#musicbrainz-devel

ianmcorvidae btw, did you get my /query the other day?

Freso: uh, possibly

Freso: I don't remember what it was though, so you might resend it :P

ianmcorvidae: Well, I was just asking if you need to reset anything if I am to work on MBS again on rika, or if I should just pull in the latest code and get cracking. :)

ianmcorvidae: inline styling is good

warp: for what?

Freso: should mostly just be pulling it in -- you may need some dependencies and/or to update DBDefs

Uh.

ianmcorvidae: inline styling is more convenient when that style isn't re-used elsewhere, so specific to that element.

ianmcorvidae: doesn't clutter up the style sheet files, and makes it easier dealing with complicated layouts where you have to muck with the html and style at the same time to get the result you want.

except for splitting up the styling so it's harder to predict a given style, moving more code into things that are harder to cache, and burying style within markup?

no, it clutters up the *HTML* instead, how is that possibly better?

(note, also, that most of these things aren't used once -- the *classes* are often only used once, but the styles are shared, since obviously we reuse images over and over again)

ianmcorvidae: I'm sure we have inline styles in our html which are bad, but it isn't inherently bad.

I disagree. Making reuse of styling harder isn't a good thing :(

exactly, the whole point is to only use inline styles when those styles should NOT be re-used, but are specific to that element.

and is it ever a good thing to have styling that's so specific it applies only to one element in an entire site?

yes

visual consistency pushed to extremes makes for boring, lifeless and clinical graphic design.

it's good to deviate in moderation

I'm sure we can deviate enough accidentally without doing it on purpose :P

haha

I'm also still not sure singleton elements shouldn't be in a central CSS file -- seems to me that's what the by-id selector is for :)

primary benefit of central CSS is that you can change things in one place, not $large_number

ianmcorvidae: it splits up stuff which for me is easier to work with together.

ianmcorvidae: a contrived simple example: let's say you have articles with paragraphs of text + some images, you typically float the images right, so you have img { float: right } in the central css.

ianmcorvidae: but for 1 particular article you need an image floated left, that is not something which should be part of the central css. that is better expressed with the content as <img style="float: left">

why not <img class="left">?

I don't think assuming you'll never need to float an image left other than "just that once" is reasonable

it is more common (i.e. more than once) that you want to float an image left, yes, then that is what you'd do.

then why make the maintenance burden of having to move it once it reaches some vague threshold of "used enough"?

but until that situation presents itself, it just clutters up the central css.

ianmcorvidae: because until the situation presents itself you don't know which parts you will need to abstract out

ianmcorvidae: if you make things flexible and general before you have the need for them, you will be making things flexible you never need flexible. and you will still have forgotton to generalize other things, because you can never abstract out everything.

I still don't think this is that hard, honestly :P but whatever

the things I'm looking at here are multiply-reused background images

which I think are an obvious thing to centralize

ianmcorvidae: the expand and collapse?

that's two of them

i.e. the buttons?

yes, those styles should be moved to .css

basically all the images on the tracklist tab of the RE

is what it's looking like :P

I'm looking at http://tickets.musicbrainz.org/browse/MBS-5044

I like the idea of serializing those small icons to data urls in a css file (which is then minimized into the large production css file)

yeah, that's the notion

that seems easier to deal with than the sprite technique

cssembed conveniently does that automagically for you

which means we need to have java on our machines, but that's no huge loss

morning

hello mr. ch

arles

arp

l

ol

rem

inder: http://codereview.musicbrainz.org/r/2166/

ianmcorvidae: Update translations from Transifex.

How do I update/pull in dependencies on rika?

carton install?

ocharles: Ping when you're awake.

adhawkins: here i am

Errr, isn't it like 3am?

no

i'm in the same country as you :)

Oh, I got confused about your Facebook status update then :)

i've only booked a hotel, not checked in :)

Aha.

Just wanted to tie up with you about a release of libcaa later.

I'll put 1.0.0 on Github and update the wiki page. Will just need someone to transclude it at the appropriate time.

Rob asked for the blog post to include appropriate references.

ok, cool, I'll try and fit them in

i completely forgot I was meant to be doing that, thanks for the reminder

Is CAA being officially announced this week?

today I believe.

Nice.

Cool. I'll try to get my stuff done early evening (say by 9pm). That way it'll be all good by 11:30

yep, today

ok, we have almost exactly 11 hours til launch there

so plenty of time

So time to get uploading cover art you say? :p

Uhm. How do I start my mbsandbox server again?

run.sh

I have run.sh in ~

:)

Is there a default DBDefs in git somewhere I can use to see what has changed?

lib/DBDefs.pm.default

should be in your checked out code too

Yep. Thanks. :)

Is the MB translation using contexts?

yes

Great.

http://tickets.musicbrainz.org/browse/MBS-5380 context requests go here :)

I made a note.

(Sorry. It's so easy to make puns with that word. :()

I get this when I run.sh on rika: http://sprunge.us/bZKi

carton install DateTime::Format::Duration ?

uk: \o/

So. What's the branch one should work from? test, beta, or master?

Project libcoverart build #8: SUCCESS in 19 sec: http://ci.musicbrainz.org/job/libcoverart/8/

andy: Up version number for v1.0.0 release

I'd probably start from beta, unless you need the schema changes.

Hm. Is there a reason carton.lock isn't .gitignore'd?

Yes, it contains the version of the relevant CPAN packages.

Freso: generally start from master unless you need something which isn't in master yet.

warp: Alright.

nikki: ping

ruaok: libcaa all sorted

adhawkins: awesome!

So can we see your big reveal online anywhere? :)

ruaok: pong

nikki: what are the latest stats? (besides those listed on the site that are a few hours old)

adhawkins: sfmusictech.com -- see if they offer session streaming.

if they do, we are in the main room. so it should be covered.

Ok cool, will take a look

hello ruaok

moin warp

they're still at http://mb.lmfao.org.uk/caa/stats , correct as of the last dump. caa stuff is still not replicated so I don't have anything more recent that than

hi CallerNo6

shag yo ass to this side of the bay, yo!

doesn't look like anyone's beaten those huge images yet anyway :P

good.

ruaok 's already in SF?

yep.

arrived shortly after obama. :)

oh, right

though I doubt he is still around

which stream is most best awesome?