[07:06] <george_e> Code hosting seems to be broken.
[07:06] <george_e> "bzr: ERROR: Unable to connect to SSH host bazaar.launchpad.net; Error reading SSH protocol banner"
[07:08] <StevenK> No shells on this server.
[07:09] <StevenK> Works for me?
[07:09] <tsimpson> I keep getting "ssh_exchange_identification: Connection closed by remote host" both from bzr and ssh
[07:09] <StevenK> Now I do too
[07:11] <tsimpson> http works fine, just ssh that's falling over
[07:11] <StevenK> It's being investigated
[07:11] <george_e> Ah, okay.
[07:16] <sakhm2> hi, I'm uploading a source package to a ppa and get no e-mail response — I signed the code of conduct, signed the package, and launchpad has my key
[07:17] <bigjools> https://answers.launchpad.net/launchpad/+faq/227
[07:19] <StevenK> tsimpson: It should be back and looking good now.
[07:19] <tsimpson> StevenK: yep, looks good now
[07:20] <persia> Could someone recommend a better way to get LP source than `bzr branch lp:launchpad`?  This keeps timing out for me.
[07:21] <wgrant> persia: How much RAM do you have, and which version of bzr are you using?
[07:21] <lifeless> persia: the export tarball thing in loggerhead.
[07:22] <lifeless> persia: I wouldn't call it 'better' though
[07:22] <persia> wgrant: bzr 2.5.1 and 1.5G in the environment I've been trying
[07:22] <wgrant> You could likely use a little more RAM, but I'd normally expect that to work
[07:23] <persia> lifeless: Why wouldn't you use "better" for that?  Should I expect the same issues with download speed, or is it that I lose all the revisions and submitting patches becomes hard?
[07:23] <lifeless> the latter, and downloading shouldn't be any faster
[07:23] <george_e_> Ah, SSH is working now.
[07:25] <lifeless> persia: it might be faster in clock time (vs faster in bps)
[07:25] <persia> lifeless: How do I get to that in loggerhead?  I don't see such a link exposed in the UI view.
[07:26] <persia> lifeless: Faster in clock time is what I seek, although I admit that sometimes the bzr pull drops to 5kB/s, which I assume to be congestion at the DC
[07:26] <persia> s/pull/branch/
[07:49]  * persia finds the tarball button
[08:23] <sakhm2> in case someone is free and could help, I submitted a question about a missing response to PPA upload - https://answers.launchpad.net/launchpad/+question/213412
[16:43] <MCR1> cate
[16:43] <MCR1> ups - wrong window focused :)
[20:20] <jtaylor> hi, you can't login to launchpad with opera 12.10
[20:20] <jtaylor> when you press login you get a page is stale page
[20:22] <czajkowski> jtaylor: our brower support is chrome and FF https://dev.launchpad.net/GradedBrowserSupport
[20:22] <czajkowski> *chromium
[20:22] <jtaylor> I know you don't support opera
[20:22] <jtaylor> but at least login should work ...
[20:22] <jtaylor> I can live with all the other issues
[20:23] <dobey> jtaylor: file a bug please. irc isn't a bug reporting mechanism. use chrome or ff to log in and file a bug please.
[20:23] <jtaylor> how should I report a bug if I can't login?
[20:23] <jtaylor> ok I could use firefox ._.
[20:23] <dobey> use a browser that you can log in with
[20:23] <dobey> or use your phone
[20:23] <dobey> or something
[20:23] <dobey> login works fine on my phone at least
[20:24] <dobey> it is webkit though
[20:27] <jtaylor> yey restored my cookies from a backup and I'm back in
[20:27] <jtaylor> I'll file a bug
[20:42] <jtaylor> figured it out
[20:42] <jtaylor> I need to enable referrer information sending
[20:43] <jtaylor> isn't that something that can be caught on the server to give a better error message?
[20:45] <tsimpson> if it's not in the HTTP standard, it can't really be done portably (if at all)
[20:53] <andol> Is HTTP referers really something which should matter alt. be trusted in a login context?
[20:57] <tsimpson> redirects allow the login process to work securely, so yes
[21:47] <wgrant> andol, jtaylor: Referer checking is a standard anti-CSRF mechanism for HTTPS sites nowadays, so blocking Referer entirely is going to become more and more problematic as sites get more secure.
[21:47] <wgrant> The error message could certainly be more useful, however
[21:49] <wgrant> I think the wording is inherited from Django
[21:59] <andol> wgrant: Ahh, thanks for the pointer.