« first day (1888 days earlier)      last day (3124 days later) » 

1:12 AM
Well, I'm experimenting with running graphical apps on my VPS, on a local X server over SSH.
Xeyes works.
Starting the Enlightenment desktop.
This slow DSL is not helping here. The desktop is drawing very slowly on the X server window.
DAMMIT! Enlightenment crashed!
> This is very bad. Enlightment SEGV'd.
 
Bob
X forwarding is poor at the best of times
Use VNC or RDP.
 
1:29 AM
I'm using the VcXsrv server.
I think I crashed E17 by resizing the X server window.
The desktop seems to be loading this time.
Enabling compression on the connection.
 
1:52 AM
The idea is that I'd have a desktop in the cloud that I can stream to any device with an X server.
Problem is that it takes a lot of bandwidth to do this.
You know what? I'm going to remove E17 and use something lighter.
Just use twm.
That way, I don't have to bog down the server with an unnecessarily heavy desktop.
 
2:11 AM
@DragonLord: I've had excellent luck with nomachine
 
For some reason, I just want to get a pet mouse, even though my mom doesn't want me to (rodents are low-maintenance, but they still need me to carve out time from my schedule).
They're adorable little critters, but not exactly long-lived (typical lifespan is 2-4 years)
 
@DragonLord Huh.. About the same as Guinea Pigs then.
Which I have four of. Cute little things!
 
2:48 AM
Removed E17 and all of the packages that were installed at the same time. Installed twm.
 
@MichaelFrank: mind if I steal a link in one of your comments for an answer?
 
@JourneymanGeek Only if you tell me which one!
 
The magnets are not going to steal your data. See this article here: kjmagnetics.com/blog.asp?p=hard-drive-destructionMichael Frank 1 hour ago
 
Oh yea, go for it.
 
3:05 AM
done ;p
 
3:19 AM
I love how the first two hits for what I was looking for was from stackexchange google.com.sg/search?q=cryptonomicon+door+electromagnet
 
3:43 AM
IceWM runs but with very high lag, probably due to the large number of small packets generated by the system.
(My server is in Newark, which helps, but latency is unavoidable.)
 
@Bob Have you noticed LTE speeds changing at all in recent years in Australia? Even though we have way more spectrum and the newer bands are supposed to be faster than the old, I'm getting speeds comparable to 3G during prime entertainment time, now. Making it difficult to continue forward with LTE as primary Internet.
 
@allquixotic, are you familiar with X forwarding?
 
For instance: it's 11:52 PM right now, which is only slightly past prime time, and I can sustain above 1 MB/s downstream. Consistently. LTE is always pretty variable and the speed tends to go up and down in an approximate sinusoidal pattern, but during prime time, it oscillates between ~100 KB/s and 600 KB/s, rather than between 1 MB/s and 2 MB/s.
@DragonLord used it years ago in the computer lab at school to display a Solaris copy of Mathematica on a Linux workstation... why?
 
Bob
@allquixotic my primary service is not lte
 
Trying to run graphical applications on my VPS.
 
3:56 AM
@DragonLord X forwarding over the Internet is very slow due to the chattiness of the protocol and lack of compression; instead, render it into an X2go or Xrfb or Xvnc server (in approximately that order of preference/efficiency, decreasing preference) and connect to the server remotely over X2go or RDP or VNC.
X2go, Xrfb and Xvnc are all basically X servers that render to an off-screen buffer, then accept TCP or UDP connections (sometimes in another process/daemon) and, after authentication, send that buffer over the 'net in a way that's at least somewhat efficient.
X2go is my current preference, but Xrdp is fine if you don't want to install custom client software (just about everything has an RDP client)
and Xvnc, well, if you have to, but VNC is significantly less efficient than even a mediocre version of RDP
 
@allquixotic x2go is nx3. And yes, its awesome (when it works) ;p
 
when you do remote X forwarding, every X client (and there can be dozens of them) will have to do a round-trip to your remote X server each time it wants to do a draw -- now, for clients like GTK with client-side rendering, that means it basically feeds the X server a buffer of pixels to draw on the screen, and that's it; but older applications especially will tend to give X (many) specific draw calls per second.
 
