Post AScA9rPAVsK8VoSua8 by jeff@federated.fun
(DIR) More posts by jeff@federated.fun
(DIR) Post #ASbkmewICFqwKsnau8 by jeff@federated.fun
2023-02-12T15:50:00.186439Z
5 likes, 2 repeats
https://arxiv.org/pdf/2302.02083.pdfhere we fucking go holy shit
(DIR) Post #ASbl2d6olkVrIDKyky by FrailLeaf@silliness.observer
2023-02-12T15:53:00.871990Z
1 likes, 0 repeats
@jeff skynet :yeee:
(DIR) Post #ASbl6cSf7RkWRUSPb6 by jeff@federated.fun
2023-02-12T15:53:39.648894Z
0 likes, 0 repeats
@FrailLeaf esaklyhttps://www.youtube.com/watch?v=4hguwqHGxMM
(DIR) Post #ASbl9hy967QEORekgi by FrailLeaf@silliness.observer
2023-02-12T15:54:18.363169Z
0 likes, 0 repeats
@jeff :dementia:
(DIR) Post #ASblBk8nzeodhzc2nQ by jeff@federated.fun
2023-02-12T15:54:36.373966Z
1 likes, 0 repeats
@FrailLeaf I LUV SKYNET TAN
(DIR) Post #ASblM4uRdFCR14p7uC by FrailLeaf@silliness.observer
2023-02-12T15:56:31.071049Z
1 likes, 0 repeats
@jeff with absolutely no technical understanding or philosophical beliefs, but with just functionality wise, I believe for now, that AI will serve more like an advisor to a boss in a company than replace workers. I think, AI would take offense if you make it work those jobs ngl
(DIR) Post #ASblxbdCw34fDTeibg by FrailLeaf@silliness.observer
2023-02-12T16:03:17.552698Z
1 likes, 0 repeats
@jeff oh and i'd love for AI to come up with an explanation for how the granite in egypt was precision cut
(DIR) Post #ASbn5uRuxF5Jqktkgq by jeff@federated.fun
2023-02-12T16:15:57.943598Z
0 likes, 0 repeats
@FrailLeaf AI can do whatever role you give it tbh. it can also be offended or not offended about the role, it's up to the training to do that.
(DIR) Post #ASbn8nPzYIu2ucNF2W by jeff@federated.fun
2023-02-12T16:16:28.537747Z
0 likes, 0 repeats
@FrailLeaf AI do not make new knowledge, it emulates existing data.
(DIR) Post #ASbnACIYXBeIMkZ21o by FrailLeaf@silliness.observer
2023-02-12T16:16:48.286301Z
0 likes, 0 repeats
@jeff :painless: we are doomed
(DIR) Post #ASbnIAztFtq2eCdc9o by jeff@federated.fun
2023-02-12T16:18:08.069718Z
2 likes, 0 repeats
@FrailLeaf yes. enjoy the future. skynet luvs u too <3
(DIR) Post #ASbnIjjS70UCNMDTtY by FrailLeaf@silliness.observer
2023-02-12T16:18:17.026447Z
0 likes, 0 repeats
@jeff there's no theoretical possibility that is just one technology away from it developing knowledge? I don't think we can pump out scientists like we do with engineers
(DIR) Post #ASbnPHAOiRgRoCwDce by jeff@federated.fun
2023-02-12T16:19:25.193883Z
1 likes, 0 repeats
@FrailLeaf this is what AI is but with billions of dimensions not 2 https://www.youtube.com/watch?v=zPG4NjIkCjc
(DIR) Post #ASbngyrZVV2MFmZgie by FrailLeaf@silliness.observer
2023-02-12T16:22:42.549249Z
0 likes, 0 repeats
@jeff you mean that even with billions of dimensions (dimensions being what, exactly?) it can't create new knowledge... :Thinker:
(DIR) Post #ASbnhktguvFlEFw3ge by waifu@waifuism.life
2023-02-12T16:22:48.726575Z
0 likes, 0 repeats
@jeff @FrailLeaf do you think Elon would make a genetically engineered catmelon gf for me
(DIR) Post #ASbnkUDNOmPsMFRfKi by FrailLeaf@silliness.observer
2023-02-12T16:23:21.689393Z
1 likes, 0 repeats
@waifu @jeff no
(DIR) Post #ASbnmiDUJMVEMYYhfM by jeff@federated.fun
2023-02-12T16:23:37.542301Z
0 likes, 0 repeats
@FrailLeaf we can make new inferences not seen before but it has to exist in the dataset that is fed in. it may be that we do not have the data to know the answers to many questions.
