#musicbrainz

/

      • pronik has quit
      • 2009-06-04 15509, 2009

      • sidd__ joined the channel
      • 2009-06-04 15543, 2009

      • luks
        aCiD2: do you have gdb installed?
      • 2009-06-04 15558, 2009

      • aCiD2
        i can install it if I don't
      • 2009-06-04 15508, 2009

      • luks
        ps aux | grep postgres
      • 2009-06-04 15521, 2009

      • aCiD2
        yup, got the pid
      • 2009-06-04 15548, 2009

      • luks
        there is many process, I'm not sure which one should be used for this
      • 2009-06-04 15500, 2009

      • luks
        +es
      • 2009-06-04 15501, 2009

      • aCiD2
        i can try them all :)
      • 2009-06-04 15507, 2009

      • luks
        ok, well..
      • 2009-06-04 15520, 2009

      • luks
        once you have the pid, run: gdb /path/to/postgres pid
      • 2009-06-04 15530, 2009

      • luks
        that should attach gdb to the process
      • 2009-06-04 15533, 2009

      • aCiD2
        ok
      • 2009-06-04 15553, 2009

      • luks
        then when it crashes, type 'bt' and it should give you a backtrace
      • 2009-06-04 15558, 2009

      • aCiD2
        attaching seems to stop the server, so I have to type "continue" in gdb, but then I never get back to the gdb prompt
      • 2009-06-04 15504, 2009

      • sonium has quit
      • 2009-06-04 15504, 2009

      • aCiD2
        ah wait
      • 2009-06-04 15537, 2009

      • aCiD2
        yea, I have to continue
      • 2009-06-04 15556, 2009

      • luks
        hm
      • 2009-06-04 15500, 2009

      • Sargun joined the channel
      • 2009-06-04 15502, 2009

      • luks
        which pid are you using?
      • 2009-06-04 15531, 2009

      • luks
        I'm trying it with the one that has argv[0] looking like a real command line
      • 2009-06-04 15535, 2009

      • aCiD2
        the main postgres one, but the others do the same (attaching prints: in __select_nocancel () from /lib/libc.so.6 then drops me back to the gdb prompt)
      • 2009-06-04 15516, 2009

      • russss
        you need the pid which your postgres session is connected to
      • 2009-06-04 15526, 2009

      • russss
        that should be in the proc title
      • 2009-06-04 15534, 2009

      • russss
        and yes, you do need to type continue
      • 2009-06-04 15547, 2009

      • russss
        then when it crashes, it will bring you back to the gdb prompt
      • 2009-06-04 15509, 2009

      • aCiD2
        so the pid of psql in this case?
      • 2009-06-04 15526, 2009

      • russss
        no
      • 2009-06-04 15538, 2009

      • russss
        the pid of the postmaster process which is serving your connection
      • 2009-06-04 15548, 2009

      • aCiD2
        ah
      • 2009-06-04 15516, 2009

      • aCiD2
        like: postgres: musicbrainz_user musicbrainz_db ?
      • 2009-06-04 15520, 2009

      • russss
        yup
      • 2009-06-04 15528, 2009

      • aCiD2
        ok, I have a backtrace luks
      • 2009-06-04 15532, 2009

      • luks
        cool
      • 2009-06-04 15541, 2009

      • aCiD2
        pasting now
      • 2009-06-04 15510, 2009

      • aCiD2
      • 2009-06-04 15527, 2009

      • aCiD2
        (error is a seg fault)
      • 2009-06-04 15548, 2009

      • luks
        hm, I wonder how to compile the extension in debug mode
      • 2009-06-04 15525, 2009

      • sidd has quit
      • 2009-06-04 15557, 2009

      • xcdfgkjhgcv joined the channel
      • 2009-06-04 15520, 2009

      • luks
        maybe just `CFLAGS="-g -O0" make` should work
      • 2009-06-04 15519, 2009

      • luks
        hm, no
      • 2009-06-04 15501, 2009

      • Rondom has quit
      • 2009-06-04 15541, 2009

      • xcdfgkjhgcv has quit
      • 2009-06-04 15529, 2009

      • xcdfgkjhgcv joined the channel
      • 2009-06-04 15534, 2009

      • navap
        So are we doing away with the separation of release group types? I.e. Instead of having separate tables, just adding a release group type column and allowing the user to sort via that.
      • 2009-06-04 15537, 2009

      • FauxFaux
        Heh.
      • 2009-06-04 15552, 2009

      • FauxFaux
        ,,,
      • 2009-06-04 15555, 2009

      • FauxFaux
        I never typed that.
      • 2009-06-04 15557, 2009

      • luks
        navap: no, we are using temporary solutions during development
      • 2009-06-04 15519, 2009

