| --- Log opened Thu Jan 31 00:00:07 2019 |
| 00:10 | | tripflag [ed@Nightstar-7pe1df.clients.your-server.de] has quit [[NS] Quit: ZNC - http://znc.in] |
| 00:15 | | tripflag [ed@Nightstar-7pe1df.clients.your-server.de] has joined #code |
| 00:16 | | Degi [Degi@Nightstar-lafbe2.dyn.telefonica.de] has quit [Connection reset by peer] |
| 01:23 | | Derakon[AFK] is now known as Derakon |
| 01:24 | | * McMartin writes a function reverse(s) which reverses the string s by turning the mind inside out, converting madness into reality and opening the door to allow the Old Ones to creep forth once more from their sunken crypt beyond time. |
| 01:29 | | macdjord|wurk is now known as macdjord |
| 01:37 | | * ToxicFrog writes a touchscreen UI in zsh using fim, imagemagick and evtest |
| 01:40 | <&McMartin> | I spent a good chunk of time today reading about NSAutoreleasePool and how it interacts with automatic reference counting in Objective-C and Swift and oh god Apple what have you done |
| 01:40 | <&McMartin> | (Answer, in part: written code that looks at the instructions around the one that called it to try to deduce what sort of code was calling it and behaving differently depending on the results) |
| 01:42 | <&McMartin> | This also has solidified my opinion that autorelease pools are a bad solution to a non-problem |
| 01:43 | <&McMartin> | But GTK's floating references are a completely different bad solution to the same non-problem, and this means I should maybe revisit my assumption about its non-problematicity |
| 01:43 | <&McMartin> | [10 seconds pass] |
| 01:43 | <&McMartin> | OK, I've revisited it. I still don't buy that it's a problem that constructor-like functions will return objects that you must free |
| 01:43 | <&McMartin> | s/free/release/ |
| 01:44 | <&ToxicFrog> | What is an "autorelease pool"? |
| 01:44 | <&McMartin> | So, normally, Obj-C objects use manual reference counting, so you call retain or release on them |
| 01:45 | <&McMartin> | In the alternative, instead of releasing an object you retain before it leaves your scope, you can "autorelease" it |
| 01:45 | <&McMartin> | This registers a delayed release to happen the next time the "drain" method is called on the most recently constructed autorelease pool, which also destroys it. |
| 01:46 | <&McMartin> | These are syntactically bound in modern ObjC code, so you'll have an @autoreleasepool { ... } block and all autoreleased objects get released when you leave that block. |
| 01:46 | <&McMartin> | This happens, in particular, every time the event loop in a GUI app does a cycle, which is generally fine. |
| 01:47 | <&McMartin> | Meanwhile, GTK has a notion of a "floating reference", and the way these work is essentially that the first time you try to retain it, it's a no-op. |
| 01:47 | <&McMartin> | Both of these are attempting to avoid the case where you have to say |
| 01:47 | <&McMartin> | Foo *foo = thing_that_makes_a_foo(); |
| 01:47 | <&McMartin> | do_thing(foo); |
| 01:47 | <&McMartin> | release(foo); |
| 01:48 | <&McMartin> | And instead let you just say do_thing(thing_that_makes_a_foo());. |
| 01:48 | <&ToxicFrog> | This seems to be solved more elegantly, to the extent it needs solving, by python's with block and similar constructions. |
| 01:49 | <&McMartin> | Right |
| 01:50 | <&McMartin> | Modern Obj-C uses a Python-like reference-counting system and the compiler automatically inserts retain/release calls as needed |
| 01:50 | <&McMartin> | The interaction with autorelease pools is that if you are returning an autoreleased object it disassembles the code around the call site to see if it was being called by an automatic-reference-counting bit of code, and if it is, it just doesn't bother getting the pool involved at all and just returns a retained object like it always should have. |
| 01:51 | <&McMartin> | This is all managed by runtime library functions, so this is implemented as sanely as it can be, but still |
| 01:51 | <&McMartin> | what the hell dudes |
| 01:52 | <&McMartin> | And yeah, my reaction overall is "I am in no way convinced that this needs solving" |
| 01:53 | <&ToxicFrog> | Yeaaaaaaaaah. |
| 01:53 | <&McMartin> | as for Python's with block etc |
| 01:53 | <&ToxicFrog> | Meanwhile, I've realized that evtest gives me the pixel dimensions of the console and stty gives me the character dimensions, which means I can map between them, which in turn means that I can make a touch-driven curses UI. |
| 01:54 | <&McMartin> | Both ObjC at the time autorelease pools were introduced to its core library, and GTK itself over its whole existence, are built on top of C, which does not provide the necessary syntactic extensions |
| 01:54 | <&McMartin> | It took Apple a long time to realize that nobody else gave a damn about ObjC and that they were thus free to evolve it aggressively in ways that made it suck less |
| 01:54 | <&McMartin> | A process that completed around the time that OS X 10.8 was released |
| 01:54 | <&ToxicFrog> | GTK has that excuse, but ObjC...aah. |
