| --- Log opened Thu Jan 07 00:00:35 2016 |
| 00:01 | | catalyst [catalyst@Nightstar-bt5k4h.81.in-addr.arpa] has quit [[NS] Quit: ] |
| 00:05 | | catadroid [catalyst@Nightstar-pie2mk.dab.02.net] has joined #code |
| 00:30 | | Derakon[AFK] is now known as Derakon |
| 01:00 | | Turaiel[Offline] is now known as Turaiel |
| 02:18 | | catadroid` [catalyst@Nightstar-7r0b49.dab.02.net] has joined #code |
| 02:21 | | catadroid [catalyst@Nightstar-pie2mk.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 02:28 | | McMartin [mcmartin@Nightstar-rpcdbf.sntcca.sbcglobal.net] has quit [[NS] Quit: testing] |
| 02:30 | | McMartin [mcmartin@Nightstar-rpcdbf.sntcca.sbcglobal.net] has joined #code |
| 02:31 | | mode/#code [+ao McMartin McMartin] by ChanServ |
| 04:09 | | Reiv [NSwebIRC@Nightstar-q8avec.kinect.net.nz] has quit [Ping timeout: 121 seconds] |
| 04:42 | <~Vornicus> | ok. superb owl decision locking works, now to make clearing work. |
| 04:48 | <@celticminstrel> | I want to know what owls have to do with this. |
| 04:48 | <&Derakon> | Super Bowl. |
| 04:49 | <&Derakon> | I'm guessing Vorn is running a bracket. |
| 04:55 | | Crossfire [Z@Nightstar-r9lk5l.cust.comxnet.dk] has quit [Ping timeout: 121 seconds] |
| 05:09 | | Derakon is now known as Derakon[AFK] |
| 05:50 | <~Vornicus> | indeed. it's more challenging than the ncaa one because of the re-seeding after the wild card games |
| 05:52 | | Turaiel is now known as Turaiel[Offline] |
| 06:25 | <~Vornicus> | okay, this just leaves reenabling predecessors, which is annoying only for wildcards. |
| 06:32 | | ion [Owner@Nightstar-gmbj85.vs.shawcable.net] has quit [Ping timeout: 121 seconds] |
| 06:38 | <~Vornicus> | but that will have to wait. |
| 06:50 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 07:00 | | Kindamoody[zZz] is now known as Kindamoody |
| 07:57 | | AverageJoe [evil1@Nightstar-chrsln.ph.cox.net] has joined #code |
| 08:04 | | celticminstrel [celticminst@Nightstar-uce74q.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 08:28 | | Kindamoody is now known as Kindamoody|afk |
| 08:40 | | AverageJoe [evil1@Nightstar-chrsln.ph.cox.net] has quit [[NS] Quit: Leaving] |
| 08:47 | | catadroid` is now known as catadroid |
| 08:48 | | catadroid` [catalyst@Nightstar-id051f.dab.02.net] has joined #code |
| 08:51 | | catadroid [catalyst@Nightstar-7r0b49.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 09:48 | | ion [Owner@Nightstar-gmbj85.vs.shawcable.net] has joined #code |
| 10:34 | | catadroid` [catalyst@Nightstar-id051f.dab.02.net] has quit [Connection closed] |
| 10:34 | | catadroid [catalyst@Nightstar-id051f.dab.02.net] has joined #code |
| 10:44 | | Emmy-zZz is now known as Emmy |
| 11:37 | | Crossfire [Z@Nightstar-r9lk5l.cust.comxnet.dk] has joined #code |
| 11:37 | | mode/#code [+o Crossfire] by ChanServ |
| 13:22 | | celticminstrel [celticminst@Nightstar-uce74q.dsl.bell.ca] has joined #code |
| 13:22 | | mode/#code [+o celticminstrel] by ChanServ |
| 15:19 | | catadroid` [catalyst@Nightstar-54qe8v.dab.02.net] has joined #code |
| 15:22 | | catadroid [catalyst@Nightstar-id051f.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 15:30 | | Crossfire [Z@Nightstar-r9lk5l.cust.comxnet.dk] has quit [Ping timeout: 121 seconds] |
| 17:30 | | catadroid` is now known as catadroid |
| 17:47 | | Crossfire [Z@Nightstar-r9lk5l.cust.comxnet.dk] has joined #code |
| 17:47 | | mode/#code [+o Crossfire] by ChanServ |
| 18:25 | | Kylo-Ren [Demolish0r@Nightstar-nelvj0.dyn.optonline.net] has joined #code |
