| --- Log opened Mon Apr 18 00:00:34 2016 |
| 00:12 | | himi [fow035@Nightstar-v37cpe.internode.on.net] has quit [Ping timeout: 121 seconds] |
| 00:17 | | himi [fow035@Nightstar-v37cpe.internode.on.net] has joined #code |
| 00:17 | | mode/#code [+o himi] by ChanServ |
| 00:46 | | Turaiel[Offline] is now known as Turaiel |
| 00:47 | | ErikMesoy1 [Erik@Nightstar-hq72t5.customer.cdi.no] has joined #code |
| 00:47 | | ErikMesoy [Erik@Nightstar-hq72t5.customer.cdi.no] has quit [Connection closed] |
| 00:56 | | catalyst [catalyst@Nightstar-bt5k4h.81.in-addr.arpa] has quit [[NS] Quit: Leaving] |
| 01:05 | | catadroid [catalyst@Nightstar-d52gc4.dab.02.net] has joined #code |
| 01:28 | | Derakon[AFK] is now known as Derakon |
| 01:53 | | ErikMesoy1 [Erik@Nightstar-hq72t5.customer.cdi.no] has quit [Connection closed] |
| 01:54 | | ErikMesoy [Erik@Nightstar-hq72t5.customer.cdi.no] has joined #code |
| 02:00 | | ErikMesoy [Erik@Nightstar-hq72t5.customer.cdi.no] has quit [Connection closed] |
| 02:01 | | ErikMesoy [Erik@Nightstar-hq72t5.customer.cdi.no] has joined #code |
| 03:02 | | Reiv [NSwebIRC@Nightstar-q8avec.kinect.net.nz] has quit [Ping timeout: 121 seconds] |
| 03:11 | | Reiv [NSwebIRC@Nightstar-q8avec.kinect.net.nz] has joined #code |
| 03:11 | | mode/#code [+o Reiv] by ChanServ |
| 03:21 | | * McMartin finishes his first draft at re-sourcing an old type-in machine-code program. |
| 03:49 | | Turaiel is now known as Turaiel[Offline] |
| 04:23 | | crystalclaw|AFK is now known as crystalclaw |
| 04:27 | | Kindamoody[zZz] is now known as Kindamoody |
| 04:46 | <&McMartin> | I salute this madman: https://github.com/mntmn/amiga2000-gfxcard |
| 04:46 | <~Vornicus> | *that* is a madman |
| 04:53 | | Crossfire [Z@Nightstar-3nmkps.tpgi.com.au] has joined #code |
| 04:53 | | mode/#code [+o Crossfire] by ChanServ |
| 05:01 | <~Vornicus> | Also: noice. |
| 05:03 | | Derakon is now known as Derakon[AFK] |
| 05:25 | <@Reiv> | ... he went and made a graphics card? |
| 05:25 | <@Reiv> | Would that involve making a whole new standard to code against? |
| 05:26 | <&McMartin> | Well, we're talking "competing with VGA" |
| 05:26 | <&McMartin> | So we're talking "here is a framebuffer; write your screen data here", I think. |
| 05:26 | <&McMartin> | So it's more "write a graphics driver" |
| 05:26 | <&McMartin> | ... and you know, print your own circuit boards and program the FPGAs one them and the rest |
| 05:27 | <&McMartin> | He's running the core pre-existing OS on top of it |
| 05:27 | <&McMartin> | The whole point was to get to play DOOM properly, after all~ |
| 05:28 | | celticminstrel [celticminst@Nightstar-q0f7bb.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 05:34 | | catadroid` [catalyst@Nightstar-p7hvmt.dab.02.net] has joined #code |
| 05:35 | <@Reiv> | Good lord |
| 05:35 | <@Reiv> | So the trick is the framebuffer, so you're no longer racing the beam? |
| 05:36 | <&McMartin> | This is an Amiga 2000 |
| 05:36 | <&McMartin> | We're already well past that point. |
| 05:36 | <@Reiv> | oh |
| 05:36 | <@Reiv> | I misunderstood, sorry |
| 05:36 | <@Reiv> | Carry on |
| 05:36 | <&McMartin> | (Arguably, we stopped racing the beam in 1981; but this is an early-90s system.) |
| 05:36 | <@Reiv> | 1981? |
| 05:36 | | catadroid [catalyst@Nightstar-d52gc4.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 05:37 | <&McMartin> | The PC's CGA card has a proper framebuffer. |
| 05:37 | <@Reiv> | oh, right |
| 05:37 | <&McMartin> | So, for that matter, does the C64, you just rarely used it because the CPU couldn't update it fast enough; you had a controlled character cell display instead |
| 05:37 | <&McMartin> | Likewise the NES, but it didn't have the full framebuffer, just the character cells |
