« first day (1498 days earlier)      last day (3680 days later) » 
02:00 - 17:0017:00 - 00:00

02:53
@DavidFreitag that looks like shell commands, not Makefile
@tylerl It's makefile when it's wrapped in a $(shell ...)
 
6 hours later…
08:45
@RoryAlsop w000t its catching on! Ya little porch monkey.
@Simon That makes no sense. What does that have to do with usability?
@AviD porch monkey? ;-P
oh - and good morning
It's okay, I'm taking it back.
good day, sir.
damn still funny
09:01
Morning All
@RоryMcCune I said Good day!
@AviD And a good day to you sir
@AviD now there, I'm busy today, none of that youtube linking!
hehe
just excited to have my proper machine back!
now to rebuild aaalll those VMs...
09:05
@AviD did it go on a trip?
@RоryMcCune naw, my HDD crashed... just replaced it.
@AviD erkk, that's not a good thing... Seagate by any chance was it?
@RоryMcCune no, WD.
the green quiet ones.
and learning to follow my own adivce about better backups.
not just data, but machines.
@AviD ah, I'd been reading a post on HDD failure rates on the backblaze blogs and they had some shocking experiences with Seagate drives
@AviD ooh nasty...
was looking at RAIDing it this time, but @JourneymanGeek pointed me at Storage Spaces - btw @JourneymanGeek thanks tons for the tip, its looking great!
09:08
@AviD ah yeah, I could see that being good. My main work system is a laptop, so I have to rely on more basic backup techniques...
its very cool. I think of it as "virtual RAID".
though they very insistently avoid the term RAID...
09:20
@deed02392 mornin'
10:06
oh England, how I missed you
(not)
@AviD I have never had one of these fail
distant sound of my entire computer setup failing
do you reckon people put fake data in their profiles?
:-)
@RoryAlsop either that or the SE generation is really popular with nonagenarians
@RоryMcCune heh heh
@kalina well, first time for me too.
@kalina @AviD I remember trying out those ones, but I wasn't a great fan of the agressive spin down they did to save power, really seemed to slow things down (subjectively)
10:19
@RоryMcCune oh I'm sure they are slower.
but the difference in speed is trivial compared to the difference between even very fast HDD and SSD - so that slowdown can be ignored.
you want fast? put it on SSD. otherwise - just hush up and dont make noise.
@RoryAlsop I find it funny that @TheDoctor is not the oldest one there. He's off by like a factor of 10.
(almost)
@AviD well of course there have been some suggestions on WD greens that the aggressive parking can increase failure rates... blog.fosketts.net/2011/04/27/…
@RоryMcCune tl;dr, but I'm pretty sure that is based on years-old information
when the greens first came out they were inferior drives, slowed down to cover for their expected failure rates - but not enough.
will read the whole thing later though
@RоryMcCune ohai
@AviD: I tend to think of it as vaguely their answer to zfs/btrfs
got far too many tabs open, hardly notice notifications here anymore unless i've got sound in
10:28
@deed02392 having fun with the super simple and straightforward web application that you're testing at the moment?
@RоryMcCune sure, a handful of static pages. Dream test to be honest
@JourneymanGeek the RAID-ish parts, yes. Probably not as flexible, but hella more useful.
@RоryMcCune never had that problem, my RAID controller doesn't let them spin down
@deed02392 yeah it's nice where there's no massively complex set of cookies and heavy use of client-side JavaScript frameworks to make app scanning painful and false-positive prone
@AviD: Oh, if you don't mind a slight speed hit, a lot more flexible IMO
10:30
saying that, when I rebooted my PC lastnight it made a scary noise
@RоryMcCune haha! :-) c'mon it's still fairly early buddy
@deed02392 true true
@kalina thats probably your startup sound.
I'm considering throwing a few VHDs on samba shares and trying them with storage spaces. Not fast, but a pretty hands off way of using storage across a few boxen
or maybe you were playing one of @Simon's tracks.
@JourneymanGeek SS is more flexible than zfs/btrfs? not that familiar with those advanced bits of those, I just assumed.
