alexene.dev
Principal Engineer @ Prime Video. I write rust every day 🦀. Fan of cats. Making https://dwarf.world in my spare time. Preciously engine guy at Ubisoft, EA, Killhouse games
https://linktr.ee/alex_ene
58 posts
106 followers
176 following
Regular Contributor
Active Commenter
comment in response to
post
So I'm back on zed now, even tho I absolutely don't like the search interface. It's significantly slower than vscode even, forgets filters between searches and also doesn't show a list of files :(
Git integration isn't fantastic either, wish they just copied lazygit
comment in response to
post
And with juniors, newbies they used to make a clear category of errors now they just generate tricky believable code that’s fundamentally wrong in ways that are harder to catch
comment in response to
post
It’s not like i can use Claude to review PRs that are frankly just garbage, Claude generated that in the first place, and what i notice is that it gets harder and harder to spot errors.
comment in response to
post
Before the ppl who didn’t care and/or don’t know much due to where they are in their career time weren’t amplified in their output as much
comment in response to
post
It’s not even that people don’t care, it was like that before and it is now but what I deeply dislike is this asymmetry where clearly ai-powered PRs make their way to my repos and require human power to reject them.
comment in response to
post
Also maybe the question should be is it worth it for the unmeasurable x% productivity boost? I honestly don’t even know what x is looking at the power users. For me, maybe 2-3% but it’s feelings based again 😅
comment in response to
post
Ultimately it’s a lot of preference-based discourse on both sides. Even this article goes into how things feel because really we have 0 to no data in how much it actually helps. Saves 1h in a script but also leads you astray on a 2h debug spree due to some slightly wrong code that looks correct
comment in response to
post
I know, we are making progress and things are progressing.
But imagine how cool it would be if we have this awesome language and instant incremental builds! How cool would that be?
comment in response to
post
At work, the thing that frustrates me the most about rust is the absolutely super slow iterative build times. I could work and iterate faster in assassin’s creed anvil engine and that thing was massive
comment in response to
post
People should write more about this since it would help. Last time i saw some rust improvement of compile time article was some team who had broken up the project in 1000 crates
comment in response to
post
I don’t really understand this. Every software I’ve worked on, of similar architecture to a c++ one, is slower to iterate when written in rust. How will bevy be faster or similar when we don’t even have dynamic libraries yet 😅
comment in response to
post
I really like the language and also wish we would acknowledge this as a big problem
comment in response to
post
I feel compelled to say this again since there’s a bunch of posts like this around build times. Build times is by far the NR 1 pain i feel on iteration times with rust when working on my game. If i would have used c or zig it would build x10-x100 faster easy especially for incremental builds
comment in response to
post
But for me, when i'm fine to build my own engine and the pain that comes with it, build times is just the 1st pain BY FAR. Similarly sized C++ would be at least 10x-20x faster, if not 100x (e.g. for incremental builds) and i'm not even exagerating.
comment in response to
post
Another one is the immaturity of the gamedev ecosystem, but that's more of a: we thought we can build everything from scratch but turns out it's hard and we would like to use an engine and simpler programming model (e.g. c#). Fine either way imho.
comment in response to
post
Build times is a rust-specific pain i'm having on iteration times in my game. This is where rust significantly fails in comparison to alternatives, even for similar/ientical code. I love the language, use it at work too, but I wish we just put build times as the number 1 priority and tackled it
comment in response to
post
It’s their software to do as they please with it I guess
comment in response to
post
www.youtube.com/watch?v=F9NF...
comment in response to
post
Day/night can be as easy as having a light color you multiply with, just tint things towards dark blue at night
comment in response to
post
This was an adventure :D
comment in response to
post
It works by carefully projecting the 2D art onto cubes and quads.
E.g this is drawn by pixel artists in aseprite and then i carefully project it on cubes so it's pixel perfect in the isometric view.
comment in response to
post
Thanks! It is not open source
comment in response to
post
I just wish on these kinds of talks the presenters would also go into what are the downsides of choice x or y