| --- Log opened Fri Dec 19 00:00:48 2014 |
| 00:23 | <@celticminstrel> | For speed, he copy-pasted a function, renamed it, and cut out parts of it. |
| 00:41 | | Derakon[AFK] is now known as Derakon |
| 03:25 | | Stumblinbear [Stumblinbea@Nightstar-jd9.012.239.67.IP] has joined #code |
| 03:26 | | Stumblinbear [Stumblinbea@Nightstar-jd9.012.239.67.IP] has quit [[NS] Quit: Leaving] |
| 03:29 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 03:29 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 03:31 | | simon_ [simon@Nightstar-2og823.pronoia.dk] has joined #code |
| 03:31 | < simon_> | hellos. |
| 03:34 | < simon_> | here's an idea: I have a set of elements with corresponding weights and quantities, e.g. [(Item,Weight,Count)], or [(ball,2,3),(lamp,3,3),(bottle,1,4)]. I am looking for an efficient, purely functional datastructure to host this set of things such that I can draw one item at randomness given those weights and subtract one from its count, eventually making it impossible to draw. |
| 03:34 | < simon_> | an inefficient solution is just a list. draw a random number between 0 and the sum of all Counts-1 and recurse the list, subtract the item from the list and return it. |
| 03:35 | < simon_> | an efficient solution would probably rely on a type of binary search tree that covers intervals rather than individual items. has anyone heard of such a thing? I see a Wikipedia article on something called an Interval tree, but the use-case is slightly different. |
| 03:36 | < simon_> | the immediate problem with such a tree is the rebalancing. |
| 03:37 | < simon_> | maybe it doesn't really need to rebalance. |
| 03:40 | <~Vornicus> | The list is pretty good. |
| 03:41 | <~Vornicus> | I mean - you have to reweigh it every time to get a good random number, but |
| 03:41 | < simon_> | yeah, but when the multiset/bag grows in number of different elements, the lookup grows. |
| 03:41 | | Reiver [quassel@Nightstar-ksqup0.co.uk] has quit [Ping timeout: 121 seconds] |
| 03:41 | <~Vornicus> | This isn't something you typically get to to anything about |
| 03:41 | <&Derakon> | I rather suspect that any kind of "jumping" search will have the same big-O as a linear search. |
| 03:42 | <&Derakon> | Because the intervals are uneven and potentially antagonistic. |
| 03:42 | <~Vornicus> | You could treeify but that involves *actually figuring out the tree* |
| 03:42 | <~Vornicus> | and that means figuring out the - ever changing - split points |
| 03:42 | < simon_> | if I have a BST that covers an interval, I'd get O(lg n) assuming that the BST remains somewhat balanced. |
| 03:42 | <&Derakon> | You're trading a ton of complexity for a minor and possibly nonexistent efficiency gain. |
| 03:42 | < simon_> | yeah |
| 03:42 | < simon_> | totally. |
| 03:43 | < simon_> | I just thought it was interesting to have a BST cover an interval of things... I haven't seen that before. |
| 03:43 | | Reiver [quassel@Nightstar-ksqup0.co.uk] has joined #code |
| 03:43 | | mode/#code [+ao Reiver Reiver] by ChanServ |
| 03:49 | <~Vornicus> | I guess all told it's not *too* terrible to have a binary tree, but doing the rebalance would be silly |
| 03:58 | <@celticminstrel> | I wish the debugger would stop breaking. |
| 04:04 | <@celticminstrel> | Whee, I just used one breakpoint to enable another one! |
| 04:04 | <@celticminstrel> | So that the first doesn't trigger until the second has been hit once. |
| 04:07 | <@celticminstrel> | It still shows as disabled in XCode though. |
