| --- Log opened Tue May 08 00:00:29 2012 |
| 00:04 | | * sshine is going to write his bachelor's project. |
| 00:08 | | Attilla [Obsolete@Nightstar-5cf8db5d.as43234.net] has quit [Ping timeout: 121 seconds] |
| 00:08 | | You're now known as TheWatcher[t-2] |
| 00:12 | | You're now known as TheWatcher[zZzZ] |
| 00:24 | <&McMartin> | ! |
| 00:25 | <&McMartin> | Since when did Fedora have an XFCE spin? |
| 00:25 | <&McMartin> | Man |
| 00:25 | | eckse [eckse@Nightstar-75493054.dsl.sentex.ca] has joined #code |
| 00:25 | | mode/#code [+o eckse] by ChanServ |
| 00:30 | | Derakon[AFK] is now known as Derakon |
| 00:51 | <&McMartin> | Ha ha yay |
| 00:51 | <&McMartin> | "I sometimes like to imagine some wingnut show taking callers to bitch about Chomsky and having someone go off about generative grammar." |
| 01:16 | <&McMartin> | Bleh |
| 01:17 | <&McMartin> | I suppose it is too much to ask for Linux distros to actually ship with the versions of SVN that don't shit all over your checkouts. |
| 01:42 | < Noah> | Don't use SVN? |
| 01:42 | <&McMartin> | Work uses SVN, since it is in fact good for centrally controlling and tracking changes. |
| 01:43 | <&McMartin> | And, as repeatedly noted, it's had working branch support since 1.6 |
| 01:43 | <&McMartin> | However, this is rarely believed because a fair number of people think "branch support" means "anyone can fork you" which is an entirely separate feature. |
| 01:44 | <&McMartin> | The solution is in fact "don't use Linux"~ |
| 01:44 | <&McMartin> | This problem has been solved in SVN for over a year now, there's really no excuse for it to not at least be in unstable |
| 02:05 | < Noah> | yea, that's true |
| 02:07 | < Noah> | That's one of my gripes with Linux, I usually have to fight tooth and nail to get anything installed that's younger than my daughter |
| 02:08 | <&McMartin> | Fedora is *usually* better about this |
| 02:08 | <&McMartin> | But not here |
| 02:09 | <&McMartin> | OS X also being naughty here |
| 02:15 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has quit [Ping timeout: 121 seconds] |
| 02:24 | <@ToxicFrog> | McMartin: what's the actual version number that fixes this? |
| 02:24 | <@ToxicFrog> | (also, git-svn supremacy~) |
| 02:25 | | Kindamoody[zZz] is now known as Kindamoody |
| 02:25 | <&McMartin> | ToxicFrog: 1.7.0 |
| 02:26 | <&McMartin> | It is now at 1.7.4 on Tortoise; all my linux distros seem to be at 1.6.17, even the one that came out like two weeks ago |
| 02:26 | <&McMartin> | 1.7 organizes the repo so that there is only one .svn directory and it's at the checkout root |
| 02:27 | <@ToxicFrog> | Yeah, even Tumbleweed is at 1.6.x |
| 02:27 | <@ToxicFrog> | Although there's probably a user-repo for 1.7 |
| 02:27 | <&McMartin> | 1.7.x has been out for a really long time -_- |
| 02:27 | <&McMartin> | It's like they have a vested interest in SVN being gratuitously awful~ |
| 02:28 | <&McMartin> | GTK, too~ |
| 02:28 | <&McMartin> | Fedora deliberately bleeding-edges GNOME |
| 02:28 | <&Derakon> | Maybe they want everyone to switch to git. |
| 02:28 | <&McMartin> | Derakon: This is my not-tremendously-subtle insinuation, yes |
| 02:29 | <&McMartin> | There was actual and genuine rage from the git community when SVN 1.6.x started serious development to fix SVN's completely broken branching mechanisms |
| 02:29 | <&McMartin> | The team that was working on it had to publish an open letter defending developing the product at all |
| 02:29 | <&McMartin> | Remember, kids, Open Source is all about Freedom (tm) |
| 02:29 | <@ToxicFrog> | Pfft |
| 02:29 | <@Tamber> | Freedom to do what you like, as long as it's what we want you to make~ |
