| --- Log opened Mon Oct 08 00:00:30 2012 |
| 00:01 | <&Derakon> | Heh, I've seen some right abominations from physicists too. |
| 00:01 | < gnolam> | Oh yes. |
| 00:01 | <&Derakon> | The problem you have here is simply that they aren't trained developers, and their interest is not in programming but in getting the data they need for their experiment. |
| 00:02 | <&Derakon> | So actually it's a good sign that he's taking the course; he could just be hacking away without any training whatsoever! |
| 00:07 | <&ToxicFrog> | I am full of hate |
| 00:07 | | Outpost [Outpost@Nightstar-798f08ac.pa.comcast.net] has joined #code |
| 00:07 | < Outpost> | ANBYONE know shared memory? |
| 00:07 | < Outpost> | Because this is annoying me. |
| 00:09 | <@TheWatcher> | Also, I really need to write the perl course for the digital biology programme, just no time ;.; |
| 00:09 | <&Derakon> | Outpost: why don't you describe your problem and then we can see if we can help you? |
| 00:09 | <&Derakon> | Don't ask to ask, just ask. |
| 00:09 | < Outpost> | I figure shared memory is one of those things not a lot of people deal with. |
| 00:10 | < Outpost> | Okay, here's the situation. |
| 00:10 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has joined #code |
| 00:10 | | mode/#code [+o himi] by ChanServ |
| 00:10 | < Outpost> | I have int *sig1get, and int *sig2get.; which are shared in a UNIX system using sys/shm.h |
| 00:11 | < Outpost> | I set up the memory correctly, it can be read between all 8 of my forked processes, and I even have spinlocks implemented for it. |
| 00:11 | < Outpost> | Now doing a statement, printf("sig 1 and sig 2 are %d %d", *sig1get, *sig2get) |
| 00:11 | < Outpost> | I will get the number of SIGUSR1 and SIGUSR2 signals sent that I've been tracking |
| 00:11 | < Outpost> | We'll say it's 5 and 9 respectively. |
| 00:12 | < Outpost> | Now if I do printf("The total signals sent was %d", (*sig1get + *sig2get)) -- OR try to save any of those to a local int. |
| 00:12 | < Outpost> | It reads as 0. |
| 00:12 | < Outpost> | I don't understand why this is, nor how to fix it. |
| 00:12 | <&Derakon> | ...hunh. |
| 00:13 | | * gnolam adds angular drag, finds the result much more stable and true to life. |
| 00:13 | < Outpost> | it is literally the one piece of this entire 9 process program that is stopping me from being able to do a write-up and submit. |
| 00:14 | <&Derakon> | So you can print them directly, but you can't do any other operations with them.- |
| 00:14 | <&Derakon> | s/-// |
| 00:14 | < Outpost> | Basically. |
| 00:14 | < Outpost> | Nor can I save them to a local int and do operations on -that-. |
| 00:14 | <&Derakon> | Right. |
| 00:14 | < Outpost> | I CAN increment them using *sig1get = *sig1get + 1; |
| 00:14 | < Outpost> | But I can't usesing *sig1get++; |
| 00:15 | < Outpost> | You know, just to be more wierd. |
| 00:17 | | * Derakon googles around, doesn't turn up much of use. |
| 00:18 | <&Derakon> | Sorry. |
| 00:18 | <&Derakon> | That is a very strange bug. |
| 00:19 | < Outpost> | Yep, making it all the more fustrating. |
| 00:20 | <@TheWatcher> | Random suggestion: tried using memcpy()? |
| 00:21 | < Outpost> | No I have not. |
| 00:22 | < Outpost> | void * memcpy ( void * destination, const void * source, size_t num ); |
| 00:22 | < Outpost> | ...so.. memcpy(totalsig1, sig1sent, ..?)) |
| 00:22 | < Outpost> | Hrm.. what size? |
| 00:23 | <@TheWatcher> | sizeof(int) |
| 00:23 | < gnolam> | Isn't dereferencing a lower priority than incrementing? |
| 00:24 | < gnolam> | (Re: *sig1get++;) |
| 00:24 | <&Derakon> | I thought * bound more tightly than just about anything. |
| 00:24 | <&Derakon> | But when in doubt, use parens! |
| 00:25 | <@TheWatcher> | gnolam: it is |
| 00:26 | < Outpost> | memcpy doesn't like totalsig1 as an int |
| 00:26 | <@TheWatcher> | &totalsigl |
| 00:26 | <@TheWatcher> | you need to pass a pointer to it to memcpy |
| 00:26 | < Outpost> | That compiled aaand. |
| 00:27 | < Outpost> | ... fixed another number |
| 00:27 | < Outpost> | Gave me 0 |
| 00:28 | | You're now known as TheWatcher[T-2] |
| 00:29 | < Outpost> | I get what I believe is the memory location. |
| 00:29 | < Outpost> | -1074213980 |
| 00:29 | < Outpost> | Or it went mad. |
| 00:30 | <&ToxicFrog> | Motherfuuuuuuuuucker |
| 00:30 | <&ToxicFrog> | \documentclass{baposter} does boxes well. |
| 00:30 | <&ToxicFrog> | Scribus does images. |
| 00:30 | <&ToxicFrog> | Impress does footers and dividers. |
| 00:30 | < Outpost> | Ooh, LaTeX? |
| 00:30 | <&ToxicFrog> | None of these do everything I need well. |
| 00:31 | | You're now known as TheWatcher[zZzZ] |
| 00:31 | <&ToxicFrog> | I may actually end up doing the divider and footer in something else, and then using tikz to load it as the background in baposter. |
| 00:31 | <&ToxicFrog> | Outpost: yes. |
| 00:31 | < Outpost> | I hate LaTeX! |
| 00:32 | < Outpost> | It's like writing a fiction novel, but every chapter you stop to do some quick coding. |
| 00:36 | <&ToxicFrog> | ...this is a problem? |
| 00:37 | <&ToxicFrog> | Seriously, death to word processors, LaTeX supremacy |
| 00:37 | <&ToxicFrog> | If I'm writing a paper, I want to write the content, not spend hours fighting with a terrible interface to get typesetting that doesn't look like complete shit |
| 00:37 | <@Tamber> | It's a problem when you're trying to hide lack of substance by shovelling on the style~ |
| 00:39 | < Reiv> | Why on earth would LaTeX be superior there, TF |
| 00:39 | < Reiv> | Surely that's like skipping on Word by writing straight HTML? |
| 00:39 | <&ToxicFrog> | No. |
| 00:39 | <&ToxicFrog> | LaTeX is not markup. |
| 00:39 | <&ToxicFrog> | It's not concerned with formatting, it's concerning with semantics - What You See Is What You Mean. |
| 00:40 | < Reiv> | So it's about formatting formulas and the like? |
| 00:40 | < Reiv> | (and presumably, code indentation etc)? |
| 00:40 | <&ToxicFrog> | In HTML, you say "this is bold text, 16pt, using Helvetica". In LaTeX you say "this is a section title" and it automatically chooses the actual typesetting parameters based on the document type. |
| 00:40 | < Outpost> | Yeah, it's handy like that. |
| 00:40 | < Reiv> | So... like using <H1> in HTML instead of hardcoding? |
| 00:40 | < Outpost> | And really good at doing any math-stuff. |
| 00:41 | <&ToxicFrog> | So I can say "this is a two-column academic paper for A4 paper size" and then just write. |
| 00:41 | | * Reiv admits he is used to HTML being done with CSS alongside. |
| 00:41 | < Reiv> | How does it handle formatting, though |