| 05:38 | <&McMartin> | But it was more the generation before those that you had to race the beam for real, or otherwise have the CPU control the display line by line |
| 05:38 | <&McMartin> | And that was late-70s tech. |
| 05:38 | <&McMartin> | (Note that the C64 still had to switch off its display when doing precisely timed serial operations. That wasn't because of beam-racing, though, it was because of bus contention with the video chip, more or less) |
| 05:39 | <&McMartin> | The original PC had a real problem there, actually, which meant it had to solve some problems that wouldn't have otherwise shown up until the late 1980s really early on |
| 05:39 | <@Reiv> | It's funny how we are in 2016 and we *still* have problems with bus bandwidth |
| 05:39 | <&McMartin> | 8-bit data bus, right |
| 05:39 | <&McMartin> | But the PC is x86-based, so it wants a 16-bit data bus |
| 05:39 | <&McMartin> | There was extra circuitry to basically buffer each "single" read the CPU thought it was doing to the two memory reads that actually had to happen |
| 05:39 | <@Reiv> | Interesting |
| 05:40 | <&McMartin> | This was the *primary* determiner of how fast your code ran, and pretty much continued to be so for PCs until the 486 introduced proper on-chip caching. |
| 05:40 | <@Reiv> | And that's why the x86 chips went with so many integrated-commands |
| 05:40 | <@Reiv> | Which I recall us discussing previously |
| 05:40 | <&McMartin> | And why coders loved them so much, yeah. |
| 05:41 | <&McMartin> | (The 286 was sixteen bit all the way down, at least, but then the 386 went 32-bit while half of the motherboards were stil 16-bit, and hiyo, there we go again) |
| 05:41 | <&McMartin> | (Turns out that's actually what the ST in Atari ST stood for too; Sixteen-bit bus and Thirty-two bit processor) |
| 05:41 | <&McMartin> | (I forget how Amigas worked, but I think both it and Macs had 32-bit processors clear back in '85) |
| 05:42 | <&McMartin> | I'm not actually convinced there's such a thing as "enough" bus bandwidth |
| 05:42 | <&McMartin> | It would stop being a bus >_> |
| 05:43 | <&McMartin> | Without that, if you have enough, that means you don't have enough peripherals attached |
| 05:43 | < Azash> | Tonight, on Hardware Hoarders |
| 05:44 | < Azash> | I'm really not good at comp arch so pardon me if it's a stupid question |
| 05:44 | < Azash> | Is device bus usage bursty or steady? |
| 05:45 | <~Vornicus> | mac and amiga are both 68k machines; the 68k is a hybrid 16/32 processor |
| 05:46 | <~Vornicus> | (...and the ST used the 68k too...) |
| 05:51 | <&McMartin> | Azash: Some of each. Depends a lot on the devices. |
| 05:51 | <&McMartin> | Video is pretty steady if you're measuring in milliseconds but bursty if you're measuring in microseconds. |
| 05:51 | < Azash> | Mm |
| 05:52 | < Azash> | So basically a difference coooould be drawn between "adequate" and "enough" considering statistical multiplexing? |
| 05:55 | <&McMartin> | Yeah, and in practice that's what it is |
| 05:55 | <&McMartin> | But the whole point of a bus is that you aren't actually using every device all the time forever |
| 05:55 | <&McMartin> | So they can share |
| 05:56 | <&McMartin> | Also, you often are *trying* to multiplex |
| 05:56 | <&McMartin> | The video card and the CPU would both like to look at memory, for instance. |
| 05:56 | <&McMartin> | Less so these days, because there's separate video ram that the video card can monopolize |
| 05:56 | <&McMartin> | That way the CPU doesn't get stunned every time the monitor has a new frame to draw |
| 05:59 | <&McMartin> | You can get clever about how you set stuff up so that you don't have to share, either... |
| 05:59 | <&McMartin> | ... I think that's what Non-Uniform Memory Access is all about, maybe? |
| 05:59 | | * Vornicus sics eric brolsma on mcm |