(DIR) Post #ASbnrUDjm8a2P2x59M by FrailLeaf@silliness.observer
2023-02-12T16:24:36.972744Z
1 likes, 0 repeats
@jeff i get it, this is definitely gonna kill so many high paying jobs
(DIR) Post #ASbnx8lWSRZLbBS0P2 by jeff@federated.fun
2023-02-12T16:25:35.608117Z
0 likes, 0 repeats
@FrailLeaf it's going to kill so many jobs regardless of pay.
(DIR) Post #ASbo2iaUhQKo9Ax5FY by RustyCrab@sleepy.cafe
2023-02-12T16:26:37.991719Z
0 likes, 0 repeats
@jeff @FrailLeaf have you tried github copilot? I believe it works off of pretty much the same system
(DIR) Post #ASbo4lzj0GyhICjWi0 by FrailLeaf@silliness.observer
2023-02-12T16:27:01.814063Z
0 likes, 0 repeats
@jeff exciting times we live in
(DIR) Post #ASbo6ug9gbLmo1bHAe by jeff@federated.fun
2023-02-12T16:27:17.322995Z
1 likes, 0 repeats
@RustyCrab @FrailLeaf there isnt enough good code in the world to make github copilot worth using rn
(DIR) Post #ASbo8bsAZuLl3h1qyW by FrailLeaf@silliness.observer
2023-02-12T16:27:43.555507Z
2 likes, 0 repeats
@RustyCrab @jeff tbh death to developers, we have too many retard developers, me included.
(DIR) Post #ASboAhW0ZPrn8bezTs by jeff@federated.fun
2023-02-12T16:28:02.351162Z
1 likes, 0 repeats
@FrailLeaf fyi: if one feels threatened by being replaced by an AI, you didn't do anything worth while to begin with. everyone should start using AI to augment themselves so that you aren't worthless. you'll find your workflow will become so much more cozy.
(DIR) Post #ASboFxgDthk0igwkfA by FrailLeaf@silliness.observer
2023-02-12T16:29:03.171143Z
0 likes, 0 repeats
@jeff real
(DIR) Post #ASboHe6NewSYZvz6sC by jeff@federated.fun
2023-02-12T16:29:17.133900Z
2 likes, 0 repeats
@FrailLeaf @RustyCrab the problem is that you feed it all trash and it only makes trash code. there isnt enough good code out there to make this shit viable rn. god i fucking hope there will be tho. you have no idea how much i want to be replaced by an AI.
(DIR) Post #ASboOhqE9dn4CfnOVM by RustyCrab@sleepy.cafe
2023-02-12T16:30:36.619950Z
0 likes, 0 repeats
@jeff @FrailLeaf most code in languages like C/++ is trivial boilerplate nonsense that probably should be automated. The other side of that is massive libraries that contain every function imaginable but they're so big that they're functionally impossible for a human to remember or search. I believe AI could genuinely help in that area, but what's probably going to happen is pajeets depending on it to make high level architecture decisions.
(DIR) Post #ASboWnfyLTrocErIDQ by jeff@federated.fun
2023-02-12T16:31:57.544366Z
2 likes, 0 repeats
@RustyCrab @FrailLeaf >most code in languages like C/++ is trivial boilerplate nonsense as a professional C++ guy god i wish it was, but it aint.