| 00:41 | < Reiv> | Given that a great many things require a specific style to formatting. |
| 00:41 | <&ToxicFrog> | Reiv: kiiiind of. Imagine HTML with all of the CSS pre-baked. |
| 00:41 | < Reiv> | So you don't get to fiddle with the CSS after? |
| 00:41 | <&ToxicFrog> | You control formatting, in the main, by choosing the CSS file for the type of document you're writing. |
| 00:42 | <&ToxicFrog> | You can, if you need to. |
| 00:42 | <&ToxicFrog> | IME, most stuff that needs a specific style provides a LaTeX documentclass file implementing that style~ |
| 00:45 | <&ToxicFrog> | Hang on, I'll see about uploading an example... |
| 00:52 | <&ToxicFrog> | Reiv: ok, here's two LaTeX files from a grad course I took, one for the slides for a presentation and one for the final paper |
| 00:52 | <&ToxicFrog> | https://gist.github.com/d34e62c15e6c47b6fa20 |
| 00:53 | <&ToxicFrog> | And here are the PDFs generated from it: http://funkyhorror.ancilla.ca/toxicfrog/detailedtalk.pdf http://funkyhorror.ancilla.ca/toxicfrog/finalpaper.pdf |
| 01:00 | | Vash is now known as Vash[ApplePieMaking] |
| 01:00 | | * ToxicFrog gnaws on Reiv |
| 01:14 | < Reiv> | Sorry, doing battle with SQL Server Management Studio :/ |
| 01:14 | < Reiv> | (Any experts handy?~) |
| 01:16 | <&McMartin> | 14:54 < gnolam> Hmm. I wonder if the values from this gamepad really are that crap or if SDL is messing up somehow. |
| 01:16 | <&McMartin> | There are two gamepad protocols for Windows, which aren't quite compatible |
| 01:16 | <&McMartin> | This is wholly Microsoft's fault and I still don't know what kind of crack they were on for it |
| 01:17 | <&McMartin> | If you have a gamepad that works like an X360 pad, though, SDL will get deeply, deeply confused about the status of the LT and RT buttons. |
| 01:17 | <&McMartin> | This is because it is correctly reflecting how DirectInput (as opposed to XInput) fucks up the reads. |
| 01:17 | < rms> | The fact that they have gamepad protocols seperate from joystick ones is awe inspiring. |
| 01:17 | <&McMartin> | Oh, they don't. |
| 01:17 | <&McMartin> | I just haven't seen an actual joystick in nine years. |
| 01:17 | <&McMartin> | It's still the USB HID game device protocol for both. |
| 01:18 | | Attilla_ [Obsolete@Nightstar-b60121d0.as43234.net] has quit [Ping timeout: 121 seconds] |
| 01:18 | <&McMartin> | But this means that, for instance, analog buttons show up as additional joystick axes. |
| 01:18 | <&McMartin> | So far, so good, really. |
| 01:19 | <&McMartin> | But then they decided "wait! If we do that, then the neutral position WILL NOT BE ZERO *scare chord*" and so, if you're using DirectInput to query a modern gamepad, The LT and RT buttons cancel each other out and it cannot distinguish "neither pressed all the way down" with "both pressed all the way down" |
| 01:19 | <&McMartin> | Or, indeed, pressing one a certain amount vs. releasing the other the same amount. |
| 01:20 | <&McMartin> | I'm pretending here that this was deliberate |
| 01:20 | <&McMartin> | By all accounts, the XInput API is actually hugely easier to use than the DirectInput one, so this might just be total neglect of the older protocol in the newer pad. |
| 01:20 | < rms> | :/ |
| 01:20 | <&McMartin> | Also, I lie, sort of |
| 01:21 | <&McMartin> | I have a Logitech Attack 3 joystick right here, but I haven't taken it out of its box yet |
| 01:21 | < gnolam> | This is indeed an X360 pad. |
| 01:21 | <&McMartin> | Yeah, this is a combination of MS's fault for fucking up DirectInput for X360 pads, and SDL's fault for stubbornly continuing to use DirectInput in its windows drivers anyway. |
| 01:21 | < gnolam> | Luckily, I won't have to worry about LT/RT for this, since I'm really emulating another controller. |
| 01:22 | <&McMartin> | OK. Everything else should work, though. |
| 01:22 | <&McMartin> | At the "should work" level of "well, I've played UQM with it" |
| 01:23 | < gnolam> | It was the centering that was off. I just had to extend the deadzone to ~20%. |
| 01:23 | <&McMartin> | Oh. |
| 01:23 | <&McMartin> | Yeah. That is the fate of all analog axis inputs. |
| 01:23 | < gnolam> | I think I've been spoiled by high-end joysticks.~ |
| 01:24 | <&McMartin> | You have - if you can get it to work with only 20% deadzone that's a pretty decent pad~ |
| 01:24 | < gnolam> | There's, uh, a slight difference in torque there though. |
| 01:24 | | * McMartin is used to 30-50%. |
| 01:27 | <&McMartin> | (Fun fact: I bought that joystick so that I could replay Wing Commander~) |
| 01:27 | < gnolam> | (You can't really use that much force to center a stick that's, what, a centimeter long?) |
| 01:28 | < gnolam> | (Nor can you keep your centering far from the fulcrum) |
| 01:28 | < gnolam> | McMartin: heh. I've been considering hooking up the full flight gear for Wing Commander 3. |
| 02:25 | < Outpost> | Humr, using Threads (not forks) is there a way to signal ALL of a parents threads? |
| 02:25 | < Outpost> | kill(0, SIGUSR1) seems to only send a single signal to the threads, which one picks up and handles; not all of them |
| 02:34 | < rms> | Signals are for processes, not threads |
| 02:35 | < rms> | KILL(2) Linux Programmer's Manual KILL(2) |
| 02:35 | < rms> | NAME |
| 02:35 | < rms> | kill - send signal to a process |
| 02:36 | < Outpost> | Whilst understandable, and good programming... |
| 02:36 | < Outpost> | And project requirements demands I send a SIGUSR1 to all my child threads. |
| 02:36 | < Outpost> | I'm trying to avoid using a for loop to do it; but I may just do that. |
| 02:38 | | RichyB [richardb@Nightstar-86656b6c.cable.virginmedia.com] has joined #code |
| 02:38 | < rms> | pthread_kill(3) Sends a signal to a specified POSIX thread in the same |
| 02:38 | < rms> | process as the caller. |
| 02:38 | < rms> | L2man |
| 02:39 | < Outpost> | Oh no, I did that. |
| 02:39 | < rms> | Clarify please |
| 02:39 | < Outpost> | See, I have three processes in that program, all three need to recieve the SIGUSR1 signal. |
| 02:40 | < Outpost> | Using pthread_kill results in only one of those threads handling the signal. |
| 02:40 | < rms> | killpg(2) Sends a signal to all of the members of a specified |
| 02:40 | < rms> | process group. |
| 02:40 | < rms> | Use a process group then |
| 02:40 | < rms> | If you have three processes |
| 02:40 | < Outpost> | Aha! That might do it! |
| 02:41 | < Outpost> | Here I was futzing with mutexs and PTHREAD_COND_BROADCAST |
| 02:41 | < rms> | Seriously though, l2man; `man 7 signal` (which I got from man pthreads) |