| 02:32 | | Syloq_Home [Syloq@NetworkAdministrator.Nightstar.Net] has quit [Connection closed] |
| 02:49 | | Syloq_Home [Syloq@NetworkAdministrator.Nightstar.Net] has joined #code |
| 02:53 | | Rhamphoryncus [rhamph@Nightstar-5697f7e2.abhsia.telus.net] has joined #code |
| 03:00 | | Thalasleep is now known as Thalass |
| 04:16 | | Derakon is now known as Derakon[AFK] |
| 04:20 | | Thalass [thalass@Nightstar-40691ad9.bigpond.net.au] has quit [Ping timeout: 121 seconds] |
| 04:24 | < Rhamphoryncus> | Blah. Much easier for me to do a dot product as a giant copy and paste blob than to figure out the correct loops |
| 04:26 | < Rhamphoryncus> | hrm. Because it's 3 dimensional and should have 3 loops. 4x4 * 4x4 = 4x4x4 before reducing back down to 4x4 |
| 04:37 | | Derakon[AFK] is now known as Derakon |
| 05:00 | | Kindamoody is now known as Kindamoody|out |
| 05:43 | | celticminstrel is now known as celmin|sleep |
| 05:52 | < Rhamphoryncus> | And after a break I refactored the giant blob into loops, one loop at a time |
| 05:56 | | ErikMesoy|sleep is now known as ErikMesoy |
| 06:22 | | eckse [eckse@Nightstar-75493054.dsl.sentex.ca] has quit [Client closed the connection] |
| 06:37 | | Derakon is now known as Derakon[AFK] |
| 07:12 | | Attilla [Obsolete@Nightstar-5cf8db5d.as43234.net] has joined #code |
| 08:15 | | Attilla [Obsolete@Nightstar-5cf8db5d.as43234.net] has quit [Ping timeout: 121 seconds] |
| 08:56 | | You're now known as TheWatcher |
| 08:59 | | * Rhamphoryncus kicks GLSL |
| 09:21 | < Rhamphoryncus> | Something in the linking process is failing. My fragment shader's "in" is being ignored, even if the name doesn't match up (should be an error) |
| 10:59 | | RichyB [MyCatVerbs@Nightstar-3b2c2db2.bethere.co.uk] has joined #code |
| 12:13 | | * TheWatcher eyes MathJax |
| 12:28 | <@TheWatcher> | Either I'm doing somethign wrong, or this thing really can't handle pages with 50+ equations on it in any sane fashion |
| 12:34 | <@TheWatcher> | Welp, that's three different browsers, same behaviour: more than a handful of equations and it takes foever, and massive amounts of memory |
| 12:39 | < RichyB> | TheWatcher, ? |
| 12:42 | <@TheWatcher> | MathJax is a client-side Tex to {various formats} javascript library. |
| 12:42 | <@TheWatcher> | It's more reliable than MathML support, and gives better accessibility than straight server-side rendered images. |
| 12:43 | <@TheWatcher> | (as it supports zooming, and formats more friendly to screen readers and suchlike) |
| 12:49 | <@TheWatcher> | When there's like 5 or 6 simple equations on the page, it's decent enough, but any of our pages with equations on them tend to have an order of magnitude more than that (one here with 80 equations has been sat chewing through processor time and memory for the last 30 minutes, and still isn't anywhere near done) |
| 13:10 | < Rhamphoryncus> | sounds like page layout doesn't scale. Used to be a major problem in browsers |
| 13:21 | | Attilla [Obsolete@Nightstar-5cf8db5d.as43234.net] has joined #code |
| 13:41 | < RichyB> | TheWatcher, I don't understand what the problem is with server-side rendered images. |
| 13:42 | < RichyB> | Er, accessibility-wise. |
| 13:42 | < RichyB> | Your TeX renderer should put an alt attribute on the rendered equation. |
| 13:47 | <@TheWatcher> | And, indeed, it does |
| 13:48 | <@TheWatcher> | But apparently some people think the lack of zooming is an Issue |
| 13:48 | <@TheWatcher> | I think the lack of fully rendered page in less than an hour is a rather bigger one, but then I have such crazy demands of things. |