| 04:13 | <@celticminstrel> | Also it seems that if it's disabled when the debugger starts, XCode doesn't even tell the debugger about it. |
| 04:13 | <@celticminstrel> | I guess that makes sense though. |
| 04:15 | | * celticminstrel is currently trying to make the text wrapping deal with words that are longer than the available space. |
| 04:19 | | Netsplit *.net <-> *.split quits: @iospace, abudhabi, @Syloq, Xires, Attilla, grindhold, @Checkmate, @Reiver, @Orthia, VirusJTG, (+18 more, use /NETSPLIT to show all of them) |
| 04:19 | | Netsplit over, joins: &Reiver, &jerith, @Orthia, &ToxicFrog, VirusJTG, &jeroud, @JustBob, @Checkmate, @gnolam, @iospace (+18 more) |
| 05:06 | | Derakon is now known as Derakon[AFK] |
| 05:10 | | Checkmate [Z@Nightstar-484uip.cust.comxnet.dk] has quit [Ping timeout: 121 seconds] |
| 05:26 | <@celticminstrel> | Trying to think if I've missed any important text-editing key controls... |
| 05:30 | <~Vornicus> | load save undo redo find replace cut copy paste gotoline -- the barest basics |
| 05:31 | <~Vornicus> | oh, and select, end&beginning of word/line/file |
| 05:32 | <@celticminstrel> | Hm. This is for an edit text field, so load and save aren't needed... not sure if there's any point in find, replace, gotoline. I accounted for all the others. (Shift plus any navigation key for select.) |
| 05:33 | <@celticminstrel> | I also have select all and delete word. |
| 05:34 | <@celticminstrel> | (They aren't all implemented yet, but I've considered them and written the code to detect them.) |
| 05:55 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has quit [[NS] Quit: Leaving] |
| 06:03 | | himi [fow035@Nightstar-dm0.2ni.203.150.IP] has quit [Ping timeout: 121 seconds] |
| 06:31 | | Kindamoody[zZz] is now known as Kindamoody |
| 06:36 | <@celticminstrel> | On a tangentially related note, it would be nice if text editors were consistent on whether shift+arrow extends or shrinks the selection. |
| 06:52 | <@celticminstrel> | Hah, integration with Mac clipboard successful. |
| 07:04 | | Turaiel is now known as Turaiel[Offline] |
| 07:17 | | Xires is now known as ^Xires |
| 07:26 | | Kindamoody [Kindamoody@Nightstar-180u8i.tbcn.telia.com] has quit [Ping timeout: 121 seconds] |
| 07:31 | | ^Xires is now known as Xires |
| 07:41 | | Kindamoody [Kindamoody@Nightstar-180u8i.tbcn.telia.com] has joined #code |
| 07:41 | | mode/#code [+o Kindamoody] by ChanServ |
| 08:14 | <@celticminstrel> | The Edit menu is stealing my keyboard shortcuts. :/ |
| 08:14 | <@celticminstrel> | And without actually calling the relevant menuitem/ |
| 08:14 | <@celticminstrel> | ^. |
| 08:21 | <@celticminstrel> | Well, by setting the menu to autoenable, I can at least get the keyboard shortcuts to work with a modal dialog onscreen. |
| 08:22 | <@celticminstrel> | The only problem is the menuitems really shouldn't be disabled when pressing their key equivalents is expected to do something. |
| 08:23 | <@celticminstrel> | But I think it's better than it not working at all. |
| 08:31 | | Orthia [orthianz@Nightstar-obo.ne6.184.203.IP] has quit [Ping timeout: 121 seconds] |
| 08:36 | | Orthia [orthianz@Nightstar-lca.o7s.224.119.IP] has joined #code |
| 08:36 | | mode/#code [+o Orthia] by ChanServ |
| 08:44 | | celticminstrel [celticminst@Nightstar-jeiu0g.dsl.bell.ca] has quit [[NS] Quit: KABOOM! It seems that I have exploded. Please wait while I reinstall the universe.] |
| 08:54 | | himi [fow035@Nightstar-v37cpe.internode.on.net] has joined #code |
| 08:54 | | mode/#code [+o himi] by ChanServ |
| 10:22 | | Kindamoody is now known as Kindamoody|afk |
