« first day (2516 days earlier)      last day (2494 days later) » 

Bob
12:13 AM
morning
 
Morning!
 
Bob
@Ramhound what type of encryption? the old pkzip one, or proper aes?
 
@Bob Would it matter if the password is guessed anyway?
 
Bob
@MichaelFrank more out of curiosity. but pkzip crypto is muuuuuuuch more crackable.
as in, it's actually crackable while aes mostly relies on brute force
 
aes
 
Bob
12:27 AM
ah
huh. this doesn't smell like a SF question o.O serverfault.com/questions/857769/…
would be appropriate on SU though
oh well
 
12:45 AM
@MichaelFrank Then don't interrupt the cows!
 
moo?
2
 
frack
i haven't done my timesheets all month :(
 
1:00 AM
@Burgi Yea... I forgot to do mine from last week... doing them now. >.>
 
i need to actually do this crap as i go...
it doesn't help that our timesheet software is crap
literally some of the worst on the planet
 
Ha... that never matters to the company.
 
heres the rub... it doesn't work properly for invoicing!
apparently there is some crazy vendor lock in that we have
 
Is invoicing your responsibility?
 
its either that or the upper management are utterly cheesed off with shitty systems they refuse to change (better the devil you know)
no but i get emails from customer services going "why have you logged 3 hours to 'updates' 6 weeks ago?"
they ask for the fracking jobs, they should be keeping track!
you can make better time management software in excel
 
1:07 AM
Yea, that sucks. You just gotta do what is required and let everyone else have a cry about the bits that aren't your problem.
 
oh while i remember... @MichaelFrank congrats on the egg chasing rugby victory
 
@Burgi Ha, I didn't even get to see it tbh.
I was at a wedding
 
my gf stayed up listening to it on the radio and promptly fell asleep literally 5 minutes before the start of the egg chasing
WE COULD HAVE GONE TO THE PUB!!!!
 
Yea, there was an announcement at the wedding by the MC to ensure that everyone kept quiet about the score as people were recording it at home for later.
 
her argument was that if we went to pub she'd be too tired to stay up to listen it the match
was not impressed
we opened a bottle of jordanian wine (who knew jordan made wine?)
 
1:12 AM
0_0
THey'll make it, not drink it?
 
i'm not sure
i'll ask
won't get an answer until tomorrow but i'll tag you in when i do
 
1:36 AM
ugh
ipv6 is still broken on apt
 
@bob @BenN I got an answer to my foreach problem I was having the other day: stackoverflow.com/questions/44688821/…
Hmm... what the heck is an Apple SIM?
 
Apple decided to not follow the standard for SIM cards
They claim the standard's size was to large or something like that
so they designed their own standard.
 
they use embedded sims and nano sims no?
and nano sims are standard
 
They might have went to the standard. I think APPLE SIM was in earlier models
 
Bob
@JourneymanGeek huh?
ipv6 apt works fine for me (scaleway)
 
1:47 AM
Oh right. I should've scrolled down like 1/2 an inch.
The new Pro, iPad and iPad Mini are all Nano + Apple SIM Compatible.
 
@Bob its never worked for me, gets stuck
 
Bob
@JourneymanGeek ...weird.
 
Its only GSA? AT&T and Sprint, in the states
 
Bob
@JourneymanGeek what's your setup?
 
1:48 AM
@Bob also do you do certbot/letsencrypt with ngnix?
 
guess not because T-Mobile is theretoo
 
@Bob roughly as you left it - just added the certbot repos and french ubuntu servers off repogen
I just disabled ipv6 for apt for now
 
Bob
@JourneymanGeek yup. I don't use the auto-install plugin.
certbot --nginx certonly
 
ah, dang
I was wondering though if it expects everything on a single config file or if I do per site config files
yeah, I can do that
 
Bob
@JourneymanGeek Your choice :)
At least under Ubuntu/Debian it defaults to sites-available/sites-enabled
 