| 02:41 | < Outpost> | In all fairness, this is my first time ever coding on a UNIX system. |
| 02:41 | < Outpost> | And I only have PUTTY. |
| 02:42 | < Outpost> | The very fact that it doesn't have a GUI is sending me completely off kilter. I already miss my KINECT coding. |
| 02:42 | < rms> | Meh |
| 02:42 | < rms> | You can google manpages BTW |
| 02:43 | < Outpost> | I did, but that led me to the less obvious pthread_cond_broadcast solution |
| 02:43 | | * rms has no idea how you'd get that, the SEE ALSO clearly has signal(7) |
| 02:44 | < Outpost> | From here - http://www.yolinux.com/TUTORIALS/LinuxTutorialPosixThreads.html |
| 02:45 | < rms> | That's not a man page :/ |
| 02:45 | < Outpost> | It leads to man pages. Sometimes. |
| 02:45 | < rms> | ... |
| 02:47 | < Outpost> | Hey, if college didn't teach me that UNIX was all man (command) and PICO; and shut off internet access on the UNIX PCs |
| 02:47 | < Outpost> | I might've been more apt to learn it on my free time better instead of going "This sucks all of the things, back in Eclipse." |
| 02:48 | < Outpost> | I'm sure they had a reason for making it as horrifically inconvienent as possible for us. |
| 02:48 | < Outpost> | They didn't even teach us emacs. |
| 02:48 | < Outpost> | Do you know how much of a godsend it was when they actually unlocked emacs after I graduated? |
| 02:49 | < Outpost> | Also killPG doesn't work. |
| 02:49 | < Outpost> | Once the signal is handled it disappears and the rest don't see it. |
| 02:50 | < rms> | Alright, to be fair, I've never actually done anything with process groups |
| 02:51 | < rms> | Looks like you'll need to loop on your processes |
| 02:51 | < Outpost> | That feels dirty to me, but I'd rather that then set up and tow with broadcast. |
| 02:51 | | * rms isn't sure what you meant by that |
| 02:52 | < Outpost> | *toy with broadcast. |
| 02:52 | < rms> | Like... USB broadcast? |
| 02:52 | < rms> | UDP* |
| 02:52 | < Outpost> | like pthread_cond_broadcast |
| 02:52 | < rms> | Ah |
| 02:52 | < rms> | Well that'd just work on threads, not processes. |
| 02:53 | < Outpost> | Yeah, my forking program works perfectly fine. I'm amazed by it. |
| 02:53 | < rms> | Which makes me confused why you have both processes and threads in this. |
| 02:53 | < Outpost> | They're two seperate programs. |
| 02:53 | < rms> | Sounds like you're doing an assignment |
| 02:53 | < Outpost> | That I am sir. |
| 02:54 | < rms> | Ah, you're not really stating which is which, hence my confusion |
| 02:54 | < Outpost> | Yeah, I should be more specific. |
| 02:55 | < Outpost> | I've been pretty fustrated with it, it's Graduate level operating systems, so the teacher assumed a fair amount of pre-existing knowledge, which I had honestly forgotten by not working with the topics in around 6-7 years. |
| 02:56 | < Outpost> | So I've been hammering it out 4-6 a day hours on top of my work schedule to learn forking, shared memory, threading, synchronization, and implementing my own spinlock method. as well atomicity. |
| 02:57 | < Outpost> | I learnt the most I've ever learnt on my own in my entire life. I also want to beer my brain until it stops doing thinky things. |
| 02:59 | < rms> | Heh, which shared memory system are you using? |
| 02:59 | < Outpost> | sys/shm.h in UNIX |
| 02:59 | < rms> | Which UNIX? |
| 02:59 | < Outpost> | That's a damn good question. |
| 02:59 | < rms> | uname -a |
| 03:00 | < Outpost> | Let me get back on it; since it likes to disconnect me. |
| 03:00 | < Outpost> | Linux CIS-Linux2 3.2.0-29-generic #46-Ubuntu SMP Fri Jul 27 17:03:23 UTC 2012 x86_64 x86_64 x86_64 GNU/Linux |
| 03:01 | < Outpost> | Looks like UBUNTU |
| 03:01 | < rms> | If it has screen, I have a screenrc that confuses sshd's auto-disconnect feature |
| 03:01 | < rms> | So that you stay connected |
| 03:01 | < rms> | PROTIP: Ubuntu isn't allcaps |
| 03:01 | < Outpost> | Tempting, but not worth it, since once this project is done I'm running back to Windows. |
| 03:02 | < Outpost> | If I don't make newbie mistakes, how will people know I'm a newbie? |
| 03:02 | < rms> | I'm sorry you value your sanity so little. |
| 03:02 | < Outpost> | To be honest, I always found windows devving ridiculously easy. |
| 03:02 | < Outpost> | More so because I can play Guild Wars 2 in the background. |
| 03:03 | <&McMartin> | If your goal is "Works on my machine", Linux is the easiest, then Windows, then Mac |
| 03:04 | <&McMartin> | If your goal is "works on every place it's likely to end up", Windows and Linux switch places. |
| 03:04 | < rms> | Opposite here, mostly because I started as a windev and found the docs impossible to find until I learned to start bookmarking the few I could. Then MS broke all my bookmarks so I refound them all, then they broke them again and I gave up, moved to Linux and barely looked back since. |
| 03:04 | | RichyB [richardb@Nightstar-86656b6c.cable.virginmedia.com] has quit [[NS] Quit: Leaving] |
| 03:04 | <&McMartin> | Mac used to be #2 there but not since 10.6 -_- |
| 03:04 | < Outpost> | What McMartin said. |
| 03:04 | < Reiv> | McMartin: I notice your ordering kinda puts Mac in last place across the board~ |
| 03:04 | < Outpost> | I've never had an issue picking up and moving my Windows code to another machine. Especially since a lot of places I work at/on are Windows. |
| 03:05 | <&McMartin> | Reiv: Yes, in that it is literally impossible to develop for stuff earlier than 10.7 without doing the desktop equivalent of jailbreaking your phone |
| 03:05 | <&McMartin> | Apple's approach to deprecation is even more psychopathic than the OSS community's. |
| 03:06 | < rms> | Which would be? |
| 03:06 | <&McMartin> | Every Day Is Flag Day, more or less. |
| 03:06 | <&McMartin> | They're bending over backwards to help you out if stuff from *one* release ago still works, or if new code can be made to work one release ago. |
| 03:07 | <&McMartin> | I may be over-bitter about OSS here right now because of my OpenGL investigations |
| 03:07 | <&McMartin> | And the 2.1/3.0 gap |
| 03:07 | <&McMartin> | Which I still, since Windows is a major platform for me, continue to think of as a major gap it's not reasonable to assume people will jump |
| 03:08 | <&McMartin> | (It's the DX9/DX10 gap, on the MS side) |
| 03:08 | <&McMartin> | And yet it is literally impossible to write code that both works in 2.1 and isn't deprecated in 3.x |
| 03:08 | <&McMartin> | There are no points of contact |
| 03:08 | <&McMartin> | Or rather, there are *some* points of contact, but they're not enough to build a complete rendering pipeline of them. |