| 13:50 | <@TheWatcher> | (context: this is MediaWiki related, specifically the Math extension for it, which introduced MathJax support recently. I've turned it on as an option in the wikis in work, and been testing it with our course development system.) |
| 14:13 | <@rms> | Suggestion: do both; send a rendered image, have the JS take over on user-demand. |
| 14:16 | < Rhamphoryncus> | eh, if it takes that long it's simply not viable in the current implementation |
| 14:16 | < Rhamphoryncus> | 30 seconds is too long for default use |
| 14:17 | <@TheWatcher> | ^-- |
| 14:17 | <@TheWatcher> | (hell, frankly, 10 is pushing it) |
| 14:23 | <@TheWatcher> | Hm, yep, appears this is a known issue. Well, sod it for now. |
| 14:57 | | Kindamoody|out is now known as Kindamoody |
| 15:27 | <@TheWatcher> | Godsdamnit, I need a 3D whiteboard. |
| 15:28 | <@TheWatcher> | We live in the future, why do I not have one?! |
| 15:37 | < RichyB> | TheWatcher, why does it take so long anyway? |
| 15:38 | < RichyB> | It sounds like it must be doing something mad and quadratic to take ridiculously long when there's more than one per page. |
| 15:39 | < RichyB> | TheWatcher, stupid workaround, but perhaps you could put them into individual iframes so that they think they're alone and hence can't run into any quadratic case? |
| 15:39 | < RichyB> | Or is it just that they take multiple seconds to run each, and the total runtime comes out unacceptable? |
| 15:42 | <@TheWatcher> | From what I've dug up, it's a combination of both: it takes some time to convert the equation from text into its target (html+css in this case), but each step involves DOM changes, and element dimension lookups (which require page reflow) |
| 15:43 | <@TheWatcher> | the latter gets hilarious when there's a lot of equations to convert |
| 15:43 | < RichyB> | Ahhh. Oh yeah, reflows are lethal. |
| 15:44 | < RichyB> | Putting them in individual iframes would probably help a lot with that. |
| 15:44 | < RichyB> | Nothing under them, less stuff to reflow. |
| 15:45 | <@TheWatcher> | And really, it's not my problem. The Math extension devs can deal with it, I'm in no rush to remove png export, and have far mor than enough to keep me occupied elsewhere |
| 15:45 | < RichyB> | Bonus, you could add a callback when it's finished and resize the iframes to fit the size of the rendered equation. :) |
| 15:45 | < RichyB> | (render them overflowing the area, then grow the iframe) |
| 15:46 | <@TheWatcher> | +e |
| 16:10 | | celmin|sleep is now known as celticminstrel |
| 16:24 | | eckse [eckse@Nightstar-22a59c0c.dsl.sentex.ca] has joined #code |
| 16:24 | | mode/#code [+o eckse] by ChanServ |
| 16:36 | | Kindamoody [Kindamoody@Nightstar-6154a72a.tbcn.telia.com] has quit [Connection closed] |
| 16:36 | | Kindamoody|out [Kindamoody@Nightstar-6154a72a.tbcn.telia.com] has joined #code |
| 16:36 | | mode/#code [+o Kindamoody|out] by ChanServ |
| 16:37 | | Kindamoody|out is now known as Kindamoody |
| 16:47 | | ErikMesoy is now known as Harrower |
| 16:51 | | SmithKurosaki [Smith@Nightstar-354b9fb9.eng.wind.ca] has joined #code |
| 16:52 | < SmithKurosaki> | Hey guys. would anyone be able to write something I feed xmls files to and have it make one file with no duplicates? |
| 16:56 | < SmithKurosaki> | *xml files |
| 16:56 | < EvilDarkLord> | I'm converting a couple of thousand lines from windows-based c++ to something that I can compile, run and debug sans Wine. Do function definitions on the level of strcpy / memcpy / strlen all work as you'd expect or did the wizards of Redmond make any modifications to function signatures (like order of source and destination)? |