1:51 AM
documentation is a bit shit there ;p
oh well
 
Bob
@JourneymanGeek as I left it there was no ipv6
 
I'll muck with it a bit.
yup
 
Bob
@JourneymanGeek ask away
I've messed around with it enough over the years to more or less know what I'm doing :P
 
Bob
@JourneymanGeek fyi, use --dry-run with certbot -- it hits the staging serer, which has higher rate limits
once it works, switch to normal
 
1:53 AM
naw, was going to set up one (test) domain first, then start moving stuff over slowly
 
Bob
@JourneymanGeek hm?
doesn't matter how much you move over, use --dry-run until it works :P
 
Oh, for ngnix
lol
 
Bob
ah
 
if one works, it should mostly work
not doing anything exotic yet ;p
 
Bob
eh, feel free to ask for anything
@JourneymanGeek I'm doing fairly exotic nginx configs recently :P
        location ~* /lxd/(?<req_upstream_host>[^/]*?)/(?<req_upstream_uri>.*) {
                resolver 10.252.212.1;
                proxy_pass http://$req_upstream_host/$req_upstream_uri;

                set $frontend_redirect_target $scheme://$host/lxd/$req_upstream_host;
                proxy_redirect http://$req_upstream_host/ $frontend_redirect_target/;
                proxy_redirect / $frontend_redirect_target/;
                proxy_set_header Host $host;
                proxy_set_header X-Real-IP $remote_addr;
 
1:55 AM
lol
as I see it I basically need a fairly simple setup that 1) does new.lupinenet.com 2) https 3) redirects http to https 4) more coffee
 
Bob
btw, @JourneymanGeek, I use mozilla.github.io/server-side-tls/ssl-config-generator modern config for nginx
that takes care of everything https
@JourneymanGeek want a step-by-step? :P
 
Bob
first, copy /etc/nginx/sites-available/default to a site-specific config, e.g. /etc/nginx/sites-enabled/news.lupinenet.com.conf
rm /etc/nginx/sites-enabled/default
ln -s /etc/nginx/sites-available/news.lupinenet.com.conf /etc/nginx/sites-enabled/
then go in and edit it
one sec, I'll dump one of mine for you
 
I assume with these changes?
 
Bob
@JourneymanGeek hm?
 
2:02 AM
I mean for ssl and all that
 
Bob
server {
        listen 80 default_server;
        listen [::]:80 default_server;

        return 301 https://$host$request_uri;
}

map $http_upgrade $connection_upgrade {
        default upgrade;
        ''      close;
}

server {
        # SSL configuration
        #
        listen 443 ssl default_server http2;
        listen [::]:443 ssl default_server http2;
        #
        # Note: You should disable gzip for SSL traffic.
        # See: bugs.debian.org/773332
        #
        # Read up on ssl_ciphers to ensure a secure configuration.
that's my try-server one
and here it is cleaned up a bit (the reverse proxy bit removed)
server {
        listen 80 default_server;
        listen [::]:80 default_server;

        return 301 https://$host$request_uri;
}

server {
        # SSL configuration
        #
        listen 443 ssl default_server http2;
        listen [::]:443 ssl default_server http2;
        #
        # Note: You should disable gzip for SSL traffic.
        # See: bugs.debian.org/773332
        #
        # Read up on ssl_ciphers to ensure a secure configuration.
        # See: bugs.debian.org/765782
 
Bob
@JourneymanGeek yea
 
I think I grok
 
Bob
@JourneymanGeek if you're using that one, things to be aware of: default_server obviously only applies once per IP address, you should probably have server_name set before trying to certbot, and don't enable ssl until you have the certificates :P
oh, and do a certbot renew --dry-run after you have everything else set up, to make sure you haven't accidentally broken the path to the ACME challenge
 
2:05 AM
lol
 
Bob
(I did that once. ownCloud setup defaulted to 404ing .well-known)
 
yeah, at this point, more coffee
I do mostly get it at this point
so in this case I'll set up a non ssl domain, try what the auto plugin does, then replace that with the mozilla suggested varient
 