| 03:09 | < rms> | Would you rather all those calls be removed entirely? |
| 03:09 | <&McMartin> | My understanding is that 3.1 did. |
| 03:09 | < Outpost> | AGH, someone infinite forked the SSH server I'm on! |
| 03:09 | < Outpost> | Then left it |
| 03:09 | < Outpost> | fuck that guy |
| 03:09 | <&McMartin> | (Also, bloodthirsty roar from behind me: "YES! DEATH TO GLVERTEX, DEATH TO CVA, DEATH TO THE FIXED LIGHTING EQUATION") |
| 03:10 | <&McMartin> | (I do admit they have a point) |
| 03:10 | <&McMartin> | (But starting at 3.0 it's really two engines and the one that actually runs on most systems is the one they're trying to erase from history) |
| 03:11 | < rms> | :/ |
| 03:11 | <~Vornicus> | cva? |
| 03:12 | <&McMartin> | Compiled Vertex Arrays. An old and widely-used extension unofficially deprecated in 1.5 when Vertex Buffer Objects replaced them completely. |
| 03:12 | <&McMartin> | They're one of the points of contact. |
| 03:13 | <&McMartin> | It's worth noting that the other big problem with developing for Linux is that the deployment target is much larger and moves much faster |
| 03:13 | <&McMartin> | Ironically, Mono is making Linux easier to deploy to now~ |
| 03:14 | <&McMartin> | Though it looks like there's enough stability in X and OGL themselves at this point that even the C++-based games on HIB run fine just distributed as binaries. |
| 03:14 | < rms> | 350MB of runtime... |
| 03:14 | < Reiv> | Mono being the Microsoft driven library, yes? |
| 03:14 | < rms> | Mono being the CIL-runner |
| 03:14 | <&McMartin> | Mono being Novell's reimplementation of Microsoft's answer to Java. |
| 03:15 | < Reiv> | ... wut |
| 03:15 | < rms> | (C# and other such cesspools) |
| 03:15 | | * Reiv tries to parse that sentence. |
| 03:15 | <&McMartin> | rms: Well, my /usr/lib is 600MB as it is. |
| 03:15 | < Reiv> | Ah-hah. |
| 03:15 | < rms> | Yeah, well I generally don't like downloading 350MB of crap for one program |
| 03:16 | <&McMartin> | Welcome to DLL hell~ |
| 03:16 | < rms> | Unless that 350MB of crap is part of the program itself |
| 03:16 | < rms> | Or the program's size dwarfs it |
| 03:16 | <&ToxicFrog> | Reiv: Mono is the open-source version of the CLR, which is to say, the .NET interpreter and libraries, basically. |
| 03:17 | <&ToxicFrog> | s/version/implementation/ |
| 03:17 | <&McMartin> | s/interpreter/compiler/ |
| 03:17 | < Reiv> | Isn't .NET now depreciated? |
| 03:17 | <&ToxicFrog> | s/compiler/compiler and VM/ |
| 03:17 | <&McMartin> | Fair |
| 03:17 | <&McMartin> | Reiv: Define "deprecated" |
| 03:17 | <&McMartin> | Windows 8 doesn't use its system calls, or would rather you didn't, yes |
| 03:17 | <&McMartin> | But (a) people still *will* because they will continue targeting Windows XP, to say nothing of Windows 7 |
| 03:18 | <&McMartin> | (This is part of why MS has to have a psychopathic attitude towards bug-for-bug backwards compatibility~) |
| 03:18 | < rms> | What's MS replacing .NET with? |
| 03:18 | <&McMartin> | They're not, they're replacing .NET's wrapper around Win32 with the Metro silliness |
| 03:18 | < Reiv> | The bug-for-bug thing is weirdass, but I kinda get it. |
| 03:19 | < rms> | Explain please? |
| 03:19 | <&McMartin> | WinForms is "deprecated" in the same way everything that looks like an ordinary desktop applicaiton is. |
| 03:19 | <&McMartin> | C# and friends are still going to be fully supported languages, though. |
| 03:20 | <&McMartin> | Also, MS's idea of deprecation is to put a little note on the function saying "please consider using this other function instead? It's only here so your existing software doesn't break" |
| 03:20 | < Reiv> | I remain baffled as to how MS expects Metro to actually succeed in its uptake. |
| 03:20 | <&ToxicFrog> | rms: they have deprecated the parts of the .NET libraries that are wrappers around the classic win32api. .NET/CLR as a whole and its languages are still fine. |
| 03:20 | < Reiv> | Hell, $my_company has already declared they won't be bothering, as Win7 is Stable Enough and the cost of retraining and redeveloping everything to Metroness would be impractical. |
| 03:20 | <&McMartin> | Quite so |
| 03:20 | < rms> | Reiv: MS is a giant monster that rolls over sanity crushing the hopes and dreams of all people who value sanity and utility in an OS |
| 03:21 | < rms> | IE: they don't give a shit |
| 03:21 | < Reiv> | rms: Bah, enough with the hubris |
| 03:21 | <&McMartin> | rms: You misspelled "Apple"~ |
| 03:21 | < Reiv> | And they /should/ give a shit |
| 03:21 | < rms> | But they don't |
| 03:21 | < Reiv> | Vista failed because people didn't take it up, and sat on XP. |
| 03:21 | < Reiv> | The same is now happening even moreso with Win7 being the Keeper. |
| 03:21 | <&McMartin> | That said |
| 03:22 | <&McMartin> | Win7 is a keeper because Vista was out there flailing, in part. |
| 03:22 | < Reiv> | I still don't get how they expect it to /succeed/ |
| 03:22 | <&McMartin> | Vista *did* break backcompat in nontrivial ways |
| 03:22 | < rms> | Vista failed becuase of a bunch of shit, notably their advertising being filled with blatant lies. |
| 03:22 | <&McMartin> | Stuff that mattered had to be rewritten ground up to work on Vista |
| 03:22 | <&McMartin> | But once it was, it Just Worked on Win7 |
| 03:22 | <&McMartin> | That means that when Win7 came out, Everything Just Worked. |
| 03:22 | < Reiv> | Because Vista had done its job as Polish Minesweeper? |
| 03:23 | <&McMartin> | Windows 8 is pretending to stay that stuff that *doesn't* matter ought to be written ground up. |
| 03:23 | <&McMartin> | Reiv: Or as "extended developer beta", yes~ |
| 03:23 | <&McMartin> | As it happens, I don't believe Windows 8 when it pretends to say this |
| 03:23 | <&McMartin> | But I'm not sure what their actual plan is here |
| 03:24 | <&McMartin> | Either they expect to see an ecosystem where entirely new software takes over everything |
| 03:24 | <&McMartin> | Or this is actually just them cynically trying to steal Apple's thunder ("Look, we can make desktops look more like tablets too") |
| 03:24 | <&McMartin> | ... or if it's just "well, shit, it all runs there, might as well ship it everywhere" |
| 03:25 | < Reiv> | McMartin: Therein lies my consernation |
| 03:25 | < Reiv> | Because Metro For Tablets /isn't/ Metro For PC. |
| 03:25 | < Reiv> | Which would be the one place I'd have otherwise given them a pass; it'd be the company trying to change direction to be Tablet First, but even then it just /isn't/ |
| 03:25 | < Reiv> | Also, uh, there's a fair bit of Inertia(tm) hanging around |
