#musicbrainz-devel

Yes, they are, but that again pulls lots of stuff, too big for the VM

well, you don't have to run it on the VM ofcourse.

but indeed the easiest is to just turn off that check for now.

http://html5.validator.nu/?out?json "works". But gives lots of (ignored) errors. Is this how it is supposed to be?

JonnyJD: probably. you can look at lib/MusicBrainz/Server/Test/HTML5.pm to see which things we ignore and why

JonnyJD: a lot of them are because our RDFa output is still 1.0, not 1.1.

Ah, okay. So that is fine. Not sure though why validator.nu is down and html5.validator.nu is not.

Mineo: that looks like the current version. (unless someone secretly transcluded it just now).

it *is* the current version, but I've updated the included pages

Mineo: included pages aren't transcluded

(unless I remember this bit wrong)

I don't know, that's why I'm asking :)

did it change now?

Well, does it show your changes?

included pages?

that's a no for both of you

I have no idea if it's just cached or actually transcluded :(

I think it was just cached

nikki: can you transclude an older version? (and then put it back to this afterward).

warp: look at the code ;)

warp: I guess it'd work, but it's not really a proper solution :P

nikki: sure

it's also annoying to do

I assume that will reset the cache.

warp: and thus we still won't know if it's just cached or actually transcluded somehow

Just wait until cache expires and check again :p

I don't understand what the problem is yet, so cannot say much else on the topic.

Is there anything I need to set up on the VM additionally to make digest auth working? The digest auth tests fail for me (401, rather than 200)

warp: that page is including the content of other pages that aren't in the transclusion table

warp: those are the pages that changed, not the main one

JonnyJD: the passwords are probably 'mb' not the real ones

No, I mean in the tests

ah

When I read the code correctly users are created in the tests, but the access fails anyways.

reosarevok: ah, nasty.

JonnyJD: I didn't say anything then :)

JonnyJD: no, I don't remember any of that code needing anything special. I wonder why those tests fail.

hm, accessing http://192.168.1.113:5000/ws/2/collection also fails (trying "mb" and the "real password") (that IP is my VM)

JonnyJD: accessing how?

ocharles: with a browser (firefox)

ok, and are you sure you have that editor in your database?

the user (my use) exists. I can login at the normal login page

hm

what branch are you running?

master d3238e5

hm

nikki has changed the topic to: 🌅 SUNRISE week | http://musicbrainz.org/#devel | note about meeting time in the US: http://bit.ly/Ym40ql | Agenda: reviews, MBS-5933 (ocharles), code reviewing non-MBS projects (ocharles/warp), MBS-2411 (ocharles)

updated from the latest VM, changed RT_STANDALONE, update scripts were all executed

try curl --digest http://192.168.1.113:5000/ws/2/collection ?

ocharles: adding --user JonnyJD and entering "mb" for the password I get "Authorization required"

Oh

Maybe you need memcached/redis

do you have memcached running?

$ /etc/init.d/redis-server status -> running, but memcached doesn't give a status and start and stop give no output, might need to reinstall

yes, that might be the problem. Aptitutde said it wasn't installed. Testing now.

Yes, curl access works fine

ocharles: Fixed, thanks a lot. Not sure where that went missing though.

the VM didn't need memcached?

JonnyJD: you can quickly get memcached stat(u)s with echo stats | nc 127.0.0.1 11211

JonnyJD: no idea, i don't maintain the vm

I read about memcached in the install, but I was quite sure that should already be working. Should have tested that explicitely.

