createuser: could not connect to database postgres: FATAL: role "xbmc" does not exist
xbmc is my linux user
ubuntu..
ianmcorvidae
-U postgres
psql/createuser/etc. connect as the role by the same name as your system user when not told otherwise
Orcie
ianmcorvidae: with what cmd?
ianmcorvidae
any of them provided by postgres, they're very consistent about their command line flags
presumably in this case for createuser
Orcie
ianmcorvidae: sorry I dont get it :)
Mineo
oh, luks and bitmap are both here: is there a planned release date for picard "bring-back-some-menus-for-german-users" 1.0.1?:)
ianmcorvidae
Orcie: http://www.postgresql.org/docs/9.1/static/refer... if you're running any of those commands, pass -U <some valid postgresql role name> or it will use the name of your system user (in your case, xmbc)
the error message you pasted was saying 'createuser:' so I assume that's what you were running
Orcie
yup
I am just confused by how the different users connect
there is a linux user and a postgresql user
ianmcorvidae
yes
by default postgresql assumes you want the postgresql user with the same name as your linux user -- but they are separate users
Orcie
what would be a good setup here
ok
so in this case I was trying to make a musicbrainz user
then the linux user should be musicbrainz aswell?
djce joined the channel
ocharles
good morning
ianmcorvidae
I don't know what convenience features mbslave provides; I know the perl server will create the user for you and you can just use a different command to get direct access that passes the user for you
these questions are all yours to answer, in any case, dependent on what you're looking for
you probably just need to learn how postgres works, which is something you should read the documentation for rather than limping along here :)
Orcie
I just want to query the database
ianmcorvidae: figures...
ianmcorvidae
or presumably mbslave has instructions, so just follow them
yeah, look at the mbslave README, it has a list of commands to follow
bitmap
Mineo: no, sadly…I forgot hack weekend is today so maybe next weekend? luks does 99.9% of the release process, so as long as he's not busy
ianmcorvidae
bitmap: still got 9.5 hours until the opening meeting!
bitmap
uh-oh, that's just enough time to sleep :P
ianmcorvidae
oh, heh, that :P
meh, I still don't know what I even want to do :P
bitmap
afaik I was going to integrate ruaok's rss stuff into picard, but we'll see
ianmcorvidae
huh, by maintaining a database of when files were last updated, or just writing last-updated stuff to tags?
nikki totally wishes picard had a database of what else she has so it would be able to figure out if she needs to disambiguate two releases
warp
ianmcorvidae: make a rename folder feature for picard for me!
ianmcorvidae
warp: rename folder?
bitmap
ianmcorvidae: I was thinking writing it to tags, but keeping a small database has some advantages
and disadvantages :)
ianmcorvidae
starting with writing it to tags is probably a good place to go for the hack weekend, I suppose
bitmap nods
nikki
how will that work?
I would've thought the idea was that you don't need to load files into picard again to find out if they've been updated or not
warp
ianmcorvidae: something which renames the folder a release is in according to some format, but doesn't otherwise move it around (nor the files inside of it).
nikki
but if the info is in the files, then...
Leftmost joined the channel
ianmcorvidae
I feel like the idea is you load the files and it'll be like nope no need to query the real WS this hasn't updated
i.e. load a file, or hit a button, or whatever you do to trigger it and it'll check the RSS feeds from the file's last-updated date until the present; if nothing relevant's in there, it says "you're good" and doesn't bother hitting the WS
and then you cache the RSS stuff so if you load any file that old or newer it doesn't need to hit *anything* to check
nikki
ah
although changing tagger script makes that awkward
ianmcorvidae
bitmap's ideas may be different, of course, but that's how I'd think about it anyway
warp
ianmcorvidae: I have the problem nikki describes. not the problem you're describing. so whatever nikki is referring to seems like a more interesting thing to solve.
ianmcorvidae
if your tagger script strips mbids it won't work, I guess, but
warp
why would you do that? the whole point of using picard is to get mbids in the files! :D
ianmcorvidae
warp: it's pretty common, I don't understand it either :/
nikki
ianmcorvidae: I mean right now I have $unset(asin), if I later remove that, my tags no longer match what picard thinks they should be if it loads the data from the real ws
ianmcorvidae
ah
that's harder, and presumably a case where you'd just need to retag everything then
nikki
or cache the ws data so you can compare it without actually requesting it
ianmcorvidae
I feel like fully caching all the WS calls picard makes is mostly just a really good way to use up disk space :P
nikki
maybe, maybe not :P
ianmcorvidae
though it would allow picard to calculate etags for webservice stuff, which can at least save bandwidth :P
nikki
I've never tried so I wouldn't know
ianmcorvidae
I mean, other than the "changing your tagger script" use case it's basically totally duplicate
I guess "picard learns how to write a new tag" would also fill in there
anything that changes the XML -> tags mapping, basically, but that's relatively stable for most use cases I think
most people decide what they want and stick with it is my impression :P
nikki
oh, and you can edit files outside of picard too
ianmcorvidae
sure, you could also open whatever database it ostensibly creates (I'd bet on either XML or sqlite :P)
or screw with the database's mapping by moving a file other than with picard
or whatever
nikki
I think it's far more likely that someone will change their files with another program than it is that they'll edit the cached data
I'm just saying that "mb data has not been changed since this file was tagged" does not mean "this file's tags match the mb data"
ianmcorvidae
true
but I think in most cases they'll both be true/false at the same times, and it's a lot easier to implement the former
(especially initially -- moving to a real cache/database/etc. system later wouldn't be impossible or anything)
totally roundabout reminder but: ocharles: at some point we should talk about varnish and the paid WS so I can take over that work (probably after the hack weekend though)
bitmap
nikki: "mb data has not been changed since this file was tagged" is what I was planning
nikki: I was thinking with a database, you'd open a window to see a list of albums that have changed
ianmcorvidae
ocharles: well, unless it's a small enough task I could just do it *for* the hack weekend, in which case I probably should just do that :)
nikki
ianmcorvidae: sure, I'm not expecting perfection immediately :P
ianmcorvidae
haha
fair enough :P
ocharles
ianmcorvidae: sure, feel free to email if you want to move the discussion offline too
nikki
bitmap: yeah, that sounds like what I'd expect. it could maybe even notify the user when they open picard if there's stuff that's changed
bitmap was planning on an icon in the statusbar for that :)
bitmap
I also wanted to store the file paths so you could double click on the album to load it and all of the files, but that's kind of iffy
ianmcorvidae
easily broken if anyone moves their files around
unless picard runs a daemon and watches everything with inotify
but that would be pretty crazy XD
er, well, inotify and equivalents on other platforms
nikki
it could just tell the user it can't find the files and ask them to locate them manually
I mean, most music players I've tried require manually updating the library if you move files around, so I don't think it would be too unexpected that picard can't find it if you've moved it
bitmap
yep, that's what I had in mind
I don't think itunes even updates your library automatically yet :P
nikki
I guess it would also need a delete option, in case they deleted it rather than moved it
bitmap nods
ianmcorvidae
ocharles: duly emailed :)
Orcie_ joined the channel
Orcie_
luks: I am getting a error on mbslave import
File "./mbslave-import.py", line 43, in <module>
psycopg2.OperationalError: fe_sendauth: no password supplied