#musicbrainz

The only Canadian beer I know is Moosehead Lager and it tastes like Moose piss.

Ya you said, I was wondering if anyone else in the channel would know what the drink was.

What did you expect? It's beer.

¬_¬

You haven't been to a Samuel Smith's I see.

I doubt they would recognize it as a drink if they saw it. Looks more like a shrubbery

You said that some of the entries on bsnpubs had the catalog numbers wrong?

107 and 108, yes.

What makes you say that?

Looking at the picture of the liner of 108 on Amazon

So they have them transposed?

The second discs, yes.

I also read your note about not trusting the listing on eBay. I appreciate what you mean, but I wonder where they got that number from. I can't believe they looked anywhere else except on the CD cover

If they show a picture of the front cover, and a valid format time-life cat #, I doubt they got the info from anywhere else

meh, I just have a feeling that its wrong, the liner on amazon was very legible and clear.

I am not sure which release we are looking at on MB

Many of them have incorrect ASINs

Just opened a new window witha search for 'The Time-Life Treasury of Christmas'

anyone bored? :)

Got 20 discs in the list

I'm very bored with this, but I want to clear it up once and for all

lol

I don't even own the damned CDs

What do you need done brianfreud?

I'm starting at the top of this list ---- http://musicbrainz.org/search/textsearch.html?q...

navap: looking for someone who's kind of familiar with the crap in freedb to take "This Is a Song, Parts 1, 3 & 5: The Subtitle" and turn it into a ton of mangled (but comprehensible) 1, 2, and 3 -part variations, with pt., pt, (part ), etc - you know the type of mess I mean

Munger: If I owned everything I edited on mb..well lol I would have quite a nice sized collection :p

I thought thats what the test was for? Doesn't it have all those types of varients?

with and without subtitles - basically mangling it by hand, or dumping a ton of examples straight from freedb

navap: it does, but I'm also somewhat influenced, probably, in writing the tests, since I also wrote the handler. Looking for a bunch of "real world" cases to add to the test.

What happens if the numbers are out of order?

I thought about writing a check for that - it'd be rather simple to do, actually.

But I think it's more likely to normally screw up intentional orders, rather than the rare times it might fix bad ordering.

Though I'm concerned that might infringe on artist intent in most cases.

lol

but yeah... I just want to make sure I'm not artificially influencing the tests by adding in a series of tests based on true life bork

I don't envy your can of worms brian

hm The subtitle makes it quite hard, there are a lot of ways a subtitle can be written on freedb et al, but the only way to know if its a true subtitle is through manual intervention.

lol, well, I have all but one nit done in handling arabic numbers - then the others are mainly just copy/paste, with slight modifications for type.

But here's the current code, just to handle PartNumberStyle: http://musicbrainz.pastebin.com/m6fb37462

Arabic uses latin numbering style, doesn't it? left-to-right

arabic numbers = 0123456789

Right, it's just the rest of the script that reads right-to-left

"Arabic" numbers are actually English numbers, "Riman" numbers are actually English letters, the world is just crazy :p

Roman*

I'm referring to the arabic number system, not arabic itself :P

I have blotted all that from my memory and never wish to remember :-)

navap: I'm also looking for various titles with the word "part" not being used as a part number.

Oh boy

So far, happily, the fingerprint I'm using (the obscenely deep if set at the top) passes -

I'm using a negative approach to identifying if it is a part number, rather than the positive match the old GC and my old part number style code used

so far, I have yet to find a non-part number string that actually makes it past that fingerprinting and gets mis-handled

or a part number that doesn't pass it

:P

take every instance of part, pt, pts, parts, etc, and only mess with it if it passes a bunch of tests, rather than messing with it if it doesn't pass a test or two

Messing with it if it passes the test is the positive match?

negative match

hm That seems backwards to me.

me too, but that's how it was always taught to me :P

negative on the Q "is this invalid for what we want"? I guess

Thats dangerous, I've had teachers who were adamant that latitude was the vertical axis'

brianfreud, A couple on here --- http://musicbrainz.org/release/8ee75e8f-0bde-4e...

ooh, nice too, they have roman numbers right after :)

hm that might not be correct how it is

The base name of the track is actually 'Spare Parts'

actually, that release hits a lot of possible trouble spots... I'll have to save it for more testing later :)

Ya, it should be "Spare Parts II: And Closing"

Nope

With a comma

nope

No?

because it's "Spare Parts", not "Spare, Part II"

Brackets?