I did run "aptitude update/upgrade" I think, so maybe it went missing there. So might not be a problem with the VM, but with me messing with the VM.

phantomjs is in debian unstable and ubuntu precise now, by the way.

ianmcorvidae: Is http://tickets.musicbrainz.org/browse/MBS-6006 a data corruption regression?

it sounds like something that should be reverted

I don't see how the fix he did would affect it

Nor does he, hence the "can't reproduce it locally"

at least, that was my understanding

because that character isn't in the list of things it removes. I'm guessing it's one of the other functions that removes unprintable stuff

I know that

but maybe the Perl that is running on production is doing something weird

it's probably just older than the version he has locally

yes

and so there is more chance it is buggy

that machine is also not configured how ian's machine is configured

not buggy. just knows about an older version of unicode :P

how do you know Perl's regular expressions on 5.10 aren't buggy?

All I'd like him to do is take a look at it

I don't know anything for sure, I just think the problem is that we're removing unprintable characters in some other function and those characters are so new that they're being seen as unprintable

yep, that sounds possible

bye!

perl -C -e 'for $char ("a", "\x{732B}", "\x{20B9}", "\x{20BA}") { print "$char\n" if $char =~ /[[:print:]]/; }' prints the first two for me but not the last two (added in 6.0 and 6.1 respectively)

(that's with 5.10.1 in three places and 5.12 in another)

I guess I run stuff that's too old to have the latest stuff :P

ocharles: when you reappear, can you rerun failed events please?

ocharles: the ticket I linked, I realized after, wasn't actually shipped to production when that was filed

ocharles: I think it's because of :print: actually, because we're using an older perl before those characters were added

which is what I said :P

ianmcorvidae: can you try that line I pasted?

nikki: all three print for me

there's four :P

yeah

I see two + the a.

I get a 猫 ₹

(the letter a, the kanji for cat, indian rupee sign, turkish lira sign)

ah, so you have 6.0 but not 6.1

v5.16.2

5.16.2 here as well

a

猫

₹

^ is what I get.

Do the tests need edits? I get tests 20 and 31 failing, last error:

Failed query:

'UPDATE edit SET status = ?, close_time = NOW() WHERE id = ?'

(2 11)

it should tell you *why* that failed, but no -- you should need the edit table, but not edits

Well I have that afterwards: DBD::Pg::st execute failed: ERROR: current transaction is aborted, commands ignored until end of transaction block at lib/Sql.pm line 107, <FILE> line 35.

and that before:

# Failed test 'all'

# at /usr/share/perl/5.14/Test/Builder.pm line 239.

I also have an error where the test gets 403 http code but excpects 401. As I understand these are 2 different failing tests (20 and 31)

current transaction is aborted should mean that's not the first failed query

well amybe to long but i was definitely peeved to see the sudden onslaught of snow again.

CatCat: feel free to change the topic.

Hm, "Child (all) exited without calling finalize()" might be the actual problem. That is the last meaningfull I get after test 29 (so that should be the failur in test 31)

freeze should have started around 6 hours ago.

warp: :O

that's because of the query failure

CatCat has changed the topic to: ❄ RANDOM SNOW IN MARCH week | http://musicbrainz.org/#devel | note about meeting time in the US: http://bit.ly/Ym40ql | Agenda: reviews, MBS-5933 (ocharles), code reviewing non-MBS projects (ocharles/warp), MBS-2411 (ocharles)

that's not a unicode character name :(

oh it went by that?

snowflake then :/

or haang on

yeah, the capitals are a convention for unicode character names

❅ TIGHT TRIFOLIATE SNOWFLAKE

sure XD

"select count(*) from edit;" -> 0 (so the table exists and is empty, for TEST and READWRITE)

CatCat has changed the topic to: ❅ TIGHT TRIFOLIATE SNOWFLAKE week | http://musicbrainz.org/#devel | note about meeting time in the US: http://bit.ly/Ym40ql | Agenda: reviews, MBS-5933 (ocharles), code reviewing non-MBS projects (ocharles/warp), MBS-2411 (ocharles)

:D

CatCat: :D

"❄ Random Snow in March week" would have worked though.

(And be very appropriate.)

dude i know

but that's not as funny :P

Yep.

and that was last week :P

I like this one too.