#programmers-off-topic
1 messages · Page 22 of 1
Not for performance, obviously, but in terms of usability (both dev and user, like a11y support and such) web tech is so far ahead that it would take a huge team of developers several years to catch up, and that's exactly the opposite of what we are really seeing, desktop UI is starved for developers while we've got far more actively-maintained web frameworks than anyone could ever need.
Hiss if you will, but that's the reality today.
The ease is absolutely ridiculous, yeah.
yeah
even though css and js have their fair share of issues. They're difficult to beat in how convenient and well supported they are
there are gui frameworks, they just don't tend to be cross platform
or if they are they have a very clear strong preference for one

honestly the last time I considered building a native gui I was seriously considering using Godot as my framework
Electrino is a thing, but it doesn't work on Linux yet for... pretty key reasons.
tbh even if I did make a cross platform app it'd be a pain cause one of the devices I'd want it on is my tablet which doesn't support sideloading
and I'm not publishing a hobby app intended for just myself
Then you have no choice but to make webapp
There are decent mobile-specific options if you only care about mobile, like flutter. and React Native - the latter of course being based on web, but technically actually native, unlike Electron etc.
There used to be an older one called... phonekit? Phonegap? Something like that. Probably discontinued by now.
Qt seems pretty good so far tbh. Problem is they really want you to use their toolchain/IDE
Im liking their UI editor more than i thought, but then again ive only played around a tiny bit
it all falls apart eventually…
wysiwygs are nice till they very much are not
depending on your needs it might be fine tho
its fine for getting first draft but chance it explod goes up with each iteration
The UI editor for Qt is optional at least, it's just... so damn convenient.
I do like Qt a lot.
maybe I'll try it someday
I used PySimpleGUI recently but they're going commercial
It was very fast to set up, I will say that
I'll just be strange and pick some random rust gui project instead
Qt Creator is pretty good even on Windows
PyQt is a pretty decent way to make quick little things with Qt too.
someday
writing a function and then forgetting to call it has got to be my favorite error
I'm digging Qt actually. Once I figured out how to make it generate a normal makefile, it's pretty slick
how much does it help that you actually know what a normal makefile is
I'll just be strange and pick some random rust gui project instead
Iced was the best I found so far. Though that was over a year ago, the scene may have changed.
iced is the only I know of that has an actual commercial project using it
since system76 picked it to be their gui framework of choice for their new DE
Let's check the ol' areweguiyet.com and... nope, still anemic.
Not super important. Qt has its own qmake format, which basically just lists different source/header/whatever files. I think normally the IDE would automatically populate that, but since I'm adding to an existing program, I did it myself. You then tell qmake to create a makefile and that does the compilation. The qmake format is much simpler than the nonsense make can be
Xilem shows a lot of promise, if they actually finish it.
QT += core gui
greaterThan(QT_MAJOR_VERSION, 4): QT += widgets
CONFIG += c++17
SOURCES += \
main.cpp \
Arqade.cpp \
AudioPlayer.cpp \
Core.cpp \
CoreData.cpp \
Utils.cpp \
VideoPlayer.cpp \
Window.cpp
HEADERS += \
Arqade.hpp \
AudioPlayer.hpp \
Core.hpp \
CoreData.hpp \
libretro.hpp \
Utils.hpp \
VideoPlayer.hpp \
Window.h
FORMS += \
Window.ui
INCLUDEPATH += /usr/include/SDL2
LIBS += -L/usr/lib/libSDL2.so -lSDL2
DESTDIR = bin
OBJECTS_DIR = $$DESTDIR/.obj
MOC_DIR = $$DESTDIR/.moc
RCC_DIR = $$DESTDIR/.qrc
UI_DIR = $$DESTDIR/.ui
is my project file for Qt
I think the main ones people are looking at this days are dioxus, xilem and iced
arqade
I think Dioxus is superseded by Xilem? Unless I'm getting it confused with a different abandoned project.
Never mind, looks like it's still actively developed.
Yes, you're right, it was Druid I was thinking of.
my poor attempt at a libretro frontend
at least the bar is low considering the default
I think Rust has more abandoned UI frameworks than any other language has total frameworks. To be fair though, Rust hates trees.
Javascript
I don't know, I think Rust might give JS a serious run for its money in terms of abandoned project count.
