#musicbrainz

/

      • Infinito_ has quit
      • 2008-02-19 05008, 2008

      • nikki
        brianfreud: word
      • 2008-02-19 05009, 2008

      • FauxFaux
        Excel 2007 doesn't even seem to do it here, dammit.
      • 2008-02-19 05013, 2008

      • brianfreud
        FauxFaux: what would it be to remove every <table>allcontents</table>? /<table*table>/?
      • 2008-02-19 05042, 2008

      • brianfreud
        ouch, Nikki, I don't think you can in word, unless you completely turn off all auto-fixers (it's seeing the new cell as a new sentence)
      • 2008-02-19 05047, 2008

      • FauxFaux
        brianfreud: Just put a .*? in the middle, * so doesn't do that. =p
      • 2008-02-19 05009, 2008

      • nikki
        that's the thing, I turned off capitalisation changing autoformats and it's still doing it!
      • 2008-02-19 05017, 2008

      • brianfreud
        --- /<table,*?table>/g?
      • 2008-02-19 05027, 2008

      • nikki
        so I don't know where to look for table-specific settings
      • 2008-02-19 05028, 2008

      • FauxFaux
        Assuming that commas is a typo, yeah.
      • 2008-02-19 05035, 2008

      • brianfreud
        did you try turning off all grammer checking completely?
      • 2008-02-19 05042, 2008

      • FauxFaux
        nikki: Mmm, do you not get a little drop-down next to the letter when it changes it?
      • 2008-02-19 05049, 2008

      • brianfreud
        lol, yes, a typo
      • 2008-02-19 05052, 2008

      • nikki
        for about 5 milliseconds :/
      • 2008-02-19 05004, 2008

      • brianfreud
        need to clear those "helpful hints" google gives you :P
      • 2008-02-19 05015, 2008

      • FauxFaux
        Comes back when you click on the word here.
      • 2008-02-19 05019, 2008

      • nikki
        ohh
      • 2008-02-19 05022, 2008

      • nikki
        thanks!
      • 2008-02-19 05034, 2008

      • nikki hugs FauxFaux
      • 2008-02-19 05043, 2008

      • FauxFaux
        \o/
      • 2008-02-19 05003, 2008

      • FauxFaux had a screenshot full of loldongs all prepared and everything. :|
      • 2008-02-19 05058, 2008

      • brianfreud may actually have found something useful about GM_xmlhttpRequest returning as string instead of object... :P
      • 2008-02-19 05040, 2008

      • brianfreud
        FauxFaux, if I want to only replace a few times - more than once, but less often than all, how would I do it? /foo/2 to replace twice?
      • 2008-02-19 05041, 2008

      • brianfreud
        or is that when I start chaining replaces :P
      • 2008-02-19 05046, 2008

      • FauxFaux
        Unfortunately not, you basically have to chain replaces, you can obviously do it in a loop. For some simple things, you may be able to get away with (abc){3}, which means "abcabcabc" (ie. abc three times).
      • 2008-02-19 05043, 2008

      • brianfreud
        ok, thanks :)
      • 2008-02-19 05059, 2008

      • brianfreud
        I think I may be able to abuse the DOM so I don't have to worry about it :P
      • 2008-02-19 05013, 2008

      • brianfreud finds something rather twisted about this:
      • 2008-02-19 05015, 2008

      • brianfreud
        a.innerHTML = a.innerHTML.replace(/<table.*?table>/,"");
      • 2008-02-19 05038, 2008

      • brianfreud
        where a is a node within the DOM :P
      • 2008-02-19 05013, 2008

      • FauxFaux
        Sloooooow.
      • 2008-02-19 05002, 2008

      • brianfreud
        meh, I had to slow this down anyhow so as to not flood servers with req's - slow here is actually ok, considering I've got 1sec delays before even sending reqs anyhow :P
      • 2008-02-19 05034, 2008

      • brianfreud
        oh, and for anyone who might try it... dealing with complex scripts + settimeout in GM is a major pain
      • 2008-02-19 05018, 2008

      • brianfreud
        and yeah! It's finally not getting caught on Google's helpfulness!
      • 2008-02-19 05005, 2008

      • brianfreud
        "would you like to see only English results?" "here's quick links 20 random other medical terms" "the temperature in kalamazoo is 20F" and other such useful random-sized DOM-mess :(
      • 2008-02-19 05009, 2008

