man fennec rocks

Tomcat|MUC: you'll need to pull from linux01 to pick up a unicode fix.

bc: yeah will do when i'm at home again

Tomcat|MUC: where are you?

MuC -> Munich :) opensource meetup

ah

but will be home in 5 hours :)

cool

Tomcat|MUC: so why is it Munich in the U.S. and Munchen there? just because the English like renaming things?

heh

jhammel: well good question :) maybe because Muenchen is too difficult for non-Germans :)

and so we needed a english name :)

Munich/Muenchen has even a different name in italy :)

Tomcat|MUC: weird :)

heh maybe its also to confuse frankonian fans like jmaher|afk ;)

what's MUC?

MuC -> Munich :) opensource meetup

Ah cool!

Have a Koingstrizer for me!

i show BOYB and fennec here

ctalbert: no beer here :)

:(

That sounds cool. You're showing fennec on your new n900?

heh not allowed here, because youth city event location :)

that was for beer :)

Ah makes sense

ctalbert:yeah on the n900 :)

Tomcat|MUC: heh, i thought you meant showing fennec ;)

that will be sweet

Tomcat|MUC: tell the MUC folks that we said "Hallo" from California

ctalbert: yeah :) also good timing, we will have some journalists here

and some press reports said we stopped fennec completly

(instead of only on Windows Mobile)

What? Did they misread Pav's blogpost?

so they can see that fennec is alive

yeah

Tomcat|MUC: I'll send them my schedule and that will show them that no we have not stopped fennec.

Tomcat|MUC: isn't that twice in the last two weeks the German press misunderstood us?

some very stange reports

from "we do not care about windows mobile users and leave them alone etc) till Mozilla dropped Mobile plans and stopped fennec

wow

jhammel: yeah , well the "first" time it was a goverment agency

Tomcat|MUC: true

ctalbert: translate.google.com/translate?js=y….de%2F1003%2F74016.html&sl=de&tl=en

"But even if Windows 7 is on the phone market, it will first be many owners of a Windows Mobile smart phones that could use a mobile Firefox. Zumal der Internet Explorer in Windows Mobile allgemein als wenig komfortabel in der Bedienung gilt. Especially as the Internet Explorer in Windows Mobile is generally considered inconvenient to use. Aber diese Kundenbasis interessiert Mozilla offenbar nicht. But this customer base, Mozilla does not seem in

err sorry for the german text ;) but that was one of the strange articles

cool, thanks

ctalbert: will inform melissa later also

good plan

btw any n900 experts here :)

i was asked if something on the n900 can read ical :)

I believe the n900's calendar application can

Tomcat|MUC: ^

ah cool

thx :)

bc, ready for meeting?

just a sec

lol. the lanikai nightly has put a white box in the middle of my screen that won't go away

ctalbert: go

ctalbert|mtg: ping?

sorry 5 mins more

ctalbert|mtg: k, thanks

Hey everyone, this is your future coop student

ahal: hey! good to meet you

ahal: when are you joining us?

welcome to #ateam ahal!

yes, welcome! and where are you coming from (for those of us not in the know)?

scrum: working skip algorithm under testing, reproducible signature generation under testing (need to make sure don't add duplicates), testing external attachments. outstanding: update bug triage during history update.

harth: next.

just got some awesome headphones, also got the backend in place for filtering crashes by branch, os, and triage, working on UI for toggling those right now. want to get a few things done by the end of the week, like limiting the number of assertions shown to speed things up.

jhammel: you're next

cleaned up the patch for Mozmill tests for buildbot and applied it; now working on getting the tests to run

also generally trying to survive the firehose; mainly focusing on figuring out the buildbot system, and to a lesser degree, mozmill

harth: and i also just got some awesome headphones ;)

I need new headphones, please do tell!

jgriffin: tag

linky linkyu

jgriffin: we remoties don't get goodies.

bc: man that makes me grumpy. I'll brb after a margarita

Grafx Bot is the new Halreftest. Clint and I met with the design team on Tues, and that's the name we cooked up to replace 'Halreftest'. They're also going to make some robot graphics for it, featuring 3-D glasses. How cool is that?

On the linux port, I've finished the C++ code, but ran into some linking problems on CentOS with OpenGL. I think I'm going to have to use the solution that Firefox uses, and that is load it dynamically via PR_LoadLibraryWithFlags and related functions, so I'm investigating that right now.

jgriffin: sennheiser h150 is what I got

oh I've seen those, I heard they're cool

ctalbert: your turn

jgriffin: mine are these, i believe store.shure.com/store/shure/en_US/pd/productID.105432500

I just ordered some new headphones

maybe next model down

:)

cool jhammel

jgriffin: the robot graphics project sounds awesome

yeah it's awesome we're going to have an actual logo for this project

:)

Let's see, I am working to finish the long standing project to get the mozmill documentation up to par. Should ahve that done today. Then I'll be working on some mozmill bugs and still trying to understand what changes our Q2 goals are going through as Bmoss is fighting thorugh the director's meetings

Also: everyone, I'm embargoed on talking about reviews and review feedback until next week, so we're going to put that off for a week.

Besides with the goals stuff I haven't had time to prepare it all. :(

Sorry about that.

np, I'll just have another margarita

hehe

jgriffin: oh we have the meeting with legal next don't we

yep

to discuss the legality of margaritas at work....I"m pushing hard for it

leagalize Margarita now!

actually, it's to discuss the privacy policy for Grafx Bot

LOL

harth: known issue: template error at qm-dashboard01 ?

ah. need port 8000

nm

bc: yeah, sorry

working off the django server to speed development

I'll reload the apache

ahal hello!

