[00:20] <lifeless> wgrant: hi
[00:21] <lifeless> wgrant: do you know of any soyuz services that do librarian uploads (directly)
[00:22] <wgrant> lifeless: process-upload.py, buildd-manager, process-pending-packagediffs.py
[00:22] <wgrant> There may be more... thinking...
[00:23] <wgrant> gina
[00:23] <wgrant> Possibly process-accepted.py, for translations stuff.
[00:24] <wgrant> Also the mirror prober, although most seem so not think that it's Soyuz.
[01:01] <lifeless> wgrant: there is an rt needs to know
[01:04] <wgrant> lifeless: Which ticket?
[01:35] <lifeless> wgrant: HA librarian upload, in the launchpad queue
[01:35] <lifeless> wgrant: we need the machines those soyuz things run on added to firewall rules
[01:37] <wgrant> lifeless: Perhaps the ticket is not visible to warthogs.
[01:37] <wgrant> I cannot see it.
[02:40] <lifeless> bah
[02:40] <lifeless> uhm
[02:40] <lifeless> it is visible to warthogs
[02:43] <lifeless> 41379
[02:43] <lifeless> wgrant: ^
[02:44] <wgrant> lifeless: Oh, it's using a custom status that's not in the default filter.
[02:48] <wgrant> lifeless: So: cesium, cocoplum, germanium, iron at least need adding to the list.
[02:49] <wgrant> I'm also somewhat surprised that loganberry isn't there.
[11:59] <pcjc2> is there any guide / examples which I could follow for writing a git commit hook to change bug statuses in LP ?
[11:59] <pcjc2> I'm figuring that I need to setup a LP account for my "commit hook", so it can authenticate as "gpleda.org Commit Hook robot" or something
[12:00] <pcjc2> The intention is to parse our git commits for "Closes-bug: lp-12345", and turn that into an API request which closes the bug
[12:01] <pcjc2> I figure it is easiest to have the robot user close the bug, but it would also be possible (in theory at least), to grab the committer / author email from the commit, look them up on LP (or via a mapping table we could maintain locally), then _pretend_ the bug closure was done by that user
[12:02] <pcjc2> (Although I'm not sure how to go about getting developers to authorise the robot to close bugs on their behalf)
[12:02] <maxb> to do that, the robot would have to have access to impersonate all your committers. Unlikely they'd accept that
[12:02] <pcjc2> "they" as in LP, or "they" as in our developers?
[12:02] <maxb> the rest ought to be pretty trivial with python's launchpadlib
[12:03] <maxb> b = lp.bugs[12345]; b.status = "Fix Released"; b.lp_save()  ---- or thereabouts
[12:03] <pcjc2> Getting the robot to authenticate might be an issue, but I've not read into it so much
[12:03] <pcjc2> maxb: That IS easy.. (easier than I expected) ;)
[12:04] <maxb> You probably want to add a comment to the bug too
[12:04] <pcjc2> of course.. with a link to the SHA1 which closed it
[12:04] <pcjc2> copy the commit message too - whatever
[12:04] <pcjc2> doesn't the robot / LP-lib client have to log on via OpenID ?
[12:05] <maxb> No, OAuth
[12:07]  * pcjc2 reads https://help.launchpad.net/API/SigningRequests
[12:08] <pcjc2> we could get our developers to opt-in and allow the robot to impersonate them I guess? From the sounds of that
[12:11] <maxb> well, you'd have to get them to opt-in to allowing your bot to impersonate them for any api action on Launchpad
[12:11] <maxb> anyone with any sort of security instinct would refuse
[12:12] <maxb> I would
[12:14] <maxb> oh, one thing I missed from my pseudo-code is that you need to change status on a bug_task not a bug
[12:15] <maxb> So you iterate the bug.bug_tasks collection to find the one that is for your project
[12:15] <pcjc2> got it.
[12:16] <pcjc2> Stupid question - how should I go about creating a LP user for my bot?