| 03:26 | < Reiv> | Did you /see/ the Microsoft Office For Tablet?~ |
| 03:26 | <&McMartin> | Heh |
| 03:26 | <&McMartin> | I think it's much more likely that they're trying to unify their development teams (which they have openly done) and would like to still keep PC first |
| 03:26 | <&McMartin> | But that also means they don't get to just kill everything else |
| 03:26 | < Reiv> | "We want you to take your PC-centric moneymaking behemoth and run it on Tablets." "...OK. We'll make the buttons bigger, I guess." |
| 03:27 | < Reiv> | "...hey guys, uh, this is pretty crap to work on!" "Eh." |
| 03:30 | <&McMartin> | But yeah |
| 03:30 | <&McMartin> | I can generally at least understand the decisions made by MS and the OSS hivemind about how stuff is supposed to work. |
| 03:30 | <&McMartin> | But, well, Apple is acting like the GNOME developers =P |
| 03:31 | < Reiv> | ?? |
| 03:31 | <&McMartin> | They live in some parallel univese |
| 03:31 | < rms> | Being all "we're going to have retaded as fuck default settings, make it nigh impossible to change them and replace godly tools with fisher price toys"? |
| 03:31 | <&McMartin> | The difference is that Apple has a userbase that is also *in* that parallel universe, and GNOME's developers are forever alone |
| 03:32 | | * rms has a high opinion of GNOME |
| 03:32 | <&McMartin> | rms: That but, in Apple's case, not having the godly tools in the first place. |
| 03:32 | <&McMartin> | XCode was at best OK |
| 03:33 | <&McMartin> | Cocoa is actually not too shabby as an OS API layer except for the part that it is the only thing in the global software ecosystem that uses its interface language, and no other ones are permitted |
| 03:33 | < Reiv> | ... what is this parallel universe in which they sit? |
| 03:33 | <&McMartin> | For GNOME, nobody's really sure |
| 03:34 | <&McMartin> | For Apple, it's one where hardware is disposable and extremely expensive, and nobody ever upgrades just the software, and every upgrade resets everything from scratch |
| 03:34 | <&McMartin> | Apple is a music-player company that also happens to make phones and computers. |
| 03:34 | <&McMartin> | Computers are now the outlier. |
| 03:35 | < Reiv> | Hence the phone comes with the software, rather than the other way around? |
| 03:36 | <&McMartin> | rms: It sounds like you had a higher opinion of GNOME2 than I did~ |
| 03:37 | < rms> | I'm a CLI guy |
| 03:37 | < rms> | Also a system admin |
| 03:37 | < Reiv> | So what kind of crazy do the GNOME guys get up to? |
| 03:37 | < rms> | GNOME actually caused downtime when I removed it from a server. |
| 03:37 | <&McMartin> | The past three releases have only removed features. |
| 03:37 | < Reiv> | I remember gnome being a pretty good Linux distro, once. |
| 03:38 | < rms> | (By shutting down the network interfaces like a psychotic girlfriend) |
| 03:38 | <&McMartin> | And they're features like "backspace moves you up the filesystem hierarchy while browsing it" |
| 03:38 | <&McMartin> | Nobody uses that! Get rid of it! |
| 03:38 | | * rms hopes that was a type Reiv made |
| 03:38 | < rms> | typo* |
| 03:38 | < Reiv> | rms: Yes. >_> <_< |
| 03:38 | <&Derakon> | I've always hated the "backspace is a navigational key" concept. |
| 03:39 | < Reiv> | Derakon: So Gnome is written for you! |
| 03:39 | < Reiv> | McMartin: Literally only removed? |
| 03:39 | <&McMartin> | Reiv: I'm sure they did stuff under the hood |
| 03:39 | <&McMartin> | But their guiding principle for all of GNOME 3 has been that options just confuse people and so we must get rid of them |
| 03:39 | <&McMartin> | This *also* means that there is no longer a dialog box for defining keybindings for doing stuff |
| 03:40 | < Reiv> | You're kidding. |
| 03:40 | <&McMartin> | Nope |
| 03:40 | <&McMartin> | As I said |
| 03:40 | <&McMartin> | They went mad |
| 03:40 | <&Derakon> | To be fair, excessive choice is a major problem with open-source UIs. |
| 03:40 | <&McMartin> | Yes, but they missed the memo where that doesn't mean that REMOVING OPTIONS IS ALL YOU HAVE TO DO |
| 03:41 | <&Derakon> | Well, yes. |
| 03:41 | <&Derakon> | You also need to make the defaults not suck. |
| 03:41 | <&McMartin> | Also, even if you don't expose stuff in the UI |
| 03:41 | <&McMartin> | Stuff still needs to be configurable |
| 03:42 | <&McMartin> | I've gotten the impression they've just started literally hardcoding shit in |
| 03:42 | <&McMartin> | Which is bad! |
| 03:42 | <&McMartin> | My GUI needs are pretty slim, really |
| 03:42 | <&McMartin> | I find that both LXDE and (as of last April) Unity both meet my needs. |
| 03:43 | <&McMartin> | Cinnamon - the project that took over from GNOME after GNOME officially went maad - is quite popular but also more heavyweight than I need. |
| 03:43 | <&McMartin> | Also also, in passing |
| 03:43 | <&McMartin> | Those of us mostly in the windows world don't get to use "fisher-price" as an insult anymore |
| 03:44 | <&McMartin> | Since one of the things Metro was planning on doing (and succeeded at doing!) was taking the Fisher-Price aspect of the UI away |
| 03:44 | <&McMartin> | It turns out people kinda liked it~ |
| 03:44 | <&Derakon> | Fisher-Price was only really an issue in XP, I thought. |
| 03:44 | <&McMartin> | Well, they were used to it by 7 |
| 03:44 | <&McMartin> | But it's still all rounded edges and shiny plastic bits. |
| 03:44 | < Reiv> | ... 7 has plastic bits? |
| 03:45 | <&McMartin> | My taskbar is full of phong shading |
| 03:45 | <&Derakon> | 7's theme was glassy transparency. |
| 03:45 | < Reiv> | I thought the fisher price bit was "Bright blue, hooray!" |
| 03:45 | <&McMartin> | Nah, 3.1 was that too |
| 03:45 | <&McMartin> | Win8 is I HOPE YOU LIKE RECTANGLES |
| 03:45 | < Reiv> | "We need another colour? OK, how about green?" |
| 03:46 | <&McMartin> | All that green on my Win8 UI is my fault. I like green, so I told it to use it places >_> |
| 03:46 | < Reiv> | McMartin: Be fair, that was the only bit of colour in the whole UI unless you stuffed about with it :P |
| 03:46 | < Reiv> | McMartin: How are you finding Win8 incidentally? |
| 03:46 | <&McMartin> | As much as possible, I'm not~ |
| 03:47 | <&McMartin> | My professional interest in it is solely that our stuff continues to work with it as it does on 7. |
| 03:47 | <&McMartin> | Which it does, because where Vista was 6.0, and 7 was 6.1, 8 is 6.2 |
| 03:47 | < Reiv> | Huh. |
| 03:47 | < Reiv> | So 8 is an update under the hood |
| 03:47 | <&McMartin> | This is supposed to mean that all kernel-level stuff is fully backwards compatible |