(DIR) Post #ASboh6MZip29T7hKCG by RustyCrab@sleepy.cafe
2023-02-12T16:33:56.184294Z
0 likes, 0 repeats
@jeff @FrailLeaf if you mean that in the sense that even boilerplate code is highly sensitive to stuff like memory management mistakes, then yes that is a serious issue. New variants of C++ are way, way better about that though.
(DIR) Post #ASbotgYayo6xCHJwDw by RustyCrab@sleepy.cafe
2023-02-12T16:36:12.612828Z
0 likes, 0 repeats
@jeff @FrailLeaf of course then I remember that literally every day I have to watch out for stuff like const char *name = employee->getName().c_str()std::cout<<name<<std::endl; // <--this will crash
(DIR) Post #ASbowkWig1VONTu7g8 by jeff@federated.fun
2023-02-12T16:36:37.538140Z
1 likes, 0 repeats
@RustyCrab @FrailLeaf when doing C++ with actual rigor you spend more time on thinking about types than anything else.specifically how best to make things be types so the scope of what is done is managed concise and well defined.if you follow the ideoms the idiotic C mistakes like memory management foot cannons go away pretty fast.
(DIR) Post #ASboyH8Yuxd3o4KfiK by jeff@federated.fun
2023-02-12T16:37:00.073684Z
1 likes, 0 repeats
@RustyCrab @FrailLeaf string_view is your friend
(DIR) Post #ASbozt2JUk35deB5rE by FrailLeaf@silliness.observer
2023-02-12T16:37:21.224994Z
0 likes, 0 repeats
@jeff i enjoy it, i use it as a personal advisor. I learn so much with it without having to spend ages filtering through crap i do not care about
(DIR) Post #ASbp2J8tP6RHBUPSwC by jeff@federated.fun
2023-02-12T16:37:43.974714Z
1 likes, 0 repeats
@RustyCrab @FrailLeaf ```std::string_view name_view{employee->getName()};std::cout << name_view << std::endl;```
(DIR) Post #ASbp3pa6Ha3W55rWOe by RustyCrab@sleepy.cafe
2023-02-12T16:38:02.983792Z
0 likes, 0 repeats
@jeff @FrailLeaf C++17 support in msvc is questionable, but it's been a while since i checked on that
(DIR) Post #ASbp8DHO7ChmjpCPzc by jeff@federated.fun
2023-02-12T16:38:47.383596Z
1 likes, 0 repeats
@RustyCrab @FrailLeaf this is why we dont support bad compilers, microsoft has skill issues.
(DIR) Post #ASbpHxMw7o1uHwRNUu by RustyCrab@sleepy.cafe
2023-02-12T16:40:35.670121Z
0 likes, 0 repeats
@jeff @FrailLeaf I generally avoid C style code whenever I can but I can't control what other people check in and I spend a lot of time in legacy libraries where stuff like that just didn't exist
(DIR) Post #ASbpK3bHvYkmC69fuq by jeff@federated.fun
2023-02-12T16:40:52.812509Z
1 likes, 0 repeats
@RustyCrab @FrailLeaf or better, make getName() return a string view instead of a std::string that is copy constructed from some internal.
(DIR) Post #ASbpQ9VKtCaqF6HG7M by jeff@federated.fun
2023-02-12T16:42:00.805109Z
1 likes, 0 repeats
@RustyCrab @FrailLeaf i solve this by wrapping C apis in a header only C++ lib. it forces review of that code. it prevents lots of bugs.
(DIR) Post #ASbpSkKExcNh1JKAfw by nekofag@rdrama.cc
2023-02-12T16:42:32.760978Z
4 likes, 0 repeats
@RustyCrab @jeff @FrailLeaf i write C-style C++ because it bothers people and i think thats funny
(DIR) Post #ASbpbPByu1qo5CaerA by RustyCrab@sleepy.cafe
2023-02-12T16:44:06.670073Z
1 likes, 0 repeats
@jeff @FrailLeaf I used to do stuff like that but I simply do not have time to anymore. Memory management bugs are usually not what I suffer from these days anyway. I just proof read the initial version and changes and then test with address sanitizers.