10:32
@AviD no, it was a "I am a moving component and I'm not spinning along the axis I'm supposed to be spinning on" sound
@kalina yeah, sounds like one of @Simon's tracks.
@RоryMcCune need to keep reminding myself they're not invincible
basically, one of my fans glued itself stopped when I turned the pc off
@deed02392 well I think that got proven yesterday :)
@AviD: If you're smart storage spaces works really well with VHDs
10:35
heh, that makes more sense.
So you can carve out bits of existing disks
@JourneymanGeek please tell me how? I am in the midst of setting this up right now...
@AviD: create and mount a VHD, don't format, add it to a storage space
@JourneymanGeek ah, you're talking about SS on top of VHD. I meant the other way around.
10:36
@JourneymanGeek does that even work though?
@JourneymanGeek no, as in not a good idea?
10:52
You mean running a VHD inside storage spaces?
(Both ways work, and I don't really see an issue. As I said, I may end up trying some nasty thing with remotely hosted VHDs if it'll work)
oh dear - too many issues:
0
Q: Android FIPS capable project

yugicoI want to create an Android FIPS Capable app (140-2 level 1). In the app I use network communications (Android SDK, HttpClient), encryption (bouncy-castle) and GCM (GooglePlayServices). How can I do it? As I understood from this post: FIPS Compliance for my Android project First step: I should...

@RoryAlsop Yay for software requests! </sarcasm>
@JourneymanGeek yes. basically since I run all my systems as VMs, there are a lot of VHDs. and basically since SS now provides a (virtual) disk to the host system - I will wind up storing my VM's VHDs on the hosts's (virtual) disks on SS.
ahh
SHOULD work. Actually probably supported on windows server
I know that could work, and I'm sure that individually there shouldnt be a substantial slowdown - just worried that accumatively there will be unnecessary slowdown due to unnecessary duplication of virtualization.
11:04
Have seperate pools and turn off redundancy?
@JourneymanGeek yeah its explicitly mentioned as a use case. I'm just curious if there is a better way, since the storage is already virtualized.
ahh, no idea there. Never really thought about it
@JourneymanGeek yeah maybe that. And do a passthrough of "physical" (i.e. SS virtual) disks straight to the VM?
hmm. I have no idea about that
so instead of running a virtual disk inside a VHD, just have create a "virtual" disk on SS, and give that straight to the VM.
though backups then become more complicated again - instead of just copying the VHD have to resort to imaging again...
11:08
Yeah.
meh, I notice that most of my problems come from when I try to be too clever. I should just stop that and go with the trivial defaults.
After all most software is designed to be best used by the sheeple anyway.
11:33
@AviD You shoudl change your avatar to a sheep then too.
@Arperum Tiggers eat sheep for lunch.
And donuts for desert.
Sometimes if noone is looking, we'll just do extra donuts for lunch. And then more donuts for desert.
11:47
@AviD you're just saying that to scare @Simon
Oh - my youngest had a query, after I was discussing the merits of houmoussaria and mentioning that the only person I knew in Israel ate doughnuts:
> Is there such a thing as a houmous doughnut?
because she would eat it
urggh, does NOT sound good
sweet / savory, not a good mix
okay not really savory, but it goes best with savory
@RoryAlsop do you often discuss merits of houmoussaria?
@AviD she loves houmous more than anything else
@RoryAlsop hahaha, brilliant, bring her here
At some point I intend to bring the whole family - just needing threat levels to reduce a little (to a point at which SWMBO will agree to a visit)
ymean at work...?
oh politically?
is there a threat level here now?
11:55
her ideal party food is houmous and carrot sticks / bread sticks etc. Unlike all her friends, who like McDonalds
dont really watch the news...
though I do get an email from the US embassy once in a while...
@RoryAlsop hehe, nice. Typical here btw
or maybe some tehini to change things up a bit
@AviD I don't either, but there's a bit more trouble up north east of you guys - and while I'd be happy to fly over, she gets nervous
for the record, you should tell her that hummous goes great on burgers too.