Yeah, way too many draw calls over the network.
 
that's what X2Go/NX/RDP/VNC all do differently than the X11 protocol: they accumulate all the clients' draw calls locally in a buffer, then, once they have a complete frame of data, they send that one frame across the network -- and often for performance/efficiency reasons they'll buffer multiple frames in one transfer, which increases input latency but decreases the bandwidth cost per frame (and allows things like delta encoding)
oh, @Bob did you see that the Surface Book by Microsoft has an Nvidia discrete GPU in a ridiculously small/light chassis?
Iris Pro would've been interesting, but Nvidia is even better
 
 
4:09 AM
not sure how fast it'll be, but my bet's on it being a semi-custom tailored version of Maxwell, basically where they sized and binned the chip so its power consumption and heat dissipation were within the limits of the device's battery and cooling system
 
Bob
Surface Book?!
 
yup.
Microsoft decided to confuse the hell out of everyone by making a new flagship that's even higher-end than the Surface Pro 4
 
so... Slightly dumb question - I'm starting on a C++ course on edx. Its by microsoft, and I'm downloading the current visual studio compiler to work on. Is visual C++ = C++ for my purposes? Anything else I should be getting?
 
Wow. Xfce is fast even over the network.
 
@JourneymanGeek Microsoft's C++ compiler supports certain subsets of the official, platform-independent C and C++ standards, and has its own set of incompatible extensions that are partially or unsupported on other compilers.
 
4:13 AM
@allquixotic: ._.
 
One notable problem with Microsoft's compilers is that they don't support the full breadth of any C standards newer than C89, so certain C99 and newer constructs won't work.
 
Still lots of draw calls, but latency is considerably lower.
 
So "yes for now, wierd shit will happen if you want to do anything newer?"
 
The GNU and LLVM approach to mixing the C and C++ standards has been to support using them more-or-less "together", meaning that if it's valid C99 code that doesn't contradict the C++ standard, they'll let it pass.
But Microsoft stopped supporting new revisions of C, and their support for the latest C++ standards has been spotty.
Basically, out of all the possible C and C++ constructs you can concoct, there are certain ones that are guaranteed to work everywhere; certain ones that work everywhere on the latest patched (bugfixed) versions of all the compilers; certain ones that only work on a couple compilers; and certain ones that work on exactly one compiler.
If you code to the absolute lowest common denominator, your feature set will be more restricted, but your code will be as portable as Java: any system that has a c++ binary will be able to produce working code.
 
;p
(I'm learning, so I'm assuming I will be coding to the lowest common denominator)
 
4:17 AM
If you use advanced features, especially features released in newer versions of the specs, your code will be cleaner, but certain compilers may not support every feature of the newer specs.
@JourneymanGeek Don't assume that! Your book may (intentionally or otherwise) teach you non-standard constructs, or aspects of the standards that don't work on certain prominent compilers or compiler versions.
And when you're writing your own code, you might be tempted to do something that you think is syntactically valid and should also be semantically valid, but your assumption about its semantics might be completely wrong.
 
@allquixotic: There's no book ._.
 
Better yet, some compilers might try to be accommodating about people who do stupid things that violate the standards, to "make it work anyway", but when you go to another compiler, it'll either fail to compile, or silently corrupt your program.
 
context!
they are using the MS compiler.
 
If writing portable C code (and hitting it with a plain C compiler, no ++ anywhere) is like trying to get a citizen of every country on Earth to eat the same meal for dinner, then writing portable C++ code is like trying to get a member of every species of life on Earth to eat the same meal for dinner.
 
;p
(now I'm vaguely confused)
 
4:22 AM
A few generic tips that should help you avoid most problems...
First, turn on all compiler warnings. Every compiler has a ton of warnings it emits when it detects code that might be fine now, but might fail on another compiler or compiler version. Don't ignore warnings! Learn why they're there, and how not to do that thing you did that made it emit the warning.
You can tell the compiler to STFU about its warnings, but that's a bad practice.
 