      • FauxFaux
        Heh.
      • 2008-02-19 05021, 2008

      • Amblin joined the channel
      • 2008-02-19 05010, 2008

      • brianfreud
        *doh*
      • 2008-02-19 05040, 2008

      • brianfreud
        yes, ok, is you have a function that checks (inputValue) against an array, you would want to pass that function the value... not the array :P
      • 2008-02-19 05033, 2008

      • Amblin- has quit
      • 2008-02-19 05022, 2008

      • FauxFaux
        Ah, JS is so horrid. :P
      • 2008-02-19 05024, 2008

      • brianfreud
        lol
      • 2008-02-19 05033, 2008

      • brianfreud
        ok, fun one - last regexp question tonight, I promise FauxFaux :D
      • 2008-02-19 05045, 2008

      • brianfreud
        is there any way to pass a variable into a js regexp?
      • 2008-02-19 05052, 2008

      • FauxFaux
        By tonight you mean this-morning, right? </5am> :p
      • 2008-02-19 05056, 2008

      • brianfreud
        lol
      • 2008-02-19 05019, 2008

      • brianfreud
        something resulting in /\%28variable\%29/ ?
      • 2008-02-19 05047, 2008

      • brianfreud
        where variable could be 'medicine', 'legal', whatever?
      • 2008-02-19 05047, 2008

      • nikki
        bleh
      • 2008-02-19 05004, 2008

      • nikki makes a note to not try doing truth tables in word again
      • 2008-02-19 05007, 2008

      • FauxFaux
        Just the shortest any-string between those two ends again?
      • 2008-02-19 05027, 2008

      • ruaok
        lie tables are perfectly suited for word
      • 2008-02-19 05031, 2008

      • brianfreud
        no, because I don't want to filter, say, (album), (band), (UK band), (label), (record), etc
      • 2008-02-19 05058, 2008

      • brianfreud
        I essentially want to be able to have an array that store possible wikipedia disambigs that are non-musical, to filter out those urls when they come up
      • 2008-02-19 05023, 2008

      • Dhraakellian ponders just declaring email-bankruptcy with the mb-style list and all the CSG discussion
      • 2008-02-19 05025, 2008

      • FauxFaux
        Oh, if you want to type out the list of alternatives to remove, you can do %28((medicine)|(legal)|(ponies))%29 etc.
      • 2008-02-19 05036, 2008

      • FauxFaux unsubscribed. o/
      • 2008-02-19 05043, 2008

      • Dhraakellian
        ...not that I've actually replied to any of it
      • 2008-02-19 05056, 2008

      • nikki
        ruaok: well, I don't know how to make them work like I want them to
      • 2008-02-19 05034, 2008

      • brianfreud
        hmm, I was hoping to move it to somewhere easier to find in the code, so I could add to it in normal array declaration notation, rather than one huge long regexp string buried in the code...
      • 2008-02-19 05035, 2008

      • ruaok
        no one does. not even M$
      • 2008-02-19 05046, 2008

      • nikki
        haha
      • 2008-02-19 05054, 2008

      • FauxFaux whaps ruaok and goes to bed.
      • 2008-02-19 05004, 2008

      • ruaok
        natta fauxfaux
      • 2008-02-19 05009, 2008

      • brianfreud
        see ya faux :)
      • 2008-02-19 05015, 2008

      • nikki
        I just hope that when I open this file on the uni computers that it doesn't screw with my alignment
      • 2008-02-19 05027, 2008

      • ruaok
        LOL.
      • 2008-02-19 05033, 2008

      • ruaok
        you're sooo screwed.
      • 2008-02-19 05044, 2008

      • ruaok
        finish and PDF it. otherwise you're doomed.
      • 2008-02-19 05050, 2008

      • nikki
        ooh, good idea
      • 2008-02-19 05053, 2008

      • ruaok
        dOOOOOOm
      • 2008-02-19 05058, 2008

      • nikki
        dooooooooome
      • 2008-02-19 05038, 2008

      • nikki collapses into a fit of giggles
      • 2008-02-19 05033, 2008

      • nikki
        actually, a better note would be to not use word in the first place
      • 2008-02-19 05012, 2008

