#musicbrainz-devel

is there a way to set times using milliseconds, other than by setting them using a disc id?

not in the current code

ah :/

Is anything put into the stash from the Root controller accessible from any other controller?

The following IF isn't working as I'd expect it to: if ($c->stash->{server_details}->{is_slave_db})

Dumper($c->stash->{server_details}->{is_slave_db}) comes up as undef, even though I can access the value of server_details.is_slave_db in the template

navap: those are set in 'sub end' in Server/Controller/Root

moin warp

time-of-day!

warp: Yeah but I'm having trouble pulling their values out of the stash in another controller.

I know they're in the stash, they're just hiding from me.

navap: yeah, obviosly. 'end' implies it is run at the end, so they only get send _after_ the other controllers are done.

navap: if you need them in the controller, I'd say just move that code from 'end' to 'begin' in Root.pm

Ahh, that also answers a question I was about to ask you :p

s/send/set/

Which syntax do you prefer? DBDefs::DB_STAGING_SERVER() or &DBDefs::DB_STAGING_SERVER?

I usually leave the () off if they're not needed.

Will DBDefs::DB_STAGING_SERVER on its own work then?

yes

does it?

I thought you had to have the & or the brackets

oh, hm. well, try it :)

Is there any reason why something would be put into the stash both in begin and in end?

navap: no.

Okay

navap: I think the idea was that you only need server_details and such if you're rendering a full page, and the controller may set c->action->attributes->{Minimal} in some way, which means server_details isn't needed.

Judging by the UK mirror, a mirror server should not user profile data?

should not have*

I'm not sure, I don't think we're actively pursuing mirror servers anymore.

Well by mirror server I mean anything that's getting replication packets.

Unless they're also in the replication packets they'd become out of date quite quickly (stats wise).

I can't think of anything :)

navap: there isn't a point, since the tables will be empty. since we do no export anything that is confidential

The confidential data will be removed, sure, but the rest of the user data is still importable from the dumps.

I was asking because I was wondering whether or not to remove user access as a part of the mirror server checks.

I was planning on removing access to the log in/registration pages, as well as trying to view any of the user pages.

hm. right.

thing is the user could view the pages, but there will be no user links at add since there will be no edits.

I personally would not make any extra effort at all in this case.

warp: It looks like data that's put into the stash in begin is not accessible from the templates, it needs to be added into both begin and end to be accessible from both controllers and templates.

navap: have you chosen a server name yet?

No

want?

Sure!

ok, go!

cartoon characters.

I'm torn between Tintin and Asterix

OOOHHHH!

:)

both good choices.

but, I would prefer tintin

easier to type.

Tintin it is

thought asterix is cooler, IMHO.

asterix

heh I can imagine a lot of typos with asterix :p

asterix

feh. asterix is cooler.

Sure, as I said I'm torn betwene them.

i can deal.

lets go with asterix.

I'm fine with either.

Okay :)

;-D

And for nikki's record keeping: http://en.wikipedia.org/wiki/Asterix_(character)

Or even just http://en.wikipedia.org/wiki/Asterix

already edited that.

cool

heh I'm first on the list :p

oh, I need to reference the chatlog too.

damn. MB is meticulous. :)

;)

and I suppose odie ought to be there too.

If I go to /account/edit I get forwarded to /user/do_login. I've added a detach at the start of do_login, but it's not taking affect because after the detach it's going back to /account/edit and then returning a 500 server error.

I thought detach would break out of the forward?

now that each part of a date is stored separated, how can I recombine them to select dates using things like now() and stuff?

it's probably faster to query them separately

because as soon as you combine them, you can't use indexes

but how would that work with things like selecting anything between one date and another date?

you have to do something like "year BETWEEN date_part('year', date1) AND date_part('year', date2) AND month BETWEEN date_part('month', date1) AND date_part('month', date2) .." :(

now that we already depend on two non-standard psql extension, we might as well just write a type for these partial dates

time to set aside MB and work on world domination?

muhahaha

aight. the second proc for asterix was just delivered.

time to get that in motion.

Djce: ping?

Alas, no.

What's the difference between $self->c->log and $c->log ?

The latter works but the former returns "Can't locate object method "log" via package "MusicBrainz::Server::Context"

See, everyone runs from my questions :p

$self->c seems to be a MusicBrainz::Server::Context instance, while $c is probably a MusicBrainz::Server instance (which extends Catalyst)

the context stuff seems different than what I remember

so I'm not sure how it creates the context object

hm, what is $self?

I'm looking at MusicBrainz::Server::Data::WikiDocIndex

seems like a leftover

I assume the code there doesn't work, right?

It does, as far as I know it's just the log->errors that don't work.

yeah, I mean if the index file doesn't exist, it crashes

Yeah

I'd just replace the call with warn and ignore it

you could fix it by changing MB::S::Model::MB to store the actual Catalyst context in MB::S::Context and then forward calls from MB::S::Context->log to the Catalyst context

but that's probably too much work for almost no benefit

Yeah.. I'll just use warn

$content holds $response->content of an LWP::UserAgent, do you have any recomendations on how to extract only a certain div out of $content?

The div has an id set.

is it a valid XML?

I think so.

It's a valid HTML document, would that make it a valid XML document?

hm

Data::WikiDoc actually already uses some HTML parser

where is this code you want to change?

I'm writing new code, but it's based off of the code in MusicBrainz::Server::Data::WikiDoc

I'm trying to get searching working from /doc/

one sec, let me read HTML::TreeBuilder::XPath's docs

Thanks

something like this should work

Thanks! That helps a lot.

/nick ruaok

heh