-Wall
 
lol. That's what the person doing the course said.
 
Very, very useful.
 
And yes, I compile stuff all the time so I do actually tend to read the errors when things break
 
Second, be aware of what C++ and C standards are enabled in the compiler(s) you are targeting.
Third, whenever you endeavor to write a program, before you actually start writing it, decide what C/C++ standard(s) you want to adhere to.
Part of that decision will involve finding out by Google what standards are supported by the compilers you have available to you.
The standards usually have a year (or month and year) after them, and there aren't too many of the standards that are still commonly used.
Big ones to be aware of: C89, C99, C11 (plain C); C++98, C++03, C++07/TR1, C++11, C++14.
 
4:26 AM
In theory, this is supposed to be using c++ 11 I think.
Not really gotten that far ;p
 
I would recommend against trying to use anything from C++14 until you're very sure of what you're doing. As for C++11, support for it is still not 100%.
Think of C++11 (or any of the other standards that go above and beyond the "original" language specification) as being a shopping list. A really, really fucking long one. And rather than spending 10 years checking off every single item on the list before delivering a single one of them to their customers, compiler developers (including Microsoft, Apple, GNU; everyone!) has released "stable" / "production" releases of their compilers with partial support for, say, C++11.
Partial support may mean that they support -- if the list is 100 items long -- two or three items, or 30, or 99. That 100th feature demanded by the spec may not get implemented until the year 2035, if it requires such a fundamental change to the compiler that they can't implement it easily now.
Hell, even the latest release of the most widely used C++ compilers (Microsoft C++, GNU C++, and Clang++) are still working on the last vestiges of support for the 2003 standard of C++.
Meaning, the stuff that was easy to do got done in 2003; the stuff that required some thought was released no later than 2005-2006; the really hard shit was out in the 2010s; and the ridiculously painful features are still not supported.
So, if you're reading a specification document, say, the C++11 spec, and you see some feature in it, and your compiler's author boldly claims "C++11 support!" -- don't assume that (1) the feature is there, or (2) that it's working properly.
(Yes, even if your compiler version is supposedly "stable".)
That said, the "older" the feature, the more likely it is to work.
New just means compiler authors have had less time to work on it, so there's inherently higher risk there that it won't be available, or won't work right.
My last tip: in order to broaden your awareness about general issues in the C/C++ landscape, and to become familiar with what will and will not work across compilers/platforms, you should frequently attempt to recompile your code on another platform and compiler.
Even if you do something as simple as run an Ubuntu VM, and use g++ to compile your code every once in a while (after successfully getting it to work on Microsoft C++), you'll learn a lot about portability issues.
A lot of portability issues go beyond the language spec, of course. Things like system calls. The core C library. The filesystem. Error handling. Exceptions. Debugging. Processes. Network sockets. I/O. File descriptors. GUIs. A lot of shit is different between operating systems. The differences are most obvious between Windows and non-Windows OSes, but even relative close cousins like OS X and BSD have many differences.
 
5:04 AM
right, one module done
@allquixotic: heh. looking at the course, it looks like something you can finish in a weekend. I'll keep all that in mind tho
 
5:24 AM
@JourneymanGeek well it all depends on how much you want to learn! a lesson that's modest in scope can be learned in a short time, but the breadth and depth of real experience won't be there.
 
naturally
But my goal is "getting comfortable with programming and making somewhat better use of time" ;p
I want to learn more eventually but getting comfortable with coding is a big thing.
 
Bob
6:00 AM
@JourneymanGeek IMO C and C++ are not the best ways toget comfortable with programming...
Yes, they do teach more about the internals/fundamentals than most other langues (along with copious amounts of reinventing the wheel).
But they're not as good at teaching the higher-level concepts.
 
6:28 AM
"reinvinting the wheel" is fundamental to learning. For example, what do you think sorting algorithms are? They've all been done before, you're just "reinventing" them in your algorithms or programming class.
 