| 16:58 | | EvilDarkLord is now known as Maze |
| 17:17 | < gnolam> | Standard C is standard C. |
| 17:18 | < ShellNinja> | And Dennis Ritchie is its prophet. |
| 17:26 | | Rhamphoryncus_ [rhamph@FFAAA3.3FB6E7.D3A0EC.7759C0] has joined #code |
| 17:28 | | Rhamphoryncus_ [rhamph@FFAAA3.3FB6E7.D3A0EC.7759C0] has quit [Client exited] |
| 17:28 | < Rhamphoryncus> | stupid kde :P |
| 17:37 | | Kindamoody is now known as Kindamoody|out |
| 17:43 | < Harrower> | What was the current way to make Google do a literal search? Quote marks and + aren't cutting it. |
| 17:45 | < SmithKurosaki> | You mean doing boolean? |
| 17:45 | < SmithKurosaki> | Also, anyone god with xml? |
| 17:45 | < SmithKurosaki> | *good, but god would work too :p |
| 17:48 | < celticminstrel> | Wizards of Redmond? |
| 19:00 | <&McMartin> | Maze: the only difference, IIRC, is that a number of functions have underscores in front. |
| 19:00 | <&McMartin> | However, _strcpy_s is not the same thing as strncopy - among other things, _strcpy_s guarantees its final null, IIRC. |
| 19:01 | < Maze> | McMartin: All right, thanks. But I guess those'd show up as compilation errors, so I will be looking at them more closely anyway. |
| 19:02 | <&McMartin> | Depending on how Windows it is, you may get wxWidgets'd, but you'd have noticed a lack of main() a lot sooner. |
| 19:02 | <&McMartin> | On Windows, wchar_t is (a) only 16-bits wide, (b) secretly UTF-16LE, (c) actually used for things (mostly system calls) |
| 19:03 | < Maze> | I was worried there would be stuff there that doesn't show up as a "name not found" error, but still does something different on Windows vs Linux. |
| 19:03 | <&McMartin> | The thing most likely to burn you there is if you're using a threading library |
| 19:03 | <&McMartin> | Linux default timeslice used to be 10ms, and is now "whatever we want" since it no longer uses a timeslice-based scheduler |
| 19:03 | <&McMartin> | Windows default timeslice is 1ms |
| 19:04 | < Maze> | I don't think there's any threading going on, luckily. |
| 19:04 | <&McMartin> | Being deadlock-free on Linux actually says very little about how many race conditions your code doesn't have |
| 19:35 | <&McMartin> | Note that in neither case is the code *wrong* |
| 19:35 | < RichyB> | McMartin, IIRC I saw a paper on FreeBSD's scheduler way back when about how timeslices get adjusted. |
| 19:35 | <&McMartin> | Er, is the *OS* wrong. |
| 19:35 | <&McMartin> | It's just that when you cross operating systems, certain kinds of bugs *stop being harmless and ignorable* |
| 19:35 | <&McMartin> | My experiences with cross-platform ports that weren't *super* OS-specific to begin with have largely burned me there |
| 19:36 | < RichyB> | McMartin, interactive processes (those that tend to yield) get given short timeslices as often as possible, while hungry programs (those that eat their entire timeslice) get given their CPU time in as large blocks as possible, to maximise cache warmth. :) |
| 19:37 | < RichyB> | Back then they had (IIRC) a fixed 100Hz clock interrupt, and timeslices would vary between 10ms and 170ms. |
| 19:38 | | * McMartin nods |
| 19:38 | <&McMartin> | Linux is doing Something Wacky Now, IIRC, so that they can have an O(1) scheduler |
| 19:42 | < RichyB> | I should look into what that is, one of these days, but it's been a while since I've really tuned anything for performance so meh. |
| 20:06 | | Kindamoody|out is now known as Kindamoody |
| 20:13 | | * McMartin writes some kernel code! |
| 20:13 | <&McMartin> | This will surely be wholesome family entertainment for all concerned |
