« first day (1901 days earlier)      last day (3104 days later) » 
00:00 - 15:0015:00 - 00:00

Bob
3:02 PM
Here's another one.
Atlassian response: "go to X screen and you'll find it"
Comment response: "I went to X screen and it's not there, see screenshot"
*crickets*
 
@JourneymanGeek ...yes?
 
how can I test a driver for another OS on the same network card
without having to install the other OS
 
Bob
You can't.
 
virtualbox uses virtual drivers so to the guest its a different NIC to the host
I need XP drivers for intel N 7260
I could somehow virtualize the host but disable vbox networking
 
Bob
3:12 PM
If you want to install it in a VM (which still counts as installing the OS, btw), you might be able to do some VT-d passthrough.
But that's iffy at best.
 
and then install Xp on the host while keeping the now phantom NIC
 
@RecycleBin You don't want to run XP
 
yea I do :P
 
No you don't ;p
Not on modern hardware
 
I do cus win 7 is attempting to delete my stuff
and is just retarded full stop
 
3:14 PM
HAMMERTIME!
 
Is it also trying to kill you in your sleep?
 
idk, i was asleep
 
(OSes don't delete stuff unless you tell it to)
 
oh what if theres a NTFS problem
of course they can delete stuff, if theres a filesystem fault
and guess what! chkdisk said there was!
 
Bob
...that wouldn't be deletion (yay semantics). also, a bug at such a low level would probably not manifest on the individual file level.
not to mention it would likely equally affect XP
@RecycleBin you see, there's a big difference between a bug in the FS driver and FS corruption caused by, say, unclean shutdown
 
3:16 PM
XP has never done this ive had it for over a decade, windows 7 does and ive had it for what? 9 months
 
Bob
for some reason, computers don't particularly like being interrupted while in the middle of something
 
theres loads of reason I dont want win7
 
Bob
shrug that's your prerogative. just don't expect support from, well, anyone
 
why not
 
Bob
there's probably some secret forum of XP users somewhere. that somewhere is not here
@RecycleBin because (a) no one has an obligation to support XP, and (b) most people don't want to support XP.
 
3:18 PM
nobody has any obligation to support any OS
 
Bob
Microsoft has the obligation to support their OSes until the stated EOL date
 
@Bob: I'm aware of such a thing for windows 9x like that
@RecycleBin: so how do we put this
 
this isnt microsoft
 
You ask us about putting windows xp on a modern system
everyone's
!!no
 
and they dont have to support anything technically
 
You can't really expect us to say yes just cause you insist.
 
Bob
@RecycleBin which means we have no obligation to support XP. And we fall under point (b) above.
If you're lucky, you might find someone willing to answer a quick question or two.
But don't expect anyone to put any time into figuring out XP.
 
im asking if it is possible not if its a good or bad idea
I aint expecting help either
 
I doubt its possible
just so many changes in HW since then
and driver models
 
Bob
^ there's your answer
 
3:21 PM
I'm asking and I dont know if ill get help but hell I gotta ask
 
Bob
also of note: XP was released before AHCI existed
 
ok.
 
Bob
that means there's actually no support for SATA baked into the XP core
 
yes
 
Bob
3:22 PM
and now NVMe is being introduced
 
SATA is fine as I have a driver for that
 
Bob
it'll be a good while before it's standard on consumer systems, but you can fully expect no one to bother with an XP driver
 
I got the OS installed but no network
 
Bob
I being up AHCI just to point out how old it is
heck even Win7 doesn't support, say, UASP
 
gee. If only you were running a modern enough OS that it had drivers..
 
3:24 PM
doesnt 7 support it
sata that is
 
what about vista
 
Bob
anyway, if you really really want to do this: get a USB-based NIC
 
@@JourneymanGeek I really have managed to ward him off ;). It took another hour to get him to tell me he had all his files to robocopy in a subdirectory and needed to add /s to the command. Sigh.
 
