| --- Log opened Tue Sep 08 00:00:17 2020 |
| 00:26 | | * McMartin considers being A Bad Person |
| 00:26 | <&McMartin> | I have a chunk of C++14-or-later code that I'm writing, and I have two ways to write it |
| 00:27 | <&McMartin> | One is with pointers all over in some ways that are tricky but traditional to reason about |
| 00:27 | <&McMartin> | The other is with std::shared_ptr<> which will be basically the same but "obviously" safer |
| 00:27 | <&McMartin> | And the final way is to make aggressive use of move constructors throughout the function and the lambdas it creates |
| 00:28 | <&McMartin> | I'm strongly tempted to do the latter just so that I can apologize for it in the code review with "I'm a little rusty with these data types" |
| 00:46 | <&McMartin> | Alas, now that I experiment with it, the third approach is in fact impossible, which makes std::shared_ptr<> the correct answer |
| 00:46 | <&McMartin> | (The need for move constructors is because some of the objects are noncopyable, but other parts of the underlying code need everything passed in to be copyable.) |
| 00:55 | < catalyst> | ahaha |
| 00:55 | < catalyst> | shared_ptr is alright I guess |
| 00:56 | < catalyst> | does it really need to be shared though? can it just be unique in the owning frame and then passed by reference? |
| 00:56 | < catalyst> | oh, lambdas |
| 00:56 | < catalyst> | nevermind |
| 00:58 | <&McMartin> | Yeah, it's a capture and the resulting function object is apparently copied by the underlying library, so capturing noncopyables (like, whoops, unique_ptr) is contraindicated |
| 00:58 | <&McMartin> | I could teeeeechnically use a raw pointer here and do a delete within the lambda at the end but I'm thinking "you know what maybe no" |
| 00:58 | < catalyst> | you know what would be really good for these kind of circumstances |
| 00:59 | <&McMartin> | Lifetimes built into the underlying system? |
| 00:59 | <&McMartin> | (Actually I'm not sure where you're going with this please proceed) |
| 00:59 | < catalyst> | also yes that sounds like the kind of optimisation that would make me cry, and if performance is a concern there then idk, write a better algorithm |
| 01:00 | < catalyst> | oh I was just going to say something throwaway about lifetimes |
| 01:00 | <&McMartin> | Yeah |
| 01:00 | <&McMartin> | Now, one option would be to dig deeper into this so that the lambdas *can* be noncopyable and everything still works |
| 01:00 | < catalyst> | rust impresses me because it casually provides a safe, single threaded shared pointer type |
| 01:01 | < catalyst> | and in implementing that in C++ I decided it would be impossible |
| 01:01 | <&McMartin> | But right now I'm trying to unfuck some unit tests so that I can implement better analytics on whether or not the system is being overtaxed or underutilized |
| 01:01 | < catalyst> | well, incredibly unwise at least |
| 01:01 | < catalyst> | honestly I would just go with the shared_ptr |
| 01:01 | <&McMartin> | And then after I have that as a baseline then I could look into "hey, maybe moves everywhere?" |
| 01:01 | <&McMartin> | Yeah, shared_ptr is 100% what I need *right now* |
| 01:02 | <&McMartin> | But we get to have C++17 here, so between mandatory RVO and move captures/initializer captures the move-based one looked way cleaner :( |
| 01:02 | < catalyst> | it's also not slow to access, only slow on copy and deallocation |
| 01:02 | < catalyst> | :< |
| 01:02 | <&McMartin> | No need for me to be coy though, I guess |
| 01:03 | <&McMartin> | It's a std::promise, it is not doing a whole lot in the first place and it's not somtehing happening thousands of times a frame |
| 01:03 | <&McMartin> | But it also means "hey, this is explicitly for crossthread synchro" |
| 01:03 | < catalyst> | sounds like you might have already overthought it I guess |
| 01:04 | <&McMartin> | But yeah, I was hoping for "declare promise auto, save its future in an auto, move the promise into a lambda that sets its value, RVO-return the future to original thread" |
| 01:04 | <&McMartin> | Yeah, Friday was spent fencing with two of the more senior guys about what needed to be done in broad terms |
| 01:04 | < catalyst> | I'm already seeing sadness in my future over going back to C++ |
| 01:05 | < catalyst> | so much of this is just simple standard types and compiler stuff in Rust |
| 01:05 | <&McMartin> | I'm definitely in that new-hire phase where I walk in with the attitude that everything will appear to be horrible and dysfunctional, and I'll be set to adapt and extend |
| 01:06 | <&McMartin> | So far I've found a number of pretty ugly bugs and seen the scars of some design battles |
| 01:06 | <&McMartin> | And tbh it's also nice to see a response to "here is my perfect geometric logic that shows that moving this one line fixes the race condition" is pretty overwhelming "no, use sledgehammer plz" |
| 01:07 | <&McMartin> | But yeah, from what I've seen so far... if this were Rust, we'd be having to use Arc<> and friends, not just Rc<>. |
| 01:13 | < catalyst> | x) |
| 01:15 | < catalyst> | pretty much, though Arc is essentially shared_ptr I suppose, to the point where i recognised the various tradeoffs in its design matching |