6:42 AM
That's it for my adventures in X forwarding. All the desktop environments and window managers (except for twm) have been removed from the server.
 
@DragonLord heh... not that the package manager wouldn't have done a good job of removing all remnants when you uninstalled packages, but I love containers for the express purpose of messing around like that without risking messing stuff up
 
Yeah. Removed some local config files.
I also had to remove some .rpmsave files.
 
you know what'd be cool, is if hosts like Linode, etc. would advertise that, within your (hypervised) VM, you can use the cgroups, etc. features of the Linux kernel to create containers -- basically, it'd let you have the security/isolation/sandboxing benefits of "VMs", but without having to suffer the onerous latency/overhead of nested hypervisors.
a container inside a hypervised guest is no slower than the hypervised guest would ordinarily be.
 
6:59 AM
I did keep track of all the dependencies zypper installed and took care to remove all of them when I was finished.
The remnants left are probably harmless configuration files; these shouldn't take up a lot of disk space.
It's more likely I'll spin up a separate instance to run Linux desktop apps.
There were quite a few packages already installed that allowed X apps to connect to a remote X server. (The X.Org Server proper wasn't installed.)
The X protocol is just way too inefficient for direct use over the Internet—way too many draw calls.
UI elements in the XEmacs interface drew one-by-one as the app loaded.
The Xfce desktop, which uses GTK+, was considerably better behaved than the "lightweight" IceWM, but still suffered from high communication overhead.
Thunar suffered from a lot of lag but remained marginally usable.
The issue isn't really bandwidth more than it is the system transferring close to a hundred packets a second (according to Wireshark).
More bandwidth can help, but not by a substantial degree.
Curiously, E17 seemed to be limited by bandwidth.
I may try again when I'm on campus tomorrow, when I have access to a faster Internet connection.
2
A: Is there anything that can be done to make X forwarding over LAN less painful?

DeltikUse Xfce. In conjunction with the Best SSH options for X11 forwarding (provided by this Super User answer), I was able to achieve surprisingly good performance (and bearable!), even from Mozilla Firefox. Test Conditions Two host machines were used for comparison: an Intel Core i7 720QM @ 1...

 
 
2 hours later…
9:45 AM
interesting..
 
is it?
 
Bob
9:59 AM
@allquixotic Holy shit. Just looked at the Surface Book, and... wow.
Convertible done right? :P
Pricey, though.
And no details on the specific GPU.
And kinda disappointingly low warranty period.
If I hadn't gotten a new laptop just a few months ago, this would've been top of the list... similar price too...
Interesting (and good) that they don't offer any with <8GB RAM
 
that hinge is awesome.
heh, temp agency that called me yesterday called me saying they had an opening "moving boxes for microfilming"
!!no
 
(besides, since they get paid if they match someone, might as well get them to work for me by narrowing down what they call me about.)
 
Bob
@JourneymanGeek I wonder how well it'll hold up in practice
Also, is there an extended battery in the keyboard? (didn't check)
@JourneymanGeek o.O
 
Yeah
Lady heard the confusion in my voice.... I was all microfilms?
 
Bob
10:07 AM
O_O the dGPU is in the keyboard?!
 
@Bob: yup
 
Bob
O_O
 
makes sense really
its not as much a keyboard as a docking bay
 
Bob
@JourneymanGeek Yea, but the whole concept of connecting/disconnecting a dGPU on the fly still feels iffy to me.
 
IIRC asus is making one with a bigass (pair?) video card and water cooling on a dock so...
driver support's probably there
 
Bob
10:39 AM
@JourneymanGeek :\
Though I get the feeling the Asus one doesn't really need hot-swapping as much as this does.
If you're connecting it to a giant static dock, restarting isn't as painful.
 
