad.net/~fta/+archive20:11
persiaIt's probably worth having a link to a precise description of the ordering of versions, so that users aren't confused by things like the odd behaviour of ~ or issues with things to the left or right of the final -20:11
cprov-lunchbeuno: https://edge.launchpad.net/~fta/+archive?field.name_filter=&field.status_filter=any results grouped by distroseries and sourcepackagename, possibly20:12
beunocprov-lunch, so, if a user deletes a package, it wouldn't show up on that list anymore, right?20:13
* fta ?20:13
beunoon +archive20:13
beunofta, nevermind, you're just a good example  :)20:14
cprov-lunchbeuno: it does, with 'any' or 'superseded' filter20:14
cprov-lunchbeuno: it's 'all there', literally.20:15
beunocprov-lunch, so if you upload a package, and then delete it, it still shows up?20:15
cprov-lunchbeuno: yes, as 'deleted'.20:15
beunocprov-lunch, so maybe we could show, just to the owner, greater than what versions of the packages they have to upload20:16
beunodoes that sound reasonable, or an overkill?20:17
cprov-lunchbeuno: sounded overkill20:17
persiabeuno, Do you think most users understand the rules well enough to benefit from greater highlighting?  I think most of the complaints are from folk who aren't using PPAs for distribution.20:18
cprov-lunchbeuno: which problem would we be solving in this case ?20:18
beunocprov-lunch, the bug where poeple have to hit their heads against re-uploading the same version after deleting it20:19
beunomaking it clearer that they can't in advance20:19
cprov-lunchbeuno: I'm afraid about trying to get too much information in a single page, the hidden area below each row is already a mystery for most users.20:19
beunopersia, well, they kinda ahve to know about versions, don't they?  As they specify them everywhere on the packaging20:19
beunothis may be just rambling on my part20:20
persiabeuno, Well, if they use dh_make or follow one of the many guides blindly, they probably don't understand why they are using -1, or why that might not be the best choice.20:20
cprov-lunchbeuno: the solution is the search for 'foo' in 'any' status, making that somehow intuitive20:21
persiabeuno, Also, I think most people don't necessarily understand that they are distributing something in a formal way, or that it's different from uploading to some web space.  The concept of "upload to replace" is fairly widespread.20:22
beunoyeah, that's kinda what I want to solve I think20:22
beunotry and convey that more clearly on the UI20:22
beunothat they're actually distributing it20:22
cprov-lunchgood point, but which UI artifact would make it clearer ?20:24
beunoquite a few20:24
persiaThat might be part of the activate-PPA interface.  Notify the user that they are creating an distribution archive for delivery to end-users, etc.20:24
beunoI think we should have a few beers at UDS  :)20:24
beunoall 3 of us20:24
cprov-lunchbeuno: :)20:24
CoMp4c7flacoste: i think i know why I get that error20:25
flacosteCoMp4c7: do tell me!20:25
CoMp4c7flacoste: no association file is created20:25
flacosteand why is that?20:25
flacostei thought FileStore should handle that20:25
CoMp4c7I don't know20:26
CoMp4c7when I login to claimid, the file is set up20:26
CoMp4c7but when login to launchpad not20:27
cprov-lunchbeuno, persia: maybe showing how many users have downloaded PPA packages might make that clear. "if you screw 1000 users will bug you."20:27
cprov-lunchalso, not in that negative way.20:27
persiacprov-lunch, No, because then the person who has a count of zero will want to know why they can't have an exception.20:27
flacosteCoMp4c7: are you sure you are hitting the same host everytime?20:28
flacosteCoMp4c7: there might be multiple /tmp20:28
cprov-lunchpersia: zero users is very likely to be a temporary status.20:28
flacostethat's why I suggested using MySQLStore20:28
CoMp4c7now i have set up the filestore in home20:28
CoMp4c7there was a problem with mysql, that is why i didnt test it20:29
CoMp4c7I have to go out20:29
=== kiko is now known as kiko-afk
cprov-lunchpersia: they will also see other PPAs heavily used and will want to reach the same status.20:29
CoMp4c7thank you very much20:29
persiacprov-lunch, Sure, but it probably makes sense to provide a counter once it's hit some threshold (e.g. 100).  Before that, just providing a generic warning avoids the case where people want exceptions.20:29
cprov-lunchpersia: right, false impression of unimportance. Anyway, just a crazy idea, we are not near the PPA-statistic task :(20:30
persiacprov-lunch, Yeah well.  Next month.20:31
RainCTbtw, can you guys remove feedback requests from blueprints? (https://blueprints.edge.launchpad.net/ubuntu-live-chat-support/+spec/ubuntu-live-chat-support got several support requests :P)20:31
=== bac is now known as bac_afk
=== hakr is now known as h[a]kr
lamalexhey guys, anyone around? i've got a question.  I'm trying to make a subproject, but my parent project isn't showing up in the list. Do I need to do anything special to flag a parent project?21:18
lamalexahh just saw this https://bugs.launchpad.net/launchpad/+bug/14933821:20
ubottuLaunchpad bug 149338 in launchpad-registry "Project Group field and search widget on project registration page need to be better explained" [Undecided,Triaged]21:20
lamalexis there anyone here who can create a project group?21:20
* paolettopn e' away (ciao a tutti!)21:22
=== paolettopn is now known as paolettopn_zzz
=== bac_afk is now known as bac
=== beuno_ is now known as beuno
Rocket2DMnAre there any LP admins available?  The user maurizio-live filed a bunch of garbage bugs which I have closed.  Further action may need to be taken to ban the user.22:33
beunoRocket2DMn, his account as already been suspended22:36
beunothanks for letting us know22:36
Rocket2DMncool, thank you beuno22:39
wgrantAh, it's good to see the Launchpad dev becoming a real wiki, with the clutter and inconsistent namespacing that normally plagues them...22:46
lamalexha22:51
=== h[a]kr is now known as hakr
lamalexany LP admins around can help me with https://answers.edge.launchpad.net/launchpad/+question/5181022:57
jmllamalex: not right now. kiko's the one who needs to deal with that.22:59
lamalexok np, that bug should be dealt with though, it's very unclear that LP needs to make a project group22:59
* wgrant checks the new bug page UI on staging, and wonders why it is offering a link to mark this bug as a duplicate when the next bit of text on the page shows that it can't be done.23:03
wgrantbeuno: ^^23:04
beunowgrant, the short answer is "it's complicated"23:09
beunoand we're working on a deeper solution23:09
wgrantbeuno: And why does the security flag hide behind a link which doesn't mention security?23:14
wgrantAnd people continue to break the actions menus...23:15
wgrantBug actions menus on staging now have an item labelled 'change', and another labelled ''. Lovely.23:15
=== kirkland` is now known as kirkland
=== hakr is now known as h[a]kr
=== jamesh_ is now known as jamesh

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