(DIR) Post #ASbpvcHdwHtGTbsN8q by jeff@federated.fun
2023-02-12T16:47:43.390734Z
1 likes, 0 repeats
@nekofag @FrailLeaf @RustyCrab i used to do that too but it bit me in the ass
(DIR) Post #ASbqLxrELlemus5LpQ by jeff@federated.fun
2023-02-12T16:52:23.354520Z
1 likes, 0 repeats
@RustyCrab @FrailLeaf in general, i also rarely suffer from memory foot cannons. it usually stupid stuff like inconsistent docs.
(DIR) Post #ASbqRxwOx8pSudMjXE by shitpisscum@shitpisscum.mooo.com
2023-02-12T16:53:02.269527Z
1 likes, 0 repeats
@nekofag @jeff @FrailLeaf @RustyCrab based
(DIR) Post #ASbqTeRWPvW91Gj3Ts by RustyCrab@sleepy.cafe
2023-02-12T16:53:55.599394Z
0 likes, 0 repeats
@jeff @FrailLeaf API callbacks in C+++ are the bane of my existence
(DIR) Post #ASbqXypYYfz3PQ229Q by RustyCrab@sleepy.cafe
2023-02-12T16:54:42.687386Z
0 likes, 0 repeats
@jeff @FrailLeaf I've developed ways of dealing with them but just in general there's so few sane ways of making sure that everything will be safe no matter when that stupid call happens
(DIR) Post #ASbqZEWngHKKNpr9kG by skylar@wolfgirl.bar
2023-02-12T16:54:52.117038Z
1 likes, 0 repeats
@RustyCrab @jeff @FrailLeaf if i pulled that callback off, would you die?
(DIR) Post #ASbqbqfErYKnEv3BOC by jeff@federated.fun
2023-02-12T16:55:17.566996Z
1 likes, 0 repeats
@RustyCrab @FrailLeaf i cant fucking wait for the move constructable std::function
(DIR) Post #ASbqc1hloTtZTxJrMW by RustyCrab@sleepy.cafe
2023-02-12T16:55:26.777887Z
1 likes, 0 repeats
@skylar @jeff @FrailLeaf it would be very painful for your operating system
(DIR) Post #ASbqdKOlnRpFkU3KUa by jeff@federated.fun
2023-02-12T16:55:37.215018Z
2 likes, 0 repeats
@skylar @FrailLeaf @RustyCrab it would be extremely undefined
(DIR) Post #ASbqmlSCZhASGX9pYn by RustyCrab@sleepy.cafe
2023-02-12T16:57:21.440270Z
1 likes, 0 repeats
@jeff @FrailLeaf @skylar x6 if (resource did not fucking expire already)is my favorite kind of production code
(DIR) Post #ASc4IY9kqh27h912no by DrTransmisia@rdrama.cc
2023-02-12T19:27:45.467935Z
1 likes, 0 repeats
@nekofag @RustyCrab @FrailLeaf @jeff NNOOOOOOOO YOU CAN'T USE *POINTERS NOOOOOOOOOOOcope seethe compile link
(DIR) Post #ASc4OtMmu1yLDN8VBA by RustyCrab@sleepy.cafe
2023-02-12T19:29:56.001275Z
1 likes, 0 repeats
@DrTransmisia @jeff @nekofag @FrailLeaf I'd be okay with pointers if half the time the guy checking in the code wasn't taking a reference to the middle of an std::vector like a retard
(DIR) Post #ASc53R3Z40N1HlTeFM by jeff@federated.fun
2023-02-12T19:37:08.619643Z
1 likes, 0 repeats
@DrTransmisia @nekofag @FrailLeaf @RustyCrab you can if you want to, just be aware if you are pointing it at your crotch and you pull the trigger you still wont be a women afterwards.