It's correct as it it is

ohh wow /me finally clues in.

that's the kind of "is it a real part number style" that routine tries to automate

damn

it fails :P

lol

first failure I've seen

Never mind. Get the album. It's really great :-)

lol, navap, it agrees with you though, "Spare, Part II: Closing"

:p

keep 'em coming Munger - that's just the kind of tricky stuff I need :)

can't be perfect, but I can try :)

Can't think of any more right now, but to make it even more fun, Amazon have chosen to add (live) to all the track names. The whole album is live with no interruptions between tracks

"The album was recorded "live"" according to wikipedia as well

ALthough I'm not sure why "live" is in quotations :/

Hard to read, but the MB data is correct

I'm guessing this one passes --- http://musicbrainz.org/track/29e84eb2-1314-49bb...

ok, i just did a raw extrat of every non-part number style "part" in the database...

this is the worst formatted track title I think I've ever seen: - 3 Etudes (from 'Methode des methodes' - part 1

(including the opening hyphen)

Hmmmm "Deck the Halls With Parts of Charlie"

I wonder if " Strokin', Parts I & II " and " Soft, Parts 1 & 2 " are actually correct. I had to re-read them :-)

:P what a mess: http://musicbrainz.org/album/8557fe59-6967-417e...

or "...pas à pas - nulle part..., Op. 36: V. Le tout petit macabre - Ligetinek"

This looks fubar --- http://musicbrainz.org/album/ba6845a1-2299-4313...

anyone know how i would escape [ in a postgres regexp?

Not just \[ ?

Invalid command \[. Try \? for help.

same for \\[

and help is no help :P

Try \\\[ ?

nope

trying to do select * from track where name ~* '^\['

\\\\\\\\[ ?

:P

Octal 133 ?

ie \133

nope, it still errors out the same

How about double quotes rathr than single. I'm only guessing here

nope, no luck either

hmm, ok, something among everything worked, even with the errors, got the results file I wanted :)

'^"["' ?

I get similar headaches in bash scripts sometimes

no clue which in there, but something worked :)

Was it "[" ?

E'\['

Okay brianfreud I'm starting to get unbored :p So here's what I've come up with so far, http://pastie.org/414251

thanks :)

and wth? http://musicbrainz.org/album/a5e2eecf-bf7b-4b66...

Do we take out funky capitalization in sortnames? For example name:KiEw, sortname:Kiew.

brianfreud, That one doesn't count. Anyone who buys an album with track names like that can't expect anything to make sense

I wouldn't

lol: [untitled] / [untitled] / [untitled] / [untitled] / [untitled] / [untitled] / [untitled] / [untitled]

lol whats the poing

point*

brianfreud, What do you get from this --- http://musicbrainz.org/album/2c34848b-7e3d-4927...

the "N. " trips a known bug: The Return of, No. EX. T, Part 2: World

Tarck 1?

track 1

세계의 문, Parts 1-유년의 끝, Parts 2-우리가 만든 세상을 보라 ------> 세계의 문, Part 1: 끝 2‐우리가 만든 세상을 보라

it assumes there's not separate subtitles - handling that's just too non-possible, imho

hi catgroove

hi mo :)

Hi CatCat

This is probably OK http://musicbrainz.org/track/4db533cf-9c29-46d1...

yup: Interview In Three Parts, Part 3