This is clint

harth, bc, we should get together and talk about what crash/valgrind/assertion/needbettercodename UI 2.0 is. I want us three to come to a set of good definitions so that we can get some of the UX team experts involved to help us design a UI for it sometime next week or the week after.

btw: remember that sshfs thing I used to mount a linux file system on a mac machine? I've got it working on Windows too, so can point people to the necessary software if anyone wants to use it

liky!

jgriffin: yeah.

er

linky!

heh

yeah, sounds good

ctalbert: tomorrow?

the mac version is : code.google.com/p/macfuse/wiki/MACFUSE_FS_SSHFS

or monday/funday

the windows version is : dokan-dev.net/en/download

make sure to read the README for the win version, there is some file copying you have to do

bc: tomorrow will work for me. Harth?

yeah, that works

so event done

Tomcat|train: how was it?

great

they love fennec

I'm free at 12, 3, or we could re-use the 10AM slot for some of this.

which ever

is 3 too late for you bc?

nah.

ok, let's do that.

awesome

and personas :)

and they realize that we're not killing it?

and interested in byob

ctalbert: yaeah seems they missed the articles :)

but /me found maybe a fennec layout bug ;)

ctalbert: care to explain your skip stuff? i don't get it. :-(

when i load mozilla.com/de on a beta 5 build on the n900 the bottom part has wrong rendering

bc: it was one of those ideas that started out ok and ended pretty convoluted.

I'm not sure it's worth it. Let me look at your latest patch first.

i think my current approach will work ok. i set the reset period to one hour. i couldn't come up with a good heuristic for that.

you're storing skip in the database?

stepped out for a bit, my names Andrew, I'll be joining you guys at the start of May and I'm coming from Ontario (uwaterloo)

ahal: awesome, I'll look forward to meeting you

I don't remember what 'viewdata' is

no, storing it in the worker's instance.

and thanks for the welcome, I look forward to joining you guys

it is just a short hand to get the data for a particular keyset into the pending jobs view. it contains the precomputed start/end keys and the skip.

instance <-> memory.

ahal, I just looked up your start date, and it coincides with our Work Week when all the folks who are remote will be in Mountain View, so you'll get a chance to meet the entire team at the very beginning of your internship. This is really cool because over half our team is remote.

harth, myself and bmoss (and soon jhammel) are the ones in Mountain View

yeah, I'm one of the "remoties"

i put that in the bug before i commented it. let me do that so it is more understandable.

cool, nice meeting you... so what exactly does 'work week' entail?

i replaced the 4 different getPendingJobs calls with a data driven loop

we all hang out in a conference room with our laptops and work on stuff, try to get collaborative stuff done while we're all together

let me finish this reftest test of the external attachment to a 0.11 db and i'll update that patch with comments.

and margaritas

that's the best part

oh wait, you probably can drink them, I forgot ):

s/can/can't

yeah... for the first two months at least

ok, I guess that's what I didn't get bc. The loop made me think that you were storing a skip for each signature

no. for each set of keys.

how long will you be with us?

it'll be a bit of a change

four months

until the end of August

cool!

so work week is just for this team then?

oh I think I get it.

yep

sometimes other teams have work weeks that coincide though

we seem to always have work week at the same time as platform

jgriffin: no julie yet, not sure where she is. I'll ping you when I find her.

ctalbert: sounds good

I think this one will just be us. But yeah we take the chance while everyone is together to do some fun things and to brainstorm and whiteboard a bunch of bigger projects

sounds fun... someone mentioned something about a conference in Vancouver?

there is...in July. It's still being planned.

what's the conference?

jgriffin: Phong is installing purify on sm-purify

it's the Mozilla Summit

coolz; do we all go? or how does that work?

yeah, we all go

awesome; i've never been to Vancouver

or wait, am i crazy? yes i have

ctalbert: cool

jhammel: that might be two different questions ;)

you might be crazy AND have been to Vancouver ;p

I've never been

but perhaps you haven't been to Whistler, which is a different ball of wax entirely

for some reason my brain transposed "Vancouver" to "Montreal"; i *have* been to Vancouver

must have been in front of the screen too long ;)

anyway, I have to take off, I'll try and hang out here once in awhile

nice meeting you all

jgriffin: whistler?

ahal: nice meeting you too!

jhammel: that's the ski resort in the mountains where the actual conference is being held

it's not in Vancouver itself, more like 1+ hr drive away

jgriffin: ah, cool; no, i haven't

here's the place where we'll be staying, I guess: fairmont.com/whistler

nice!!!

So, bc, my idea was to try to tie the skip to specific signatures so that it would simplify the math of trying to map a given number to the set of things we shoudl skip. But as you can see in my comment, it got complicated quick. I am starting to understand this latest patch though and I think this is a better solution.

i was thinking that if the view was keyed on signatures it would be possible to do what you wanted, but it is keyed on priority, os, cpu, os version, url count.

right. I was thinking it was keyed on signature

k. i'll add comments and fold it into the first patch for review.

sweet

i couldn't figure out a good way to recheck to see if new signatures have been added or priorities changed. i figure an hour allows me to cut down the overhead and not take too long to respond to changes in job queue. with more machines in different stages of work, the effective time will be lots less.

jgriffin: we're calling in now

ok I will too

The ship's time is now 3PM. Ding Ding Ding.

ctalbert: ping

bc: true

saving an external external attachment wasn't even noticiable in terms of cpu/memory. 350K for a 5.6M log. not bad.

bc: want me to call you directly?

what ever works best.

that is fine.

bc: update deployed with no problem so far

Tomcat: cool. there is a problem but you won't see it until you get a crash.

bc: ok and thanks for the preference hint :)

i'm working on the crash issue. it is with the extra signatures for reproducing crashes.

cool