      • Muz_
        Ocks.
      • 2009-06-04 15520, 2009

      • Muz_
        Lolc.
      • 2009-06-04 15537, 2009

      • navap
        Okay, so does that mean that the artist page will still have multiple tables, or is it a "we'll worry about it when we get to it" kind of thing?
      • 2009-06-04 15520, 2009

      • luks
        it means that the plan is still to do this - http://oxygene.sk/lukas/mb/
      • 2009-06-04 15535, 2009

      • sonium joined the channel
      • 2009-06-04 15545, 2009

      • navap
        Ah right, okay :)
      • 2009-06-04 15523, 2009

      • luks
        aCiD2: ok, so there are two problems. I'm not handling libunac errors, and libunac for some reason can't unaccent the string
      • 2009-06-04 15552, 2009

      • aCiD2
        So maybe my libunac installation is broke?
      • 2009-06-04 15504, 2009

      • lancerocke has quit
      • 2009-06-04 15545, 2009

      • outsidecontext has quit
      • 2009-06-04 15507, 2009

      • outsidecontext joined the channel
      • 2009-06-04 15523, 2009

      • luks
        aCiD2: can you open unaccent.c, add PG_RETURN_TEXT_P(CStringGetTextP(GetDatabaseEncodingName())); below the unac_string line in unaccent(PG_FUNCTION_ARGS)
      • 2009-06-04 15532, 2009

      • luks
        then make && sudo make install
      • 2009-06-04 15550, 2009

      • luks
        and try to run unaccent(')
      • 2009-06-04 15505, 2009

      • aCiD2
        yup, ok
      • 2009-06-04 15511, 2009

      • luks
        this time without gdb, it shouldn't crash
      • 2009-06-04 15509, 2009

      • aCiD2
        no crash, returns SQL_ASCII
      • 2009-06-04 15513, 2009

      • luks
        hm
      • 2009-06-04 15521, 2009

      • luks
        aha, that's probably because you have DB in the C locale
      • 2009-06-04 15559, 2009

      • aCiD2
        should it not be in that locale?
      • 2009-06-04 15540, 2009

      • luks
        according to InitDb.pl it should be
      • 2009-06-04 15546, 2009

      • luks
        can you try one more thing?
      • 2009-06-04 15557, 2009

      • lancerocke joined the channel
      • 2009-06-04 15510, 2009

      • aCiD2_ joined the channel
      • 2009-06-04 15531, 2009

      • aCiD2_
        luks: yup, what else should I change?
      • 2009-06-04 15551, 2009

      • luks
        revert the last change
      • 2009-06-04 15559, 2009

      • luks
        and replace the unac_string line with unac_string("UTF8", input, input_len, &output, &output_len);
      • 2009-06-04 15509, 2009

      • aCiD2_
        ok
      • 2009-06-04 15513, 2009

      • luks
        and then run e.g. unaccent('á');
      • 2009-06-04 15501, 2009

      • aCiD2_
        that seems to work
      • 2009-06-04 15517, 2009

      • aCiD2_
        yup
      • 2009-06-04 15558, 2009

      • luks
        damn
      • 2009-06-04 15513, 2009

      • luks
        I wonder how does it manage to store 8-bit characters in ASCII :)
      • 2009-06-04 15523, 2009

      • luks
        but I guess it's not really ASCII
      • 2009-06-04 15522, 2009

      • luks
        I'm not sure what now
      • 2009-06-04 15549, 2009

      • luks
        I really don't like the idea of hardcoding UTF8 there
      • 2009-06-04 15536, 2009

      • aCiD2_
        unless it is just a problem with my database
      • 2009-06-04 15503, 2009

      • luks
        I doubt it is
      • 2009-06-04 15519, 2009

      • luks
        you told it to use the C locale, it does so
      • 2009-06-04 15503, 2009

      • Clint
        what does using ASCII instead of UTF-8 buy you?
      • 2009-06-04 15532, 2009

      • luks
        I'm not sure
      • 2009-06-04 15555, 2009

      • luks
        the check in the database initialization script was there for ages
      • 2009-06-04 15508, 2009

      • luks
        I don't think *anybody* knows why it's there :)
      • 2009-06-04 15542, 2009

      • russss
        do SHOW client_encoding()
      • 2009-06-04 15546, 2009

      • russss
        in psql
      • 2009-06-04 15551, 2009

      • russss
        with a semicolon
      • 2009-06-04 15513, 2009

      • russss
        maybe also SHOW server_encoding();
      • 2009-06-04 15539, 2009