 
2 hours later…
Bob
1:01 PM
@HackToHell I'm not surprived.
Was actually expecting it at some point.
Now this is impressive O_O
 
The x86 arch helped i suppose
Even then, phones are damn powerful these days
 
Bob
@HackToHell Eh... it runs doesn't mean it runs well :P
Probably on par with a $200 tablet.
 
1:39 PM
you don't need a very good CPU to run Windows 95....
 
He meant win 7
 
x86 + KVM I guess
tho he probably rolled his own kernel
 
1:59 PM
yes, it's on a custom kernel with kvm enabled
@Bob it's 100$ more :/
 
2:14 PM
Looks pretty darn cool
 
Bob
@HackToHell Ya, apparently they paused at 0:55 and talked about it as a laptop before springing the tablet surprise :P
 
Oh, lemme find that
And it packs a pretty good GPU
That seems like the part ;p
 
2:38 PM
what exactly is the GPU in it?
 
Graphics

i5: Intel® HD graphics 520

i5/i7: NVIDIA GeForce Graphics Processor
 
> GeForce Graphics Processor
specifics needed
stupid Microsoft and their playing coy with details
DETAILS MATTER!
if it's a cut-down Fermi GT 520, that's a very different performance level and desirability for gamers than if it's a miniaturized GTX 970M.
 
The word from our Nvidia spokesperson is this: “Microsoft has announced a new laptop named the Microsoft Surface Book. The new laptop includes an NVIDIA GeForce GPU.” Microsoft has teased us in a couple of different ways, referring to the dGPU as an “optional NVIDIA GeForce Graphics Processor with dedicated 1GB GDDR5 high-speed memory”, with “the full power of hardware-accelerated graphics” — but there’s no more detail than that, at least for now. Nvidia does say that the GPU is new, and as yet unannounced, suggesting that it might be a semi-custom chip developed specifically for the Surfac
 
well if they wouldn't make the screen such a stupidly high res, maybe the GPU would be just fine for gaming!
 
Looks like nobody knows yet ...
 
2:47 PM
so both Microsoft and Nvidia are playing coy; that's usually an indicator that it's a pretty damn low-end chip, and they want people to go with brand recognition and then be disappointed when it's slow
 
Microsoft said it would go on sale on 26 October
 
I'm actually pretty hopeful about a possible refresh of the Surface Book coming out next year with Nvidia Pascal -- that architecture should be on a 14nm FinFET process, so we can expect greatly improved performance per watt
Maxwell is STILL stuck on 28nm, so yeah, with such a thin and light laptop, it's amazing they got a dGPU in there at all.... definitely a low-end part
 
Everything, apart from the Nvidia graphics chip, is contained within the screen. Kind of limits it ...
 
I'm one of those people who would love to have a thin and light laptop suitable for gaming when traveling, so that really got my hopes up... now I won't be buying it
maybe next gen tho
 
3:08 PM
louwrentius.com/zfs-on-linux-monitor-cache-hit-ratio.html @Bob an interesting script for ZFS tuning...
In 26 days of uptime, my server has had 104,493,479 ARC hits, 10,376,665 ARC misses, 268,479 L2ARC hits and 10,108,126 L2ARC misses. That means 90.9% of my I/O reads are coming out of RAM. If it doesn't come out of RAM, though, 2.5% of the time (only!) it will hit the SSD cache, and the rest of the time it goes to the HDDs.
 
@allquixotic Didn't they specify it as Fermi or something o0
 
@HackToHell no, I was just throwing that out there
if it's Fermi, I'll throw things
32 mins ago, by DavidPostill
The word from our Nvidia spokesperson is this: “Microsoft has announced a new laptop named the Microsoft Surface Book. The new laptop includes an NVIDIA GeForce GPU.” Microsoft has teased us in a couple of different ways, referring to the dGPU as an “optional NVIDIA GeForce Graphics Processor with dedicated 1GB GDDR5 high-speed memory”, with “the full power of hardware-accelerated graphics” — but there’s no more detail than that, at least for now. Nvidia does say that the GPU is new, and as yet unannounced, suggesting that it might be a semi-custom chip developed specifically for the Surfac
> Maxwell
 
4:01 PM
Got SanDisk Extreme PRO SDHC but do I use it out of the box in my mirrorless or I format it in PC? I actually heard that flash card should be formatted in device it will be used in but also heard this called myth...
 