| 12:26 | < abudhabi> | Hmmm. Fast box. It takes only a second or so for an errant loop to fill up the Java heap with integers. |
| 12:39 | <@Tarinaky> | Remind me, I think I'm doing something stupid. |
| 12:39 | <@Tarinaky> | In C++, can you have function calls in the body of a unit of translation, i.e. to populate some global? |
| 12:39 | <@Tarinaky> | *global state |
| 12:41 | < abudhabi> | Unit of translation? |
| 12:41 | <@Tarinaky> | .cpp file |
| 12:41 | < abudhabi> | I don't see why not? |
| 12:42 | <@Tarinaky> | It appears to be getting confused and thinking it's a function declaration not a call |
| 12:44 | < abudhabi> | Are you trying to call a function directly from the top-level of the document? |
| 12:44 | < abudhabi> | I mean, outside of a function? |
| 12:44 | <@Tarinaky> | Yes. |
| 12:44 | < abudhabi> | Because that, I think, you can't do. You can declare and assign, but not call functions. |
| 12:44 | | EvilDarkLord [jjlehto3@Nightstar-evu5hu.org.aalto.fi] has quit [Operation timed out] |
| 12:44 | <@Tarinaky> | I think I might have been writing too much Ruby |
| 12:44 | < abudhabi> | Best do it in main(). |
| 12:44 | < abudhabi> | Or a dedicated setup function. |
| 12:46 | | Kindamoody|afk [Kindamoody@Nightstar-180u8i.tbcn.telia.com] has quit [Ping timeout: 121 seconds] |
| 12:47 | | EvilDarkLord [jjlehto3@Nightstar-evu5hu.org.aalto.fi] has joined #code |
| 13:04 | <@Azash> | āTo make matters worse, we saw distinct traces of local mist around several of the Puppet developersā |
| 13:05 | | * Tarinaky headdesks. |
| 13:05 | <@Tarinaky> | Why are DLLs so smegging hard to use :/ |
| 13:09 | <&McMartin> | Two reasons: (a) their init/deinit order is not fixed, much less topologically sorted by dependencies, (b) you're trying to use a compiled language like a scripting one, (c) if these are Windows DLLs, symbol table management is a pain |
| 13:10 | <&McMartin> | Three, three reasons |
| 13:10 | <&McMartin> | among the various reasons |
| 13:10 | <@Tarinaky> | I'm trying to import some functions from a DLL I wrote a month ago. |
| 13:11 | <&McMartin> | Did you do the WHATEVER_EXPORTS #define trick? |
| 13:11 | <@Tarinaky> | Yes, but it's giving me unresolved external symbol for the functions that are defined in the DLL |
| 13:11 | <@Azash> | 14:10 <&McMartin> Three, three reasons < Ah, ah, ah, ah |
| 13:11 | <@Tarinaky> | I know the DLL 'works' because I can access the functions with Ruby and FFI |
| 13:12 | <&McMartin> | Which compiler? |
| 13:12 | <@Tarinaky> | Microsoft Visual Studio's C++ compiler... cl.exe I think? |
| 13:12 | <&McMartin> | OK |
| 13:13 | <&McMartin> | you are both (a) linking the associated .lib file and (b) ensuring that when the *client* uses it those functions are marked __declspec(dllimport) instead of __declspec(dllexport) like when you wrote it, right |
| 13:13 | <&McMartin> | The (b) thing being what WHATEVER_EXPORTS is designed to hide |
| 13:13 | <@Tarinaky> | (b) Yes. |
| 13:13 | <@Tarinaky> | (a) I /think/ so, but it's all hidden behind GUIs so I don't actually know what VS is doing |
| 13:15 | <&McMartin> | This one is not; there should be a property named something like "additional library dependencies" under Linker/Input if this DLL isn't part of the stuff you build when you hit Build, or under "Project References" if it *is* part of that stuff. |
| 13:15 | <@Tarinaky> | This DLL is part of the stuff build when I hit build. |
| 13:15 | <&McMartin> | OK. Make sure it's listed under References in the config. |
| 13:16 | <&McMartin> | That should suffice to have the .lib show up at link time (which you can check by going to Properties -> Linker -> Command line) |
| 13:16 | <@Tarinaky> | I'm not seeing it. |
| 13:16 | <&McMartin> | Add it |
| 13:17 | <@Tarinaky> | Cannot open input file. |
| 13:19 | <&McMartin> | Are you *sure* you're building it? |
| 13:19 | <&McMartin> | Now you've got "I'm trying to link in the code that calls the DLL but it isn't there" |
| 13:19 | <@Tarinaky> | Well the compiler output says it's building it. |
| 13:20 | <@Tarinaky> | And the actual .lib file is in the Release folder after a failed build. |
| 13:20 | <@Tarinaky> | I'm not sure I'm adding the lib to the dependencies in the right way though |
| 13:20 | <@Tarinaky> | brb a sec. |
| 13:21 | <&McMartin> | Ah |
| 13:21 | <&McMartin> | Yeah, you don't want to add it to "additional input libraries"; you want to add it to "Project references" which is actually in the other top-level properties section |
| 13:21 | <&McMartin> | (Assuming this is MSVS2013. If it's 2008, you have to go to a different dialog called "project dependencies" and check the box for the other project) |
| 13:24 | <@Tarinaky> | 2012 I think |
| 13:24 | <@Tarinaky> | Wait so... |
| 13:24 | <@Tarinaky> | What the fuck am I doing? |
| 13:28 | <@Tarinaky> | I cant find anything for either additional inpuit libraries or project references |
| 13:28 | | * Tarinaky is totally lost. |
| 13:31 | <&McMartin> | It should look vaguely like this: http://blogs.msdn.com/blogfiles/visualstudio/WindowsLiveWriter/FrameworkMultiTar getingforVCProjects_E3BC/clip_image002_2.jpg |
| 13:31 | <&McMartin> | (But not exactly, those are C# things) |
| 13:33 | <@Tarinaky> | Okay, I have that dialog open. |
| 13:33 | <@Tarinaky> | Then it's somewhere under Linker? |
| 13:33 | <&McMartin> | No, it's up at the very very top above everything |
| 13:34 | <@Tarinaky> | Above the dialog? |
| 13:34 | <@Tarinaky> | o.O |
| 13:34 | <&McMartin> | Er, no |
| 13:34 | <&McMartin> | There's a list of stuff on the left part of the dialog |
| 13:34 | <@Tarinaky> | Yes. |
| 13:34 | <&McMartin> | The "Framework and References" option is near the top of that list. |
| 13:35 | <@Tarinaky> | Ohhhh |
| 13:35 | | * Tarinaky derps. |
| 13:35 | <&McMartin> | You need to go to it (or whatever the C++ name is, I forget) and select "Add Reference" or similar, and then add a reference to the project that produces the DLL you want to link in. |
| 13:35 | <@Tarinaky> | Still getting unresolved external symbol. |
| 13:36 | <&McMartin> | I'm sort of running out of useful things I can do from a distance, so here's the MSDN walkthrough: http://msdn.microsoft.com/en-us/library/ms235636.aspx |
| 14:18 | | Checkmate [Z@Nightstar-484uip.cust.comxnet.dk] has joined #code |
| 14:18 | | mode/#code [+o Checkmate] by ChanServ |
| 15:32 | <@Tarinaky> | Update. |
| 15:32 | <@Tarinaky> | I needed to be using extern "C" __declspec(dllimport) and not _declspec(dllimport) |
| 15:52 | <@gnolam> | http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=690b05 43a813b0ecfc51b0374c0ce6c8275435f0 |
| 15:56 | <@Tarinaky> | Best patch ever. |
| 16:12 | <&McMartin> | Tarinaky: Ah, yes. I had missed that the library itself was written in C (or declared extern C). C++ symbols for function names include type info in them because foo(int x) and foo(double y) are different functions in C++ and can coexist. |
| 16:13 | <&McMartin> | gnolam: hooray |
| 16:23 | | macdjord|slep [macdjord@Nightstar-r9vt2h.mc.videotron.ca] has quit [[NS] Quit: Tekeli-li! Tekeli-li!] |
| 16:33 | <@Tarinaky> | McMartin: I hadn't mentioned that because I didn't realise either. |
| 16:33 | <@Tarinaky> | Coworker spotted it |
| 16:33 | | * McMartin nods |
| 17:38 | | Reiver [quassel@Nightstar-ksqup0.co.uk] has quit [Connection closed] |
| --- Log closed Fri Dec 19 17:38:51 2014 |
| --- Log opened Fri Dec 19 17:39:02 2014 |
| 17:39 | | TheWatcher [chris@Nightstar-ksqup0.co.uk] has joined #code |
| 17:39 | | Irssi: #code: Total of 35 nicks [16 ops, 0 halfops, 0 voices, 19 normal] |
| 17:39 | | mode/#code [+o TheWatcher] by ChanServ |
| 17:39 | | Irssi: Join to #code was synced in 38 secs |
| 17:39 | | Reiver [quassel@Nightstar-ksqup0.co.uk] has joined #code |
| 17:39 | | mode/#code [+ao Reiver Reiver] by ChanServ |
| 17:46 | <&McMartin> | Incidentally, if you use git on Windows or Mac there's a security flaw that was patched yesterday. You should upgrade to 2.2.1 or 1.9.5. |
| 17:46 | | celticminstrel [celticminst@Nightstar-jeiu0g.dsl.bell.ca] has joined #code |
| 17:46 | | mode/#code [+o celticminstrel] by ChanServ |
| 17:47 | <&McMartin> | http://www.securityweek.com/apple-microsoft-github-release-updates-fix-critical- git-vulnerability |
| 17:50 | <@celticminstrel> | Well, I can update git, but not XCode. |
| 18:15 | | Alek [omegaboot@Nightstar-c8t.a00.36.73.IP] has quit [Operation timed out] |
| 20:00 | | Alek [omegaboot@Nightstar-c8t.a00.36.73.IP] has joined #code |
| 20:00 | | mode/#code [+o Alek] by ChanServ |
| 20:48 | <@celticminstrel> | ...so I now have two layers of binding. |
| 20:48 | <@Tamber> | kinky. |
| 20:48 | <@celticminstrel> | ie, I call std::bind, get a function out of it, then pass that into std::bind again. |
| 21:06 | | Kindamoody|autojoin [Kindamoody@Nightstar-180u8i.tbcn.telia.com] has joined #code |
| 21:06 | | mode/#code [+o Kindamoody|autojoin] by ChanServ |
| 21:06 | | Kindamoody|autojoin is now known as Kindamoody |
| 21:10 | <&McMartin> | That's dangerously close to currying |
| 21:12 | <@celticminstrel> | Dangerously? |
| 21:17 | <&McMartin> | It's the adverb I'm used to applying to close |
| 21:18 | <@celticminstrel> | But that makes it sound dangerous! |
| 21:19 | <&McMartin> | Currying *is* dangerous. Once you've gotten to use it, every language that doesn't have it is a thorn in your side~ |
| 21:19 | <@celticminstrel> | XD |
| 21:21 | <&McMartin> | (I suppose if I'm picking nits, it's not the curry that is so addictively delicious but instead the ability for partial evaluation, which of course std::bind totally is) |
| 21:22 | <@celticminstrel> | By which you mean when you bind some parameters but not others? |
| 21:22 | <&McMartin> | And get a callable back out of that, yeah |
| 21:23 | <&McMartin> | This gets completely nuts when you're using a language that makes heavy use of higher-order functions (Haskell, Clojure, the ML family) but it's Not Shabby in C++ either |
| 21:23 | <&McMartin> | I've been known to write my C++ "with a functional accent" though |
| 21:26 | | * celticminstrel likes the new lambdas. |
| 21:29 | <&McMartin> | I haven't looked to closely at the new syntaxes |
| 21:29 | <&McMartin> | *too |
| 21:30 | <@celticminstrel> | One major difference from other languages I use with lambdas (like Python, JavaScript, or Java) is that the closure has to be specified explicitly. |
| 21:30 | | * celticminstrel is counting anonymous inner classes as lambdas here. |
| 21:55 | | * McMartin nods |
| 22:02 | | Kindamoody is now known as Kindamoody[zZz] |
| 23:34 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 23:34 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| --- Log closed Sat Dec 20 00:00:02 2014 |