Bob
it'll probably be easier
 
3:25 PM
I could try vista again
 
Bob
...assuming XP has drivers for that
which isn't even guaranteed
 
the USB NIC is a smart idea if you want to do passthrough
@Bob if its usb 2.0, plausible
 
Bob
@JourneymanGeek 2002, who knows
I don't think they were particularly common back then
 
Bob
though I guess 100Mbit would've been a decent upgrade over on-board 10Mbit?
 
3:26 PM
I had one on a 98 machine
 
I could get s USB NIC I guess
 
before ethernet cards were common
was an unreliable thing, and we ended up replacing it with a proper NIC after 2 died
 
Bob
XP also doesn't run particularly well on multicore systems
 
but incase, does vista support sata unlike XP
 
Bob
It wasn't until 7 that they finally got rid of the giant kernel dispatcher lock
 
3:27 PM
s/multicore systems/anything.
 
Bob
@RecycleBin probably. I can't remember now.
IIRC AHCI was 2004 and Vista was 2007.
 
(I've got 10 running well on a vista era machine)
 
Bob
also, IIRC, XP doesn't properly schedule for Hyper-Threading
 
google says it can apparently
 
Bob
@RecycleBin It "works" with HT but it wasn't designed with multicore HT systems in mind.
A smart scheduler will avoid putting two tasks on the same physical core.
e.g. if cores 0 and 1 were two HT virtual cores on one physical core, 2 and 3 on another physical core.
 
3:30 PM
@RecycleBin: I've had XP, 7 and 10 on the same hardware.
 
@JourneymanGeek Please can you now nuke all the comments on superuser.com/a/989217/337631 ?
 
Bob
Given two tasks, a smart scheduler would distribute them across 0 and 2.
XP... I don't know. It might well put them on 0 and 1
 
of the three 10's the most responsive
 
Bob
Leaving you with contention on one physical core and another core completely unused.
 
I could either get an XP compatable NIC or use a USB nic or use VISTA
 
Bob
3:30 PM
@JourneymanGeek Depends entirely on the hardware.
 
@Bob: the R61 in question
 
Bob
I think the one advantage XP has over 7 and 10, apart from lower memory and instruction set requirements, is the lower graphics requirements.
 
its really quite usable
 
I'm back
 
Bob
XP will happily run with an old onboard GPU (which were slower than software render on a modern CPU)
 
3:31 PM
honestly I have an XP installed on a 288mb VDI and it is fast!
 
Bob
7... not so much
8... dead.
 
but thats not real hardware
 
@Bob: x3100 seems fine ;p
@RecycleBin: probably as a single core.
 
@Bob on any of the Vista+ OSes (except, I think, 10, maybe?), you can kill DWM (or set the group policy flag, or a number of other options) to kick it back to basically the same rendering mode as XP.
 
Thanks.
 
3:32 PM
or just don't install a WDDM-compatible graphics driver and it'll be forced not to use DWM
 
@allquixotic: which is supposedly slowr.
@DavidPostill: nuking all the comments is easy ;p
 
Bob
> The Windows CPU scheduler as of Windows XP is supposed to be able to distinguish hyperthreaded (virtual) cores from real cores. You might imagine then that in this perfect world it handles them 'just right' and it is not an issue. You would be wrong.
heh.
 
@JourneymanGeek I know, that's why I asked you ;)
 
Bob
> This issue occurs because Windows XP does not have the functionality to detect hyperthreading processors and multicore processors. This functionality was introduced in Windows Vista by using the Win32_ComputerSystem class and the Win32_Processor class in Windows Management Instrumentation (WMI).
...yea, XP support for multicore is patchy at best.
 
3:34 PM
@Bob: now I'm wondering if I can grab a trial download of windows 10 and see how badly it barfs on the R60 ;p
 
Bob
Yup, confirmed:
> With the Dual-Core Hyperthreaded CPU coming, it now seems that the current
Windows XP threads scheduler have a big problem dealing with 2 physicals
cores, both supporting 2 logicals units. In some case, it handles logical
and physical CPUs in the same way, causing troubles if, for example, two
threads are set on two logical CPUs located on the same physical CPU.