| 20:14 | <@TheWatcher> | Everybody, down! |
| 20:15 | <@TheWatcher> | (what're you writing?) |
| 20:15 | | * celticminstrel dives under a table. |
| 20:19 | | Kindamoody is now known as Kindamoody[zZz] |
| 20:20 | < RichyB> | Bet you it's a device driver. |
| 20:21 | <&McMartin> | Modifying an existing one. |
| 20:21 | <&McMartin> | it isn't compatible with certain other drivers, so my code makes it try to get out of the way if it detects that it's there. |
| 20:25 | | * McMartin throws that at the build machine, goes to get lunch. |
| 20:26 | | RichyB [MyCatVerbs@Nightstar-3b2c2db2.bethere.co.uk] has quit [[NS] Quit: Leaving] |
| 20:30 | < Rhamphoryncus> | Harrower: "more search tools" at the bottom of the left column reveals a verbatim option |
| 20:30 | < Rhamphoryncus> | But you have to reapply it every time |
| 20:30 | < Rhamphoryncus> | I did try a greasemonkey script to make it automatic but it just cocked everything up |
| 20:43 | | Harrower is now known as ErikMesoy |
| 20:43 | | Maze is now known as EvilDarkLord |
| 20:46 | | Attilla [Obsolete@Nightstar-5cf8db5d.as43234.net] has quit [Ping timeout: 121 seconds] |
| 20:49 | | Attilla [Obsolete@Nightstar-aef331e5.as43234.net] has joined #code |
| 20:53 | | Noah [maoranma@Nightstar-0b54271b.pools.spcsdns.net] has quit [Ping timeout: 121 seconds] |
| 20:54 | | maoranma [maoranma@Nightstar-0b54271b.pools.spcsdns.net] has joined #code |
| 21:03 | <@ToxicFrog> | SmithKurosaki: question is underspecified. No duplicate whats? |
| 21:03 | < SmithKurosaki> | entries |
| 21:04 | < SmithKurosaki> | I'm trying to take all my eve fitting files and put them into one file so I can change to pyfa from evehq since my os wont support the new version |
| 21:04 | < SmithKurosaki> | But, pyfa doesnt filter out doubles properly |
| 21:06 | < SmithKurosaki> | And it will only import one xml file properly at a time |
| 21:07 | < SmithKurosaki> | SO i need something to feed my xml fitting files to and end up with a unified file with no duplicate fittings |
| 21:08 | < SmithKurosaki> | Does that make sense or are you still missing something? |
| 21:10 | <@TheWatcher> | My first reaction would be to load all the files into one hash, which should filter out the duplicates, and then save that hash out... but *shrug* |
| 21:12 | < SmithKurosaki> | I know nothing about xml, so I came here |
| 22:06 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has quit [Ping timeout: 121 seconds] |
| 22:06 | | celticminstrel [celticminst@Nightstar-5d22ab1d.cable.rogers.com] has quit [Connection reset by peer] |
| 22:06 | | celticminstrel [celticminst@Nightstar-5d22ab1d.cable.rogers.com] has joined #code |
| 22:12 | | ErikMesoy is now known as ErikMesoy|sleep |
| 22:14 | | Rhamphoryncus [rhamph@Nightstar-5697f7e2.abhsia.telus.net] has quit [Client exited] |
| 22:42 | <&McMartin> | XML processing is kind of a pain for that sort of thing, though TheWatcher's first thought matches mine |
| 22:42 | <&McMartin> | But that's as someone who's used to dealing with XML-reading libraries, which are pretty nontrivial |
| 22:43 | <&McMartin> | There might be some cunning way of doing it with XSLT |
| 23:34 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 23:34 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 23:36 | | himi [fow035@D741F1.243F35.CADC30.81D435] has joined #code |
| 23:36 | | mode/#code [+o himi] by ChanServ |
| 23:42 | | You're now known as TheWatcher[T-2] |
| 23:48 | | You're now known as TheWatcher[zZzZ] |
| 23:58 | | Attilla [Obsolete@Nightstar-aef331e5.as43234.net] has quit [[NS] Quit: ] |
| --- Log closed Wed May 09 00:00:44 2012 |