@AviD oh - that could work. I could make venison burgers with a houmous topping
@RoryAlsop fair enough. it's not as peaceful as, say, NYC's Central Park. ;-)
Or even Paris :-S
11:58
@AviD just looked at that user - hahahaha
@AviD I know - but there's only so much risk enumeration I can talk through
@RoryAlsop hehe
Bruce Schneier actually does talk about the validity of security theatre in some scenarios. Like tourism.
@AviD yup
the truth is that tourists really ARE at increased risk, no matter where they go.
(if you discount the 80% of accidents happen within a mile of the home statistic...)
@AviD yup, but the actual risks of us holidaying in Tel Aviv when compared to the risk of holidaying in San Francisco are so close
slightly increased here
yknow, with all the risks of sunburn from laying out on the beach so much.
12:04
@AviD lol
12:27
@l0b0 the question is offtopic, as basically you have a quick rant about two sites with poor SSL and then infer this means they are dodgy companies. In reality CREST approval is considered the 'Gold Standard' for pen test companies in the UK, and Tiger provides approval to a wide range of pen-test courses. The threat models for them certainly won't have SSL as an essential
Sure, it would be better if they did sort it out, but it doesn't matter
It's not a security issue
@RoryAlsop ENCRYPT ALL THE THINGS!
3
@RoryAlsop although it is poor form.
@RoryAlsop and it would be astonishingly funny if someone MITMed one of those sites at a conf and re-wrote key sections. Not particularly high threat, but amusing nonetheless
12:44
@RоryMcCune I disagree. Since their sales model is based largely on their reputation, and this can potentially lead directly to a big hit on their reputation - if not the very existence of this issue - I would say that for them this should be high on their risk model.
@AviD hey it's t'other rory you're disagreeing with, he said it didn't matter :op
That said, pentesting companies don't necessarily sell to orgs that are particularly well-versed in security issues.
@AviD crest do
@RоryMcCune "Not particularly high threat"
@RоryMcCune I said "not necessarily"
@AviD yeah as an overall org, that's still more than "doesn't matter" ;op
12:45
@RоryMcCune and even CREST is mostly for gov type sites, no?
@AviD not really CREST co's are all (or at least most) of the big UK pen test companies, so do stuff for all sectors
e.g. CBEST is largely focused on the FS market
@RоryMcCune kinda proving my point for me here ;-)
@AviD I didn't say it wasn't important at all!
@RoryAlsop It wasn't meant as a rant, and it certainly wasn't about their threat model, but rather the threat model of the companies accessing their web sites. I've clarified the question to include this.
@AviD Woooooow!
12:56
Yay for social security site using SSL3 and browser telling it to go to hell.
@Simon TWSS
@Arperum they said social security. nothing about computer security.
I mean, its right there in the name.
and for that matter, they suck at the first, why wouldnt they at the other...
13:23
@AviD Yup. And it doesn't help that there are two different things with nearly the same name.
two?
to?
I get soooo many requests on LinkedIn from people in so many different unrelated fields: "Oh I see you're also in [X] security..."
@Simon donuts.
I should add a picture to my LinkedIn account.
@AviD way more actually but two that are relevant to my case (there are at least 5, based on politic preference, Belgium is wierd)
But thetwo relevant ones are basically called vnz and nzvl, where the second one was earlier known as vnz. And to add insult to injury: I've already got a letter from one of them in an envelop from the other one.
both of them are supposed to be politically and religiously neutral.
13:28
oh - I thought you were talking about "security".
I'm famous!
@paj28 you the bearded guy with the plaid shirt right in front of the camera?
@AviD Wait? You think Belgium has anything remotely "secure"?
@Arperum YOU were talking about "social security"
@Arperum Yo momma.
13:35
@AviD - yep, that's just before my talk starts
@paj28 oh ha! I was kidding, assumed you were the guy talking
is eyefi a commercial product?
yeah, SD card with integrated wireless
quite cool... and hopefully now secure :-)
sounds cool but pleeeeeze tell me thats not a marketing talk :-)
it's an OWASP talk... but with branding on my slides
@Simon I think @ScottPack has competition in the glorious beard department
13:39
make of it what you will :-)
@paj28 Look at you looking good.