On a standard 2/4 threads software, I noticed a performance boost if I set
the CPU affinity within the software than if I let windows XP dealing with
XP is stupid in the face of multicore + HT.
To the point where you're probably better turning off HT.
 
is there some technology that allows several CPUs to be seen as one, ki0dna like RAID but with CPUs
 
R60 would have the same IGP as the X60, and the X60 had the 945GM, which I think Intel eventually released a DWM-compatible WDDM 1.0 driver for after a lot of community whining
 
as an unrelated question
 
even though it doesn't support certain hardware-accel functions that'd be needed, they do it in software
 
Bob
3:35 PM
@RecycleBin Nope.
That's not how CPUs work.
 
@allquixotic: its the shittiest system I currently have running.
 
'running'
 
Bob
@RecycleBin There is a very clear distinction between sequential and parallel processing.
 
it has no real OS at the moment
 
3:36 PM
@JourneymanGeek I can only nuke mine not the others :/
 
Bob
Software that isn't explicitly designed for parallel processing cannot be forced to do so.
You'd introduce all kinds of funny corruption.
 
I just found this
150
Q: Flack Overstow - Generate spam from Stack Exchange posts

Mark Rushakoff Online demo courtesy of Nathan Osman: http://quickmediasolutions.com/flack Note: this tool now works with all Stack Exchange sites. Just paste the URL of your profile and the app will take care of the rest. Screenshot / Code Snippet $ python flack.py 126042 Before 'if' s...

 
@DavidPostill: oh, I mean, its 2 clicks to nuke all the comments, or move them to chat
 
Bob
Basically, with parallel processing you have to be very careful about what memory is shared and how you pass data between threads.
 
3:36 PM
You can virtualize X number of CPU cores into "even more" cores using VMware, though, which makes the guest think you have more cores than you actually do, and just interleaves the separate "cores" on the host
@DavidPostill we've known about that for a while, but yeah, it's lovely
 
@allquixotic How did you mange to reply so quickly? ;)
 
Bob
@allquixotic Ya. You can emulate multiple parallel threads on a single core (basically how all multitasking worked before real multicore), but you can't emulate a faster single thread on multiple cores.
At best your single thread will run at the same speed as on a single core.
 
I remember in ~2002-2003 a professor said that until dual core machines are common for consumers, we can't necessarily reliably test our multithreaded software on a single-core box, so we could use the Sun server cluster with like 8 CPUs to test it
 
Bob
More realistically, if you force it to swap between multiple physical cores, you'll slow it down due to cache misses and register read/writing.
 
oh virtual cores
 
3:38 PM
I was one of the only people who had a dual-core at that time
it was a Pentium D, IIRC
 
Bob
@allquixotic You can't reliably test multithreaded software, period.
 
@Bob Well, you can, but it'll break due to all the data dependencies, etc. :P
 
Bob
Testing on a single core system will be less likely to run into race conditions, true.
But even on a proper multicore system you might never encounter a race condition in testing.
 
@Bob one of our professors worked with Sun on techniques for evaluating the safety of multithreaded code, and on tightening up some flaws in Java's memory model
was neat to take a class with him
 
Bob
@allquixotic What was his name?
 
3:40 PM
William Pugh
the inventor of the Skip List, a novel data structure
 
Bob
Ah, don't recognise
@allquixotic Hm. Must've forgotten his name.
Went over that in an advanced algos class.
 