Well, other than rebooting as a first step, I think second step for some tickets is now, officially, "Have you plugged it in ALL THE WAY?"
 
4:17 PM
@CanadianLuke There's no need to reboot for that man.
`Formatting in the camera does more than formatting in the computer (e.g., using a card reader). The camera adds the directory structure and ancillary files needed by the image processor. These files should be automatically added if insert a computer-formatted card - but it's better to be sure.

A new card should always be formatted in the device in which it is to be used.`
 
4:32 PM
I have just been restarted! This happens daily automatically, or when my owner restarts me. Ready for commands.
 
finally got around to documenting for myself a thorough procedure for restarting the server - it's not too bad, just involves a series of commands
 
4:48 PM
> Last night Larry Wall unveiled the first development release of Perl 6, joking that now a top priority was fixing bugs that could be mistaken for features.
 
5:16 PM
@allquixotic It took them this long.
Looks like it'll be a long road ahead, just as Python has a long migration path from 2 to 3.
Given that Perl 6 is defined by a formal specification and not by a hard-to-understand reference implementation, it should be easier for third parties to come up with their own implementation.
I do not know Perl, though, and now might be a good time to start learning another language.
I could grab the latest version of Strawberry Perl for my laptop; alternatively, I could do it in the cloud using the Perl interpreter supplied with my Linux distribution.
@Boris_yo Formatting in the camera ensures that the data structures are exactly as the camera expects them. This minimizes the potential for errors.
I've been doing this for years with the cards I use for my Pentax DSLRs.
 
6:21 PM
@DragonLord And because the camera has formatted the card itself it usually sets up the cluster size and cluster alignment to what is most efficient for it to use itself, rather than what the computer thinks might be best. Back in the days of PDAs I had an SD card that formatted on the computer could get 2-3MB/s, but formatted on the PDA got 5MB/s...
 
6:37 PM
@CanadianLuke Ah, experience
I also started asking things like "yes, ok, it is plugged in". "Is it plugged into an extension cord and is that extension cord plugged into something which is not itsself?"
 
I thought the step after "Have you plugged it in?" was "Have you pressed the power button?"
 
Why assume that the user knows what the power button is?
(and no, I am not joking)
I have spoken with too many people who pushed the button on their monitors
 
so step 3 is "Have you pressed the correct power button?"
 
steo 4: could you do that again, just to make sure?
 
7:03 PM
Step 0: Check there isn't a power cut ;)
 
7:29 PM
I've always taken memory cards out of my camera and inserted them into my laptop's media reader whenever I need to transfer files to the computer.
This ensures that the card can operate at its highest speed, not getting bottlenecked by the camera—even though my Pentax K-3 II has USB 3.0, it can't transfer files at the 80-90 MB/s the card can do through the media reader.
 
7:46 PM
I have limited space on my MP3 player and most songs I have are 256Kbps, 320Kbps.

I want to know what is the best way to make them 128Kbps to win more space.
 
Continuing the X forwarding experiments. Xfce is running once again.
On the college connection, I'm seeing Wireshark report about 500 packets a second.
I just need a faster Internet connection at home :(
I'm honestly surprised that a desktop can run at least somewhat smoothly over X forwarding.
 
any recommendations for a good miniITX case?
 
8:13 PM
@tereško what use case? what kind of peripherals are you going to have in it? do you anticipate any risk of a long discrete GPU being so long that it runs into the disk bay and thus prevents you from filling it up with disks (and if so, would that even be a problem you care about)?
 
it is supposed to be a mid-range gaming box
also, the environment is kinda dusty, so filters on intakes would be really good (if not mandatory)
 
so it's going to have a dGPU, but what one in particular are you looking at?
 