Bob
@JourneymanGeek if you want -- it largely just shoves in this bit:
ssl_certificate /etc/letsencrypt/live/try.vulpin.com/fullchain.pem;
ssl_certificate_key /etc/letsencrypt/live/try.vulpin.com/privkey.pem;
#include /etc/letsencrypt/options-ssl-nginx.conf; # managed by Certbot
I commented the last line out
 
Bob
If you have multiple sites you might want to centralise the SSL config similarly
I didn't particularly like their default -- the Mozilla one scores higher on ssllabs :P
('course, the letsencrypt one is more compatible... but I'm fine with the Mozilla list)
> Oldest compatible clients: Firefox 27, Chrome 30, IE 11 on Windows 7, Edge, Opera 17, Safari 9, Android 5.0, and Java 8
 
ugh...
 
Bob
@JourneymanGeek oh, don't forget to play with lxd :P
@allquixotic I just bought mobaxterm
 
nice
 
@Bob eventually just 4-5 sites
openid.lupinenet, ttrss.lupinenet...
tbh saturday is the best day for it
mom keeps distracting me ;p
 
Bob
@JourneymanGeek lxd! lxd! :P
 
2:21 AM
@Bob eh, I'd rather throw stuff like miou, which runs its own webserver anyway there ;p
I want to get my core stuff, over working and documented by this month ;p
Well next month
 
Bob
lol
lemme know if you want any help... if nothign else I can give you working examples :P
 
well, part of the fun here is working it out ;p
 
Bob
true
 
I mean if all else fails, I do have a working config on lighttpd
.... which looks oddly similar
just... kinda going eh. and then wandering off to do other stuff
not my day I guess? ;p
 
I feel like doing that at the moment... since now that my script works how I want, I'm trying to implement pipeline input.
 
Bob
2:36 AM
DVI-D is digital and analog. However, I do believe DVI-I is digital only, which you are out of luck. It's extremely expensive to convert VGA to a digital DVI signal.Ramhound 1 hour ago
@Ramhound ...did you just ... invert my comment?
@JourneymanGeek the default config on nginx covers most bases :P
 
And -I stands for..."Iverything"?
 
Bob
The only problem is it doesn't support .htaccess so you'll have to convert those to nginx config directives. (@JourneymanGeek)
 
Integral
also NOT expensive
@Bob I don't use .htaccess
 
Bob
@JourneymanGeek re: your answer... that tablet should support digital in
@JourneymanGeek some software does
 
ok.... firewatch is.... odd...
 
Bob
2:43 AM
it's common to have one blocking php execution inside a cache directory, etc
 
anyway, nn
 
Bob
@Burgi 'night
 
@Bob but I'm assuming they tried DVI - so its another option really
and I've used one of those on a probably 15-20 year old 4:3 monitor at work soooo
 
Bob
@JourneymanGeek I'd rather figure out why DVI didn't work than suggest active conversion.
Active conversion is horribad
 
Bob
2:44 AM
Also, it costs some $40 for the adapter
 
0_0
Pretty sure they're like in the 10-20 dollar range
 
Bob
@JourneymanGeek Maybe down to $30 depending on whether you can catch a sale. That's for a decent/cheap one.
The $10-$20 range is for chinadodgy adapters... would not recommend.
 
I think I spent about 20 dollars on mine, reasonably decent, active one with all the cables.
 
Bob
In fact, I'd recommend running far, far away from anything chinacheap when dealing with displays.
(nightmares with dodgy DP cables and dodgy USB => VGA adapters come to mind)
 
admittedly my DP to VGA adaptor was free
disposed of with an old machine at job -1 ._.
also USB => VGA is naff
 
Bob
2:54 AM
@JourneymanGeek Remember that motherboard with no expansion slots?
 