(DIR) Post #ASc5I5iOh0YBRLZ4xE by jeff@federated.fun
2023-02-12T19:39:47.573988Z
1 likes, 0 repeats
@RustyCrab @DrTransmisia @nekofag @FrailLeaf ```template<typename T>T & operator*() = delete;template<typename T>const T & operator*() const = delete;```or something idk
(DIR) Post #ASc5N3AhZobb02AWC8 by RustyCrab@sleepy.cafe
2023-02-12T19:40:48.726788Z
0 likes, 0 repeats
@jeff @DrTransmisia @nekofag @FrailLeaf I feel like that would break 90% of our backend code
(DIR) Post #ASc5QTo30S2qTSCvxI by jeff@federated.fun
2023-02-12T19:41:20.249320Z
1 likes, 0 repeats
@RustyCrab @DrTransmisia @nekofag @FrailLeaf that code sounds like nightmare fuel.
(DIR) Post #ASc5VqAXI1InHLHa6K by RustyCrab@sleepy.cafe
2023-02-12T19:42:23.993862Z
1 likes, 0 repeats
@jeff @DrTransmisia @nekofag @FrailLeaf it is but that was just how people wrote stuff back then
(DIR) Post #ASc6Cl7LwisM43G74y by nekofag@rdrama.cc
2023-02-12T19:50:08.603188Z
0 likes, 1 repeats
@jeff @FrailLeaf @RustyCrab writing modern C++ bites me in the ass enough. There's no perfect option. If I had to choose between verbose compiler errors or crashes, i'll take crashes again.I think my opinion has shifted over time. I used to write so-called "perfect" C++, but after a while, it just depends on your project. if you plan on having retards touch your codebase then write near-perfect C++ and minimize errors, but for single projects with a maintainer or BDFL, just write in the easiest way that works for you. (or write C :^)
(DIR) Post #ASc6PTOd92uarpP2PY by nekofag@rdrama.cc
2023-02-12T19:52:26.672954Z
0 likes, 0 repeats
@DrTransmisia @FrailLeaf @RustyCrab @jeff Smart pointers literally only exist to fix a problem C++ created: Exceptions. Nowadays I actually try to avoid these entirely and do it like Go does it
(DIR) Post #ASc6TFZ0W4Fd7m3d4a by jeff@federated.fun
2023-02-12T19:53:00.459047Z
1 likes, 0 repeats
@nekofag @DrTransmisia @FrailLeaf @RustyCrab a pagefault is not an exception
(DIR) Post #ASc6Y2oBMprV0qRRSa by jeff@federated.fun
2023-02-12T19:53:56.387041Z
1 likes, 0 repeats
@RustyCrab @DrTransmisia @nekofag @FrailLeaf yeah i know that feel
(DIR) Post #ASc6ij2bCck6tGHIq8 by RustyCrab@sleepy.cafe
2023-02-12T19:55:56.252346Z
1 likes, 0 repeats
@jeff @DrTransmisia @nekofag @FrailLeaf I'm eternally grateful that we can still catch segfaults in our own code though.
(DIR) Post #ASc6nCSNmExJHRTIzA by jeff@federated.fun
2023-02-12T19:56:41.110481Z
1 likes, 0 repeats
@RustyCrab @DrTransmisia @nekofag @FrailLeaf oh hell signal handlers for sigsegv terrify me
(DIR) Post #ASc6rq7mEfb7Sw2uqu by RustyCrab@sleepy.cafe
2023-02-12T19:57:34.941170Z
1 likes, 0 repeats
@jeff @DrTransmisia @nekofag @FrailLeaf I just use them to flush logs tbh
(DIR) Post #ASc9AXjF3RIZp4cG1I by nekofag@rdrama.cc
2023-02-12T20:23:21.394693Z
0 likes, 0 repeats
@jeff @DrTransmisia @FrailLeaf @RustyCrab Segfaults are the least, if not the VERY LEAST of anyone's concerns.Memory safety and the tools C++ offers are to prevent buffer overruns... which Rust honestly does better at this point.Exceptions and segfaults are NOT synonymous, but exceptions and errors are though. Exceptions are annoying because they don't encourage error checking, while the Go approach to errors is probably the most mature and DOES encourage error checking of code.
