Datasets:

showvikdbz

/

code-reviews

like 0

Are there any additional concerns about the volume of trace data being collected here, compared to baseline tracy usage? I.e. Should developers be more careful about long-running traces? (I'm thinking of cases like with WTF where if you enabled WebGL tracing and included texture/buffer data the files could get much larger than if they just contained collected metrics) Were we ever considering turning on tracing in release builds? Maybe that was on a different project, and with WTF.

Yep - this is the nonsense: https://github.com/pantsbuild/pants/blob/c1eebc96fdcc73fbdfbb0df7805bca4f4d8026ae/build-support/bin/_generate_all_lockfiles_helper.py#L196

oh, crap, thanks for spotting that.

What's up with this shit?

why does not this metavar have i noticed neutron net id below also does not have

if it can't find a blobstart that's a pretty big map issue. Anyway this is specifically not my shit and this is not the pr to give it its own spawn landmarks.

let us move this new code up above dump with the other new code i like dump being at the bottom

it s important that we are super clear about this when an integration is mixing both async and sync api

cover the situation that all soft deleted rows will be purged

what is the root because of this problem you need to run unstack sh before running stack sh ovs will be stopped and uninstalled there

Crap sorry, needs to be `if(isliving(target))`.

use decorators skip because

this one is actually a pretty good idea that scratchaddons can make. so it would list some templates which would b obtained from a studio and then when you click one it would create the project? cloud variables would be wacky to add though

also found this problem in cinder tests api contrib test admin actions py cinder tests api contrib test backups py cinder tests api contrib test cgsnapshots py a lot of places there are bugs tracking this problem and

firestacks is an awful 'feature', if not bug. It means you're basically fucked if you don't have an extinguisher/non roll method of putting the fire out.

ugh what a stupid option the above is it should diaf also it has nothing to do with database configuration setting this to false causes new service records to be marked disabled upon their creation

HOLY JESUS. WHAT IS THAT? WHAT THE FUCK IS THAT?

can we kill the base classes too

why use jsonfield if it is serialized and not versioned just dump it into a textfield

need to be removed

no longer a draft methings

Okay, but for someone complaining about single-character variable names. ;) At least name this `e` or `ex` or something. I was like "wtf is c?"

here should be used following validators packstack installer validators validate multi regexp regexps are given as values of option list

dammit, it's really misleading.

What I expected: To toolbox the clown while calling him a faggot. What I got: Having to click on the chat bar between tool boxing his sorry shitty ass face in

it is direct check of ssh executor if it does not work then it will fail can be any command here this one is just one neutral command it is executed on start of a driver

egor i think the registry container is not on master node it is on each minion slave node

we want to remove the dict compat mixin from instance objects so please use dot notation for accessing fields on an instance object

Seems like this is 4xx: you screwed up, you haven't confgured things yet?

these params should be nudged to the right a single space

Ahh yes should be fine. sorry i should have re-reviewed before commenting rubbish

cyberboss fix ur shit.

Still 10, must have been accidentally reverted in one of your commits or you forgot to commit it. Shit happens.

actually this is in a unit test so that is kind of a silly concern nevermind

export libvirt vol pool tripleo

you should be using a mock instead of implementing you are own fake

Removing strict mode just for the sake of a button is ridiculous. This is a legit problem with antd that needs to be fixed.

chris i think this key alue should be string string because of l sorry if i miss something

i put them into two unittest for mediamonitor i emitted the signal to simulate insert remove events

ah lmao nvm this is a classic pathing fuckup, see op about the bar/atrium thing

you do not actually need these haschildnodes calls the dom will return a shared empty node list for you avoids extra nesting

if we have write failure we waste seconds checking not sure if i have an issue with that i do not know how remount ro done can pass if this command did not make it into the kernel do we want to consider a runtime force remount on the list marking the ones that fail so we do not attempt the fsck on those

create bitmap true

i have no problem with cutting down on the verbosity but do we really want to remove all logging of the request rest client py does do some logging

just remove the tab indents at this point, they're hideous

@ilovezfs because I suck at typing, lol?