@AviD Damn right.
@ThomasPornin Hello. I have a crypto question for you. :) Are you aware of any general way of constructing a secure, authenticated streaming encryption API?
@paj28 SELLOUT! ;)
@paj28 I got that, branding is fine but are you trying to sell it in your talk, or talking about the cool tech? :-)
I wanted to grow a similar beard but I trimmed it earlier this week :(
@AviD - I'm talking about how to hack it
13:40
2 months of work, almost all gone.
@paj28 aaahhhh okay cool then.
There was a directory traversal flaw in the component you put on your laptop
wanted to know what I'm gonna watch before I watch it.
That lets someone on the same wireless network take control of your PC
@paj28 ouch!
13:41
@paj28 lmao
@paj28 whoops.
@paj28 whoa
was also worried about the integrity of OWASP Manchester
see the video for more info
@TerryChia At API level, I usually use a basic stream API (like java.io.InputStream+OutputStream, or .NET System.IO.Stream) with a couple of callback functions (or objects) to handle authentication.
13:42
thought maybe they do things differently over there ;-)
@ThomasPornin Perhaps I should explain what I mean more clearly; for large datasets it may be useful to pass data into the encrypt or decrypt function in a streaming fashion instead of buffering it and passing it in all at once. Problem with this is that data decrypted in this fashion is unauthenticated until the last byte is received.
@TerryChia Ah I see.
This is a protocol issue, not an API issue.
If you want to have authentication for partial data, you need to split the data into individual records, each with its own MAC, and the MAC should also cover the sequence number.
That's what SSL does.
@ThomasPornin Yeah, that's what I thought. Is there something out there that describes implementing this for general encryption instead of a transport protocol like TLS?
It seems simple but I'm sure there are a lot of pitfalls. :)
@TerryChia It is full of pitfalls.
In general, it cannot be done securely in abstracto.
Because it must rely on an initial negotiation that establishes the algorithms and keys, and details matter.
In practice, you have two existing models, SSL and SSH. Both do it wrong, but in different ways.
To do it somewhat right, you should have at least:
- A definite maximum record size
- Encrypt-then-MAC for each record (ideally, GCM or something similar)
- The MAC also covers an implicit record number (to detect missing, duplicated and reordered records)
- There is a definite, unambiguous, MACed "end-of-stream" record (the SSL "close_notify").
14:06
@ThomasPornin I found a related crypto.SE question: crypto.stackexchange.com/questions/6008/…
Cheers for the knowledge as usual. :)
@ThomasPornin could be worse -- you could end up with WEP when you're trying to make a cipher system.
14:38
@AviD @paj28 I saw a pre-note of the talks they've got lined up for the next one in Feb, look really cool!
Yeah, OWASP Manchester, in the global heart of intercepting proxies!
@paj28 Manchester, UK i assume?
@paj28 actually I'd not thought of it like that, but it is!
@Ohnana is there another?
@RоryMcCune urggh, I really need to get to set up the next meeting here too...
@AviD 'tis cool, some have taken to livestreaming the meeting with things like google hangouts!
14:50
@AviD I live in a state that has a Bethlehem. I have to ask these things
@AviD It's an American behavior to believe that everything exists in 'MERICA.
@RоryMcCune we should try that...
though its all in hebrew anyway, so wouldnt help you - but others here might like it.
Manchester is the largest city in the state of New Hampshire, the tenth largest city in New England, and the largest city in northern New England, an area comprising the states of Maine, New Hampshire, and Vermont. It was first named by the merchant and inventor Samuel Blodget (after whom the Samuel Blodget Park in Manchester North is named). Blodget's vision was to create a great industrial center similar to that of Manchester in England, which was the world's first industrialized city. It is located in Hillsborough County along the banks of the Merrimack River, which divides the city into eastern...
@AviD yeah IIRC you can log to Youtube at the same time...
we're quite a large country
14:51
@Simon and, if it exists in 'MERICA - THATS THE ONLY ONE THAT COUNTS
also we have ties to the UK, so we ripped off a LOT of names
New York, anyone?