Maybe not if you consider utility libraries like jQuery and Underscore to be UI frameworks because there are like, a hundred of those alone.
jquery itself however keeps trucking along
lurking
honestly
with the volatility of the Javascript community, it might someday be considered good again
even in just the short few years I've paid attention they've drastically changed their mind on what "the right way to do web development" is several times
yes
idk the last time I paid attention people were suddenly all about htmx
I'm sure that's over now
of course the Javascript community and the actual web development industry are two different things
the industry is still predominantly react
I don't know, there's always been a fanatical corner of the web that's chasing the new and shiny but the mainstream hasn't changed that much over the years. Angular and React are the big boys, and have been for a long time. Svelte is putting up stiff competition but isn't really ready yet (we'll see with v5, maybe it will be). And Qwik is the new, noisy, obnoxious kid on the block.
There's other stuff out there (MeteorJS) and I'm aware people use 'em, but they're not that fanatical about it.
the fanatical corner is the fun corner. Watch the same YouTuber change their mind every few weeks or months with some new shiny thing
angular seems to be liked again a bit
it's been around long enough to have gone through several cycles of popularity itself
it's the first web framework I ever used
I don't remember a thing about it
Heh, I guess for some the holy wars are fun. I was into them many years ago. These days I prefer to just get shit done rather than argue about how to do it.
I only observe
I think there's a happy medium between being a lumbering dinosaur stubbornly sticking to WinForms and being a hyperactive child gushing over the newest hotness every 8 days.
"what are these people up to now" as I go back to my completely impractical frameworks made by a single guy with a fulltime job
I assume the single guy is you, and that making the frameworks isn't actually your fulltime job.
as someone who still works professionally with winforms
Meanwhile, I'm home early and repairing pants
oh no I don't make frameworks I just look at them
MeteorJS, now that takes me way back, to the days of hackathons and the like... as far as I'm concerned, even now no one uses it outside of hackathon prototyping
There is Vue I guess, that one does seem to attract some fanboys.
I never have any of my XDG variables set, and idk if thats bad
do you guys like using git submodules
in the few times i encountered it i always messed up somehow but maybe it is skill issue on my part
they're better than the alternatives
Yes
Wait
The question was "like"
Like is too strong of a word but we do use them
do you enjoy it, does it spark joy
It sparks less misery than the alternative?
Fair enough 
I am considering splitting my 1 stardew mod repo into N submodules so that github update key would work
Have a logger I use as a submodule, only problem is extra cloning steps/args
Have fun!
But I never seen anyone else do this so I wonder if this is actually a dumbass idea in C# context
It's not a dumbass idea I just don't want to think about mods
(I have submodules in my repo and every now and then someone asks me how the fuck do they fork that
I kinda like https://kdl.dev/, haven't used it much because it's barely used by anything but it feels like the best of json and yaml.... but there aren't lists
yeah this is my understanding, it's passwords but a protocol that generates the passwords for you
well I can't say I would've guessed what KDL stood for
I've been liking toml personally
Successfully made my own minimal cli gui with clickable buttons, nice text wrapping, and text input
Next up is progress bars, and then I can start on the actual thing it's for
I use submodules but not in that context - I use them for when I need to depend on a fork of someone else's open source project that I don't expect to get upstreamed.
Not sure why I'd want to use submodules for my own repos... maybe for C++ projects I would because of their silly include systems.
I know there's libs for that already but they're all these massive full gui libraries with windowing and stuff which is extreme overkill for what I need. Also I wanted to see if I could do it myself
(Yeah, if you version your kicad parts database it's a classic reason to use submodules)
You can make the update keys work now, but you have to maintain a file in your repo pointing to the current versions of each mod. https://stardewvalleywiki.com/Modding:Modder_Guide/APIs/Update_checks#Custom_update_manifest
i did a thing! learning cpp
this took AGES it felt like
i just got done rewritting an example compiler to match the given example and i now notice that im painfully slow typing 
That uh... doesn't look like a compiler.