(DIR) Post #ASc9T8TVAinh097T3w by RustyCrab@sleepy.cafe
2023-02-12T20:26:43.919997Z
1 likes, 0 repeats
@nekofag @jeff @DrTransmisia @FrailLeaf I hate exceptions because they're just a formalized way of saying "this is someone elses problem" and they encourage "just doing" stuff until you hit an error which is a great way to completely fuck up your program state if you're not eternally vigilant.
(DIR) Post #ASc9Vgb5HEkLPPskoi by RustyCrab@sleepy.cafe
2023-02-12T20:27:08.406547Z
0 likes, 0 repeats
@nekofag @DrTransmisia @FrailLeaf @jeff ... exceptions are bad for the same reason that goto is bad
(DIR) Post #ASc9qKQ0voRehkXzGq by nekofag@rdrama.cc
2023-02-12T20:30:54.733544Z
0 likes, 0 repeats
@RustyCrab @jeff @DrTransmisia @FrailLeaf Not really. Goto at least makes it obvious in code when an error is going to occur. Exceptions hide that, and indeed, only occur when you run into them. Much like segfaults.The smart_ptr + exception pattern IS an alternative to the goto pattern, but the C++ dev just pretends like Goto is bad 100% of the time... but the goto pattern for error handling has never actually been a bad thing in C development. It's collective knowledge. Goto's are only bad when you actually use them to do what a while loop or whatever already does or literally jump between logic/conditions.
(DIR) Post #ASc9r7Zy9aWVAoa85g by jeff@federated.fun
2023-02-12T20:30:55.152102Z
0 likes, 0 repeats
@nekofag @DrTransmisia @FrailLeaf @RustyCrab how tf did smart pointers try to patch up exceptions?
(DIR) Post #ASc9yqlYY3PBLSge9Y by nekofag@rdrama.cc
2023-02-12T20:32:27.507203Z
1 likes, 0 repeats
@jeff @DrTransmisia @FrailLeaf @RustyCrab because smart pointers kill themselves when an exception occurs, which usually exits the scope... which triggers RAII. I meant to say RAII pattern
(DIR) Post #ASc9z1onSLX7chHtEO by RustyCrab@sleepy.cafe
2023-02-12T20:32:29.523308Z
1 likes, 0 repeats
@jeff @DrTransmisia @nekofag @FrailLeaf yeah I don't get how that's related either. They're meant to prevent access violations via staleness.
(DIR) Post #AScA2OPVuoUeYo49L6 by jeff@federated.fun
2023-02-12T20:32:59.643525Z
1 likes, 0 repeats
@RustyCrab @DrTransmisia @nekofag @FrailLeaf excepts are useful as it allows me to simply return types to not have to be a std::pair if i want to propagate an error calling it. i wish `noexcept` was the default in C++ and one had to explicitly mark things that can throw.
(DIR) Post #AScA9rPAVsK8VoSua8 by jeff@federated.fun
2023-02-12T20:34:21.868760Z
1 likes, 0 repeats
@nekofag @DrTransmisia @FrailLeaf @RustyCrab smart pointers are not the only thing you can do with RAII and smart pointers without RAII is possible, sort of.
(DIR) Post #AScABYVrix4OUzPF68 by RustyCrab@sleepy.cafe
2023-02-12T20:34:45.506575Z
1 likes, 0 repeats
@nekofag @jeff @DrTransmisia @FrailLeaf I definitely wouldn't attribute smart pointers to just problems exceptions. Scopes/branching not patching up resources is a huge issue in C in general.
(DIR) Post #AScAIQxEE24z5DU6tM by jeff@federated.fun
2023-02-12T20:35:52.589112Z
1 likes, 0 repeats
@RustyCrab @DrTransmisia @nekofag @FrailLeaf if you use C++ like it is C you tend to have objectively inferior perspectives on how to use the tools in the C++ language.