| 01:16 | < catalyst> | it's more the idiomatic nature of the language supporting it rather than giving you a bunch of get outs that will just break when you look at it from the wrong place |
| 01:17 | <&McMartin> | Yeah |
| 01:17 | <&McMartin> | It's more "if this had been in Rust in the first place, there's a good chance that the move-based approach would have just worked" |
| 01:18 | < catalyst> | aye |
| 01:18 | < catalyst> | it certainly has a much stronger idea of lifetimes via lambdas |
| 01:19 | < catalyst> | or at least that's my impression |
| 01:30 | | himi [sjjf@Nightstar-1drtbs.anu.edu.au] has joined #code |
| 01:30 | | mode/#code [+o himi] by ChanServ |
| 01:39 | | Kindamoody is now known as Kindamoody[zZz] |
| 01:45 | | Pink [user1@Nightstar-g7hdo5.dyn.optonline.net] has joined #code |
| 01:46 | | Pinkhair [user1@Nightstar-g7hdo5.dyn.optonline.net] has quit [Ping timeout: 121 seconds] |
| 01:48 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 02:17 | < catalyst> | I am feeling very competent |
| 02:17 | < catalyst> | that's kind of new, honestly |
| 02:18 | < catalyst> | I know I always was but I don't think I've ever felt so innately like I know what I'm talking about |
| 03:03 | <&McMartin> | Yay, whatever the opposite of imposter syndrome is |
| 03:03 | <&McMartin> | ... poster-girl syndrome? |
| 03:04 | < Mahal> | Ahah |
| 03:04 | < Mahal> | I like that |
| 04:30 | | Degi [Degi@Nightstar-pkig1n.dyn.telefonica.de] has quit [Ping timeout: 121 seconds] |
| 04:31 | | Degi [Degi@Nightstar-dd9i1i.dyn.telefonica.de] has joined #code |
| 05:14 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has quit [Connection closed] |
| 05:14 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has joined #code |
| 05:14 | | mode/#code [+ao VirusJTG VirusJTG] by ChanServ |
| 05:50 | | Vorntastic [uid293981@Nightstar-ks9.9ff.184.192.IP] has joined #code |
| 05:50 | | mode/#code [+qo Vorntastic Vorntastic] by ChanServ |
| 06:01 | | celticminstrel [celticminst@Nightstar-8kptuo.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 06:01 | | celticminstrel [celticminst@Nightstar-8kptuo.dsl.bell.ca] has joined #code |
| 06:01 | | mode/#code [+o celticminstrel] by ChanServ |
| 06:03 | | celticminstrel [celticminst@Nightstar-8kptuo.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 07:42 | | himi [sjjf@Nightstar-1drtbs.anu.edu.au] has quit [Connection closed] |
| 09:23 | | Kindamoody[zZz] is now known as Kindamoody |
| 09:31 | | Emmy [Emmy@Nightstar-l49opt.fixed.kpn.net] has joined #code |
| 09:38 | | Emmy [Emmy@Nightstar-l49opt.fixed.kpn.net] has quit [Ping timeout: 121 seconds] |
| 10:27 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has joined #code |
| 10:27 | | mode/#code [+o himi] by ChanServ |
| 10:49 | | catalyst_ [catalyst@Nightstar-vsirs3.dab.02.net] has joined #code |
| 10:51 | | catalyst [catalyst@Nightstar-v6lb30.cable.virginm.net] has quit [Connection reset by peer] |
| 10:51 | | catalyst [catalyst@Nightstar-v6lb30.cable.virginm.net] has joined #code |
| 10:53 | | catalyst_ [catalyst@Nightstar-vsirs3.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 11:30 | | JustLurk [justbob@ServerAdministrator.Nightstar.Net] has joined #code |
| 11:30 | | JustBob [justbob@Nightstar.Customer.Dissatisfaction.Administrator] has quit [NickServ (RECOVER command used by JustLurk)] |
| 11:30 | | JustLurk is now known as JustBob |
| 11:30 | | mode/#code [+o JustBob] by ChanServ |
| 11:49 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has quit [[NS] Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] |
| 11:50 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has joined #code |
| 11:50 | | mode/#code [+o Reiv] by ChanServ |
| 11:53 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has quit [[NS] Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] |
| 11:53 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has joined #code |
| 11:53 | | mode/#code [+o Reiv] by ChanServ |
| 12:06 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has quit [[NS] Quit: Leaving] |
| 13:35 | | celticminstrel [celticminst@Nightstar-8kptuo.dsl.bell.ca] has joined #code |
| 13:35 | | mode/#code [+o celticminstrel] by ChanServ |
| 15:15 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 15:15 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 16:31 | | catalyst [catalyst@Nightstar-v6lb30.cable.virginm.net] has quit [Connection closed] |
| 16:34 | | catalyst [catalyst@Nightstar-v6lb30.cable.virginm.net] has joined #code |
| 17:11 | | Emmy [Emmy@Nightstar-l49opt.fixed.kpn.net] has joined #code |
| 17:29 | | Vorntastic [uid293981@Nightstar-ks9.9ff.184.192.IP] has quit [[NS] Quit: Connection closed for inactivity] |
| 21:47 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has joined #code |
| 21:47 | | mode/#code [+ao VirusJTG VirusJTG] by ChanServ |
| 22:49 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has quit [Ping timeout: 121 seconds] |
| 23:29 | | Emmy [Emmy@Nightstar-l49opt.fixed.kpn.net] has quit [Ping timeout: 121 seconds] |
| --- Log closed Wed Sep 09 00:00:19 2020 |