      • mikemorr has left the channel
      • 2008-02-19 05050, 2008

      • brianfreud has quit
      • 2008-02-19 05057, 2008

      • brianfreud joined the channel
      • 2008-02-19 05052, 2008

      • brianfreud figures out a way to do it :)
      • 2008-02-19 05022, 2008

      • cooperaa
        I'm looking for some opinions
      • 2008-02-19 05028, 2008

      • cooperaa
        Elvis has soundtracks for his movies
      • 2008-02-19 05045, 2008

      • cooperaa
        they were reissued on CD with two soundtracks per CD
      • 2008-02-19 05056, 2008

      • luks joined the channel
      • 2008-02-19 05025, 2008

      • cooperaa
        on the combo reissued CDs they say "Elvis Double Features" \n "Title #1 / Title #2"
      • 2008-02-19 05035, 2008

      • cooperaa
      • 2008-02-19 05000, 2008

      • brianfreud
        ok
      • 2008-02-19 05027, 2008

      • brianfreud
        the last few like this I ran into we decided just to title "Spinout / Double Trouble"
      • 2008-02-19 05042, 2008

      • cooperaa
        okay but there are several with "Double Feature:"
      • 2008-02-19 05045, 2008

      • cooperaa
        I don't think we should
      • 2008-02-19 05050, 2008

      • brianfreud
        think they were Ella combos - same company/idea
      • 2008-02-19 05059, 2008

      • cooperaa
        I'd argue for OriginalConsistentData
      • 2008-02-19 05046, 2008

      • brianfreud
        yeah, I suggested series style, but panda/etc argued that "Double Feature: " was descriptive, not titular
      • 2008-02-19 05026, 2008

      • cooperaa
        cool
      • 2008-02-19 05051, 2008

      • brianfreud
        hey cooperaa, offhand, any wiki (disambig)s you can think of that we wouldn't care about? 'medical' and 'legal' are the two I can think of offhand
      • 2008-02-19 05049, 2008

      • cooperaa
        nope, not very familiar with any other than (band) and (album)
      • 2008-02-19 05038, 2008

      • brianfreud
        ok - trying to filter out those we don't care about :)
      • 2008-02-19 05033, 2008

      • brianfreud
        btw, for luks or ruaok or whoever might be able to say, here's what I came up with, maybe you'd see logic flaws in it. It's being fed theValue, which is the url string being tested:
      • 2008-02-19 05034, 2008

      • brianfreud
        var bogusity = new RegExp(); for(checkValue = 0; checkValue < checkBork.length; checkValue++){ bogusity.compile("%28"+checkBork[checkValue]+"%29"); if(bogusity.test(theValue)) return true; }
      • 2008-02-19 05026, 2008

      • ruaok
        ya know brianfreud, I've been meaning to talk to you about your variable names....
      • 2008-02-19 05054, 2008

      • brianfreud
        lol
      • 2008-02-19 05043, 2008

      • ruaok
        what's the goal, BTW? :)
      • 2008-02-19 05023, 2008

      • brianfreud
        having determined it's not a url we already have, or one the script already has seen in a prior iteration, filter out http://en.wikipedia.org/wiki/Vertigo_%28medical%29 as not possibly useful
      • 2008-02-19 05014, 2008

      • brianfreud
        http://en.wikipedia.org/wiki/Vertigo_%28label%29 or http://en.wikipedia.org/wiki/Vertigo_%28album%29 might be, but not the page for Vertigo, the medical condition
      • 2008-02-19 05037, 2008

      • ruaok
        do you know the desired words like label and album?
      • 2008-02-19 05027, 2008

      • brianfreud
        well, that's why I figured I'd go with a blacklist instead of a whitelist - normally it's album, band, label, record, UK band, etc
      • 2008-02-19 05044, 2008

      • brianfreud
        but it seemed much better to try and filter false positives than to possibly filter real positives
      • 2008-02-19 05047, 2008

      • ruaok
        either way it would seem easier to parse the word between %28 and %29. then look it up in a list or disqualify it.
      • 2008-02-19 05006, 2008

      • brianfreud
        that's what that snippet should be doing... :)
      • 2008-02-19 05039, 2008

      • ruaok
        by constructing the URL you're hard coding more constants into your code. makes it less flexible
      • 2008-02-19 05004, 2008

      • ruaok
        var parse = new RegExp();
      • 2008-02-19 05016, 2008

      • brianfreud
        isn't that only filtering for a (match) within the url?
      • 2008-02-19 05038, 2008