| 06:00 | <&McMartin> | Computers only actually make sense for a brief period in the middle of their lifespans |
| 06:00 | <&McMartin> | I even spelled it out =P |
| 06:02 | <&McMartin> | looks like NUMA is more for many CPUs each with associated memory chunks |
| 06:02 | <&McMartin> | Though at some point the CPU stops being "C", doesn't it. |
| 06:11 | | Reiv [NSwebIRC@Nightstar-q8avec.kinect.net.nz] has quit [Ping timeout: 121 seconds] |
| 07:04 | <@himi> | NUMA is about having memory that isn't accessed directly, or more specifically about memory that doesn't have uniform access cost/timing |
| 07:05 | <@himi> | Almost universally implemented by having memory attached to CPU cores with memory attached to a non-local core accessed across the CPU interconnect |
| 07:05 | <&McMartin> | Yeah |
| 07:06 | <&McMartin> | That still complicates the notion of "the memory store" for a computer, though |
| 07:06 | <@himi> | Totally |
| 07:06 | <@himi> | "memory" hasn't been simple for some time, though, since the advent of memory mapped IO |
| 07:07 | <@himi> | That wasn't too much of a pain until people started attaching graphics cards with as much on-board memory as the host system had |
| 07:07 | <@himi> | NUMA's been around for ages, though, since it's the only way to make a single system image scale to something interesting |
| 07:09 | <&McMartin> | Hm. Doesn't Memory Mapped I/O go back at least to the PC? |
| 07:09 | <&McMartin> | The IN and OUT instructions ultimately write to segment zero |
| 07:09 | <@himi> | McMartin: yeah, but that's more of an implementation detail |
| 07:10 | <@himi> | Memory mapped IO is higher level, with specialised hardware that manages the mapping of device registers/memory/whatever into the CPU's logical memory map |
| 07:10 | <@himi> | It's the /hardware support/ that makes it interesting, because it means you can generically talk to hardware the same way you deal with any other chunk of memory |
| 07:13 | <@himi> | Also, to respond directly to Azash, there are many buses in a computer, and the behaviour can be pretty much arbitrarily complicated |
| 07:13 | <@himi> | So the answer to the question gets down to "how are you using it?" |
| 07:42 | | Kindamoody is now known as Kindamoody|afk |
| 08:10 | | crystalclaw is now known as crystalclaw|AFK |
| 08:43 | | catadroid` [catalyst@Nightstar-p7hvmt.dab.02.net] has quit [[NS] Quit: Bye] |
| 08:51 | < abudhabi> | Anyone know how properties files work in Java? |
| 08:56 | <&McMartin> | Oh man, that's been a long time |
| 08:56 | <&McMartin> | Last time I used them they were basically one-level INI files though |
| 08:56 | <&McMartin> | foo=bar |
| 08:56 | <&McMartin> | longer.foo="yeah, whatever" |
| 08:56 | <&McMartin> | etc |
| 08:58 | | * Vornicus replaces the computer buses with computer subways. |
| 08:59 | < abudhabi> | McMartin: Right. How do I use them to specify a configuration for a web app so that it doesn't need to be redeployed when I need to change the file dump directory, other configuration file directories, etc? |
| 09:02 | < abudhabi> | Specify the path to the properties file in the web.xml, and specify the other paths and shit in the properties file? |
| 09:02 | <&McMartin> | Mmm. That's a question one level up. I'm probably an entire decade out of date, there. |
| 09:02 | <&McMartin> | (.WAR! What is it good for?) |
| 09:02 | <&McMartin> | I'd expect that to be in the app container's docs |
| 09:15 | <@froztbyte> | abudhabi: I found http://stackoverflow.com/questions/13956651/externalizing-tomcat-webapp-config-f rom-war-file last time I was looking into this |
| 09:15 | <@froztbyte> | (for a Groovy mess someone handed over) |
| 09:21 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 09:47 | < abudhabi> | Thanks. |