| 03:48 | < Reiv> | The real difference being the tabletisation of the UI? |
| 03:48 | <&McMartin> | Yeah, all the changes in 8 are to the UI and wrappers around the kernel. |
| 03:48 | <&McMartin> | Mostly. |
| 03:48 | <&McMartin> | We've been hit by a couple of places where our kernel-level stuff looks a little too much like a rootkit and is getting thwarted. |
| 03:48 | < Reiv> | ... that would be an excellent explanation of why $my_company isn't bothering to update, then. |
| 03:49 | <&McMartin> | So we'll need to defeat its security^W^W^Wupdate our OS interaction layers to be more compatible. |
| 03:50 | < Reiv> | "Same engine, different UI" is just about the worst possible argument to try and sell a corporate who is not immersed in the Tech industry. |
| 03:50 | <&McMartin> | Or who is. |
| 03:50 | <&McMartin> | We have less than zero interest in making our interface Metro-capable, though making it so 'do the one action you actually care about' is a Start Menu^WScreen item is a nice-to-have. |
| 03:50 | < Reiv> | Whozawhutnow? |
| 03:51 | <&McMartin> | Our product is designed to manage other people's virtual machines, more or less, and make them interoperate with IT environments that have not sold their soul to VMware |
| 03:52 | <&McMartin> | That means that while there's a pretty sizable client-side component, the operation they probably really want is "get me into that desktop environment now please" |
| 03:52 | <&McMartin> | The challenge is to make that play right with the many and varied and variously absurd authentication mechanisms out there in the wild, and to sensibly handle cases where there are updates to the environment &c |
| 03:53 | < Reiv> | right |
| 03:53 | < Reiv> | (Selling ones soul to VMware is bad?) |
| 03:53 | <&McMartin> | (Well, they get your soul) |
| 03:53 | <&McMartin> | (Also, server-side VDI infrastructure is not always a good fit for offices.) |
| 03:53 | <&McMartin> | (And by "not always" I mean "seriously, basically never except for four or five use cases") |
| 03:54 | <&McMartin> | (We cover a few other ones they don't, and are bad at ones they do.) |
| 03:54 | <&McMartin> | (Almost nobody in this space actually *directly* competes with one another) |
| 03:54 | <&McMartin> | (The closest it gets is "Oh! You're overhauling your entire IT infrastructure. You should do so in a way that makes our product work for you." |
| 03:55 | <&McMartin> | "No! You should do so in a way that makes our product work for you.") |
| 03:55 | <&McMartin> | (And these will be *different ways*.) |
| 03:55 | < Reiv> | Ah-hah. |
| 03:55 | | * Reiv muses. |
| 03:55 | <~Vornicus> | The only thing I've heard about metro that I've actually kind of agreed with is "the start menu is now a whole screen" |
| 03:55 | < Reiv> | Call centre with a couple dozen PCs with Thin Clients would be...? |
| 03:56 | <&McMartin> | Doesn't Need VDI At All, but VMware will happily take your money for an ESX server. |
| 03:56 | <&McMartin> | Also, Not Us |
| 03:56 | < Reiv> | Vornicus: And then they deleted the search function. |
| 03:56 | <~Vornicus> | okay that's unsmooth |
| 03:56 | <&McMartin> | wat |
| 03:56 | < Reiv> | Which was the bestest bit /ever/ ;_; |
| 03:56 | <&McMartin> | You just start typing and it goes |
| 03:56 | < Reiv> | Well, OK |
| 03:56 | < Reiv> | But telling this to a user over the phone...! |
| 03:56 | < Reiv> | "HELP I PUSHED A BUTTON WHAT HAPPEN" |
| 03:57 | < Reiv> | Needs textbox for typings. Then what you pushed to start typings in makes sense. |
| 03:57 | <&McMartin> | Anyway, our version of VDI is thick - in fact, fully local - clients, with intermittent connections to the back office for updates and leases on the client. |
| 03:57 | < Reiv> | hm |
| 03:57 | < Reiv> | The ideal use case for this? |
| 03:58 | <&McMartin> | People with crew that travel a lot like our version better because they don't need to remote-desktop across half the world to get stuff done. |
| 03:58 | <&McMartin> | Sales forces, etc. |
| 03:58 | <&McMartin> | People who already have a lot of stock hardware sitting around but no million dollar mainframe and no heavy network infrastructure |
| 03:58 | < Reiv> | Hm |
| 03:59 | < Reiv> | You then use the thing for..? |
| 03:59 | <&McMartin> | Running the desktop environment |
| 03:59 | <&McMartin> | The one that we put on the glossy brochures is "you're doing BYOC, but people are bringing in Macs" |
| 03:59 | <&McMartin> | "So now your corporate environment can be under your control, but run in their copy of Fusion" |
| 04:00 | < Reiv> | ... aha |
| 04:00 | <&McMartin> | And then we've got various other things to make that work |
| 04:00 | < Reiv> | Then the Desktop Enviroment is Just Another Application. |
| 04:00 | < Reiv> | Running on $Whatever. |
| 04:00 | < Reiv> | Yes? |
| 04:00 | <&McMartin> | I really loathe that way of putting it, but yes~ |
| 04:00 | < Reiv> | Hey, it's a good thing~ |
| 04:01 | < Reiv> | Wheras for my job, I, uh |
| 04:01 | < Reiv> | Help make the systems talk to each other. So that we can tell the computer how much power someone used. |
| 04:01 | <&McMartin> | Apparently - and I don't really have numbers for it - there's also certain scales of deployment where stock hardware running VMs and only checking in with ordinary web servers for updates occasionally is less TCO than doing back office with continuous remote desktop |
| 04:01 | <&McMartin> | And at that point you can use it to provision all your secretaries and similar clerical workers |
| 04:02 | <@himi> | What virtual environments do you support? |
| 04:02 | < Reiv> | With intermittent interruptions to tell the nice ladies in the office what the /old/ system said that someone used, 'cuz they've noticed that there's something Fishy with a power meter and want to backcheck for how long. >_> |
| 04:02 | <&McMartin> | himi: Officially, only Windows in the guest, and if you want All The Prizes, only Windows 7 |
| 04:02 | <@himi> | And on the host? |
| 04:02 | <&McMartin> | Unofficially, Linux works and isn't difficult to set up to work the way people expect, though handling updates cleanly is a bit of a pain. |
| 04:03 | <&McMartin> | Host is any Windows post-XP, as many Macs as we can get to accept it (this has been where much of my hatred for Apple Flag Days is coming from), and we do a Linux-based product that controls the entire hardware set that is laughably called "Bare Metal" but which is for the cases like the clerical workers |
| 04:04 | <@himi> | But no 'normal' VMs on Linux machines? |
