« first day (3 days earlier)      last day (150 days later) » 

8:02 AM
@jokerdino blame the different spellings and the dirty english.
 
the code is fine. they want tests and really rigorous ones
 
and I've no idea how to do that.
though I've a feeling they'll use unit tests python module.
 
maybe
 
give me20 min. I'm hungry.
 
okay
 
8:07 AM
I see how awesome github is when I start one public repo on bitbucket. :|
 
totally
 
bb gives me free unlimited private repos.
take that away, and no reason to look at bb.
 
8:30 AM
summons @AmithKK and @jokerdino here
 
sup
 
why are you using doc strings when not needed?
 
Nothing much :D
 
who?
 
""" something """ and ''' something ''' are strings, that can go multiline
you @jokerdino
 
8:33 AM
i just want to throw those parts.
 
then comment them.
 
bah
 
wait.. someone says multiline strings are comments? thats ridiculous.
 
kk
what do you want?
pushed update :P
 
66
A: Comment out a python code block

John FeminellaNo, Python does not have such a mechanism. Block comments are made by prepending each line with a pound sign. For more info, see good ol'-fashioned PEP 8. In any case, almost all Python IDEs support a mechanism to do the block-commenting-with-pound-signs automatically for you. For example, in IDL...

okay. confirmed that doc string style comments are not official.
now, do we have sample tests for rest of unity?
 
8:37 AM
we have some tests in the unity trunk
 
@jokerdino \o/ I understood these tests
 
excitement, you call it.
 
We need to make another function that is verbose
as it is run, you'll get outpul like
 
so, I bzr branch what? lp:unity?
 
8:39 AM
succeed: Deleting key
 
@Mahesh yesw
 
@Mahesh yeah
 
@AmithKK my tests are not legit. i just wrote them as i wished
 
man, its going to take forever...
 
@Mahesh agrees
this sucks
 
8:41 AM
whoa.. bzr is not bad at compressing revisions.
 
bzr is amazing. Well after you know it :P
 
okay. I'll try bzr for a while then.
 
Can't we just have manual tests?
 
no.
 
documented on a fresh install of 12.10?
 
8:43 AM
i am not going to compile it. just no
 
make takes a looong time
 
we need automated unit tests.
 
it takes 1 hour +
 
unless your on an i7
 
what?
 
8:44 AM
i am on i3
 
one hour?????
 
one hour to compile
 
@jokerdino did you use -j8 ?
 
i used -j8
jeez. yes
 
hmm... looks like I have a reason to ask for a VM on my institute cloud.
 
8:45 AM
I could have ninjask to compile it
:P
 
ninjask??
 
ninja ask
 
name of my amazon instance :P
 
oh. okay.
wait.. what do you use that amazon instance for? /just curious/
 
Still... 1 hour?
 
8:47 AM
to the release?
 
@AmithKK DNS fail
 
I know
 
okay.
 
Apache is down
 
8:49 AM
what does it cost?
 
@jokerdino you're not using the unity-test suite
@Mahesh nothing for one year
 
@AmithKK i am not using anything atm
 
@AmithKK o.O
 
@Mahesh yes, it's pretty cool. Not bad for free
 
How do I get one? :P
 
8:51 AM
You do need a credit card
 
@amith, what unity-test are you talking about?
 
@jokerdino autopilot
 
ok write some tests thx
 
using it after one year also doesn't incur much cost. I'll sign up for one.
 
@Mahesh \o/
<3
I have some work to do :D
 
9:03 AM
like what?
 
back
homework :P
Done
 
okay.
let's figure out what we should be doing :P
 
trying to do that
 
@jokerdino am allowed to read a log?
 
no idea
 
9:10 AM
@AmithKK log of what?
 
log made by our script
 
Ideally, you should have a lot of output and provide sample output to whatever test module.
it then compares.
 
@Mahesh that was my idea
but you then need a new verbose parameter
 
yup that.
 
9:12 AM
well I can do that
let me finish my packet of lays
 
so i was trying to add a launcher and i failed
it is stored in a string array and the gio and glib were too messy
who wants to work on the next patch?
adding the ability to toggle the show desktop button in launcher :)
@amith, also, did you fix the if loop in your proposed merge request?
 
I should remove the merge request right?
or can I just modify the code?
 
update it.
 
@jokerdino how do you do that?
 
@jokerdino /me /me /me
 
9:19 AM
modify your local bzr branch and push it to your branch
 
@jokerdino it was there, right?
 
so, that's it?
 
@Mahesh not in 12.10 :)
@AmithKK yeah
 
oh god, why? 12.10 looks like a mess.
 
once we are done, we can then work on another regression
the mounts right now will show up automatically.
we should add setting to determine if it is mounted and then show, etc
 
9:21 AM
We should be called the unity patch team :P
2
 
when we are done with one, then yeah :P
We should be the Unity regression police.
 
@jokerdino get a list of all unity regressions and dump them in a list.
 
haha, i would
but most of them are too technical for my taste
 
I would have never tjought i'd be able to do what we did, if you hadn't pulled me in.
so, why not try the others too?:P
 
okay, we shall go after them too.
 
9:25 AM
lets get the unit tests straight for now. :P
 
yep :P
 
pushed
Should it just check if the command worked or should it also check where it failed?
 
i think just the former?
but more info never hurts
@amith, s/favourites/favorites
 
amith@amith-desktop:~/unity.cmake/unity$ bzr commit
bzr: ERROR: invalid header line: ''
O.o
 
you sure you are at the right branch?
 