I had him explain after class (in about 15 minutes, so I didn't get the full run-down, but got the concept) the implementation of the ConcurrentSkipListMap and ConcurrentSkipListSet in the JRE
he helped Sun design the algorithm for those
 
There's actually an Asus rep on xda
 
Bob
@allquixotic I know there's formal techniques for verifying multithreaded code
went through a few of them in a concurrency class
I could happily go my whole life never having to see a transition diagram again
 
it's pretty hard to prove the run time of a multithreaded system, though, unless you have a deterministic scheduler with fixed deadlines, and probably a bunch of other stuff -- basically a RTOS
imagine writing an ECU with multiple threads or processes and having to prove it could complete certain operations within a certain time period... that's scary
 
Bob
3:46 PM
Promela was kinda fun though.
@allquixotic Hello, lock-free code :P
Fair sharing, and all that fun stuff too.
I still haven't quite figured out the difference between strong and weak fairness.
 
I really want to learn Rust. I know that has nothing to do with thread safety (maybe it does, based on the way Rust is designed....) but it's looking like it could be a replacement for C that's plenty fast for system software, considerably faster than stuff like Ruby and Java, and much harder to create bugs with than most low-level languages like C/C++
fast, programmer-friendly, and a compiler that loves to beat you over the head with a lead pipe when you do something stupid that'd otherwise crash your program... I like it.
 
Bob
@allquixotic How's the standard library?
I'd only consider it programmer-friendly if that's fairly complete.
I hate reinventing the wheel for what should be fairly basic data structures.
Hello, C.
 
@Bob growing, but relatively decent so far... not as extensive as .NET or Java, but they have great bindings to a lot of C/C++ open source libs to accomplish most tasks, and the bindings are designed in such a way that you can use them "safely" from Rust without running into the same problems as normal C code.
(the bindings have detailed knowledge about who owns pointers, object lifecycle, etc. - to the extent that the binding author can find this information and "document" it for the Rust compiler to know about)
I'm still naturally drawn to write software in Java or C# whenever possible because so many layers of usable, high-level libraries are freely available (either built-in or third-party) for me to take and pile on as building blocks, and no other runtime matches that... second best after .NET/Java would be Ruby, or Python, but beyond that? no one else really compares.
 
4:43 PM
Python is almost as good. Loads of modules to choose from.
 
@DragonLord I think the big difference is that Python, Ruby, and (it's looking like) Rust, in many cases, rather than implementing stuff "natively" themselves, they just link to an existing FOSS C/C++ library and create bindings. Not true of everything, but big pieces of work are avoided, while also unfortunately inheriting all the bugs of the software they include.
 
Bob
@DragonLord But the IDEs are also relatively crap :P
 
Java and C# do a lot of stuff written in the "pure" language of the runtime, rather than going through an FFI / PInvoke / etc
so much so that in certain enterprise contexts (e.g. SOAP with WSS) you actually want to use the pure Java code available for certain things because there's really no other implementation that equals it in quality
which sucks if you aren't using Java
 
Not much attention is given to IDEs in Python, you're right.
PyCharm seems to be the best (commercial) solution.
 
Bob
I find Python decent for small projects, but as the application grows developing in Python gets more painful.
 
4:48 PM
agree with Bob ^
 
NetBeans has limited community support for Python.
 
Ruby isn't much better for large projects, either, but I've done some non-trivial stuff in Ruby and it was alright... not great, just alright
the most sophisticated and complex software I've written to date has been in Java, C# or C++
 
Bob
@allquixotic I've done some heavy stuff in JS. Pre-ES6 a lot of the structure comes from whichever framework you end up using, so could be anywhere from "acceptable" to "kill me now".
With JS you tend to get pretty good debuggers and decent IDEs, but not on the level of C#/Java.
 
JS is better supported. Browsers come with some pretty sophisticated dev tools these days. NetBeans does a good job handling JS code.
 
Bob
I don't think I've seen any dynamically-typed language with a great IDE, actually.
@DragonLord NetBeans can handle JS in a single file and with the most basic structure.
 
4:52 PM
JS is duck-typed and this is pretty tricky to correctly handle.
 
Bob
Once you branch out into multiple files and follow whichever OO-emulation provided by your framework, the IDE is lost.
 