Paris
Venice
...
@TildalWave Paris USA is a disambiguation page lol
@Ohnana Try Paris, TX
hmmm actually there's loads of Parises in US
@AviD Absolutely.
14:55
@TildalWave I know! that's why it's a disambiguation page. Paris is in 17 states
@Ohnana 18, you forgot about the one in France :P
@TildalWave france is a country not a state
unless they also chunk the country into states
is that a common division? i know canada does provinces
15:08
networks people: is there a substantial difference between static IP and fixed DHCP?
specifically for a AD DC, which is running a DNS server but not the DHCP - the router is doing that.
damn where is @ScottPack when you need him
@ScottPack probably got insulted because of what @Simon said about his beard.
@Ohnana Well France does have other territories throughout the world too, so I'd imagine it's ... convoluted
sowie
translate: sowie
(from German) as well as
thought as much :)
@TildalWave "sorry" (from Canadian)
translate: [pretty much anything at all]
(from English) * [pretty much anything at all] *
@AviD you mean as well as sorry? :)
15:16
(from Canadian) "Sorry"
(well, except for "donut")
translate: donut
(from English) donut
bah
@AviD i believe that the configuration for static IPs reside on the clients. fix DHCP means that the router assigns static ips based on the MAC it sees
static ips are also very finicky and evil
@Ohnana dur
I know THAT
I mean, I know what they ARE, and how they work. I meant, is there a substantial difference in effect, when dealing with DC / DNS server.
@Ohnana why do you think they are finicky?
if anything the fixed DHCP is more finicky. easier to manage, but also easier to fudge up.
dur
@AviD i've tried to do one on my linux server. occasionally it'll just change it's ip address or just not have hostname to ip translation on boot
15:19
and if you want to rely on the IP for anything - even say IPsec - youre sqrewed
@Ohnana that has to do with the DHCP client / DNS registration.
@AviD i see. I can fix the problem with dhclient so I figured it was something to do with that
although the ip migration is really the painful part
@Ohnana yeah, thats the one.
@Ohnana Funny, there's also a Manchester, Connecticut. It's where I was born.
so I realized that my DC was set up with fixed DHCP. That was all fine, until I had to rebuild the VM, and got allocated a different MAC - leading to a different IP. Took me a while to figure out why things were screwy...
oh thank god the us senate passed the climate change amendment. there is light in the darkness
15:25
lemme guess, they made it illegal to change the climate?
no they said it's real
woo.
@Ohnana What, so you believe everything anyone says about some mythical thing up in the sky?
@DavidFreitag i see your bait. I see it :P
@Ohnana :D
@Terry It turns out that git is just really really slow on windows.
Either that or something is wrong with my HDD.
Nah, it's #window.
15:32
I made the mistake of getting a velociraptor.
Yeah, what's up with that, you never watched Jurassic Park?
@DavidFreitag not really a problem for me, though I havent used it too much on windows yet.
but when I did it was near immediate.
what are you using?
@AviD It's easily 5-6x faster in a lube VM.
@DavidFreitag sure, its all the lube.
when I was using on windows it was mostly smallish projects, so that may have been a factor.
15:34
@DavidFreitag i don't want to google that
then again I was doing it almost completely through VS, so they may have just put in a ton of optimizations.
@Ohnana Lubuntu VM
@AviD How many files in your repo?
@DavidFreitag k.
@DavidFreitag not many, maybe a few dozen. dont really remember.
@Ohnana You never lubricate VMs? Premature replacement incoming.
15:36
@AviD Yeah mine have a few hundred ~100-1000 line source files in them. It makes the windows git choke a little bit
I use mostly hg, though I am slowly starting to adopt git for some dev projects.
mainly thanks to the VS integration.
Remember when VS used to cost 10k?
Good old days.
@DavidFreitag yeah, I only used it on small ones. I did use the Mac client on larger ones, but then I would have to use a Mac.
@AviD I think that VS has adopted the good old windows delayed write tactics and made the entire process very transparent
@DavidFreitag how do you mean?