Bob
Guess why we used USB :P
Side note. It also did not have USB 3.0.
 
but active HDMI is likely less dodgy than USB 2.0 to VGA
 
Bob
Literally the only available expansion interface was USB 2.0.
 
3:12 AM
The old IT guy was... not an IT guy
so essentially everything at work except the laptops are VGA
 
3:52 AM
roar
Just finished an assembly job.
 
What did you make?
 
Propane grill.
(note that in NYC, such grills are illegal on residential properties except 1- and 2-family residences, which means that most apartments cannot have these kinds of grills; obviously, we're not living in this sort of apartment. can't go into further detail)
> For a one- or two-family home, a propane barbecue grill is permitted ā€” with no more than two 20-pound LPG tanks.
 
4:24 AM
Good morning!
@allquixotic lmao
 
Bob
Oh god.
The answers on that CPU cores running at different speed question.
Most of the answers there are terrible.
Maybe a decade ago...
 
Protected.
 
Bob
@bwDraco Why?
The terrible answers are from high-rep users.
Heck. The topmost terrible answer is from harrymc.
 
Undone.
So, how does my answer look?
 
Bob
@bwDraco haven't gotten to it yet
so far I've seen answers asserting that it isn't worth it
answers talking about big.LITTLE, which is more of an aside from the context of desktop CPUs (see question)
oh at some point Matteo edited and added the frequency scaling bit
@bwDraco I would suggest that you make the distinction between clock scaling and HMP more obvious
The question seems to be asking under the assumption of clock scaling.
In which case, the answer is "that's already happening", "but HMP is an alternative approach for varying speeds"
"Windows isn't really designed to run on processors that have cores with wildly varying performance" => not really performance that's an issue; Windows copes fine with that (that's one of the effects of clock scaling)
I mean, screenshot I just took:
Now, that said, Windows currently doesn't really try to optimise for it, since it has no way of knowing which processes should be operating in a low-power mode.
But it copes more or less fine.
*shrug*
@bwDraco Your answer's mostly alright, I'm just nitpicking and a bit grumpy after reading the others :P
I'm not entirely sure the OP knows what they're asking either...
> The question is about lots of big cores vs a couple of big cores and lots of small cores - the merits of the two scenarios.
Except... the examples called out in the question are the same cores just binned/clocked differently.
Or maybe I misread the question. *shrug*
Eh, in the world of configurable clocks, a "big" core can just as easily be a single BIOS-overlocked core. And a "small" core can be, similarly, BIOS-underclocked/locked.
 
4:49 AM
Expanding...
 
Bob
@bwDraco nah, just ignore my comments :P
unless you want to anyway, but pretend I didn't say anything
 
Just expanded the answer.
 
Bob
@bwDraco Nice.
Though, that last bit - I think GPU acceleration is actually more often seen in Intel CPUs these days :P
Or at least it was qhile QuickSync was a thing. It's not as much of a thing anymore.
Haven't seen a decent "APU" in a while.
 
We're still waiting for Raven Ridge APUs that use Zen and Vega.
 
Bob
5:05 AM
@bwDraco you still around?
 
Not for much longer.
 
Bob
@bwDraco FYI, Win10 supports HMP: i.imgur.com/RFEcmBG.png
 
I'll see if I can edit that in.
 
Bob
@bwDraco It's not that important, but it's an interesting note. And I actually happened to be on that page when I opened that book to look something else up :P
 
The fact is, Goldmont is so much slower than Kaby Lake that I'm not sure if 8 Goldmont cores are as fast as two Kaby Lake cores with Hyper-Threading enabled. (Assume perfect parallel scaling.)
 
Bob
5:11 AM
@bwDraco Depends on clock speed.
 
Realistically? Let's say 3.5 GHz on the fast cores, 2.5 GHz on the slow cores.
 