R9 380x or GTX 970
no particular brand, ATM
the list of considered options currently are: Phanteks EVOLV ITX, BitFenix Prodigy, Corsair Obsidian 250D, Fractal Design Node 304
if 250D didn't cost €100, there would be not contests :(
.. but it's damned expensive
(also looked at Node 804, but it's kinda big)
 
Eur 100 for a case is expensive?
I think 80-100 Eur is about avarage for cases I bought.
Mind you, I still use my sturdy dragon case which came with my dual socket AMD 2200 board
 
8:22 PM
the build has to be in €800-1000 range .. I don't mid picking a more expensive case, but I will need to adjust to have a good build in the specified range
 
Which was about 2007 (aka P4 era)
 
it would be nice to have €400-ish just for the GPU
so ... any options?
 
I'm glad that happened!
Owning a patent on a human gene? Fuck that!
@DragonLord Needs an FPS counter!
@tereško Lifehacker were saying that case if the best for home servers just a couple days ago.
 
@MichaelFrank Doesn't really work that way. The X11 draw calls are sent over the network, so a frame-rate counter is going to produce misleading results at best.
 
@MichaelFrank lifehacker is filled with pointless bullshit
 
8:37 PM
@tereško They are now. I remember discovering them like... 12 or so years ago. Sooo much better then.
 
yeah
maybe not 12 years ago .. more like 6
 
Wikipedia reckons they launched 10 years ago. I know it was when we still have a Win98 computer. It was very early on anyway...
@DragonLord I meant Wikipedia reckons Lifehacker launched 10 years ago.
 
so .. about those miniITX cases .. ideas, suggestions, opinions?
 
9:08 PM
@tereško I was looking for a Mini-ITX case for a friend a few months ago, didn't end up buying the build but it looked promising
oh, I was looking at MicroATX, nevermind
 
I am actually in almost the same situation
only I am looking to build a box for my brother
 
hehe
 
he's currently suffering on a laptop
 
poor boy :(
 
Microsoft just out iPad Pro'd Apple.
On the Surface Pro 4.
 
9:22 PM
:P
 
Bob
@allquixotic cat: /proc/spl/kstat/zfs/arcstats: No such file or directory
...I'm an idiot
was running it on the wrong server
:\
need to install bc too
root@debian:~# ./zfs_stats_efficiency
hits.value  94.85
misses.value  5.14
actual_hits.value  93.63
data_demand_efficiency.value  94.91
data_prefetch_efficiency.value  99.55
@allquixotic How did you get those numbers?
 
10:27 PM
@Bob that file exists for me O_O
@Bob division
hits.value 80.70
misses.value 19.29
actual_hits.value 77.58
data_demand_efficiency.value 94.60
data_prefetch_efficiency.value 2.58
 
10:44 PM
but that's after a fairly recent reboot
 
Bob
1 hour ago, by Bob
was running it on the wrong server
:P
@allquixotic ...huh?
I'm at 84 days uptime now
Division? Of what?
 
11:16 PM
heh, the microsoft ads feel... old school apple.
 
Bob
Full circle.
 
Yup ;p
and blah, I need to look at those when I get my next job
The surfaces finally hit the threshold of OOOH.
 
Bob
ha
It's very tempting.
 
Even if my dream laptop's still the blade.
 
Bob
Entirely unnecessary, but tempting :P
I wonder how well that hinge holds up compared to traditional designs.
@JourneymanGeek Would you go for the SP4, the Book, the mid-grade or high-end? :P
Hm. I just realised. "MacBook". "Surface Book". sigh
 
11:21 PM
;p
Mid range Book, I'd guess
I prefer a proper keyboard and a moderate amount of heft. And that hinge.
One thing that strikes me about the book. Those vents.
ooh, so much crazy shit in the news today...
theverge.com/2015/10/7/9471381/… this is strange, cool, and in about a week, someone will build an open source equivilent of it.
 
11:59 PM
I have just been restarted! This happens daily automatically, or when my owner restarts me. Ready for commands.
 

« first day (1888 days earlier)      last day (3124 days later) »