| 09:48 | | Emmy [NSkiwiirc@Nightstar-esfu0j.dynamic.ziggo.nl] has joined #code |
| 13:07 | | Emmy is now known as Emmy-AFK |
| 13:08 | | * TheWatcher readsup |
| 13:11 | <@TheWatcher> | McMartin: when I was installing jenkins, gerrit, sonarqube and some others, I made sure to do a `find . -iname '*.war' | xargs chmod -w` on them. |
| 14:35 | | celticminstrel [celticminst@Nightstar-q0f7bb.dsl.bell.ca] has joined #code |
| 14:35 | | mode/#code [+o celticminstrel] by ChanServ |
| 15:44 | <&McMartin> | https://github.com/yberreby/rgo |
| 15:47 | <@TheWatcher> | ... why |
| 15:53 | <&McMartin> | Practice? :shrug: |
| 15:53 | <@TheWatcher> | Fair enough |
| 16:24 | | Netsplit Deepthought.Nightstar.Net <-> Krikkit.Nightstar.Net quits: @PinkFreud, @ToxicFrog |
| 16:34 | | Netsplit over, joins: @PinkFreud |
| 16:36 | | ToxicFrog [ToxicFrog@ServerAdministrator.Nightstar.Net] has joined #code |
| 16:36 | | mode/#code [+ao ToxicFrog ToxicFrog] by ChanServ |
| 16:41 | | Crossfire [Z@Nightstar-3nmkps.tpgi.com.au] has quit [Ping timeout: 121 seconds] |
| 16:45 | | Netsplit Deepthought.Nightstar.Net <-> Krikkit.Nightstar.Net quits: @PinkFreud, @ToxicFrog |
| 16:47 | | Netsplit over, joins: @PinkFreud, &ToxicFrog |
| 17:23 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 17:23 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 18:16 | | Emmy-AFK [NSkiwiirc@Nightstar-esfu0j.dynamic.ziggo.nl] has quit [[NS] Quit: ] |
| 18:59 | | catalyst [catalyst@Nightstar-bt5k4h.81.in-addr.arpa] has joined #code |
| 19:58 | <&McMartin> | http://arstechnica.com/security/2016/04/flashback-declassified-1970-dod-cybersec urity-document-still-relevant/ |
| 19:58 | <&McMartin> | In the "have we learned anything?" department and also in the "actually this stuff has been relevant for longer than you'd think" department |
| 19:58 | <&McMartin> | Both of which are kind of hobby horses of mine >.> |
| 19:59 | < [R]> | Is there actually a way to defend against that? |
| 19:59 | <&McMartin> | Which "that"? |
| 20:00 | <&McMartin> | "Security being hard and something you have to design in from the start"... probably not. Hard problems do tend to stay hard, and humans change more slowly than machines. |
| 20:00 | <&McMartin> | "Can we get better about expanding the kinds of access that *can* be done securely"... probably, but I'm not an expert and hesitate to answer "yes" conclusively |
| 20:00 | <&McMartin> | This bit jumped out at me: |
| 20:00 | <&McMartin> | "Another set of conclusions of the Ware Report also rings true today: while "contemporary technology" could make a closed system (one without network connections, locked in a room in a building) "acceptably resistant to external attack, accidental disclosures, internal subversion and denial of use to legitimate users," it couldn't do so for an "open environment, which includes uncleared users working at phy |
| 20:01 | <&McMartin> | sically unprotected consoles connected to the systems by unprotected communications"" |
| 20:01 | <&McMartin> | I suspect that "open environment" as described is intrinsically unsecurable to a degree an IT guy would consider actually secure |
| 20:01 | <&McMartin> | As opposed to "eh it'll probably be fine" |
| 20:34 | | crystalclaw|AFK is now known as crystalclaw |
| 20:37 | | Kindamoody|afk is now known as Kindamoody |
| 21:57 | | Reiv [NSwebIRC@Nightstar-q8avec.kinect.net.nz] has joined #code |
| 21:58 | | mode/#code [+o Reiv] by ChanServ |
| 22:29 | | Kindamoody is now known as Kindamoody[zZz] |
| 23:43 | | catalyst [catalyst@Nightstar-bt5k4h.81.in-addr.arpa] has quit [Connection closed] |
| 23:46 | | Turaiel[Offline] is now known as Turaiel |
| 23:59 | | catadroid [catalyst@Nightstar-5dp6v9.dab.02.net] has joined #code |
| --- Log closed Tue Apr 19 00:00:50 2016 |