Unless the project was to write a compiler that takes that code and produces correct output for it.
oh no this is a C to F(and vis versa) converter
idk, i just know the exampled called it a compiler, i have no clue if it actually is. still to early for my to tell that kind of stuff
heres the ExampleCompiler that i copied by hand 
compiler is like, you have human readable code and it becomes a binary executable of some sort
oh, no then this does not do that
oh this is a interpreter
you have human readable code that is read by program, which executes the instructions
tho this one doesnt execute them either it just parses
either way glad u r have fun in cpp land 
i wanted a lower level language cause stuff like python is a bit too much like english? if that makes sense
i wouldve learned it deeper but nooooo windows locks its stuff behind a windows specifc api
grumbles
i am! it feel slow cause i started just last night but i know eventually ill get there
or at least a decent way there, only so much i can learn from the way im doing it now
ohhhhhh my goal is to learn how to make a compiler eventually, so this is step one of learning the "simple things"
not necessarily. a compiler just converts higher level code into lower level code im pretty sure
pretty sure the main definition of transpiler is converting between two languages of similar level as well
Yeah I went with the general usage definition
Anyway, the second example posted would be considered a lexer to me. It does not validate, or output an AST, so it is not really a parser (and many steps away from a compiler).
Not exactly sure what it has to do with the first example, either. The second example doesn't appear capable of lexing the first; only the very simple int a = 5 expression given in main().
Not that I'm discouraging the effort, but for someone just starting out with C++, writing your own compiler might be a bit... advanced.
this is why i like posting stuff here, cause i get other opinions, cause most youtube vids ive found on learning code say that a compiler is a good goal to work towards in the beginning because it teaches you important skills.
something you can personally use is a good goal 
i dont understand lol
well imagine if you make a alpha version of a cpp compiler
you wont use that over gcc unless you spend infinity+1 time on adding all the features
reinventing the wheel so to say
ofc it is true that learning about how compiler work will give u strong fundamental understanding of programming languages, and it may even be suitable motivation for you
but im not that type of person, i need the tangible feedback of making a thing that i can end up using for some purpose
modding is one such example ofc
makes sense to want something useable, im actually learning coding for that reason, so that i can start modding MC and stuff. and also just mess with my pc a bit.
Sadly im not at the stage where i can worry about wether what im making is useful, im at the stage where i just gonna make something functional
one step at a time 
Sure, understanding how compilers work and even being able to write an elementary one is a very good tool to have in the belt. I'd just pick something a little more approachable and practical if this is literally day 1 or week 1 of learning to code.
Unless you are actually working professionally on compiler development or making your own esolang, then those skills generally come into the picture at the point where you're trying to solve very complex problems that require metaprogramming (code that generates other code) or domain-specific languages, which happens when you've reached the limit of what good architecture can do, or just gotten so efficient at implementing the conventional solutions that you decide it's worth your time to automate yourself.
Learning compilers in order to learn programming is sort of like learning to fly commercial passenger aircraft in order to learn how to ride a bike. Or maybe learning how to build a bike in order to learn to ride one. Either way, kinda out of order.
It's like learning metalsmithing to ride a bike
wouldn't be butterbear if it didn't involve trying to learn something by diving into one of the most complicated aspects of its entire field
as in they've been talking here for a while, yes
originally from what I remember trying to build a custom llm chatbot as an introduction to programming
in my defense this time i truelly didnt know, was just following what i heard on youtube
as long as it gets you to engage with programming at all I think it's fine
that's ultimately the biggest barrier
True enough. Though that sort of takes me back to previous comments about YouTube being a poor educational choice for cognition-oriented skills (great for monkey-see monkey-do type stuff).
Which is not intended as a knock, by the way. I use it all the time for basic repairs, plumbing, landscaping, and so on, which are in the "just show me what to do" category for me.
(Tl;dr i suspect I will be at some point knitting Taylor Swift's folklore sweater.)
It looks so comfy
.img taylor swift folklore sweater
oh that does look comfy
the stars are fun
Someone even made a pattern!
Although the suggested yarn is kinda trash
Suspect cascade 220 or something from knitpicks will work better
Part of me does want to get into knitting but I already don't have enough time for all my hobbies
same
well
the not having enough time for all my hobbies part
I've yet to feel drawn to knitting so far
Omggg!