Bob
@bwDraco I have 8 Avoton cores here, if you want to benchmark.
Lemme know which (Linux) benchmark to run and I'll do it.
Clocked at 2.4 GHz
 
(Avoton is Silvermont, which is about 20-30% slower than Goldmont)
Something purely multithreaded?
 
Bob
@bwDraco one sec lemme pause my VMs
@bwDraco I don't have Kaby available right now, so you'll have to test that part yourself..?
hang on
 
Skylake (other than the server SP or HEDT X variants) has the same IPC.
 
Bob
5:16 AM
@bwDraco do you have it?
 
Unfortunately not.
 
Bob
crap, suseri was swapped out
this could take a bit
 
brb in 10-15 minutes.
 
Bob
2 GB @ 2MB/s
 
@Bob What's this?
 
Bob
5:20 AM
@bwDraco amaterasu is my dedibox test server, susanoo is my win2016 server/vm, suseri is my win10 vm
 
Heh.
 
Bob
> Intel(R) Xeon(R) CPU E3-1245 v5 @ 3.50GHz
@bwDraco ^ will that do?
 
@Bob That's okay, this is Skylake-S.
Again, will be back in a moment.
 
Bob
@bwDraco 4x Avoton approx equal to 2x Skylake, for 64-bit integer (prime number) calculations.
E3-1245 v5:
# sysbench --test=cpu --cpu-max-prime=20000 --num-threads=2 run
sysbench 0.4.12:  multi-threaded system evaluation benchmark

Running the test with following options:
Number of threads: 2

Doing CPU performance benchmark

Threads started!
Done.

Maximum prime number checked in CPU test: 20000


Test execution summary:
    total time:                          9.8672s
    total number of events:              10000
    total time taken by event execution: 19.7306
    per-request statistics:
         min:                                  1.96ms
C2750:
# sysbench --test=cpu --cpu-max-prime=20000 --num-threads=8 run
sysbench 0.4.12:  multi-threaded system evaluation benchmark

Running the test with following options:
Number of threads: 8

Doing CPU performance benchmark

Threads started!
Done.

Maximum prime number checked in CPU test: 20000


Test execution summary:
    total time:                          5.8546s
    total number of events:              10000
    total time taken by event execution: 46.6695
    per-request statistics:
         min:                                  4.55ms
The Avoton cores are maybe 2.25x slower.?
Of course, for all I know this could be a pathological case for Skylake :P
 
5:58 AM
hello
 
So... 8C Goldmont will most likely beat 2C/4T Kaby Lake in heavily-threaded workloads with better performance per watt. That pretty much completes the discussion.
 
Bob
@bwDraco Oh, remember I'm testing two threads.
I don't have an easy way to test 4 threads on 2 cores only.
Probably won't matter much for a heavy integer load.
 
Good night, then.
 
Bob
@bwDraco 'night
 
6:23 AM
@satibel 'lo
 
@Rahul2001 'o
 
Bob
:38368526
 
@satibel there
 
Bob
dammit
@bwDraco Sorry, could you not link to that convo above? I'm thinking I probably misinterpreted :P
 
6:31 AM
:38368543
 
Bob
:38368607
 
oh, fun it shows as an answer but then, it is removed.
 
Bob
@satibel ļ»æ
 
test ' '
 
Bob
\o/
 
6:38 AM
@satibel try now
 
I'm keeping that cookie :p
 
@satibel haha, I can force clear them ;p
@satibel It's still working?
 
@Rahul2001 the cookie
 
@satibel I'll flush it, wait...
man, mum is having another terrible mood swing
 
hug her?
 
6:50 AM
@satibel She just left
@satibel Still works?
 
nope
@Rahul2001 WP
 
@satibel hm?
I just added this handy JS to expire all cookies to the transcript:
<script>(function(){C=document.cookie.split("; ");for(d="."+location.host;d;d=(""+d).substr(1).match(/\..*$/))for(sl=0;sl<2;++s‌​l)for(p="/"+location.pathname;p;p=p.substring(0,p.lastIndexOf('/')))for(i in C)if(c=C[i]){document.cookie=c+"; domain="+d.slice(sl)+"; path="+p.slice(1)+"/"+"; expires="+new Date((new Date).getTime()-1e11).toGMTString()}})()</script>
 
works for what I've tried.
 
