14 Jun 2023 |
kvark | kvark at fastmail com | 03:10:05 |
Ralith | sent | 05:00:52 |
kvark | got it, thanks! | 05:17:53 |
| Aditya Suresh joined the room. | 23:11:41 |
15 Jun 2023 |
| suidieni joined the room. | 10:35:22 |
suidieni | is there a repos for the wgpu website? https://wgpu.rs/ | 10:36:42 |
David Huculak | Seems like: https://github.com/gfx-rs/wgpu-rs.github.io | 12:57:11 |
David Huculak | Oh someone already answered | 12:57:35 |
17 Jun 2023 |
David Huculak | anyone here know of a way to view git changes between branches in a GUI app where it follows renamed files like the regular git diff command does? git difftool --dir-diff doesn't seem to be working | 23:30:16 |
David Huculak | github is actually able to do it but I don't like their UI.. I want tabs or a tree | 23:31:56 |
David Huculak | maybe something that can read patch files (direct output from git diff command) | 23:35:44 |
David Huculak | * maybe something that can read patch files (direct output from git diff command) so I can control the -M flag | 23:40:01 |
20 Jun 2023 |
m4b | so is most of parking_lot in rust std now or did i misremember that? | 06:51:10 |
cwfitzgerald | In reply to @m4b:matrix.org so is most of parking_lot in rust std now or did i misremember that? No, I think they're still separate | 10:37:56 |
cwfitzgerald | Rust mutexes got faster recently | 10:38:16 |
m4b | Ah. Bummer. I seem to recall the mutex work being folded in :/ I know their are api differences thst can’t change eg panicking vs not but I was under impression the impl was being swapped out for parking lot | 19:59:54 |
21 Jun 2023 |
| Joe O'Connor joined the room. | 14:54:50 |
23 Jun 2023 |
| Timo ⚡️ changed their profile picture. | 18:57:35 |
25 Jun 2023 |
m4b | has anyone done any ragdoll stuff with rapier by any chance? i don't see much of anything at all searching for ragdoll on crates.io either | 21:02:48 |
8 Jul 2023 |
Ralith | figured it's just a matter of joints | 16:40:34 |
14 Jul 2023 |
cwfitzgerald | man, xcode's UX is so bad | 03:54:03 |
18 Jul 2023 |
m4b | so, i updated to rustc, it shows me that a drop(&mut Foo) was not doing what i wanted in some very old, very unsafe rust code i had. and so then i embarked on an odyssey into unsafe rust land, where i just spent like literally hours trying to fix the memory leak of basically a [MaybeUninit<Box<T>>; SIZE] that i just gave up and rewrote it all with [Option<*mut T>; SIZE] and now valgrind reports no leaks, all tests pass, etc. The fact that I couldn't rely on Box being null is the major pain point there (and in fact the code ported to Option<*mut T> just creates a Box and then does into_raw and during dealloc does same thing | 05:53:16 |
m4b | its still not clear to me how i could even have gotten it working with MaybeUninit tbh; it's a very complicated api, with lots of gotchas | 05:54:19 |
m4b | also, valgrind is so awesome; thankfully i generally don't ever have to use it, but it's sooo good | 05:54:59 |
cwfitzgerald | In reply to @m4b:matrix.org its still not clear to me how i could even have gotten it working with MaybeUninit tbh; it's a very complicated api, with lots of gotchas What are you finding difficult about it | 05:59:18 |
cwfitzgerald | It does have a lot of gotchas, but such is life on the edge | 05:59:33 |
m4b | well it's api is super annoying for one, in various ways; it feels extremely verbose, and way less typesafe than the Option<*mut t> version, and depending on what i did, would maybe/maybe not segfault in release, etc. i rewrote it to option version and it compiled and passed tests and valgrind on frist try 🤷 | 06:01:08 |
m4b | of course that i had option to rewrite as Option<*mut T> was nice, there are certainly cases where MaybeUninit would be required | 06:01:54 |
m4b | (no pun intended) | 06:02:02 |
cwfitzgerald | I mean that makes sense that using uninitialized memory is much harder than just using option :) also why not just option box? | 06:02:08 |