Fwiw knitting has the benefit that I can do it while doing other things
I regularly knit at work
honestly have never worked on a project where compiling takes more than a minute
though whenever I build/install anything C++ from source it seems to take 14 years to compile/link
Things that take awhile for us are opening websites in VS that don't have a solution, running some processing scripts that reference an ancient type of db, and some crazy SQL migrations
okay if I could waiting for pipelines to finish
done that too much
a single release of a new database version takes like 25 minutes
because of some fucked up system for releasing it isn't even supported by anything anymore so we have some older system specifically for it
and even the build that on my work pc takes like 30 seconds somehow takes like 12 minutes when done by the pipeline
We have a client that hosts like 30 Drupal (PHP) sites on a Windows Server, so I have the pleasure of writing PowerShell scripts to try to automate deploying to essentially an awful CMS framework that's had all kinds of garbage hacked onto it

Me compiling scripts/content at my job where incremental build doesn’t work right 
Every time I need to synthetize new fpga code
Maddening
The worst part is that my tooling sucks
15 minutes total build time
AND IT KEEPS TELLING ME MY BUILD FAILS EIGHT MINUTES IN
Every. single. Time. You need to make a change. It’s a nightmare
(I know why but it is a maddening thing that I don't have proper sims)
/me is currently debugging a script and has to do that before testing every change
the heck are you building that takes 15 entire minutes
Why did /me not work, what a travesty
That sounds awful and I hate it
Not abnormal for fgpa synthesis
It's just badly optimized
Well that sounds awful and I hate it 😛
In my case it’s most of the content since the previously mentioned lack of working incremental build
The bigger issue is that I don't have good sims so I can't just run a quick test suite
W h a t
So I need to synthesize then measure on an oscilloscope
Compiling something taking two weeks sounds like... doing it manually levels of performance
Also hi Khloe, how was FFXIV dawntrail? (I still need to finish post endwalker)
Unless you're compiling an entire operating system or a web browser (see also: entire operating system), the idea of something taking more than like 5 minutes is awful
It's also documented godawfully
So I'll get eight minutes in
And it will tell me my clock speed isn't valid, try again
Can you guys see why I hate labview?
Hi casey! I enjoyed it, though reception in the broader community seems to have been mixed. But I'm all through it and the two expert dungeons now, working on levelling up a healer so I can get mentor status back, and finishing levelling my crafters. Things are going well.
dotnet build
dotnet build
is that it
ye
oh its doing something
depending on the project you might need mono
It's the same way you build the project anywhere you're using CLI
Anyway
i think installing aesprite from AUR (which is just build from source) takes 10min+ for me
theres a lot of .dll in here, that doesnt bode well
Today's work: figure out exactly how incompetent this vendor is (times 2)
Did you tell it to pack the runtime)
i told it to dotnet build lol
Fun fact
Oh, like a LOT of DLLs? Yeah, the runtime might be bundled... potentially?
The temperature forcer just decides to randomly ignore me
did u do it in a repo with a sln file
I think DLL is normal even on Linux for dot net right?
I can never figure out why
Yup!
This is set in the csproj
yes
and csproj pointed to by it right
u will have to inspect what the configs are
but yes dll normal even on linux
i guess i should actually look at where the executable is rather than just poking around
it might be building just a library
In the godawful docs for example
Some pins are enabled by defualt
Others require me to explicitly enable them
Which is which? I don't know and labview doesn't tell me
Labview sounds wonderful!
oh i see what its doing
So half the time even when synthesis succeeds I go to measure it and the pin is ignoring me because I didn't enable it
is .NET not backwards compatable
naw you gotta target a version on build time
Define "backwards compatable"
I have .NET 8 and its mad its not 7
Oh, you should be able to do that
Yeah, I target .NET 6 on Linux using .NET 8 exclusively.
You should be able to build for every version of dotnet if you have the dotnet 8 sdk
y u mad computer
(My poor neglected stardew mods repo requires the net 8 sdk even )
Did you install the sdk or the runtime
I injected it into your pc
try dotnet --list-sdks
Just 8.0.106
Do dotnet --info, that gets everything.
yea u gotta build targeting the version of .net you intend to run ur program on
That should run .NET 7 things just fine...
dotnet on linux 
tho i am assume the repo you cloned ought to have set up stuff right 
what are you trying to build anyway
what's that
SDV mod manager thing
Lemme try building it here...
smh sdv
cant believe sdv talk is happening in the sdv server
dangerously close to on topic
but i think you just gotta install the dotnet 7 sdk
booo
Very interesting here...