15:37
But checking out 500 files on windows takes a little under two minutes.
wat
thats not even on pull, right?
No that's git checkout some-branch
A pull takes not much more time since we have a pretty nice business class connection and it's really just the download speed on top of the checkout
@DavidFreitag yeah, thought maybe there was some lag there.
15:39
i've always wondered how git checkout works
@Ohnana I imagine git creates a patch for each commit starting with the most recent common ancestor and just rebases to the common ancestor and applies them.
@DavidFreitag i know some of those words
@Ohnana When you create a branch it basically creates an exact copy of the original repo, in this case the head of both branches is the common ancestor. If you add commits to the head of each branch (possibly very different commits) that commit is still the common ancestor
And I don't know if "rebase" is the right word for it, but basically git would have to reverse each commit past the common ancestor, then apply the patches for each commit on the branch being checked out.
@DavidFreitag oh I see. so it cascades changes to change the repo status
@Ohnana I wouldn't quote me on that, but it's one way it could be done.
15:47
@DavidFreitag it doesnt actually make a copy, it hardlinks to the original.
@AviD That makes more sense since there will always be a common ancestor.
@DavidFreitag always-ish. git does let you do some wonky things.
@AviD Oh, trust me, I know. I have done my fair share of ruining git repos
heh
one way to ruin git repos: expect them to behave like hg repos.
Like rebase to a commit before the common ancestor of several branches, push to remote and suddenly wonder why none of my branches work
raz
raz
15:49
@AviD As an April Fools joke my friend removed all the white space from an entire repo and checked it back in.
The main guy didn't know how to revert.
Quite hilarious.
hahaha wat
raz
raz
Hey the code compiles.
@raz That's so messed up
raz
raz
lol
@raz is he still alive
raz
raz
15:50
Best way to teach someone how to revert a repo without destroying functionality :D
heres another fun way: change all EOLs to LF only. nice way to screw with Windows guys.
raz
raz
hahaha
How do you revert a repo properly? Everytime I do it everything breaks
raz
raz
@DavidFreitag Then you done effed up.
@raz Yeah especially when you accidently push to remote and need to nuke the whole repo
15:51
@DavidFreitag in the whitespace case, you dont really need to revert the repo, just checkout the previous commit, and base all future commits on that.
@AviD So git checkout <sha> then just commit from there?
sure. so what if there is an extraneous commit just hanging off the tree...
'course a purist would want to cleanse it, but so what?
Heh, I do recall asking how to do it properly...
with all the things that could go wrong with a revert, why bother
@DavidFreitag oh I wouldnt know much about that. Ask @TerryChia.
I would consult google, but everytime I do that something goes horribly wrong and ends with me deleting a repo from Github.
16:05
I've flattened entire repos before. it's a really, really long and ugly command
you just keep stomping on git until it loses its marbles and then go from there
16:22
I've stayed away from git because it seems to be a damn mess.
I'm gonna have to learn it soon though.
@Simon a mess??
Refer to the previous discussion.
the thing is it's half easy enough to use without reading up on it properly
I always hear people talking about git issues.
raz
raz
@Simon Git is the easiest source repo of all time.
16:27
but if you haven't taken the time to read about it, you can easily mess things up quite badly
Then I guess I should just read all about it and then I'll be fine.
i think source management is a generally difficult problem to solve though
True.
yea i think so @Simon, a little doc reading goes a long way with Git
raz
raz
Pretty much. It's about 10x better than SVN. And I generally like SVN.
16:27
Yeah, SVN is what I use.
raz
raz
But git will change your life.
yea it's far, far easier than SVN when you've got to grips with it imho
:o
raz
raz
The best SVN client is Git.
You can actually trick SVN and use Git to manage SVN repos. It's fantastic.
cheeky
16:28
Nice.
glorious
raz
raz
And it takes up no space. You can do crazy branching, and merging is made so much easier because most of Git is just pointers to previous commits.
Unless there are merge conflicts which there always seems to be
02:00 - 17:0017:00 - 00:00

« first day (1498 days earlier)      last day (3680 days later) »