9:34 AM
yep
 
okay. try again
make sure you don't have any ~ files
 
Hmm
my dirstate file is empty
 
where should I be looking? I'm totally lost.
 
look what?
 
@Mahesh ?
 
9:38 AM
it is inside tests/ folder
 
its all cpp
 
inside automake
 
got it
 
@Mahesh just add print("ERROR") statements and use subprocess.popen() inside the test to check the output line by line for ERROR
all those examples are for autopilot, which is gui testing AFAIK :P
 
we can still use autopilot
or I may try unittest.
wait.
 
9:43 AM
yesterday, by Amith KK
9:28 PM <amithkk> didrocks: can you elaborate a bit on the tests
9:28 PM <didrocks> amithkk: sure, so all code entering unity now needs tests
9:28 PM <didrocks> for testing the wrapper, I would say:
9:29 PM <didrocks> import the unity python file (not sure how it works when you have no .py at the end)
9:29 PM <amithkk> oh, it works
9:29 PM <amithkk> :D
9:29 PM <didrocks> waow, never tried that TBH :p
9:29 PM <didrocks> then change some keys in the unit test module in the Setup() phase
9:30 PM <didrocks> calling each of your functions
 
9:55 AM
once lost, forever lost.
 
?
I am writing a unittest.
 
wrote any tests yet?
oh ok
 
where's the unity.cmake ?
 
in tools
 
@Mahesh /tools/
 
10:02 AM
I pulled lp:unity
 
yes.
 
How do I get amith's branch on top of this?
 
pull his branch
 
how exactly?
 
or get the diff and apply it
 
10:03 AM
good idea.
not needed.
 
okay.
 
bzr merge does the trick
 
good idea.
 
@AmithKK what did you import earlier?
in the chat with didrocks?
 
@Mahesh for?
@Mahesh nothing
 
10:10 AM
I need to import the unity wrapper as a module.
or find a way to call functions in unity wrapper from another python script.
which actually means the same. :P
 
heh
 
and I see that no unit tests exist for unity wrapper yet.
or did I go blind?
 
they only started implementing mandatory tests a couple of months ago
before the original wrapper was written
 
then I need to mess with the entire wrapper.
 
This is mucky
 
10:14 AM
sure.
 
@Mahesh import /usr/bin/unity ?
 
nothing we have changed in that
 
that should work, but I dont know....
 
hm, we didn't modify that binary file. ??!?
 
nope. you need to place the module in the search path and then import.
@jokerdino thats plaintext. no binary
 
10:17 AM
@Mahesh write a basic one and i'll just follow it.
 
got it.
I'm dumb.
 
yep
 
you need a .py extension to be able to import it.
 
looks at @amith
 
O.o
It did work once
 
10:23 AM
I've been headbanging for halfan hour to import without .py
okay. no probs. lets move ahead.
 
okay
 
facepalm that was with imp
 
?
 
okay.
subprocess.Popen(["killall", "unity-panel-service"])
subprocess.Popen(["killall", "compiz"])
why dont we just call?
 
10:31 AM
i was following the existing format in that
 
ok.
 
there was another Popen in there
 
@Mahesh We should
 
I'm making some changes.
 
BBL
Good Luck
 
10:38 AM
cya
 
should I write unittest fr the final unity wrapper or just our snippet?
 
just our snippet first
after that, we can consult with didrocks
do you want to talk to didrocks about call and Popen?
 
yup.
we should use call whenever output is not needed
 
i see
 
10:59 AM
any idea whom to contact to get this mess corrected?
@jokerdino ^
 
11:16 AM
holy
i guess you ping one of the people in ubuntu-devel
 
11:32 AM
get ready for a surprise.
 
bang
 
checkout the repo.
 
i see a unittest file
 
look at the unity-reset.py
I changed its shape almost totally completely.
 
jeez
 
11:38 AM
:D
test that for me, will ya?
 
okay @_@
\o/
but there is a significant time lapse
and a bug with the launcher
logging out. brb
 
two subprocess calls. I expect a Quarter second lapse at max. anymore than that, no idea why.
 
it takes about 10 seconds.
last time (the old wrapper) it took a couple of seconds.
 
thats ridiculous.
I haven't changed much, did I?
what getss executed is all the same.
 
hm not sure
 
11:47 AM
retry please : |
 
i don't know why we have such a long time gap
 
ten seconds is like a lot lot lot of time.
 
it's much more than that.
ok just slightly less than 10 seconds
do you want to sync the settings?
.sync()
 
.apply() is ok.
we dont need to sync.
 
it got removed somewhere
 
11:49 AM
nope. its still there.
line 56
 
ok but it is very slow. why not just sync it
 
comment out 39 and 40 and try again.
I am eaget to blame subprocess.
 
no diff
 
:(
actually, we can get a significant speedup.
 
yeah
 
11:54 AM
comment 50,51,52.
we got the schemas from Gio. no need to recheck if they are there.
 
using sync slows down the reset but result is pretty fast
 
wait.. who added 65? I remember removing it.
 
o.O
brb. give me 3 minutes
 
okay.
 
gsettings.apply() is only if you are also using .delay()
 

« first day (3 days earlier)      last day (150 days later) »