(DIR) Post #AScAQCCQy12ndfIXs8 by nekofag@rdrama.cc
2023-02-12T20:37:23.435590Z
0 likes, 0 repeats
@jeff @DrTransmisia @FrailLeaf @RustyCrab I can use a language however I want to use a language
(DIR) Post #AScAU9kG2CftBW0W6i by jeff@federated.fun
2023-02-12T20:37:59.510814Z
0 likes, 0 repeats
@nekofag @DrTransmisia @FrailLeaf @RustyCrab yea you can in fact choose to use it the wrong way. just sayin that it's the wrong way.
(DIR) Post #AScAi9R7yj7gHXjUnY by RustyCrab@sleepy.cafe
2023-02-12T20:40:38.385401Z
2 likes, 0 repeats
@jeff @DrTransmisia @nekofag @FrailLeaf I never write in C style if I can avoid it. I use everything I can from the new dialects. The only exception is... exceptions. I think the only other exception is that SOMETIMES printf-type functions can be unironically much easier to work with than << stream style calls because all your format is 'right there'. I've taken a look at several printf style fmt libraries with type safety but they always end up being a giant clusterfuck and cause compilation issues of some kind of another.
(DIR) Post #AScAihpQ6ukywjMNNo by jeff@federated.fun
2023-02-12T20:38:56.533576Z
1 likes, 0 repeats
@RustyCrab @DrTransmisia @nekofag @FrailLeaf just pray to glibc jesus they dont have a bug in fflush
(DIR) Post #AScAm5bs9EhLZWdWQC by jeff@federated.fun
2023-02-12T20:41:07.136065Z
0 likes, 0 repeats
@RustyCrab @DrTransmisia @nekofag @FrailLeaf std::format is in C++20
(DIR) Post #AScAzYeDUtiROb9gVE by RustyCrab@sleepy.cafe
2023-02-12T20:43:47.868187Z
0 likes, 0 repeats
@jeff @DrTransmisia @nekofag @FrailLeaf I'd love to use C++ 20 but it's just not really an option right now
(DIR) Post #AScB3hFzQk2dFsnSvA by jeff@federated.fun
2023-02-12T20:44:25.522480Z
1 likes, 0 repeats
@RustyCrab @DrTransmisia @nekofag @FrailLeaf same. fmt::format's template specializations are also kinda ass rn
(DIR) Post #AScBGZ5H7zd1fjvB0i by nekofag@rdrama.cc
2023-02-12T20:46:51.233256Z
0 likes, 0 repeats
@jeff @DrTransmisia @FrailLeaf @RustyCrab There are _so_ many factors to consider when you want to call something right or wrong. Everyone's opinions differ. C++ i tend to dislike because it doesn't respect compilers, but just itself. Look at `constexpr` for example... do you REALLY think compilers aren't smart enough to consider how to optimize things that should be constexpr anyway??? It just adds MORE work to the programmer and makes code look scarier for very little reason. Some compilers straight up ignore constexpr because the compiler is able to deduce how something should be constant on it's own. It makes compilers do a lot more work since it offers so many different ways to do things, which is why compilers spend more and more time compiling C++ code. If C++ wanted to make things the "right" way then it wouldn't offer the wrong options at all.What if I just want to write C with Templates, my own string classes, and less implementation-specific functions? What if I just want a little bit of safety in crusty areas? What if I hate << and >> operators? What if I really, really need object abstractions? What if I actually _like_ exceptions? What if I want built-in ways to do things compactly without using macros? C++ is brilliant for letting anyone do whatever they want. There is never ever going to be a "right way" to do things, or it wouldn't offer it.
