#musicbrainz-devel

ok, bed time

nn

nikki: pong

ultimately we found it XD were looking for the CSS customizations on the wiki (in monobook.css, as we found out, originally we'd thought it was common.css)

heh common.css doesn't have much does it?

yeah, not really XD we looked at it and roughly went "hm, that's not nearly big enough, there must be somewhere else"

ianmcorvidae: ping

Ian McEwen: Silence a few warnings about uninitialized hash values.

ruaok: morning

morning.

sadly this was yak shaving morning here. I'll be in soon.

but, you should see the shiny yaks!

cool. i'm working on fs anyway

nice

can;t wait

hi there !

thanks ianmcorvidae for translating my werido‐english→useful‐english in some tickets ! ;)

i have spotted an very light “issue”, when we use entity name to sort what we show on screen, we don’t use the comment at all (it seems)… maybe each time we use the name, we should use the comment right after as they are both appearing as being one text in the same column, they appear even like if only one thing, so we should treat them at the same time… ach, bad said

maybe i’m incorrect but it seems to be the why in https://musicbrainz.org/artist/88d8f38f-adb4-48... 風神界逅 (通常盤) appears AFTER 風神界逅 (11th album) but 雷神創世 (通常盤) appears BEFORE 雷神創世 (12th album)

and sometimes release with 通常盤 appears before the one without comment and sometimes after (in those two pages)

the name sort appears after many other sorts in that page, it’s very good already (after date, catalogue, etc.)

otherwise

They seem to be sorted correctly for me.

i’ll take a case more obvious then. on page 2 魔王戴天 (通常盤) then 魔王戴天

on page 1 鳳翼麟瞳 then 鳳翼麟瞳 (通常盤)

(all attributes the same except package and comment)

you see ? the comment seems having no effect on the sort

I never said they did. Just that those two you highlighted (11th vs. 12th album) sorts correctly for me. (Probably due to cat. no.'s.)

Well, unlike RG problem (MBS-6796 12 before 11) indeed they appear as 11 THEN 12. but what i said was

Freso: for the 11th album 風神界逅… we have 11th album THEN 通常盤

for the 12th album, it’s the opposite… we have 通常盤 THEN 12th album comment… or maybe it’s random each time you display page and you don’t even see same as me ?

(11 appears befroe 12 in label page and releases page, thanks to catalogue number indeed)

Freso: even if this was fixed, maybe if we sort by name+" ("+comment+")" would be even more stylish than sort by name then comment : so that name:“release”/comment:“version 1” and name:“release (version 2.0)”/comment:“only sold on concerts” would sort 1 then 2 instead of 2 then 1 if we would use name sort then comment sort

stylish : “release (version 1)” then “release (version 2.0) (only sold on concerts)” would be, without manual concatenation before sort : “release (version 2.0) (only sold on concerts)” then “release (version 1)” because “only” comes before “versuion” in the alphabet :)

i don’t want to make a ticket because i’m flooding the email notifictaions of jira already enoughly and also i don’t want this to interfere with the unsorted RG ticket (it’s in another part and is unrelated but as I explain so bad, things could get confusing)

If you don't make a ticket, it's not going to get "fixed".

we knew that, Freso :)

hi ruaok ! i had something to say but i forgot :)

:)

maybe just “hello”… anyway i’m going to lunch, see you soon, of course

is logging

Users on #musicbrainz-devel: Sebastinas mat_ luks _Dave_ warp Muz djinni` Kovensky muesli kloeri

ijabz1: you about?

hi

I've finally finished debugging the transfer of indexes between servers using netcat.

which will transfer all of the indexes in just a few minutes.

so, I am going to proceed to copy out the indexe built yesterday.

thats great

then we'll take one server out of rotation and swap in the new indexes.

and then I could use your help to test that new server.

so, hang tight, we should have something to play with in a few minutes.

ok, please remind me how i connect I cant remember

one sec

ok, copy started. clearly some fine tuning is still needed there.

you have an account on hobbes, yes?

log in to that for starters.

wget -O - 'http://dora.mb:8080?type=recording&query=love'

is how you can query dora once I rotate the new indexes in.

pushed to dora. now roobarb.

ive pmed u

oh sorry.

ok, now that we've got that sorted... the second copy should be done soon.

nearly 8 minutes to copy to a live server. beats 45. :)

ive stopped being able to connect to dora !?

yep

hang on.

ok, ready now.

alll indexes seem present and returning data.

not sure if the data is correct, but. :)

this is right on dora but wrong on roobarb wget -O - 'http://roobarb.mb:8080/?type=artist&query=blink-182&fmt=jsonnew'

dora returns 'artists' and roobarb returns 'artist'

has roobarb got latest search code ?

no

we're testing one server at a time.

dora is not in roation, roobarb is.

oh right

well it seems okay with the tests ive done

good.

let me put this server back into roation then.

ok, roobarb is up.

help me check, ijabz1

Are they both being used from musicbrainz.org now or just dora ?

just dora.

how is the going ijabz1 ?

Yep it seems okay

swet.

got the items for a blog post ready?

yup, http://tickets.musicbrainz.org/secure/ReleaseNo...

ijabz1: did anyone else help with this release?

no I dont think so

the text has version TBD.

should that just be "2014-10-08" ?

yep

ok, looks like we're done.

thanks ijabz1

and I've restarted the index building. which should be even faster now. :)

So Ive now tested all eight issues fixed from the website and added comments to issues

Six work perfectly, two work but need a bit more tweaking to be be optimum

they will be improved some time in the (not-near) future

:)

non complex releases are easy.

ok, now lunch.

Its the nature of the beast Im afraid, in that any fix that improves how the text is analysed means the same analyser has to be used at both index and search time

Other type of fixes such as fixes to json,and indexing of brand new fields are okay for simple release

New post: blog: Search server release: 2014-10-08 <http://blog.musicbrainz.org/2014/10/08/search-s...;