I think I'm going to just merge this. You could try rebasing again as we've put a few stabilizing patches in recently but the amount of retries you need here is ridiculous.

i think should be removed

kill extra line

i can do that are these redundant or actually useless though

of course i do... brain fart

we have machine flags for these behaviors -- you want screwtoggle and crowdestroy (probably wrenchanchor too)

you should use objects consistencygroup get by id since you already use consistencygroup object here

I know you didn't change this part but wtf

Well that is a ridiculous solution. What is going on that will make me have to update 200+ files with boilerplate like this, just for updating Typescript and installing the Redux types? Anyone else think this is asinine?

it is entirely worthless and everyone knows it, besides it's not like you're gonna need it in fuckin' space

instead of putting in this jank ass while loop, why not put it in some lambda ?

Thanks for the review @anomiex - you picked up a bunch of clumsy mistakes.

I'll close it, but I still think it sucks pretty hard that the installer situation is so incomplete for Matplotlib. It's almost like the developers want mac users to complicate their lives with esoteric installation methods, rather than just a simple binary installation. Sure, it's not that hard to compile, but to have to compile things to install in 2012 is a bit oldschool. Even linux has good binary package managers these days...

do not add a format here this is deprecated in favor of the gbm format below

i think these below two methods should be some utils calls not the baseclass for datasourcedrivers

is 0.4 meth ok? (1 tick) and meth isnt a lategame chem, just look at shit like rdx, tatp, life and then look at meth made without ever swapping a beaker and an igniter

thanks very helpful we will use this convention

profileid profileid adapter type adapter type extra config extra config

Our embedded file is much smaller than what's shipped with Hellfire, at only 899 bytes vs 11.5 KiB, so that's a nice advantage of shipping our own as well

Looks like the issue is with the ```window.zoomLevel: -1``` when the value is 0 or higher, ghosting disappears. Kind of sucks since on 1440p, the UI is a bit too big for me. that's why I zoomed it out.

why was tunnel update missing earlier

Ya I started a new gatsby project (around my 4th one, but havent had a new one in months), updated everything and then all shit went to hell. Does anyone know when there will be a npm fix or do we all just have to start using yarn

you almost never want to clean up the way teardown works instead you probably want something like this at the end of setup def cleanup servers for i in self nova servers list self addcleanup cleanup servers the main reason is that teardown does not get called on failures

why use this won t the socket be closed by the garbage collector and closed as soon as it goes out of scope

updated at should be nullable false because upon insertion it should and will be set

`brainsucker` - really?

wow github fucked up the diff hard on this whole file

end all help strings with a add the that the privsep daemon

~sigh~ I upgraded to 0.45.0 from 0.42.0 and noticed this problem crop up. However, I had also upgraded to Node 6.9.1 at the same time. Downgrading to Node 4.6.1 has alleviated this issue. It's quite ridiculous that the current LTS release can't be used.

I spent so much time debugging this locally with a client verb that I never actually did the server implementation. Conclusion: I am actually stupid, and wondering how the fuck I missed this.

do you know the difference between dump and dumps

Too much crap in cu_users to not have it on when Travis is running tests. I still don't think cu_users should be a dependency of cu_atlas. At this point, it might as well be placed in cu_core so that it is always enabled no matter what.

the fuck are these procs? looks like some LRP BS, consider removing it?

really it s still in most brillo mk files afaict

nit it would be nice to have this rename ndb utils as db utils so that it is consistent across all the files in this patch

capitalist faggot

ok maybe put them to lib compress cros image sh

shut the fuck up goof

nitpick the os here is an author hint to oliver specht usually it is nice to remove those too thats what dscms are for

damn this is very very clean. very cleaver.

this should be marked for translation

damn, nice catch. I keep expecting babel will sort out this stuff for us. @willeastcott, do you know if it's possible to get babel to handle this for us automatically?

alright fikou confirmed 1575 works, i guess i fucked up testing that one

how about having some uuid here

these are not needed i will completely remove them

fmt self fmt

new unit tests should use mock instead of mox the above mock ified would look like this def test list instance ids empty self driver libvirt driver libvirtdriver fake fakevirtapi true with mock patch object driver get num instances return value as mocked self assertequal driver list instance ids

Why the fuck are these arguments never consistent, is that seriously armor and not shield block?

That's what happens when you make shitpost additions at 3am I guess.

With that setup, you'll get THIS horrible mess instead of the usual flickering... 

i think we would better to get rid of on the head of func name because this will be called by process py

it's still awful but at least it fixes the bug

super nit end comment with period here and elsewhere

damn my IDE overdo on this one