| 04:05 | <@himi> | Just the 'bare metal'? |
| 04:05 | <&McMartin> | Not sure what you mean by "Normal" here. |
| 04:05 | <&McMartin> | We don't have a .deb or .rpm for the player/creator, no |
| 04:05 | <&McMartin> | No customer has ever been willing to pay us enough for this to make it worth queueing up as a feature. |
| 04:06 | <@himi> | So if you replaced the Macs in the brochure with Linux boxes there wouldn't be an option to deal with it |
| 04:07 | <&McMartin> | Not yet, no. |
| 04:07 | <&McMartin> | As soon a customer actually has this as a problem, it would probably get scheduled for two releases from the first point a sizable prospective customer brought it up. |
| 04:07 | <&McMartin> | *as a |
| 04:08 | | * himi nods |
| 04:08 | <&McMartin> | If I may wax cynical momentarily, the kind of people who bring Linux boxes when told to Bring Their Own Computer refuse to work for companies where the company desktop environment is Windows-based~ |
| 04:08 | <@himi> | heh |
| 04:08 | <@himi> | You haven't dealt with my organisation, then |
| 04:09 | <&McMartin> | Well, it's also worth noting that if your company desktop environment *isn't* Windows-based there's a lot less we offer you. |
| 04:09 | <&McMartin> | You're basically getting an easy way to autoinstall VMware and keep it updated, and maybe push updates to some of the disks in an image if you set it up to do that. |
| 04:10 | <&McMartin> | (On Windows, we have a cool unionfs-like thing except it actually works) |
| 04:10 | <&McMartin> | (One of the few places where Windows being insane worked out for us - merely being registry-aware lets you do the equivalent of merging arbitrary /etc conf files) |
| 04:13 | | Vash[ApplePieMaking] is now known as Vash |
| 04:31 | < Reiv> | Windows being insane by dint of having a registry? |
| 04:32 | <&ToxicFrog> | I actually feel a lot more charitable towards the registry, now that I'm comparing it not to /etc, but to /etc, plus .local, plus .config, plus two or three DE-specific registry reimplementations backed by XML. |
| 04:34 | | * rms still doesn't get why some of the keys in the registry are freaking GUIDs |
| 04:36 | <&ToxicFrog> | Those are application/hardware identifiers, I thought? |
| 04:38 | < rms> | The ones I know of deal with My Computer, My Documents and such icons somewhere |
| 04:39 | <@himi> | Possibly user identifiers, then |
| 04:39 | < rms> | Dunno |
| 04:40 | < rms> | I try to avoid accumulating Windows knowledge so that mundanes don't ask me to use it for any reason |
| 04:42 | <@himi> | I normally tell people I charge extra for Windows stuff |
| 04:43 | <&ToxicFrog> | Oh. Yeah, windows internally uses GUIDs as user IDs. |
| 05:05 | <&McMartin> | TF: Man, I would kill for an XML-based registry format over .reg and the internal UTF-16-BE one >_< |
| 05:11 | < Reiv> | hrn |
| 05:11 | < Reiv> | Math query: I am sure there's a relatively simple formula for the following progression. |
| 05:12 | < Reiv> | 1: 0, 2: 4, 3: 12, 4: 24, 5: 30 |
| 05:12 | < Reiv> | Possibly something about multiplying the previous n or something |
| 05:15 | < rms> | McMartin: GNOME does it in a fairly retarded way (namely using a mix of a directory structure and multiple xml files) |
| 05:16 | < rms> | I can see some benefit to that model, but really they'd be best to chose one and not do both |
| 05:19 | <&McMartin> | Yeah. what the format *is* matters much less than that you have one and stick to it. |
| 05:26 | <&ToxicFrog> | McMartin: yeah, it's not so nice when you have multiple mutually incompatible registries and it's a total crapshoot which one any given program uses |
| 05:27 | <&ToxicFrog> | Assuming it only uses one! |
| 05:27 | <&ToxicFrog> | This is not universally the case! |
| 05:28 | <&ToxicFrog> | (in fact it's pretty rarely the case if you assume transitivity, since, for example, anything that uses the network is hitting /etc/hosts and /etc/resolv.conf, at minimum, in addition to whatever registry it uses for user settings) |
| 05:30 | <&McMartin> | (At that level Windows is no better, since it has a secret /etc/hosts of its own deep under %WINDIR%) |
| 05:35 | <~Vornicus> | http://oeis.org/search?q=0%2C4%2C12%2C24%2C30 no results. |
| 05:49 | <~Vornicus> | As for an interpolating polynomial, bad news: it's a straight 4th order one, (-5x^4+30x^3-31x^2+54x)/12, next number in the sequence is 10, and all other numbers in both directions are negative. |
| 05:49 | <~Vornicus> | now 6 12 20 30 is a known sequence, x(x+1) |
| 05:53 | < Reiv> | hrm |
| 05:53 | <~Vornicus> | 0,4,12,24,40 is also a known sequence, 2x(x+1) |
| 05:53 | < Reiv> | oh shits, I typoed the sequence >_< |
| 05:53 | < Reiv> | Your sequence is correct! |
| 05:53 | <~Vornicus> | hee |
| 05:54 | | * Reiv hang head in shame. |
| 05:54 | < Reiv> | (I /knew/ it was a real sequence, honest boss) |
| 05:55 | < Reiv> | 0,3,9,18,30 is therefore 1.5x(x+1) correct? |
| 05:55 | <~Vornicus> | yes |
| 05:55 | < Reiv> | 3,5,9,15,23? |
| 05:56 | <~Vornicus> | .oh that one's cute, hang on |
| 05:56 | <~Vornicus> | n^2 + n + 3 |
| 05:56 | < Reiv> | Cute! |
| 05:57 | | * Reiv suspects this one actually doesn't work: |
| 05:57 | < Reiv> | 3,4,7,12,19 |
| 05:57 | <~Vornicus> | that one never hits 0. |
| 05:57 | <~Vornicus> | You would be Wrong, sir! n^2 + 3 |
| 05:57 | < Reiv> | ... duh. |
| 05:58 | < Reiv> | Of course. |
| 05:58 | <&ToxicFrog> | Vornicus: wait, how do you get (n^2 + n + 3) for [3,5,9,15,23]? |
| 05:58 | < Reiv> | 0,12,26,42,60 |
| 05:58 | <&ToxicFrog> | Oh wait |
| 05:59 | | * ToxicFrog headdesks |
| 05:59 | <~Vornicus> | n(n+11) |
| 05:59 | <&ToxicFrog> | You're meant to define them non-recursively, aren't you |
| 05:59 | < Reiv> | ToxicFrog: Buh? |
| 05:59 | <~Vornicus> | yeah, if i were doing recursion it'd be x_n = x_{n-1} |
| 05:59 | <~Vornicus> | or in that form. |
| 05:59 | < Reiv> | Vornicus: Just to query, are these starting with n is 0, or n is 1? |
| 05:59 | <~Vornicus> | So far these are all starting at 0. |
| 06:00 | < Reiv> | OK! |
| 06:00 | < Reiv> | So for starting-at-one, you just subtract 1 from n beforehand. |
| 06:00 | <~Vornicus> | something like that, yeah. You'll find easier forms if you simplify. |
| 06:00 | < Reiv> | Cute. |
| 06:01 | < Reiv> | And with that, Vornypie, you just algebraically stated the entire XP costs system for Exalted. |
| 06:01 | <~Vornicus> | Bitch please, I did that for 3e D&D |
| 06:01 | < Reiv> | Which has polynomial costs for abilities rated 1-5. |
| 06:01 | < Reiv> | Yeah but did you do that in five minutes~ |