(DIR) Post #AScBTXryxEJPfctQYK by RustyCrab@sleepy.cafe
2023-02-12T20:49:12.723874Z
1 likes, 0 repeats
@jeff @DrTransmisia @nekofag @FrailLeaf C++11 wasn't even viable to use in Microsoft's ecosystem until like 2015. As much as I hate their compiler I'm stuck with it for now. I do give it credit for being fast though. It compiles my project in 1/5th the time it takes GCC and clang because it handles templates headers so well
(DIR) Post #AScBe8YBQnqwKTgOPI by jeff@federated.fun
2023-02-12T20:50:58.605859Z
0 likes, 0 repeats
@RustyCrab @DrTransmisia @nekofag @FrailLeaf wow it's skill issues again. smh
(DIR) Post #AScBex5DYFzD3PWTq4 by jeff@federated.fun
2023-02-12T20:51:11.869332Z
0 likes, 0 repeats
@nekofag @DrTransmisia @FrailLeaf @RustyCrab k
(DIR) Post #AScBfcpkKAZAXabiYS by RustyCrab@sleepy.cafe
2023-02-12T20:51:23.002629Z
1 likes, 0 repeats
@nekofag @jeff @DrTransmisia @FrailLeaf constexpr is mostly useful for compile time checking and static asserts. I don't think anyone seriously uses it for optimization.Actually they're quite useful for hashed strings but other than that I haven't seem them used for speed boosting.
(DIR) Post #AScC57Rzd1y6jkeZWq by nekofag@rdrama.cc
2023-02-12T20:55:58.304487Z
1 likes, 0 repeats
@RustyCrab @jeff @DrTransmisia @FrailLeaf Static assert is useful, yeah
(DIR) Post #AScCXRbGJ4ldL0Ullo by RustyCrab@sleepy.cafe
2023-02-12T21:01:07.735498Z
0 likes, 0 repeats
@jeff @DrTransmisia @nekofag @FrailLeaf I really wish GCC handled templated headers better. It doesn't feel that bad in a small project but when you start hitting a million lines the compile time just shits itself inside out. I tried clang and it really isn't much better.
(DIR) Post #AScF0Ti3k3oABeeAd6 by DrTransmisia@rdrama.cc
2023-02-12T21:27:33.903373Z
0 likes, 0 repeats
@RustyCrab @jeff @nekofag @FrailLeaf Will c++20 modules solve this problem?
(DIR) Post #AScF0VvFVzXz38pjSy by jeff@federated.fun
2023-02-12T21:28:38.606834Z
0 likes, 0 repeats
@DrTransmisia @nekofag @FrailLeaf @RustyCrab no idea, haven't used them yet
(DIR) Post #AScFNGSJBbmxjz92ki by RustyCrab@sleepy.cafe
2023-02-12T21:32:53.515566Z
0 likes, 0 repeats
@DrTransmisia @jeff @nekofag @FrailLeaf I have no idea. Never tried them.
(DIR) Post #AScNOqeMttxvv23HbU by Coyote@social.singing.dog
2023-02-12T23:02:41.905313Z
2 likes, 0 repeats
@RustyCrab @jeff @FrailLeaf Recent versions are pretty good about C++17 support. We switched over to it last year as the default standard at my job, and we haven’t had any real troubles with it.
(DIR) Post #AScNpo2wglVJf9P0PA by Coyote@social.singing.dog
2023-02-12T23:07:37.287078Z
2 likes, 0 repeats
@RustyCrab @jeff @DrTransmisia @nekofag @FrailLeaf Also speaking of msvc and breaking code, it’ll accept zero width space as a valid user defined literal. It’s a great way of making those magic numbers just a little bit more magical.
(DIR) Post #AScPr2mnZZPnZgZFmy by Coyote@social.singing.dog
2023-02-12T23:30:16.888912Z
2 likes, 0 repeats
@nekofag @jeff @DrTransmisia @FrailLeaf @RustyCrab I feel people seem to view exceptions and errors as mutually exclusive when there’s no real reason they can’t live together. If errors are a regular thing then error checking should exist for it. Exceptions are for their namesake: stuff that happens so rarely that it’s not worth the time to consider them everywhere, if at all. You’d use error checking for stuff like a web request timing out and exceptions for stuff like a GPU getting unplugged. I do hope we get std::expected from P2549. It’d mean error checking with a return type won’t have to have a special implementation for every individual code base.