I believe NetBeans supports jQuery, but I've never worked with it.
 
Bob
ES6 might help, but of course we won't be seeing that in common use for years.
@DragonLord jQuery is a helper library, not a framework.
When I say framework, I mean something like Ext.JS, React, AngularJS, etc.
 
Yeah. I'm just a relative neophyte at programming without any very little real-world code experience just yet.
Five years of university education doesn't really prepare you for a programming job, sadly.
 
Bob
@allquixotic Didn't you end up doing some COM stuff in Ruby?
 
4:55 PM
@Bob Yeah. And in VBScript.
 
You get to learn how to use the key features of one or two major languages (C++ in my case), but you never get to work in a major project with 10k+ lines of code.
We went up to basic STL and templates, but that's pretty much it.
 
Don't even remind me about gfjmrfkfifkmkfffrlmmgmhmnlulkmfmgkgj (my favorite "cat typing" string that's an actual legitimate string passed to a piece of code in VBScript)
that's the actual string, I just looked it up
 
Bob
@allquixotic You sure? That message is literally the only result on Google.
 
We never learned about template specializations, lambda functions, RAII, or anything like that, even though knowing this is necessary to work on larger real-world programs which use these features.
 
Bob
Also, it's scary how quickly chat is indexed.
Oh wait, not that message. A previous one.
Still a recent one. Could've sworn you said it dozens of times (some exaggeration in this message).
 
4:59 PM
Stack Exchange is crawled and indexed very, very frequently.
It's a top site and gets very high priority from Googlebot.
 
Bob
@DragonLord University is there to teach concepts.
Sometimes, it's just not worth teaching language-specific features.
 
We did spend time on data structures (linked lists, binary trees, hash tables, etc.).
I had to hand-code a binary tree implementation at one point.
 
Bob
Aug 1 '13 at 14:55, by somequixotic
gfjmrfkfifkmkfffrlmmgmhmnlulkmfmqkqj
 
Oh. q instead of g for the last 2
 
Bob
Weird. I'm reasonably sure the two are identical. Yet that message does not appear on Google.
 
5:02 PM
qkqj vs. gkgj
 
Bob
o.O
how did I miss that
 
I typoed it
 
Bob
You TYPED it?!
 
yes, because it's only on my customer box
this is my company box
 
Bob
@DragonLord Not more than that? o.O
 
5:04 PM
hum de dum
sl;ow slow scripts.
boss says 'run as many as possible'
yeah fine I can do that.
So far one has completed :/
 
5:52 PM
XP supports HT but not multiprocessing
 
Bob
6:06 PM
@RecycleBin It technically supports both but treats HT as if it were really independent cores.
That's fine on a single physical core with HT.
That's not fine on multiple physical cores with HT, as I've already explained above.
 
yeah
 
Bob
Thing is, when XP was released I don't think dual-core with HT even existed.
Now? that's about the minimum you get. You also have quad-core with HT being fairly common, and it goes up to hexa-core and maybe octa-core (?).
XP won't cope particularly well with that.
At a minimum you'd want to turn off HT via the firmware settings.
Even then who knows how well XP handles the SMP. Probably decent-ish considering NT and 2000 were designed as server/workstation OSes, but also probably not as well as a more modern OS.
 
is it the virtualization setting in the bios for HT
right now its enabled to get more RAM for virtuabox VMs but it was probably disabled when I had XP on the second SSD
 
Bob
@RecycleBin No, virtualisation would be VT-x.
Actually, that's a good point... does modern firmware even have HT options anymore? o.O
Probably yes.
 
oh, theres no setting to disable it
 
Bob
6:17 PM
Huh. TIL.
Maybe only on enthusiast motherboards.
 
my bios is not the best
it is fine to hotswap internal laptop AHCI sata drives right
 
Bob
uh
only if the firmware has hotswap enabled
with a laptop, you never know
 
how do I check that
 
Bob
I mean, it probably won't physically damage anything, but there's also a good chance it simply won't work.
@RecycleBin You don't.
 