[12:16] <maxb> Just sign up as a new user, using a dedicated email address for the bot
[12:17] <pcjc2> (And presumably I can manually go through the steps of authenticating the bot to that user, then store the auth keys in the bot)
[12:17] <maxb> indeed
[12:17] <pcjc2> hmm - that is not so good... if I need a dedicated email for the bot
[12:17] <pcjc2> I don't have one - and I can't think where to set one up which isn't amaturish
[12:17] <pcjc2> mybot@gmail.com or whatever
[12:18] <pcjc2> Any way to have a bot without an email address?
[12:19] <maxb> LP users are always identified by their email address
[12:19] <wgrant> An active Launchpad account requires an email address, unfortunately.
[12:22] <pcjc2> Imported users from SF don't always have email addresses - is that a special case?
[12:22] <wgrant> They're not active.
[13:31] <pcjc2> I've asked our project admin to set up an email address (If possible) on the domain hosting out git repository
[16:02] <maxb> pcjc2: Oh, by the way, if you want to track people closing bugs, you might choose to change the assignee of the bug task (if it's not already set) in your robot
[17:15] <pcjc2> maxb: good idea
[17:15] <pcjc2> A friend of mine set up an email alias with his domain provider, so... I now have:
[17:15] <pcjc2> https://launchpad.net/~gpleda-launchpad-robot
[17:16] <pcjc2> (Now just have to write the code to make it do my bidding ;))
[20:35] <pcjc2> https://login.launchpad.net/+forgot_password
[20:36] <pcjc2> Does that require any lp-dev interaction - there are a few of our users who want to do bug triage, but they have forgotten their LP passwords
[20:42] <pcjc2> apparently they haven't received any email with confirmation code
[20:50] <pcjc2> anyone about?
[20:50] <lifeless> hi
[20:51] <pcjc2> Apparently the email sent by https://launchpad.net/people/+requestmerge?field.dupe_person=djdelorie is not going through for that user
[20:51] <lifeless> lp no longer stores passwords
[20:51] <lifeless> thats really an alias to login.ubuntu.com
[20:51] <pcjc2> I got the wrong end of the stick with the original complaint
[20:51] <pcjc2> he wants to merge ~djdelorie and ~dj-delorie
[20:51] <lifeless> ah ok
[20:52] <pcjc2> (~dj-delorie is the registered on LP account, but he wishes to rename it to ~djdelorie (the one which will be merged), once the merge is done)
[20:52] <lifeless> sure
[20:52] <lifeless> so what happens when he merges
[20:52] <pcjc2> DJ is a smart guy (developer of DJGPP and all that), so I trust him when he says the mail doesn't get through - he's been watching his server logs
[20:52] <pcjc2> the email alias from the bug import I had is djdelorie@users.sf.net
[20:52] <pcjc2> but it would appear the confirmation email is getting lost in the process
[20:53] <lifeless> ok
[20:53] <lifeless> so possibilities are:
[20:53] <lifeless>  - our outbound mail is on the blink (low, very reliable service)
[20:54] <lifeless>  - greylisting is in place at one or more of the mail providers chaining through users.sf.net
[20:54] <lifeless>  - the mail chain is incorrect and mail to djdelorie@users.sf.net doesn't end up where you'd want it to
[20:54] <lifeless> AIUI the usual thing is greylisting
[20:54] <pcjc2> strange thing is my bug export has djdelorie@users.sf.net
[20:55] <pcjc2> but he says when he tries to claim it, Launchpad says the email is going to djdelorie@users.sourceforge.net
[20:55] <pcjc2> any chance we can cheat this one, or does that need a LOSA?
[20:55] <lifeless> there is a link on the page to file support tickets
[20:55] <lifeless> under 'note:'
[20:56] <lifeless> also this is a email address search bug
[20:56] <lifeless> sorry
[20:56] <lifeless> spam protection bug
[20:56] <lifeless> shouldn't be able to get the email address for an arbitrary user by requesting a merge
[20:56] <lifeless> pcjc2: please let him know I clicked through myself to verify, so one of the confirmations he may receive is to merge with me... I advise against doing that :)
[20:57] <pcjc2> he just got YOUR email lifeless ;)
[20:57] <pcjc2> but he never got his own
[20:57] <lifeless> interesting
[20:57] <lifeless> uhm
[20:57] <pcjc2> he's going to try again
[20:57] <lifeless> does he have (in his active account) his email address hidden?
