#musicbrainz-devel

/

      • the_metalgamer joined the channel
      • kepstin-work joined the channel
      • kurtjx joined the channel
      • Leftmost joined the channel
      • navap is fed up with our anti-narrow-browser footer.
      • navap
      • bitmap
        looks nice
      • two lines feels like the max for a footer though, so as long as it doesn't creep beyond that :)
      • ianmcorvidae
        the open licenses note is new, yeah?
      • navap
        Yes
      • ianmcorvidae
        however, yeah, that looks better, it doesn't do a gross wrap thing at the narrowest it can get. +1 :)
      • that schwag link is sorta terrible too
      • in that it only has terrible cafepress crap :P
      • navap
        I added the data license link to balance the two sides out, but I'm thinking what if we just had that as the left side?
      • The only important link that we'd miss is the donate link. Everything else is already on the page or..schwag..
      • nikki joined the channel
      • ianmcorvidae
        hm
      • bitmap prefers keeping the links
      • I like the links to wiki/forums/tracker/etc. being not buried in menus
      • bitmap
        I hate using the menus
      • navap
        I would expect someone like you to skip the links and just type the urls in :p
      • ianmcorvidae
        well, yes, that's also true
      • I like having the links not buried in menus because menus suck, not because I use *either* :P
      • hm. the wiki link should also be scheme-independent
      • I should see about getting https on our other subdomains too
      • solve the "the default goes to the wrong place omg!" ticket with "okay fine, now you can't get to the default because everything's https" :P
      • navap changes the wiki link to //wiki.musicbrainz.org
      • cool, thanks
      • navap
        Are we putting the blog, jira, and forums on https as well?
      • ianmcorvidae
        they aren't yet
      • however, that's the intention
      • we have a *.musicbrainz.org cert, we should use it :P
      • navap
        Yeah
      • ianmcorvidae
        I haven't for those because blog/forums are apache and jira's it's own weirdness -- so I don't know the incantations already :)
      • meh, its*
      • twitter can also be scheme-independent FWIW
      • kurtjx joined the channel
      • ianmcorvidae idly wonders if the metabrainz site will become HTTPS at some point
      • obviously the donation stuff is through other sites, so that much isn't an issue, but
      • navap
        A single domain cert isn't too expensive I think
      • ianmcorvidae
        no, indeed not, I just don't know if that's something that's wanted :)
      • we'll probably want one for caa.org at some point, to silence warnings about insecure content
      • but we also need to pester the IA, because at present their https stuff is set up wrong
      • e.g. https://ia701207.s3dns.us.archive.org/mbid-189d... works, but there's a cert warning because it's for *.us.archive.org, not *.s3dns.us.archive.org
      • navap
        Shouldn't *.us.archive.org encompass *.s3dns.us.archive.org?
      • ianmcorvidae
        no
      • SSL wildcards only support one level
      • navap
        I see
      • ianmcorvidae
        yeah
      • it's somewhat annoying :/ but that's how it is
      • navap
        Is there a good reason for not using https as default or even forcing https via an nginx redirect?
      • ianmcorvidae
        not forcing people into it immediately
      • especially because userscripts are by and large not updated to account for https yet
      • kepstin
        navap: temporarily limiting exposure for testing purposes
      • ianmcorvidae
        also that :)
      • kepstin
        but yes; I do think that all logged-in access should be going through https; maybe even all website access.
      • on the other hand, doing non-authenticated webservice access unencrypted might be a good idea.
      • ianmcorvidae
        the webservice is the other thing, we don't have a way to guarantee our clients have stuff
      • we don't even redirect www.musicbrainz.org to musicbrainz.org for the webservice
      • which incidentally means it works better over HTTPS than the rest of the site, since our www -> non-www redirect is currently broken, but :P
      • kepstin
        that shouldn't be hard to fix in the nginx config tho
      • ianmcorvidae
        no, there's already a codereview up for it :)
      • kepstin
        yep, that looks about right :)
      • Prophet5 joined the channel
      • the nginx config for musicbrainz is rather complicated :/
      • ianmcorvidae
        the rewrites file is the biggest pain
      • because legacy URLs :?