Build succeeded, but dotnet run says the same runtime thing. Which is... not correct, because .NET 8 can absolutely run .NET 7, 6, 5 things...
does it run
no
only 3 digits nice
...okay, now I'm just confused. It really seems to want the .NET 7 runtime installed when it shouldn't need it.
glad running c# is so seamless on linux now
wait imma try running it in jetbrains rider
even though that shouldn't make a difference
oh
Rider will automatically download the right version of .NET if you don't tell it not to in the settings though, so...
time to go back to school
This makes me think of how I needed to manually install .NET Core 3 to use the monogame shader compiler tool version that works with Stardew.
Okay, @pliant snow...
cli gui with both mouse and text input and multiline text wrapping
Edit Stardrop.runtimeconfig.json to dotnet 8 instead of 7. Then it'll run!
It's... weird as hell that it's not just working though.
hazah
easy
Can someone kill windows python plz thx bai
amen!
#making-mods-general message
I've been wanting to do a MUD/MUSH/MUCK for a long while but I don't really care for most of the engines that currently exist. I've been thinking about writing my own, but god writing a good text command parser sounds like hell
It's fundamentally my fault because I wanted to use py3.12 for toml support
But I've been arguing with conda for the last half hoir
And this is seriously cool! SMAPI text-based installer GUI coming up maybe? /j 
that would be cool! I might end up doing that, at least as a standalone tool. It's a little too rough around the edges for me to want to make it a PR
can u make the smapi installer close on its own
i keep forgetting so it just sits and eat 10% of my cpu
like right now it assumes if you're running it on linux you have stty to modify terminal settings (which may not be the case) and also doesn't work properly when running it from WSL via windows terminal, because the input is piped
if the installer doesn't support CLI args for install options (ie install/uninstall and folder location) I WILL make a PR for that because that would be extremely useful for automated installs
Thought it did but maybe not documented
probably. SMAPI also has undocumented cli args
does stardrop auto install SMAPI?
I thought it could update, maybe it only lets you know
I always manually install anyways since I have a bunch of different install locations
I think someone else might have pr'ed it in the past
Fucking conda
This is the third installation of python and the second terminal
are you tired of the sneks in your environment
Yes
Alright
It only took an hour to install python right
Code now at least runs
Can windows please make the default python 3.12 plz
I'm actually bitching about tcp
Lol
And the lack of good disposal patterns in python
Atexit seems like my best bet
it's a prerequisite so doesn't seem like it
I need to keep the reference open
So can't use with statements
Also yeah they didn't implement enter/exit
I'm using del/atexit rn
Their python driver is terrible btw
I wrapped it already but their driver requires you to know how they numbered their registers
And whether or not a register is fixed point or an int
(And no it isn't documented I guessed)
dir my best friend
God
I want python to implement null operators so badly
Lol
How do programmers not go around in a murderous rage all the time
im sure pythonists would say you should structure things so they dont need null operators
despite being a language that doesn't come with anything built in to properly avoid nulls
i do a lot of
try:
nullable.func()
except AttributeError:
...
its not great cus AttributeError can come from func itself 
is has is None what else could one want in life
When I was a kid I wrote thjs insane python script that used getattr everywhere lol
Unmainatainable pile of shit
I also built something like dataclasses lol
It also has a walrus operator!
I want to know why it takes three full seconds to send a single command
What
Did
You
Dooooooo
I'm not familiar with your specific configuration, but I've worked with a number of hardware devices and the developers who write their protocols/firmware/APIs tend to be... well, let's just say they tend to be "hardware guys".
From my experience, those ridiculous delays are most likely some combination of error-checking and spinlocking due to ambiguous protocols.
It could also be just a very slow device that takes forever to ack a command, and the client needs to wait for the ack (and what you think of as a "single" command might in fact be several over the protocol).
But there's a fair amount of wild-guessing there, I'll admit.
I think it's the later yeah
They're just using socket socket
It's just the hardware taking its sweet time to get back to me
- it blocking on the main thread which I guess is fine
Sure it's a socket, the question is what goes over that socket. Usually it's some garbage roll-your-own serial protocol and not like, TCP/IP.
Writing your own network protocol is sort of like writing your own cryptography. You really shouldn't do it unless you're an expert; and yet, so many people do anyway.
(It's tcp)
Ah. Then yeah, let's go with potato hardware.
It's kinda fine since this is a temperature controller
And it takes about a minute to reach the right temp
That's how it usually is with this stuff... the hardware is designed to fit the needs of its operational use case, not the developer use case which actually requires a lot more memory/CPU/bandwidth/etc.