[20:58] <pcjc2> yes
[20:58] <lifeless> it can be set to show-always, show-loggedin-users, hide-completely
[20:58] <lifeless> what From: and sender: did the one he got, have?
[20:58] <pcjc2> he's now whitelisted the @sf.net email address
[20:58] <pcjc2> Turns out it liked your email, but not his own
[20:58] <lifeless> his spam filter?
[20:59] <lifeless> what From: and sender: did the email from me have?
[20:59] <pcjc2> yes, apparently
[20:59] <lifeless> and what ones did the one from him have, if he's gotten one now
[20:59] <pcjc2> hi gde (@lifeless, gde was the other person who is stuck with the merge)
[21:01] <lifeless> pcjc2: I'm wondering if mail-confidentiality settings are preventing correct sending of the mail
[21:01] <gde> lifeless: pcjc2 Hi there
[21:01] <lifeless> pcjc2: so if you can get those questions answered, that would be great
[21:01] <pcjc2> lifeless: He got it in the end, but I've asked him for the info
[21:01] <lifeless> pcjc2: also say, hi, long time no see ;)
[21:02] <pcjc2> (21:02:03) djgpp: To: djdelorie@users.sf.net
[21:02] <pcjc2> (21:02:04) djgpp: From: Launchpad Account Merge <noreply@launchpad.net>
[21:02] <pcjc2> (21:02:10) djgpp: Sender: bounces@canonical.com
[21:02] <lifeless> pcjc2: thats after he whitelisted?
[21:02] <lifeless> pcjc2: I'm guessing the one from me had my email address in the from: ?
[21:05] <pcjc2> All emails came from the same From address
[21:05] <lifeless> interesting
[21:05] <pcjc2> and it turns out, were only received after he whitelisted the @sf address
[21:05] <pcjc2> So the thing which was misleading was the address which LP claimed to be sending email to
[21:05] <lifeless> how so?
[21:05] <pcjc2> The lack of merge request was apparently just due to some spam filter / server setting
[21:05] <lifeless> djgpp: hi, thanks for dropping in
[21:06] <djgpp> hi
[21:06] <djgpp> when I tried to recover a lost password on "djdelorie", it said it was sending a confirmation token to djdelorie@users.sourceforge.net
[21:06] <djgpp> (I still haven't gotten those mails yet though)
[21:06] <lifeless> thats very odd
[21:06] <lifeless> when I tried it it said "An email message was sent to djdelorie@users.sf.net. Please follow the instructions on that message to complete the merge."
[21:06] <pcjc2> ahh.. I guess that is the leak
[21:07] <djgpp> the two merge requests came through to "To: djdelorie@users.sf.net"
[21:07] <lifeless> we don't have a dba/sysadmin handing around in the weekends, and our staging copy won't have your data in it yet.
[21:07] <lifeless> but I can arrange to look next week and see if there are two emails on the ~djdelorie account
[21:08] <djgpp> and I can get my icon to change...
[21:08] <djgpp> well, there aren't *now*
[21:08] <pcjc2> so DJ - what you did was try to login with the old ~djdelorie account (the SF import one), and it leaked out / gave you the wrong address
[21:08] <lifeless> oh, click - I see
[21:08] <djgpp> I tried to log in to the *new* ~djdelorie (from the sf import), but yes.
[21:08] <pcjc2> and presumably, it is a bug for that to leak _any_ email address
[21:08] <lifeless> so, I think I know where the friction is
[21:08] <djgpp> dj-delorie was my existing LP account (auto-generated from "DJ Delorie" I suppose)
[21:09] <lifeless> password recovery is a separate system, separate db - its the openid database
[21:09] <pcjc2> I'm still curious where it got the incorrect email from
[21:09] <lifeless> account merge is purely launchpad
[21:09] <lifeless> the isd folk may know
[21:09] <pcjc2> should we (or someone) file a bug about the email address being given out for that account?
[21:09] <djgpp> s/can/can't/ get my icon to change
[21:10] <lifeless> (Information Services Development)
[21:10] <lifeless> pcjc2: please file three bugs ...