| 01:55 | <&McMartin> | But yeah, I hit a corner case where This Interaction Matters and now I'm digging into how to make sure the spiders stay in their holes and well yeah this is all madness |
| 02:22 | | celmin|away is now known as celticminstrel |
| 04:39 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has joined #code |
| 04:55 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has quit [Ping timeout: 121 seconds] |
| 05:03 | | Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has joined #code |
| 05:03 | | mode/#code [+qo Vorntastic Vorntastic] by ChanServ |
| 05:16 | <&McMartin> | Man, the late-1980s systems were a weird mishmash |
| 05:16 | | * McMartin uses the CP/M interface on the Atari ST to do MS-DOS-inspired system calls, but finds that they fail because he wasn't being UNIXy enough in his usage of them. |
| 05:16 | <&McMartin> | But now that I've gotten that sorted out, file redirection actually works. |
| 05:16 | <&McMartin> | (It turns out that stdout and CON: are not actually the same thing.) |
| 05:19 | <&McMartin> | (Though they are in MS-DOS! But CP/M doesn't have a stdout, so the ST followed that half the time. >_<) |
| 05:29 | | Derakon is now known as Derakon[AFK] |
| 05:49 | | celticminstrel is now known as celmin|sleep |
| 06:03 | | macdjord is now known as macdjord|slep |
| 06:15 | | himi [sjjf@Nightstar-1drtbs.anu.edu.au] has quit [Ping timeout: 121 seconds] |
| 07:24 | <&McMartin> | That feeling when you rewrite a big chunk of code with super-dumb brute force and it ends up saving nearly 50% in space |
| 07:26 | <&McMartin> | Though it also involves replacing a bunch of actual multiply operations with dozens of shifts and adds |
| 10:53 | | Degi [Degi@Nightstar-qs95gn.dyn.telefonica.de] has joined #code |
| 10:53 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has joined #code |
| 11:25 | | Kindamoody[zZz] is now known as Kindamoody |
| 13:06 | | celmin|sleep is now known as celmin|away |
| 15:19 | | Degi [Degi@Nightstar-qs95gn.dyn.telefonica.de] has quit [Connection reset by peer] |
| 15:41 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 15:41 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 16:42 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 17:25 | <@iospace> | well |
| 17:25 | <@iospace> | due to logistical issues |
| 17:25 | <@iospace> | intel looks like it'll be falling through |
| 17:25 | <@iospace> | (the contractor role) |
| 17:30 | | Alek [Alek@Nightstar-o723m2.cicril.sbcglobal.net] has quit [Ping timeout: 121 seconds] |
| 17:38 | | Alek [Alek@Nightstar-o723m2.cicril.sbcglobal.net] has joined #code |
| 17:38 | | mode/#code [+o Alek] by ChanServ |
| 17:42 | | Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 18:33 | <@iospace> | I'd rather have it fail there, but even the fact they said "hmm... this resume looks good" gives me a nice ego boost xD |
| 18:40 | <&[R]> | That's always good |
| 18:58 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has quit [[NS] Quit: Leaving] |
| 19:16 | | McMartin [mcmartin@Nightstar-rpcdbf.sntcca.sbcglobal.net] has quit [Operation timed out] |
| 20:20 | | VirusJTG [VirusJTG@Nightstar-q9r0g8.clt.hostedsolutions.com] has joined #code |
| 20:20 | | mode/#code [+ao VirusJTG VirusJTG] by ChanServ |
| 20:42 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has joined #code |
| 20:42 | | mode/#code [+o himi] by ChanServ |
| 20:44 | | Degi [Degi@Nightstar-qs95gn.dyn.telefonica.de] has joined #code |
| 20:45 | | Alek [Alek@Nightstar-o723m2.cicril.sbcglobal.net] has quit [Ping timeout: 121 seconds] |
| 20:49 | | Alek [Alek@Nightstar-o723m2.cicril.sbcglobal.net] has joined #code |
| 20:49 | | mode/#code [+o Alek] by ChanServ |
| 21:43 | | Kindamoody is now known as Kindamoody[zZz] |
| 21:53 | | VirusJTG_ [VirusJTG@Nightstar-42s.jso.104.208.IP] has joined #code |
| 21:55 | | VirusJTG [VirusJTG@Nightstar-q9r0g8.clt.hostedsolutions.com] has quit [Operation timed out] |
| 21:55 | | VirusJTG_ [VirusJTG@Nightstar-42s.jso.104.208.IP] has quit [[NS] Quit: Leaving] |
| 21:55 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has joined #code |
| 21:55 | | mode/#code [+ao VirusJTG VirusJTG] by ChanServ |
| 23:03 | | Alek [Alek@Nightstar-o723m2.cicril.sbcglobal.net] has quit [Ping timeout: 121 seconds] |
| 23:07 | | Alek [Alek@Nightstar-o723m2.cicril.sbcglobal.net] has joined #code |
| 23:07 | | mode/#code [+o Alek] by ChanServ |
| 23:17 | | Alek [Alek@Nightstar-o723m2.cicril.sbcglobal.net] has quit [Ping timeout: 121 seconds] |
| 23:21 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has quit [Ping timeout: 121 seconds] |
| 23:25 | | Alek [Alek@Nightstar-o723m2.cicril.sbcglobal.net] has joined #code |
| 23:25 | | mode/#code [+o Alek] by ChanServ |
| 23:46 | | crystalclaw [crystalclaw@Nightstar-12q9ui.xyz] has quit [Ping timeout: 121 seconds] |
| 23:48 | | crystalclaw [crystalclaw@Nightstar-12q9ui.xyz] has joined #code |
| 23:48 | | mode/#code [+o crystalclaw] by ChanServ |
| --- Log closed Fri Feb 01 00:00:09 2019 |