Bob
@Rahul2001 ...JS probably isn't the correct way to expire cookies.
In fact, your cookies probably should be httponly
 
@Bob Meh, I know it isn't, I only had to do it for @satibel's session to get rid of a pesky thing that let him use my username, without him reloading the page ;p
 
Bob
6:56 AM
@Rahul2001 should make it httponly in the first place
 
@Bob yeah, reading it
 
@Rahul2001 nuh
 
@Bob enabled :)
 
Bob
@Rahul2001 cool
 
@satibel you devil
 
Bob
7:00 AM
note that it protects against JS access (XSS, mostly) but not against malicious users*/*clients where you can still edit browser cookies
 
@satibel What are you doing? -_-
 
can you discard cookies server side?
@Rahul2001 using the old cookie.
@Rahul2001 by seting PHPSESSID
 
Bob
lol
 
@satibel huh.
 
Bob
#JustPHPThings
 
7:04 AM
I don't think I can destroy a certain SESSID
 
@Rahul2001 nuke the server o/
 
Bob
> Cleanup $_SESSION array
PHP pls
> When session.use_strict_mode is enabled. You do not have to remove obsolete session ID cookie because session module will not accept session ID cookie when there is no data associated to the session ID and set new session ID cookie. Enabling session.use_strict_mode is recommended for all sites.
Lemme guess, it's off by default
You really have a couple levels here...
 
setup a filter that drops packet with this cookie.
 
Bob
Delete the params (e.g. username) associted with the session
Delete the session itself (...which apparently doesn't mean much on top of deleting the data)
And prevent session fixation by not letting the client choose session IDs
 
ah, I got it
@satibel Is your cookie dead yet?
 
7:09 AM
@Rahul2001 no
 
@satibel OMG I JUST ADDED A NEW FEATURE! RELOAD RELOAD RELOAD!
 
@Rahul2001 he he
 
@satibel You didn't reload, did you :/
Okay, I didn't want to play this card...
 
@Rahul2001 I did
not the Rahul session
 
@satibel hehehehehehe
 
7:14 AM
finally managed it @Rahul2001, here's a cookie: dfdaa972a8d11c3bb0d4332df8203cba
 
@satibel Since you were refusing to reload, I added session_destroy(); to the transcript, which reloads automagically ;p
 
@Rahul2001 I could've blocked the reloads, but was lazy.
so I'd have a write only session.
 
@satibel Then the thing wouldn't have worked for you
Yeah
 
teh-he
 
@satibel wtf is going on
 
7:25 AM
@Rahul2001 Find out, and I'll tell you :p
 
@satibel I hate you
 
thanks
I hate you too :3
 
Prison Architect is pretty darn good
 
@satibel wail
 
@Rahul2001 hint: my name is RŠ°hul, what's yours?
 
7:37 AM
@satibel -_-
 
@Rahul2001 curse me right?
 
@satibel wot
 
@Rahul2001 that's a hint too ^^
 
@satibel I'll come back to this later, gtg do homework :/
 
ok.
 
7:41 AM
@satibel I hate you
 
@Rahul2001 thanks.
 
@satibel wait
oh
I
 
@Rahul2001 what I would suggest is that you make your name in another colour.
 
@satibel You're using hex
I got it
 
@Rahul2001 not exactly, but if you look at both names, you'll see.
 
7:49 AM
morning
 
@satibel wha
 
@satibel rahul obviously has work to do. Can't you just say what's problematic?
 
rahul should get to work then.
 
@Avery lol, it's no problem
I'll get back to it later
 
@rahul go to work è.é
 
7:52 AM
today's reddit logo is great.
 

« first day (2516 days earlier)      last day (2494 days later) »