| 06:01 | <~Vornicus> | No more like an hour and a half. |
| 06:01 | <~Vornicus> | But |
| 06:01 | < Reiv> | (Well, not /all/ of it.) |
| 06:02 | < Reiv> | (But the rest are linear, and even I am not *that* hopeless ??) |
| 06:03 | <~Vornicus> | (yeah well) |
| 06:08 | <~Vornicus> | (also quadratics are not exactly amenable to recursive descriptions because they require you to know what n is, as well as x_n.) |
| 06:28 | < Orthia> | Reiver? |
| 06:29 | | Derakon is now known as Derakon[AFK] |
| 06:36 | | * ToxicFrog boxes up Reiv, sends him to Orthia |
| 07:00 | | Syloq_Home [Syloq@NetworkAdministrator.Nightstar.Net] has quit [Client closed the connection] |
| 07:42 | | Netsplit *.net <-> *.split quits: simon_, Orthia, @Tamber, Tarinaky, Outpost, rms, @ToxicFrog, AnnoDomini |
| 07:43 | | Netsplit over, joins: @Tamber, &ToxicFrog, Orthia, AnnoDomini, Tarinaky, simon_, rms, Outpost |
| 07:56 | | Vash [Vash@Nightstar-3ba4108e.wlfrct.sbcglobal.net] has quit [[NS] Quit: I lovecraft Vorn!] |
| 08:00 | < AnnoDomini> | TheWatcher[zZzZ]: Gratz. |
| 08:12 | | Attilla [Obsolete@Nightstar-b60121d0.as43234.net] has joined #code |
| 08:42 | | Derakon[AFK] [Derakon@Nightstar-a3b183ae.ca.comcast.net] has quit [Ping timeout: 121 seconds] |
| 09:10 | | You're now known as TheWatcher |
| 09:48 | | RichyB [mycatverbs@A08927.DD15B2.DD7982.099181] has joined #code |
| 10:11 | | RichyB [mycatverbs@A08927.DD15B2.DD7982.099181] has quit [[NS] Quit: Aaaaaaa, run away!] |
| 11:23 | | io\dis4ea is now known as iospace |
| 11:33 | | RichyB [richardb@7D8456.14C5E0.08BE20.B646E9] has joined #code |
| 11:52 | | * TheWatcher trundles around this code in the refactor tractor |
| 11:52 | <@TheWatcher> | So messy, so little time to clean it up |
| 11:53 | | * jerith rewrites some sandbox setup code. |
| 12:41 | <&jerith> | Yay! It works! |
| 12:41 | <@TheWatcher> | \o/ |
| 12:44 | | RichyB [richardb@7D8456.14C5E0.08BE20.B646E9] has quit [Ping timeout: 121 seconds] |
| 12:45 | <&jerith> | I suddenly stopped being stupid and started using envvars to pass config to my sandbox setup process. |
| 12:46 | <&jerith> | We used to send it through stdin and then worry about readahead before the execvpe() to the actual sandboxed executable. |
| 13:07 | | RichyB [richardb@7D8456.14C5E0.08BE20.B646E9] has joined #code |
| 13:52 | | RichyB [richardb@7D8456.14C5E0.08BE20.B646E9] has quit [Ping timeout: 121 seconds] |
| 14:02 | | RichyB [richardb@7D8456.14C5E0.08BE20.B646E9] has joined #code |
| 14:32 | | simon_ [simon@842092.D5D4B4.CE904E.AD841A] has quit [Ping timeout: 121 seconds] |
| 14:42 | | RichyB [richardb@7D8456.14C5E0.08BE20.B646E9] has quit [[NS] Quit: Leaving] |
| 14:42 | | RichyB [richardb@7D8456.14C5E0.08BE20.B646E9] has joined #code |
| 14:42 | | RichyB [richardb@7D8456.14C5E0.08BE20.B646E9] has quit [Connection closed] |
| 15:08 | | Syloq_Home [Syloq@NetworkAdministrator.Nightstar.Net] has joined #code |
| 16:55 | | Derakon[AFK] [Derakon@Nightstar-a3b183ae.ca.comcast.net] has joined #code |
| 17:06 | | RichyB [richardb@7D8456.14C5E0.08BE20.B646E9] has joined #code |
| 17:07 | | RichyB [richardb@7D8456.14C5E0.08BE20.B646E9] has quit [[NS] Quit: Leaving] |
| 18:04 | | Derakon[AFK] [Derakon@Nightstar-a3b183ae.ca.comcast.net] has quit [Ping timeout: 121 seconds] |
| 18:39 | | Derakon[AFK] [Derakon@Nightstar-a3b183ae.ca.comcast.net] has joined #code |
| 19:16 | | Vash [Vash@Nightstar-3ba4108e.wlfrct.sbcglobal.net] has joined #code |
| 19:16 | | mode/#code [+o Vash] by ChanServ |
| 19:17 | | celticminstrel [celticminst@Nightstar-05d23b97.cable.rogers.com] has joined #code |
| 19:29 | | * Vornicus wonders how TOrchlight et al manages to do its "character hiding behind terrain" trick, where you instead get a blue or red xray image |
| 19:30 | < celticminstrel> | Sounds useful... |
| 19:32 | <~Vornicus> | it is |
| 19:39 | < iospace> | it's probably two sprites, one the outline, the other the normal one. depending on the positioning (and since the view is fixed in combat), just a simple check |
| 19:39 | < iospace> | is the object at this point on the screen? Then use outline. otherwise use the filled in sprite |
| 19:39 | < iospace> | i think that's how they do it anyway ^^;; |
| 19:40 | | * iospace isn't a game designer |
| 19:43 | <~Vornicus> | it's not a sprite |
| 19:44 | <~Vornicus> | it's a 3d model, and a dynamic one at that, with things attached to it like gloves. Some things - railings, for instance - you can't see behind but that's all right because they're really short. |
| 19:44 | < iospace> | true |
| 19:44 | < iospace> | two models then |
| 19:45 | < iospace> | not two sprites :P |
| 19:45 | < iospace> | either way, i still think there's position checking involved |
| 19:47 | <~Vornicus> | That's also unlikely. The transition is smooth - if part of your body is visible and the other part is not, the visible part is not x-ray |
| 19:48 | < gnolam> | Failing/passing the depth test. Or via the stencil buffer. |
| 19:48 | < celticminstrel> | But the invisible part is? |
| 19:50 | < iospace> | also, i hate anti-static footstraps |
| 19:50 | < iospace> | just saying |
| 19:53 | < gnolam> | E.g. |
| 19:53 | < gnolam> | glDepthFunc(GL_GEQUAL); |
| 19:53 | < gnolam> | xrayModel()->Draw(); |
| 19:53 | < gnolam> | glDepthFunc(GL_LESS); |
| 19:53 | < gnolam> | normalModel->Draw(); |
| 19:54 | < gnolam> | (After drawing the terrain) |
| 19:59 | <~Vornicus> | Ladies and gentlemen, the graphics expert. I didn't know you had that ability at all |
| 20:07 | <&McMartin> | It's Old Testament, so therefore it must be PURGED FROM THE STAGE OF HISTORY |
| 20:18 | < gnolam> | Actually, it should be the same even in OpenGL 4. |
| 20:19 | < gnolam> | Since failing the depth test means that no fragment is generated. |
| 20:19 | < gnolam> | (So you can't just say "stick it in your shader", like with everything else that was hacked off in 3.x and onwards) |
| 20:20 | < gnolam> | ... why am I rotating the torque vector? The local torque unit vector is the same as the local lift unit vector. |
| 21:09 | | Rhamphoryncus [rhamph@Nightstar-cc6253d6.abhsia.telus.net] has joined #code |
| 22:32 | < Tarinaky> | Vornicus/iospace: Edge Detection, surely? |
| 22:33 | < Tarinaky> | I think that's what edge detection is, anyway >.> |
| 22:59 | | Vash [Vash@Nightstar-3ba4108e.wlfrct.sbcglobal.net] has quit [[NS] Quit: I lovecraft Vorn!] |
| 23:00 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has quit [Operation timed out] |
| 23:59 | <~Vornicus> | I don't think you can do 'edge detection' live |
| --- Log closed Tue Oct 09 00:00:45 2012 |