      • the frontend configs aren't even public, that's another layer of complexity
      • kepstin
        something went wrong in my libvirt, so my musicbrainz vm has decided to disappear without a trace
      • I need to rebuild it now :/
      • onlything left is the frontend config in my nginx :)
      • ianmcorvidae
        haha
      • well, the *rest* is all in the git repository
      • it's just the frontend stuff that's in a separate private repo
      • kepstin
        admittedly, my frontend nginx config looks like this: https://raw.github.com/gist/3856052
      • very impressive ;)
      • ianmcorvidae
        looks about right, yeah
      • kepstin
        (that address is on a split dns setup, and can only be seen in my house)
      • ianmcorvidae
        if you want https, set X-MB-https to on/off accordingly :)
      • kepstin
        where are you terminating the https?
      • ianmcorvidae
        frontend
      • and then passing X-MB-https so the server instances know how to format links
      • kepstin
        and then you pass a header to the backend to indicate whether or not https was used?
      • ianmcorvidae
        yup :)
      • kepstin
        makes sense.
      • navap
        This http vs https link business is a bit complicated. e.g. links in emails
      • kepstin
        links in emails should be always https, imo.
      • since you only get an email if you've logged in at some point, and logged in users should always be using ssl :)
      • navap
        But what about non-musicbrainz.org servers?
      • kepstin
        fun.
      • I guess you'll have to add another parameter to DBDefs.pm :/
      • navap
        Shhh don't let ocharles hear you
      • I think I'll just leave this scheme-independant link fixing for later
      • kepstin suspects that most non-musicbrainz servers don't have email delivery configured anyways.
      • kepstin
        although the sandbox ones are a special case, i guess.
      • kepstin joined the channel
      • there we go, finally fixed my automounting authentication issues properly
      • ooh, fun. just set up my mail client with a startssl certificate, so I can now send signed emails :)
      • it's just a class 1 cert tho, so all it does is show that the person sending the email has control of the email address being sent from... which is still something.
      • btw, if anyone uses HTTPS-Everywhere, rules file for musicbrainz: https://gist.github.com/3856165
      • kepstin-work joined the channel
      • ianmcorvidae
        kepstin-work: the rule for beta has been 50% completion or better -- you're probably right but you do need to go through and check them :)
      • bitmap
        yay, the picard-daily ppa is finally building again
      • ianmcorvidae
        yay :D
      • hawke_1 joined the channel
      • hobbes is so sloooow
      • I wonder if test's database should be on rika and just the server code on hobbes
      • or something
      • Leftmost joined the channel
      • Prophet5 joined the channel
      • warp: if musicbrainz_db_warp on hobbes can be dropped, either do it or tell me; test is now using a different DB
      • (specifically: musicbrainz_db_20121008)
      • jdamcd joined the channel
      • warp, ijabz: related, if the test search server needs updating, someone should make that happen :) either by doing it yourselves or informing me how to do that
      • warp
        ianmcorvidae: I am not personally using musicbrainz_db_warp , was it being used by test?
      • (also, goodmorning! :)
      • djce joined the channel
      • RootWyrm_ joined the channel
      • ianmcorvidae
        yes, it's what test was using until a bit more than an hour ago
      • ijabz joined the channel
      • warp: also, are the "Show all pending edits" changes in the schema change branch?
      • warp: I don't see the script in upgrade.sh
      • which understandably worries me, given I thought testing for schema change stuff had already started
      • warp
        ianmcorvidae: could be
      • ianmcorvidae: I didn't merge it in, and if it's not in there apparently ocharles didn't either :)
      • (I have a shippit from you, not from oliver)
      • ianmcorvidae
        yeah
      • I had expected that the upgrade script, if nothing else, would be there
      • the RG coverart schema change scripts are in there
      • anyway
      • I guess it should at least be on test
      • warp nods.
      • Freso joined the channel
      • ijabz joined the channel
      • RootWyrm_ joined the channel
      • warp
        hm, lots of warnings in t::MusicBrainz::Server::Controller::Statistics
      • ianmcorvidae
        yeah
      • a lot of those are the result of it making sure the pages don't crash horribly when there are no statistics
      • ocharles has a patch up that silences lots of warnings, IIRC
      • warp
        I'm validating the html
      • RootWyrm_ joined the channel
      • but it looks like I'm going to have to silence another error
      • (silence as in ignore it, instead of fixing the generated html)
      • djce joined the channel
      • oh no, this is an error, not a warning. grmbl.
      • djce joined the channel
      • ijabz joined the channel
      • stefans joined the channel
      • ianmcorvidae
        so much inline styling in root/release/edit/tracklist.tt D: