| --- Log opened Wed Jun 19 00:00:29 2013 |
| 00:03 | | Netsplit *.net <-> *.split quits: @Reiv, @jerith, @PinkFreud |
| 00:04 | | Netsplit over, joins: @PinkFreud, &jerith, @Reiv |
| 00:19 | | Derakon is now known as Derakon[AFK] |
| 00:24 | | Harlow [Harlow@8E7DA3.5F0415.4D5E13.93A5E9] has joined #code |
| 00:55 | | himi [fow035@D741F1.243F35.CADC30.81D435] has joined #code |
| 00:55 | | mode/#code [+o himi] by ChanServ |
| 01:07 | | Derakon[AFK] is now known as Derakon |
| 01:20 | | Orthia [orthianz@3CF3A5.E1CD01.B089B9.1E14D1] has joined #code |
| 01:20 | | mode/#code [+o Orthia] by ChanServ |
| 01:32 | | Harlow [Harlow@8E7DA3.5F0415.4D5E13.93A5E9] has quit [[NS] Quit: Leaving] |
| 01:40 | | ktemkin[avol] is now known as ktemkin[work] |
| 01:52 | | * McMartin wtfs at his laptop, which is signaling his program at various points where malloc() is called inside Xlib or libpulse. |
| 01:52 | <&McMartin> | Which then hard-locks it because SDL eats its own tail |
| 01:53 | <@Azash> | What kind of signal? |
| 01:53 | <&McMartin> | I can't tell from the backtrace |
| 01:53 | <&McMartin> | It just says "<signal caught>" |
| 01:54 | <@Azash> | Hm |
| 01:54 | <~Vornicus> | cute. |
| 01:54 | <&McMartin> | It's Mix_CloseAudio -> SDL_QuitSubsystem -> SDL_AudioQuit -> pa_xrealloc -> malloc -> <signal> -> SDL_Quit -> SDL_QuitSubsystem -> SDL_AudioQuit -> hang |
| 01:55 | <&McMartin> | Alternately, I get that from a malloc that happens inside Xlib on calls to SDL_Flip() |
| 01:55 | <&McMartin> | And this hangs it so hard that it needs to be kill -9'd, *and* gdb can only attach to it if I sudo gdb. |
| 01:56 | | * McMartin powers down that laptop for a bit |
| 02:02 | <&McMartin> | Hm, yeah. After a nap it seems happy again |
| 02:02 | <&McMartin> | wtf |
| 02:02 | <@Azash> | RAM corruption? |
| 02:02 | <&McMartin> | Could be |
| 02:03 | <&McMartin> | Or it's running too hot |
| 02:06 | | RichyB [RichyB@D553D1.68E9F7.02BB7C.3AF784] has quit [[NS] Quit: Gone.] |
| 02:08 | <@Azash> | Could overheating really cause that, though? |
| 02:09 | <&ToxicFrog> | Yes. |
| 02:09 | | RichyB [RichyB@D553D1.68E9F7.02BB7C.3AF784] has joined #code |
| 02:09 | <&McMartin> | The ventilation wasn't great, and as summer wears on that becomes a bigger deal |
| 02:09 | <&McMartin> | I should get one of those bed-desks |
| 02:11 | <@Azash> | Huh, I didn't know that |
| 02:12 | <@Azash> | Well, I've clocked in 0x29 hours now, so.. Good night! |
| 02:46 | | Typherix is now known as Typh|offline |
| 02:51 | | Kindamoody[zZz] is now known as Kindamoody |
| 03:14 | | Typh|offline is now known as Typherix |
| 03:26 | <~Vornicus> | Herp. Okay, have a method that can't be allowed to be const -- it's operator[], and I need to be able to use its results as an lvalue; however, it also gets used on const objects (as an rvalue). This is problematic. |
| 03:26 | <&McMartin> | You can define it twice, once const, one not. |
| 03:26 | <&McMartin> | const-ness is part of the signature and you can do overloads. |
| 03:27 | < RichyB> | Really? Handy. |
| 03:27 | <&McMartin> | Pretty sure std::vector and friends all do this, in exactly this use case |
| 03:28 | <~Vornicus> | So it'd be const int& operator[](...) and int& operator[](...)? |
| 03:29 | <&McMartin> | int& operator[](...) and const int& operator[](...) const |
| 03:29 | <&McMartin> | That last const is the marker that "this is const in this function" |
| 03:29 | <~Vornicus> | You can put that "const" word in a lot of places. |
| 03:29 | <&McMartin> | C++ had a mortal fear of introducing new keywords, so it recycles C's. |
| 03:29 | <~Vornicus> | What's the one at the start do? |
| 03:30 | < ktemkin[work]> | Modifies the return value. |
| 03:30 | <&McMartin> | Modifies the return type |
| 03:30 | < ktemkin[work]> | I guess I should have said "return value's type". |
| 03:30 | <&McMartin> | There isn't really a reason to return const int& instead of just int, though. |
| 03:31 | <~Vornicus> | Okay so it'd be "int operator[](...) const" and "int& operator[](...)" |
| 03:31 | <&McMartin> | Yeah, that should work |
| 03:31 | <&McMartin> | And in the former, "this" will be a constant pointer to your class. |
| 03:31 | <&McMartin> | So constness should be neatly enforced. |
| 03:41 | < ktemkin[work]> | How did I go this long without knowing how Bash's tilde expansions really worked? |
| 03:41 | < ktemkin[work]> | I had just always assumed ~ was shorthand for the user's home directory; I didn't know you could do ~ENVVAR. |
| 03:42 | <&McMartin> | ~ENVVAR? I thought it was ~otheruser. |
| 03:42 | <&McMartin> | What does ~ENVVAR do? |
| 03:42 | < ktemkin[work]> | If you have an environment variable ENVVAR that contains a path, ~ENVVAR is equivalent to that path. |
| 03:43 | < ktemkin[work]> | So you could do CD ~ENVVAR; and it'd be quivalent to CD $ENVVAR. |
| 03:43 | < RichyB> | uh |
| 03:43 | < RichyB> | That's not standard Bourne shell semantics |
| 03:44 | < ktemkin[work]> | Ah; it looks like that's a zsh thing; not a Bash thing. |
| 03:44 | < ktemkin[work]> | </wrong> |
| 03:48 | < ktemkin[work]> | Thinking about it more, I can't see why that'd be /useful/, other than giving a succinct way of expressing it in prompt. |
| 04:08 | | ktemkin[work] is now known as ktemkin |
| 04:12 | <~Vornicus> | okay having learned how I'm supposed to be using const now I have to go hunting to see what else I'm up to. |
| 04:18 | | * Vornicus tries to use initializer list on a fixed-size array in a class, fails miserably on both attempts. |
| 04:21 | <~Vornicus> | I tried x[0](a), x[1](b), x[2](c), and x({a, b, c}) and neither worked, so looks like I'm again in a position of either not being able to guess the right syntax, or just not being able to do it at all. |
| 04:24 | | Kindamoody is now known as Kindamoody[zZz] |
| 04:25 | < Syka> | https://pbs.twimg.com/media/BNELF1GCUAExynU.png:large -- "What the hell have you built." |
| 04:27 | <&Derakon> | Vorn: I wouldn't be surprised if that's something you just can't do. |
| 04:36 | <~Vornicus> | No fun. |
| 04:38 | <~Vornicus> | yay, no more compile errors because I cannot into C |
| 04:38 | <~Vornicus> | nap. |
| 05:21 | | Orthia [orthianz@3CF3A5.E1CD01.B089B9.1E14D1] has quit [Ping timeout: 121 seconds] |
| 05:45 | <~Vornicus> | denap. |
| 05:57 | | Typherix is now known as Typh|offline |
| 06:14 | | Derakon is now known as Derakon[AFK] |
| 06:16 | | ErikMesoy|sleep is now known as ErikMesoy |
| 06:28 | | Kindamoody[zZz] is now known as Kindamoody |
| 06:36 | | Turaiel is now known as Turaiel[Offline] |
| 07:03 | | Orthia [orthianz@3CF3A5.E1CD01.B089B9.1E14D1] has joined #code |
| 07:03 | | mode/#code [+o Orthia] by ChanServ |
| 07:05 | <&McMartin> | Dang |
| 07:05 | <&McMartin> | Remote Desktop at shitty DSL speeds really has no right to be this good |
| 07:06 | <~Vornicus> | Every time I've played with remote desktop I've been very impressed at its performance |
| 07:10 | <~Vornicus> | Ah, /finally/ the static typing gets around to coming home to roost, making my code bigger by its existence. |
| 07:11 | <&McMartin> | C++ iterator type names are The Worst |
| 07:11 | <~Vornicus> | (as opposed to just making me describe the data types first) |
| 07:12 | <~Vornicus> | eh. typedef defeats all |
| 07:12 | <&McMartin> | If you're replicating code with just the typenames changed, you may want to consider templates. |
| 07:12 | <&McMartin> | Or not, since templates are a pain in the ass and you're still kind of new at this~ |
| 07:12 | <~Vornicus> | Well, I do have, uh, two template functions, but what it's actually doing here is forcing me to construct pieces before the full object. |
| 07:13 | <~Vornicus> | or rather, I do have two functions where templates will help. |
| 07:14 | <~Vornicus> | but that's a little bit later. |
| 07:17 | <&McMartin> | One of C++'s actively weirder bits is the way it lets you have unboxed objects |
| 07:17 | <&McMartin> | Which does tend to put in super-constrained ordering restraints. If you want to break those, consider using pointer members under the hood. |
| 07:18 | <&McMartin> | But if you can avoid it, and stick to internal value types, it's often harder to fuck up the unboxed ones. |
| 07:18 | <~Vornicus> | 'unboxed objects'? |
| 07:19 | <~Vornicus> | (you're talking to a guy who's worked in dynamic languages basically exclusively for 7 years -- the only reference point I have for 'boxed' is what java does to basic types to put them on the heap) |
| 07:34 | <&McMartin> | Right; that's the opposite of that. These are treating the objects as if they were basic types that you can just jam onto the stack or into the middle of some other structure |
| 07:35 | <&McMartin> | So there are strong requirements on when exactly construction and destruction must be done as a result of that |
| 07:49 | < RichyB> | Vornicus: I'm letting a hobby horse get ahead of me here, but please do not judge all static type systems by C++. There do exist some much nicer ones which involve far fewer explicit type names laboriously tapped out everywhere. |
| 07:50 | <~Vornicus> | That's not really the issue. |
| 07:52 | | celticminstrel [celticminst@Nightstar-e83b3651.cable.rogers.com] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 07:52 | <~Vornicus> | it's more like, this is now complicated enough that it's no longer possible to do on one line the kind of bullshit I pulled in Python taking advantage of all the crazy type literals. |
| 08:17 | | himi [fow035@D741F1.243F35.CADC30.81D435] has quit [Ping timeout: 121 seconds] |
| 08:46 | <&McMartin> | Blargh |
| 08:46 | <&McMartin> | Heap corruption bugs! |
| 08:46 | <&McMartin> | How I haven't missed you |
| 08:46 | <&McMartin> | Those malloc failures from this morning, revealed as secondary damage. |
| 08:47 | | * Syka adds some bugs to the heap |
| 08:47 | < Syka> | like needles in a haystack |
| 08:48 | < Syka> | don't know theyre there until you dive in and suddenly you've got a rusty needle in your spleen |
| 08:49 | | * McMartin did a revision tree search, found the faulty commit |
| 08:49 | <&McMartin> | It's been Busted (tm) |
| 08:49 | <&McMartin> | I also found a non-fatal memory leak~ |
| 08:49 | < Syka> | heapbusted |
| 08:52 | <&McMartin> | Yeah |
| 08:52 | <&McMartin> | Sorted it out, I think |
| 08:52 | <&McMartin> | As part of loading PNGs I was instructing SDL to start freeing random integers that could be valid pointers but weren't~ |
| 08:53 | <~Vornicus> | |
| 08:54 | <@TheWatcher> | Wups |
| 09:00 | <&McMartin> | https://github.com/michaelcmartin/monocle/commit/0737dd30775a440497e1c54b2a24aa5 8c7bcc404 |
| 09:02 | <&McMartin> | OK, just ran the earthball demo 20 times in a row with no issues |
| 09:05 | | * TheWatcher eyes this code |
| 09:06 | <@TheWatcher> | On the other hand, I'm apparently reliably killing dromed trying to pull stim intensities out of messages. And I have no idea why. |
| 09:07 | <&McMartin> | Dromed is the Thief level editor? |
| 09:07 | <@TheWatcher> | Indeed so. |
| 09:15 | | ktemkin is now known as ktemkin[awol] |
| 09:20 | | Kindamoody is now known as Kindamoody|out |
| 09:24 | | Typh|offline is now known as Typherix |
| 09:33 | | Typherix is now known as Typh|offline |
| 09:50 | <~Vornicus> | whee, terrible math |
| 09:52 | <&McMartin> | What is it you're unleashing with this C++ project? |
| 09:53 | <~Vornicus> | You may remember some time back I was working on some brute force optimization problem. |
| 09:55 | <~Vornicus> | My work in Python turned out to be not nearly fast enough. |
| 09:55 | <&McMartin> | Ah |
| 09:56 | <&McMartin> | I had in fact forgotten that |
| 09:56 | <~Vornicus> | So now I'm working on it in C++, and in the process discovering that about 95% of the work is remembering how C++ works~ |
| 09:56 | <~Vornicus> | ...more like 98% |
| 10:23 | <&McMartin> | https://twitter.com/codinghorror/status/347070841059692545/photo/1 |
| 10:28 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has joined #code |
| 10:28 | | mode/#code [+o himi] by ChanServ |
| 10:36 | | * TheWatcher ...s at this |
| 10:43 | <~Vornicus> | Anyway now I'm finally getting into the proper meat of the problem itself |
| 10:44 | <@TheWatcher> | Well, that's interesting |
| 10:48 | <@TheWatcher> | This is crashing trying to call GetName() on messages, even though GetName should be perfectly safe to call |
| 10:50 | <@TheWatcher> | (Also, I should actually be working on, y'know, work... but this is bugging the hell out of me) |
| 11:12 | | Typh|offline is now known as Typherix |
| 11:21 | | Typherix is now known as Typh|offline |
| 13:00 | | Typh|offline is now known as Typherix |
| 13:09 | | Typherix is now known as Typh|offline |
| 13:15 | | Typh|offline is now known as Typherix |
| 13:53 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has quit [[NS] Quit: Leaving] |
| 14:20 | | * McMartin rocks out, does some work while his brain still functions. |
| 14:27 | | celticminstrel [celticminst@Nightstar-e83b3651.cable.rogers.com] has joined #code |
| 14:27 | | mode/#code [+o celticminstrel] by ChanServ |
| 14:31 | <@Azash> | McMartin: At least you figured that bug out |
| 14:32 | <&McMartin> | It's a relief to know it's my fault and not some kind of implausibly selective hardware failure... |
| 14:32 | <&McMartin> | ... and also to know that it isn't a library I'm relying on being even more unreliable than I had feared |
| 14:34 | <&McMartin> | It was the audio routines that were doing the mallocs that were SIGSEGVing |
| 14:35 | <&McMartin> | And SDL_Mixer is, as noted, horribly finicky and distressingly prone to fucking everything up at inopportune moments unless handled with a biohazard suit and six-foot tongs |
| 14:41 | <@Azash> | I guess there are no better alternatives? |
| 14:41 | <&McMartin> | Not for what I want, not without a shitload of work I really don't want to do |
| 14:42 | <&McMartin> | I'm mimicking a third-party API here, and SDL_Mixer mimics it almost exactly as is. |
| 14:42 | <@Azash> | Right~ |
| 14:46 | | celticminstrel [celticminst@Nightstar-e83b3651.cable.rogers.com] has left #code [] |
| 14:46 | | celticminstrel [celticminst@Nightstar-e83b3651.cable.rogers.com] has joined #code |
| 14:46 | | mode/#code [+o celticminstrel] by ChanServ |
| 14:46 | | celticminstrel [celticminst@Nightstar-e83b3651.cable.rogers.com] has quit [[NS] Quit: KABOOM! It seems that I have exploded. Please wait while I reinstall the universe.] |
| 14:46 | | celticminstrel [celticminst@Nightstar-e83b3651.cable.rogers.com] has joined #code |
| 14:46 | | mode/#code [+o celticminstrel] by ChanServ |
| 14:47 | <&McMartin> | I really need to stop messing around at this point and start writing the event system |
| 14:48 | <@Azash> | What was it you are working on again? |
| 14:51 | <&McMartin> | 2D game engine |
| 14:51 | <&McMartin> | Want it to be trivially drivable by pretty much any other language |
| 14:53 | <@Azash> | Hm hm |
| 14:54 | <&McMartin> | I'm trying to replicate the main functionality behind the Game Maker engine, but without the reliance on some of the things it wants (like modern graphics cards) |
| 14:54 | <&McMartin> | I'm giving things up (like rotozoom) to get that, but that's fine, the games I want to write don't need that. |
| 15:21 | <@TheWatcher> | Could such things be added via a plugin system, should they ever be needed/desirable/itchscratchworthy? |
| 15:27 | <&McMartin> | There'd be some fairly major yakshaving to do so "properly" |
| 15:27 | <&McMartin> | One of the things I'm backdooring into the API is neutrality between a framebuffer (which I'm using now) and an OpenGL backend |
| 15:27 | <&McMartin> | To really do rotozoom properly you need the opengl backend |
| 15:46 | | himi [fow035@Nightstar-5d05bada.internode.on.net] has quit [Ping timeout: 121 seconds] |
| 15:48 | | Orthia [orthianz@3CF3A5.E1CD01.B089B9.1E14D1] has quit [Ping timeout: 121 seconds] |
| 16:20 | | Turaiel[Offline] is now known as Turaiel |
| 16:29 | <@Azash> | https://twitter.com/vmw_workstation/status/347017072812908544 |
| 16:29 | | * Azash gets confus |
| 17:16 | <&McMartin> | New versions of Fusion (and thus Workstation) tend to come out around when new versions of OSX do, I guess? |
| 17:17 | <@Azash> | I was more amused with making it Father's Day themed |
| 17:17 | <@Azash> | "Grats dad, I got you a copy of VMWare Workstation" |
| 17:21 | <&McMartin> | I dunno |
| 17:21 | <&McMartin> | If I were a dad I'd rather have that than a tie |
| 17:21 | <@Tamber> | Socks. |
| 17:34 | <@Azash> | McMartin: Sure but you would know what to do with it at least :P |
| 18:22 | | Turaiel is now known as Turaiel[Offline] |
| 19:40 | | Typherix is now known as Typh|offline |
| 21:19 | | Derakon[AFK] is now known as Derakon |
| 21:24 | | ktemkin[awol] is now known as ktemkin[work] |
| 21:24 | | ktemkin[work] is now known as ktemkin |
| 21:41 | | Kindamoody|out is now known as Kindamoody |
| 22:07 | | ErikMesoy is now known as ErikMesoy|sleep |
| 22:14 | | Kindamoody is now known as Kindamoody[zZz] |
| 22:31 | | gnolam_ [lenin@Nightstar-b2aa51c5.cust.bredbandsbolaget.se] has joined #code |
| 22:31 | | gnolam is now known as NSGuest52050 |
| 22:31 | | gnolam_ is now known as gnolam |
| 22:31 | | mode/#code [+o gnolam] by ChanServ |
| 22:33 | | NSGuest52050 [lenin@Nightstar-b2aa51c5.cust.bredbandsbolaget.se] has quit [Ping timeout: 121 seconds] |
| --- Log closed Thu Jun 20 00:00:44 2013 |