      • Clint
        maybe without parens
      • 2009-06-04 15552, 2009

      • russss
        yeah
      • 2009-06-04 15556, 2009

      • russss
        maybe that
      • 2009-06-04 15506, 2009

      • russss
        long day :/
      • 2009-06-04 15517, 2009

      • aCiD2_
        brb
      • 2009-06-04 15522, 2009

      • aCiD2_ has quit
      • 2009-06-04 15518, 2009

      • aCiD2_ joined the channel
      • 2009-06-04 15544, 2009

      • aCiD2 has quit
      • 2009-06-04 15558, 2009

      • aCiD2_ has quit
      • 2009-06-04 15536, 2009

      • xcdfgkjhgcv
        I think Piicard may have crashed...
      • 2009-06-04 15543, 2009

      • xcdfgkjhgcv
        It's not doing anything.
      • 2009-06-04 15509, 2009

      • sonium has quit
      • 2009-06-04 15521, 2009

      • xcdfgkjhgcv
        Yay it's come back to life!
      • 2009-06-04 15508, 2009

      • outsidecontext has quit
      • 2009-06-04 15515, 2009

      • sonium joined the channel
      • 2009-06-04 15546, 2009

      • outsidecontext joined the channel
      • 2009-06-04 15511, 2009

      • rfzero joined the channel
      • 2009-06-04 15543, 2009

      • sonium has quit
      • 2009-06-04 15547, 2009

      • odyssey joined the channel
      • 2009-06-04 15551, 2009

      • ruaok joined the channel
      • 2009-06-04 15524, 2009

      • Alan_New joined the channel
      • 2009-06-04 15514, 2009

      • outsidecontext
        luks: hi
      • 2009-06-04 15554, 2009

      • luks
        hi
      • 2009-06-04 15507, 2009

      • ruaok waves
      • 2009-06-04 15536, 2009

      • Sargun has quit
      • 2009-06-04 15512, 2009

      • outsidecontext waves at ruaok
      • 2009-06-04 15535, 2009

      • outsidecontext
        luks: do you think you can provide me with the necessary stuff for a windows build of picard next week?
      • 2009-06-04 15550, 2009

      • luks
        outsidecontext: I think I can't :(
      • 2009-06-04 15556, 2009

      • brianfreud
        hi ruaok, outsidecontext, and everyone else :)
      • 2009-06-04 15502, 2009

      • ruaok
        wn brianfreud
      • 2009-06-04 15503, 2009

      • luks
        but I think I should be able to build it myself
      • 2009-06-04 15511, 2009

      • ruaok
        wb, that is. :)
      • 2009-06-04 15528, 2009

      • outsidecontext
        luks: ok, that's another solution
      • 2009-06-04 15532, 2009

      • brianfreud
        :)
      • 2009-06-04 15510, 2009

      • outsidecontext
        luks: i think if you could give me some directions i could figure out most stuff myself. but without anything i'm lost
      • 2009-06-04 15529, 2009

      • outsidecontext
        luks: and hey, how was paris?
      • 2009-06-04 15556, 2009

      • luks
        paris was great. I really needed to take a few days off :)
      • 2009-06-04 15515, 2009

      • outsidecontext
        i can imagine that :)
      • 2009-06-04 15522, 2009

      • luks
        regarding picard, the trickest part is to build ffmpeg
      • 2009-06-04 15529, 2009

      • luks
        you need msys for that
      • 2009-06-04 15530, 2009

      • brianfreud
        Personally, I think I need a vacation from my "vacation". :D
      • 2009-06-04 15536, 2009

      • brianfreud
        Driving to Florida and back was... exhausting
      • 2009-06-04 15541, 2009

      • luks
        but it didn't work out of the box last time I tried
      • 2009-06-04 15558, 2009

      • luks
        but there is a whole site about that somewhere
      • 2009-06-04 15540, 2009

      • luks
      • 2009-06-04 15544, 2009

      • outsidecontext
        how is that whole python stuff build on windows?
      • 2009-06-04 15557, 2009

      • luks
        well, that's easy
      • 2009-06-04 15510, 2009

      • outsidecontext
        luks: the last time i was trying to build the picard C modules they wouldn't run on windows. but i guess my compiler was not compatible with the binary python i had installed
      • 2009-06-04 15526, 2009

      • luks
        you can use mingw, so that's fine
      • 2009-06-04 15540, 2009

      • outsidecontext
        ok
      • 2009-06-04 15544, 2009

      • luks
        I was using msvc for picard builds
      • 2009-06-04 15549, 2009

      • luks
        but mingw should work without problems