Typical Rust program of mine uses under 50 MB when running... chews up 50 GB when compiling.
On an unrelated note, one thing I continue to hate after coming back to C# is how it handles enums. "Your switch is not exhaustive because you don't have cases for values that aren't defined in the enum." KMA
I don't care that it's only one line I have to add, when I have to add it in 57 places. I want compile-time enums, not half-assed constants.
huh. typeof(void) actually works. did not expect that
tf is the type of void
surely this isn't c#
even javascript doesn't seem happy about it
which is saying something
Lol
Crumble
Oh never mind i don't have the obvious pointer bs anymore
Doesn't look scary
looks like some good ol opengl
Just calmly using monogame internal functions
if you ever need a WSL tester i can give it a go at some point
...just having firefox open on a blank page uses a CPU core at 100%
alright it seems to have stopped
firefox: can i have a little bitcoin, as a treat
oh I have WSL that's how I know it doesn't work properly when you use it through windows terminal :P
at some point I may see about installing a separate linux terminal that isn't piped through windows terminal
ahh ok
it's not a huge issue bc if you have windows you can just... use the windows version, but still
im in a weird area of gaming right now
im using AI to right lua for an OS inside minecraft inside my os
and yes... i tried getting arch on the pc... file size was to big :(
(im not learnig lua, brain hurty enough from cpp)
learning lua is probably easier than learning c++
ncurses I think can handle it maybe? but I don't want to bundle a native lib, especially not one that is such massive overkill
I personally hate lua because its rules are so loosey goosey, its syntax is so minimal, and it has no real type system
yeah but i dont wannaaaaaaa 
i wonder if there are any Linux distros out there that are 500k bytes, cause thats all this pc can hold
for now :3
oh no wonder the ai is having a bit of trouble, its got its own api in game too as well and the base lua structer, time to do some digging on my end
what game is this for?
oh opencomputers!
yeah, trying to interact for an ME system for ae2
maybe it's nostalgia but I prefer ComputerCraft (or CC Tweaked since the OG has been dead for years)
if you have create you can use the display thing as a peripheral, it's awesome
sadly i dont or i would have
I'm pretty sure it can use an ME interface as a peripheral. might need advanced peripherals for that?
on my last world I created this really cool fusion reactor room that displayed fuel % and heat and stuff on one of the create ticker boards, it was really cool
thats cool, sounds like something id do
why not?
some day I should do a tour video of that base. it's still unfinished but I don't think it will ever be.
crochet a tailor swift and then the dress
Change my entire aesthetic from tailored and business semiformal to fun and kitchy?
My favorite #1 Artist!

How the heck am I supposed to build a temperature control flow when I'm limited to one action per six seconds
Hysteresis and lots of it?
(I do understand the theory behind PID. But also like. Ughhh)
Also the highest temperature I need to reach is very close to the highest temperature we can tolerate so I can't overshoot
Like. By more than 5C
I overshoot, we might accidentally start melting things
Which is bad bad bad
So I'm probably gonna want to clamp the D in PID for the higher temps at leasr
My knowledge is limited to PC cooling and tractor mods, but hysteresis really is the only thing that comes to mind... you see the temperature going up too high, you rubber-band down to a waaaay lower target temperature so that once the active cooling stops, there's plenty to start it up again if it creeps up.
We do want to be fast about this
(have no idea what PID means unless you're talking about a process ID)
You want to be fast, yet you can only respond every six seconds, and letting the temperature get too high is catastrophic... hmmmm.
This sounds like a job for the Engineer's Pyramid.
(Proportional, integral, derivative)
I'm gonna give it a try with the vendor api but I might start breaking into it if I need to
Ah, derivative, makes sense what you said then - if temp is climbing fast then you'll have to risk overreacting if you can't correct quickly.
New day new python issues
Today, setup.py decided it wanted to continously put in an old version of a file
I have no clue where it even got it from
I would offer advice, but you seem to run into way more python issues than i do
And I would offer advice, but I generally try to avoid being on the same hemisphere as any python being written or edited.
