| 14 Apr 2021 |
 Dmitrii Okunev (at work) | Here's the gopanic function: https://github.com/golang/go/blob/master/src/runtime/panic.go#L951 | 19:16:49 |
 diamondburned | 276ns/op is pretty fast, actually | 19:17:04 |
| diamondburned | I still would never do this, but that was a lot less than I expected | 19:17:16 |
| diamondburned | I was expecting a number around 20000ns/op | 19:17:30 |
 Discordian | They made it way faster recently | 19:22:29 |
| Discordian | Maybe the errors.New should be inside the loop | 19:23:35 |
| diamondburned | making global the errors.New thing is pretty common | 19:23:57 |
| diamondburned | people probably just don't do it because it's effort, like me | 19:24:04 |
| diamondburned | you should add errors.Wrap/fmt.Errorf %w though, that'd be interesting | 19:24:21 |
| Dmitrii Okunev (at work) | In reply to @diamondburned:matrix.org
you should add errors.Wrap/fmt.Errorf %w though, that'd be interesting I prefer to avoid fmt.Errorf in places where performance is really important. | 19:46:42 |
| Dmitrii Okunev (at work) | In reply to @xaionaro-guest:matrix.dx.center
Here's the gopanic function: https://github.com/golang/go/blob/master/src/runtime/panic.go#L951 Added one more experiment: what if there's a stack of 100 calls between panic and recover. In this case it takes 3300ns against 700ns. | 20:11:20 |
| 15 Apr 2021 |
|  Moritz joined the room. | 06:41:17 |
| Moritz | question: has anyone had an actual use for full unicode support apart from deliberately cursed code like this: https://github.com/Ta180m/Idiot-Code-Golf/blob/main/gcd/gcd.go ? | 06:42:06 |
| Moritz | * question: has anyone had an actual use for full unicode support apart from deliberately cursed code like this: https://github.com/Ta180m/Idiot-Code-Golf/blob/main/gcd/gcd.go ? | 06:42:27 |
| diamondburned | yeah, some math codes use mathematical symbols in Unicode | 06:45:56 |
| diamondburned | Go also uses middot internally to delimit package name and function/type names in symbols afaik | 06:46:15 |
| Moritz | i can get behind the latter, but the first seems to make mathematician code even harder to read and understand | 06:47:07 |
| diamondburned | math is already full of symbols anyway | 06:48:04 |
| diamondburned | you should've seen the stuff that formal proof people write | 06:48:33 |
| Moritz | I have… that's why I am strongly in favour of union over ∪ | 06:49:38 |
| Dmitrii Okunev (at work) | In reply to @moritz:moritz.sh
question: has anyone had an actual use for full unicode support apart from deliberately cursed code like this: https://github.com/Ta180m/Idiot-Code-Golf/blob/main/gcd/gcd.go ? Code could be obfuscated and without unicode. If a programmer wants to obfuscate then you won't stop them. | 09:39:14 |
| Moritz | sure, but obfuscation in a compiled languages isn't that useful, right? | 09:41:55 |
| Dmitrii Okunev (at work) | In reply to @moritz:moritz.sh
sure, but obfuscation in a compiled languages isn't that useful, right? Somebody may find it fun. Moreover there are obfuscation contests with prizes, where people use C. | 09:45:31 |
|  Phillip Crandall joined the room. | 10:32:04 |
| 16 Apr 2021 |
|  @01010011e:matrix.org joined the room. | 13:45:21 |
| @01010011e:matrix.org joined the room. | 13:47:09 |
| @01010011e:matrix.org joined the room. | 13:48:35 |
| @01010011e:matrix.org joined the room. | 13:48:37 |
| @01010011e:matrix.org left the room. | 13:48:37 |
|  nil joined the room. | 19:32:20 |