[21:10] <lifeless> 1) that the email addresses via the two routes were inconsistent
[21:10] <lifeless> 2) that the password recovery handed out the email willy-nilly
[21:10] <lifeless> 3) that the merge form handed out the email
[21:11] <lifeless> there are some tensions between diagnosing faults and preventing disclosure to spammers
[21:11] <lifeless> but we should at least take a look at it
[21:12] <lifeless> gde: you were having trouble too ?
[21:12] <pcjc2> ok. I bet it confuses them because DJ renamed his existing LP account dj-delorie to djdelorie (the merge one) after the merge ;)
[21:12] <djgpp> and deleted the SF email from it, too :-)
[21:13] <lifeless> that will be tricky to diagnose
[21:13] <pcjc2> sounds like something we should recreate on staging
[21:13] <lifeless> possibly
[21:13] <lifeless> possibly not ;)
[21:13] <pcjc2> (but the OpenID bit is production only, right)
[21:13] <lifeless> its complicated ;)
[21:14] <lifeless> I have to finish packing, flying to dallas this afternoon
[21:14] <lifeless> you're all sorted, right ?
[21:14] <djgpp> my account seems OK now, although I should log out and back in again...
[21:15] <pcjc2> gde?
[21:15] <djgpp> yup, I can log out/in
[21:15] <djgpp> still doesn't have my icon...
[21:16] <lifeless> branding?
[21:16] <djgpp> "change details" mugshot
[21:17] <lifeless> it hasn't saved it? or its not showing it on some page you're looking at ?
[21:17] <djgpp> if I go there, it has my photo, but the homepage (/~djdelorie) still has the generic blue-shirt icon.
[21:17] <djgpp> ah!  That's a branding logo.  GRRR!  That was NOT obvious.
[21:17] <pcjc2> gde's import was this account https://launchpad.net/~gareth-uk
[21:17] <pcjc2> I don't know what his actual LP account is
[21:18] <pcjc2> API couldn't find the email address I have for him
[21:18] <lifeless> if its hidden api cannot see it
[21:19] <lifeless> djgpp: yeah, I would like us to auto-resize by default and allow overrides
[21:19] <lifeless> djgpp: but noone has contributed a patch as yet :)
[21:21] <pcjc2> lifeless: gde's issue was that he's forgotten his LP login credentials
[21:22] <pcjc2> but we can't figure out what his proper LP account is called
[21:22] <lifeless> ask a question on answers.launchpad.net/launchpad
[21:22] <djgpp> I mean it *says* "that will be displayed on your home page in Launchpad" but it doesn't, and doesn't say that the icon that *is* displayed is set via branding, not details.
[21:22] <lifeless> will need losa assistance
[21:24]  * djgpp must get back to other projects...
[21:26] <pcjc2> I've pointer gde at https://login.launchpad.net/+forgot_password
[21:26] <pcjc2> So apparently the merge process is confusing for non-logged in, non-regular Launchpad users.
[21:29] <pcjc2> If you click the "I am <user>?" link when logged out, you get forwarded to the login page, where the "forgot password" option is pretty tempting
[21:29] <pcjc2> OOPS: (Error ID: 1834carambolalaunchpad18)
[21:30] <pcjc2> That is me playing about, Probably LP shouldn't "Oops!"
[21:31] <gde> lifeless: that link doesn't work for me - no auth email arrives
[21:32] <pcjc2> @gde: Check spam folder / whitelist?
[21:32] <pcjc2> I tried it for my account and the email came from 	Launchpad Login Service <noreply@launchpad.net>
[21:33] <pcjc2> But to avoid information leaks, it is possible that LP doesn't tell you if you enter an email address it doesn't know of an account for
[21:33] <pcjc2> it may just silently pretend to have emailed you, and then get on with its business
[21:33] <pcjc2> (to stop people probing for accounts)
[21:33] <pcjc2> (lifeless can probably confirm / deny)
[21:36] <lifeless> pcjc2: thats another bug to file
[21:36] <pcjc2> (which?)
[21:36] <pcjc2> the OOPS?