| 18:25 | | Kylo-Ren [Demolish0r@Nightstar-nelvj0.dyn.optonline.net] has left #code [] |
| 18:59 | < ToxicFrog> | "If we're afraid to change a thing because we don't understand it, then we are duty bound to dig that thing up by the roots and burn it because it's an existential threat." -- SRE |
| 19:00 | <@Tamber> | Wise words indeed. |
| 19:00 | | mode/#code [+ooooooooooooo abudhabi Attilla grindhold Xon ion himi VirusJTG Azash Turaiel[Offline] ToxicFrog catadroid [R] Emmy] by Tamber |
| 19:05 | | catalyst [catalyst@Nightstar-bt5k4h.81.in-addr.arpa] has joined #code |
| 20:20 | | Reiv [NSwebIRC@Nightstar-q8avec.kinect.net.nz] has joined #code |
| 20:20 | | mode/#code [+o Reiv] by ChanServ |
| 20:33 | | Derakon [chriswei@Nightstar-5mvs4e.ca.comcast.net] has joined #code |
| 20:33 | | mode/#code [+ao Derakon Derakon] by ChanServ |
| 20:33 | | * Derakon glares at his program at work. |
| 20:33 | <&Derakon> | I had a memory leak. |
| 20:33 | <&Derakon> | Then I worked on the code a bunch. |
| 20:34 | <@Tamber> | Now it's sprung more leaks in different places? |
| 20:34 | <&Derakon> | Now I have a memory leak that only reproduces itself if I don't use the memory profiler. |
| 20:34 | <@Tamber> | Ouch. |
| 20:36 | <&Derakon> | This does suggest that the references are due to a race condition or something...how the hell that works is gonna be fun to try to figure out. |
| 20:42 | <&Derakon> | Hm, though I reproduced it by hand, probably at about the same speed the profiler uses. ;_; |
| 21:09 | <&Derakon> | Fun fact: if you hide an application in OSX, and that application subsequently creates a new window, then it will unhide itself and raise itself to the top, just under the window you currently have open. |
| 21:10 | <&Derakon> | If you are trying to debug a process that involves creating like 10 windows slowly over the course of a couple minutes, then this makes it very hard to get anything else done in the interim. |
| 21:47 | <&Derakon> | "As a novice macro writer I hope I'm still allowed some dumb questions. My programming history is learning Basic in the 70s and 80s, not much more than that since. I am really missing my old "GoTo (line number) command from those Basic days. So, my question is: is there a command similar or equivalent to my old, Basic, "Goto." "Function" is the closest I've found." |
| 21:48 | <&Derakon> | Just a quick reminder of how far languages have come in the past half-century or so. |
| 21:48 | <@Tamber> | Now you get to give the target of your goto a name, not just a line number! |
| 21:49 | | catadroid` [catalyst@Nightstar-6ke4d9.dab.02.net] has joined #code |
| 21:49 | <&Derakon> | There's a little more involved in invoking a function besides that. :p |
| 21:50 | <@Tamber> | I suppose... |
| 21:50 | <@Tamber> | Push everything onto stack, then call... |
| 21:50 | | catadroid` [catalyst@Nightstar-6ke4d9.dab.02.net] has quit [[NS] Quit: Bye] |
| 21:51 | | catadroid [catalyst@Nightstar-54qe8v.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 21:55 | | Kindamoody|afk is now known as Kindamoody |
| 22:08 | <&McMartin> | Functions are GOSUB. =P |
| 22:08 | <&McMartin> | (Actually, by later basics, functions are CALL.) |
| 22:23 | | * Derakon creates and destroys a few displays, incurring the memory leak, and then manually nullifies every object he can think of that keeps track of displays; no good. |
| 22:23 | <&Derakon> | Where the hell are these references being kept. |
| 22:28 | | himi [fow035@Nightstar-v37cpe.internode.on.net] has quit [Connection closed] |