      • ruaok
        parse.compile("%28(.*?)%29")
      • 2008-02-19 05052, 2008

      • ruaok
        ... and then the grouping result 0 should give you the parsed word.
      • 2008-02-19 05002, 2008

      • ruaok
        then check it in a whitel|black list
      • 2008-02-19 05003, 2008

      • brianfreud
        that would match everything within the (), including (band), (UK band), (album), etc
      • 2008-02-19 05008, 2008

      • brianfreud
        seems to work, anyhow - here's the latest attempt - 9 1st position hits, 4 post-filter: http://i2.photobucket.com/albums/y48/brianfreud/f…
      • 2008-02-19 05057, 2008

      • brianfreud
        it's only loading release (and label, unless you fixed that) url-rels at the moment, don't have it snagging real release or artist url-rels yet
      • 2008-02-19 05057, 2008

      • ruaok is only working on search right now
      • 2008-02-19 05004, 2008

      • brianfreud
        only checking for wiki links while I tighten the code, but that's 4 wiki links - one at least new to us, since it is filtering against release url-rels
      • 2008-02-19 05035, 2008

      • luks
        wikipedia, not wiki... :(
      • 2008-02-19 05024, 2008

      • brianfreud
        lol, yes, that's what I meant
      • 2008-02-19 05053, 2008

      • brianfreud
        and very nice, the type-check I just put in works now, so the "Metallica_%28album%29" comes up as a possible release match, not an artist match :)
      • 2008-02-19 05001, 2008

      • luks
        ruaok: any plans for releasing DB dumps of raw tags for users who have set in the preferences that the tags are public?
      • 2008-02-19 05038, 2008

      • ruaok
        handn't given that much thought.
      • 2008-02-19 05007, 2008

      • ruaok wonders what the privacy issues are
      • 2008-02-19 05020, 2008

      • ruaok
        even if users are ok with sharing, are we sure we have no issues?
      • 2008-02-19 05029, 2008

      • brianfreud
        so long as the user has the option to opt-out, I wouldn't see anything wrong with it...
      • 2008-02-19 05006, 2008

      • luks
        well, there already are dumps with user data with only removed password and email
      • 2008-02-19 05018, 2008

      • luks
        so I think tags should be ok
      • 2008-02-19 05030, 2008

      • ruaok
        I *think* I agree with you.
      • 2008-02-19 05039, 2008

      • ruaok
        lemme sleep on it to make sure.
      • 2008-02-19 05041, 2008

      • brianfreud
        out of curiousity, in the feed for releases, language and script are two different entities, but why are status and release type mixed within the same attribute?
      • 2008-02-19 05057, 2008

      • luks
        'feed'?
      • 2008-02-19 05008, 2008

      • brianfreud
        webservice
      • 2008-02-19 05011, 2008

      • luks
        oh
      • 2008-02-19 05012, 2008

      • ruaok
        xml.
      • 2008-02-19 05025, 2008

      • luks
        because that's the way they are stored in the database, I guess
      • 2008-02-19 05027, 2008

      • ruaok
        status and type are mixed in the DB like that.
      • 2008-02-19 05047, 2008

      • brianfreud
        ah
      • 2008-02-19 05053, 2008

      • ruaok
        mattfr also had some other reasons.
      • 2008-02-19 05053, 2008

      • tedrock has quit
      • 2008-02-19 05005, 2008

      • ruaok
        er yalaforge
      • 2008-02-19 05019, 2008

      • ruaok
        but I didn't fully agree and thus can't remember ;-)
      • 2008-02-19 05020, 2008

      • brianfreud
        is that (partially) why I can search for new add release edits by language, but not for new soundtrack add release edits?
      • 2008-02-19 05053, 2008

      • ruaok would have to look
      • 2008-02-19 05030, 2008

      • brianfreud
        just had jumped out at me as the one place in the xml where 2 fields seem to be mixed together
      • 2008-02-19 05018, 2008

      • luks
        the idea is to allow more than one type there
      • 2008-02-19 05027, 2008

      • ruaok
        oh luks.. the improved search XML results now match the old XML results.
      • 2008-02-19 05030, 2008

      • luks
        "Album Live"
      • 2008-02-19 05036, 2008

      • ruaok
        just a little tweaking, but all is well.
      • 2008-02-19 05057, 2008

      • ruaok
        the web site is a little different, but humans are better at adapting. :)