[21:36] <gde> pcjc2: nothinhg in spam, and it's hosted gmail so there's no explicit whitelisting afaik
[21:36] <lifeless> 10:29 < pcjc2> If you click the "I am <user>?" link when logged out, you get forwarded to the login page, where the "forgot password" option is pretty tempting
[21:36] <lifeless> 10:29 < pcjc2> OOPS: (Error ID: 1834carambolalaunchpad18)
[21:36] <pcjc2> ah..
[21:36] <lifeless> I can't debug now, still packing
[21:37] <pcjc2> I've no idea what the OOPS was caused by, but I can try to reproduce
[21:37] <lifeless> just file the bug
[21:37] <lifeless> we get a backtrace in the oops
[21:37] <pcjc2> Filing
[21:37] <lifeless> and request variables etc; its usually enough to figure out the defect
[21:38] <pcjc2> its not related to https://bugs.launchpad.net/launchpad/+bug/644824 is it?
[21:38] <_mup_> Bug #644824: User created a second Launchpad account; now gets mixed success/denied to various services <canonical-losa-lp> <defect> <lp-foundations> <Canonical SSO provider:Confirmed> <Launchpad itself:Fix Released by stub> < https://launchpad.net/bugs/644824 >
[21:38] <lifeless> related yes, dupe no
[21:38] <pcjc2> Will point you at the bugs I file shortly
[21:39] <lifeless> no need :)
[21:39] <lifeless> I have ~ 0 time today
[21:39] <lifeless> (sorry)
[21:42] <pcjc2> NP, thanks for helping us!
[21:42]  * pcjc2 goes back to bug filing and leaning LP API
[21:43] <gde> lifeless: I think I'm good now- I needed to use an old email address
[21:43] <gde> lifeless: thanks for the help
[22:05] <pcjc2> any way to search for all bugs in a project with a remote bug watch?
[22:06] <pcjc2> (Short of me re-importing our bug data on a local instance and running SQL?)
[22:06] <pcjc2> The old bug watches auto-added based upon comments are no good because the SF trackers are shut down
[22:09] <lifeless> so you want third party bug watches? not sf ?
[22:09] <lifeless> may need a tweak to bugtask.py & the search interface
[22:09] <lifeless> but we do have a field in the backend thats (IIRC) indexed)
[22:12] <pcjc2> I just wanted to seek out bugs which had external watches (incorrectly) added, so I could delete the bugwatches
[22:13] <pcjc2> I've already removed the couple I spotted by hand
[22:49] <pcjc2> ‏‫‏, #70048
[22:49] <_mup_> Bug #70048: Python-crash while using Deluge-Torrent-Client <deluge-torrent (Ubuntu):Invalid> < https://launchpad.net/bugs/70048 >
[22:50] <pcjc2> Lifeless, 5x Bugs filed against LP, Bug#700483, Bug#700490, Bug#700491, Bug#700493, Bug#700496
[22:50] <_mup_> Bug #700483: Email address leaked by account merge request <Launchpad itself:New> < https://launchpad.net/bugs/700483 >
[22:50] <_mup_> Bug #700490: Incorrect email address presented when merging accounts <Launchpad itself:New> < https://launchpad.net/bugs/700490 >
[22:50] <pcjc2> comeon mup, and the rest?
[22:50] <pcjc2> Bug#700491, Bug#700493
[22:50] <pcjc2> Bug#700496
[23:09] <lifeless> pcjc2: you may wish to file a request for the search you couldn't do
[23:18] <pcjc2> gah... don't want to make work for you guys
[23:18] <pcjc2> more appropriate would be a way to disable auto-adding them for imported bugs (no point linking to the tracker you're just about to replace)
[23:19] <pcjc2> but I have run out of energy for tonight
[23:24] <pcjc2> https://bugs.launchpad.net/launchpad/+bug/700507
[23:24] <_mup_> Bug #700507: Feature request: Search for bugs with an external bugwatch <Launchpad itself:New> < https://launchpad.net/bugs/700507 >
[23:43] <lifeless> pcjc2: polish isn't work ;)
[23:43] <lifeless> pcjc2: its just polish