#musicbrainz

Knio is now known as Knio-dinner

Knio-dinner is now known as Knio

{global notice} Hi all! The hub that services was on apparrently was having problems again, and we dont know what went wrong yet. I'm still currently looking into the situation, but the backup of services I just started unfortunately has a db that's 2 days old. if we have reset your password, and you no longer have it because of the slightly older db, please contact a staff member and we will help you. I'm very sorry forthe inconvenience, and thank yo

{global notice} Hi all! We are in need of a hub for Europe. If you think that you may be able to sponsor a server, please take a look at 'http://www.freenode.net/sponsoring_servers.shtml' Thank you for using freenode!

Knio is now known as Knio-sleep

icepick-away is now known as icepick

icepick is now known as icepick-away

elinenbe_ is now known as elinenbe

Knio is now known as Knio-school

morning/evening!

hey

another long weekend.

back to work finally.

tunepimp hacking :-)

ruaok: Can I check something with you?

I need someone to double-check my reasoning....

shoot

ok, load up Discid.pm

and find _FindFreeDBEntry()

I think I understand what it's doing, but the problem is in the two "update" statements in that routine.

got it

The two "updates" hit two different tables, using the same "id=$id" value.

but AFAICT the PKs of those tables don't match up - there's no reason why they would really.

I see your point.

or, to put it another way: you might /hope/ they match up, but they don't.

And I wonder if this code still gets called...

it's further up in this module, yes

but as to whether it actually /really/ gets invoked, I don't know.

the idea is that if we import a partial freedb id, it would go and fill it out with the right leadout/id.

agreed.

I tried this: select count(*) from discid d, toc t where d.disc=t.discid and d.album!=t.album;

that gets 0 rows, which sounds good.

so I'm suspecting that the aforementioned bug has fortunately not bitten us.

does that sound convincing?

not reassuring.

:-)

I guess the question is this: do we still import freedb data TOCs that are incomplete?

If not, then the updates can go away.

yes, I think we do, for sure. don't we?

If so, we need to make the updates more robust.

I'm not sure anymore....

I'll give it a try.

ok

I had an interesting chat with this guy on a chairlift over the weekend

he was bitching over how his old company had to pay Gracenote license fees

ahahahahahah.

they do a hard-drive MP3 car stereo thingy

I forget their name, 'fat' something

phatnoise/phatbox.

I spoke to one of their engineers last weekend.

that sounds right

passed on a bunch of my cards. :-)

they got a cease and desist when they were looking to move elsewhere (freedby perhaps)

standard operating procedure.

they caved in due to potential legal costs and conservative clients/backers(?)

no surprise.

they're still one of Gracenote's bitches fttb

what's the upcoming suit that Gracenote are in?

ruaok: A "freedb import" sets the track durations, but doesn't touch the discid/toc tables.

ahh.

no wonder we haven't gotten bitten by that bug. :-)

so can that logic be removed completely for good? Or does it need to be fixed ?

lets nuke it.

I'm hoping to find someone who wants to work with us on CD lookup stuff.

... dunno if this is related but importing from freedb is such a pain sometimes cos artist /tracks sometimes get imported other way round and to cut paste and fix name :P takes ages

ok, will do.

Someone who can help us pay for a patent lawyer to look at exactly what we need to do to circum vent GN's patents.

If they are willing to do that, then I'd we willing to revamp our id system to be GN safe.

Until that time, lets leave things be.

I thought there was a case coming up that was going to test their patents?

Maxxxxx: there's a "swap" button to swap over the track/artist names. It should take that long.

There is, but I don't think its about the patent unfortunately.

how many patents do they hold?

google: gracenote patents [i'm feeling lucky]

that'll be 'more than one' then ;-)

bout a dozen or so.

real is now known as real|afk

can you be issued a patent if you buy the IP that was the prior art?

I would not think so.

*I* have written prior art code at CD Index.

wasn't CDDB itself the prior art?

Our original CD Index code is prior art and was checked in befor they filed the patent.

that tool

it was public and therefore should be considered prior art.

bizarre, they got a patent for which a large part of their application must have described the prior art that meant it sholdn't have been issued

yup./

hi

hey mattias

how is your monday coming along?

oh, I started with some exam preparations, but that's done

nice

I'm currently reading the mailing list

lots of stuff there

how's your still young monday?

I went and did webjay evangelism to a netlabel

got a really good reception, so I'm feeling good.

...experimenting with explicit evangelism to try to get the ball rolling.

webjay is still very immature and buggy, but there should have been more uptake than there has been.

hm, do you have a link for me?