| 22:49 | | * Derakon engages the time machine, i.e. runs git bisect. |
| 22:49 | <&Derakon> | I've got to find where it all went wrong! |
| 22:51 | | Kindamoody is now known as Kindamoody[zZz] |
| 22:55 | | Turaiel[Offline] [Brandon@Nightstar-7mqsi0.mi.comcast.net] has quit [Ping timeout: 121 seconds] |
| 22:55 | | Turaiel[Offline] [Brandon@Nightstar-7mqsi0.mi.comcast.net] has joined #code |
| 23:07 | <@Reiv> | what did you do |
| 23:11 | <&McMartin> | Leak memory, it sounds like |
| 23:14 | <&Derakon> | Yeah. |
| 23:14 | <&Derakon> | Every time you create and then destroy an image-display window, a little more memory gets seized by some greedy bit of code. |
| 23:15 | <&McMartin> | ... do you have a display list or event listener somewhere that isn't letting go? |
| 23:15 | <&McMartin> | Event listeners in particular have *strong* references, which makes life exciting |
| 23:16 | <&Derakon> | Yeah, it's probably something like that. |
| 23:16 | <&Derakon> | I've scoured every one that I can think of though. |
| 23:16 | <&Derakon> | Also, I just wasted the last half-hour, because what I thought was a good commit wasn't. |
| 23:16 | <@Reiv> | Obviously, the solution is to never actually destroy an image-display window, and instead merely turn it invisible until the next time you load it with a new picture! |
| 23:16 | <&Derakon> | So git bisect was searching the wrong part of history. |
| 23:16 | <&Derakon> | Ha ha ha no |
| 23:23 | <&Derakon> | Cripes, I'm back in last April. |
| 23:25 | <&Derakon> | ...it's not my code. |
| 23:25 | <&Derakon> | It's code I depend on. |
| 23:25 | <&Derakon> | That has to be it. |
| 23:25 | <&Derakon> | Yeah, we use ImageJ for our data visualization and it has its own resource tracking. |
| 23:25 | <&Derakon> | And if I just use pure ImageJ to open and close windows, there's still a memory leak. |
| 23:30 | <@Reiv> | Hooray! |
| 23:30 | <@Reiv> | [quiet sobbing] |
| 23:30 | | Emmy is now known as Emmy-zZz |
| 23:38 | | macdjord [macdjord@Nightstar-r9vt2h.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 23:39 | <&Derakon> | I give up for today. |
| 23:39 | | Derakon [chriswei@Nightstar-5mvs4e.ca.comcast.net] has quit [[NS] Quit: leaving] |
| 23:39 | <&McMartin> | Congratulations! It's not your fault: Bad news: It's not your fault |
| 23:47 | <@Reiv> | Clearly the new solution |
| 23:48 | <@Reiv> | Is to roll your own! |
| 23:48 | <@Reiv> | That way you have full autonomy of the code. |
| 23:48 | | * Reiv nods sagely. |
| 23:48 | <&McMartin> | Sometimes that *is* the answer |
| 23:48 | <@Reiv> | (If you do, it is Exceedingly Polite to actually comment *why*) |
| 23:49 | | * McMartin in fact just finished ripping the horrifically buggy boost::network out of this codebase and replaced it with raw iOS syscalls. |
| 23:49 | <&McMartin> | This has made the code smaller, clearer, and more reliable |
| 23:49 | <@Reiv> | (So in a decades time some poor bastard goes bug hunting and stares incomprehensibly at why in the name of everloving fucks you would have written your own goddamn custom module instead of just using the stock one) |
| 23:50 | <&McMartin> | I'd say "less portable" but since any alteration of the code using boost::network would corrupt the heap in unusual ways - and this included "storing them in a list instead of a vector" - it's not clear it's actually *less* portable. |
| 23:58 | | macdjord [macdjord@Nightstar-r9vt2h.mc.videotron.ca] has joined #code |
| 23:58 | | mode/#code [+o macdjord] by ChanServ |
| --- Log closed Fri Jan 08 00:00:51 2016 |