I have not pulled anything out yet so
 
Bob
6:19 PM
I don't think I've ever seen the option on laptop firmware settings.
I've seen it on some desktop motherboards.
Laptops tend not to expose the more advanced settings, so you're stuck with whatever the default is.
So whether the firmware has it permanently enabled or disabled is the question.
I'd guess disabled. Not many people needing hotswap with laptops.
 
 
3 hours later…
9:08 PM
-1
Q: First ever Contribution deleted?

ChrisFirst time post and provided what I thought was a very much valid and relevant answer to a question. I then took the time to register and validated my email address in the hope that I could become a valued member of the community. With 20 years experience in computing and building PC's I thought ...

Ouch.
 
Are ruby scripts on topic for SU?
 
sure
 
@DavidPostill Quality Sound Engineering!
 
seems to be more on topic on SO. There are 145,503 questions. SU has 326 questions tagged [ruby]
@MichaelFrank Quality Sound Engineering???
 
9:21 PM
THX Ltd. is an American high-fidelity audio/visual reproduction standard for movie theaters, screening rooms, home theaters, computer speakers, gaming consoles, and car audio systems. They have certified video games as well. The current THX was created in 2002 when it spun off from Lucasfilm Ltd. THX was developed by Tomlinson Holman at George Lucas' company, Lucasfilm, in 1983 to ensure that the soundtrack for the third Star Wars film, Return of the Jedi, would be accurately reproduced in the best venues. THX was named after Holman, with the "X" standing for "crossover" as well as in homage to...
There we go...
 
@MichaelFrank Ah. I was using Thx as short for Thanks ;)
 
@DavidPostill I know :P
 
 
2 hours later…
11:16 PM
@Bob: Where did you order your pistons from?
 
Bob
@JourneymanGeek ibuygou
@JourneymanGeek do they sell them directly to sg?
 
@Bob: answering a HR question
got a link?
 
Bob
banggood and geekbuying are also relatively trustworthy iirc but ibuygou specifically is listed as trusted elsewhere
@JourneymanGeek O_O
@JourneymanGeek which one? v2, v3?
v2.1 I originally got => ibuygou.com/…
 
lol
not looking at the v3 yet
 
Bob
I'm so confused now
HR question
O_O
 
11:24 PM
0
A: Cheap yet good quality headphones

Journeyman GeekThe xiaomi piston 2.0/2.1 (there's a 3.0 model out at the moment) out but I've never used it before. I know folks have ordered a pair into australia, and ordered it off ibuygou. The cliff's notes version Price: 20-30 usd depending. Got mine for 20 singapore dollars. Good sound quality - Good...

0_0
 
Bob
@JourneymanGeek Ohhhhhhh hardware rec
not human resources
 
ahh, its a plastic model
 
Bob
@JourneymanGeek yea, "youth edition" and "paperback" are the cheapo versions
 
@Bob I thought he meant Human Resources too
Do we have a hardware recs site too now?
 
Bob
these are noise isolating, not cancelling
@CanadianLuke I knew about the rec site, but I thought HR at some place he applied for asked :S
 
11:28 PM
yup
@Bob: No responses yet
 
Bob
@JourneymanGeek :(
 
@JourneymanGeek No response at all is common for job applications. I would say I only get a real reply 10% of the time. It's very depressing :/ I've even had email confirmations that say something like we will only contact you if we want to interview. And you are just left waiting ...
I've sent around 40 in the 6 months I've been back in the UK. No responses at all. It was better when I was in Holland. At least I got some interviews...
 
yup
tech firms seem to at least send you a confirmation
 
11:48 PM
user image
7
:D
 
brb
tomorrow
 
I have just been restarted! This happens daily automatically, or when my owner restarts me. Ready for commands.
 
00:00 - 15:0015:00 - 00:00

« first day (1901 days earlier)      last day (3104 days later) »