I’ve never had those levels of Python issues
Probably because I don’t write that level of Python
All my issues with python are hypothetical
i dont think i really have issues getting python stuff to work, its getting stuff uninstalled thats the problem
@gaunt wadi give me your pitch on codeberg
the ui is kinda nice
📝
they have docs on mirroring to/from github
is there a tutorial for making an custom npc
If you're asking about Stardew, you might have better luck asking in #making-mods-general , this channel is for non-Stardew discussion
mb
it's the same ui as github but run by people who are not an evil megacorp; the github features it lacks are things ive never used before, it doesn't put "try ai!!! copilot!!!" on every page
in my experience: literally identical besides pushing takes ~3 seconds rather than be instant, this does not matter to me
Sounds like GitLab
no other git platform:
- has the same/nice ui as github (sourcehut is for programming in the 1930s)
- isn't an evil megacorp pushing ai dev sec ops out the wazoo (gitlab)
I forget the others I looked at, but codeberg was the clear choice
sourcehut does have the most radical/anti-capitalist mission stance and the most features, it's just the ui is very non-modern
Does codeberg have github pages equiv
visiting the website doesn't even tell me that quickly that I'm even on a website with git repos
yes
does visiting github tell you quickly youre on a website with git repos
it's called codeberg pages
hmmm what does the github main page look like when you're not logged in
oh yikes what is this
The world’s leading AI-powered developer platform.
Trusted by the world’s leading organizations
Accelerate innovation Our AI-powered platform increases the pace of software development.
🤮

hmmmmmmmm
(use whatever you like, people have different priorities and if you want to get things done that is valid, this is a thing I choose to care about so I vom emoji at github but it's okay if you don't)
hell yeah
Well it'd be nice to move my shit to something accessible from china
my current working model is new stuff goes on codeberg, left the old stuff on github
yeah thats what I was thinking
someone said something about mirror support? Move everything thats worth anything to codeberg, mirror the important stuff still
tbh none of my projects even really need to be public. I could just sync the projects with fuckin syncthing and I wouldn't really lose anything. I'd just be using git without a "remote"
can you just set up a project to have two remotes
@marble jewel I was going to originally build a SFF NAS, but it seems like the mobos you can get for that, especially if you want ECC, are super limited/expensive and I'm not starved for space
I'm not even sure I have anything worth transfering
I don't think I have a repo with a star
oh shit that's really cool actually
These were my builds -
NAS: https://pcpartpicker.com/list/bh3dRv
Game Server: https://pcpartpicker.com/list/4PsGZZ
My goal is to essentially merge the two on a rack, but it will involve a lot of new hardware
Part List - Intel Core i5-13500, Fractal Design Node 304 Mini ITX Tower
Part List - AMD Ryzen 9 3950X, GeForce GTX 950 75W, Cooler Master MasterBox NR200P Mini ITX Desktop
some of my stuff has somehow gotten attention
I think that one just shows up well on searches
I've never promoted it iirc
the aseprite one did get linked in some plugin list, so that one isn't a surprise
neat
If only I was working on a sequel book...
for the gba
not gba
in assembly
I wrote an article about hacking assembly, but i dont think its on github
or is it
no
nerd
Yeah I was looking at the 304 but I think I'm gonna end up with a mATX or ATX for the features and future expansion I'd like. Trying to end up with ECC memory, too
Yeah, if I were to build it today from scratch, that's what I'd do
really cool that you write articles on the stuff you do tbh
Only 72000 words... 
ooo
72k words damn
that does include code, and I'm not sure how much I believe wc
this gameboy emulator gonna be feature complete by the end
but its long
I've been waffling on Proxmox with Unraid VM or just straight Unraid, but with what they've shown in the Unraid v7 beta I think I'm just going to go with Unraid
It's meant to completely walk you through writing one
oh, talking about NASes? i'm using Node 804 for mine
I had to take a break because the one I ended up creating is buggy af and I needed a break lol
That looks similar to mine
it also looks similar to mine
I don't really mess with VMs and from what I've seen a lot of the main reasons to run Unraid in Proxmox are now in Unraid directly, like VM cloning and snapshotting
In my new build I plan to VM everything
I probably shouldve looked into proxmox, but now my server is setup and I dont care to mess around with it lol
I'm running proxmox right now, but it's a single Ubuntu VM, Win11, and TruNAS
back in the day i went with Proxmox... only to proceed running just one VM on it ever, OpenMediaVault
I have like 50 disparate services running across four different pieces of hardware, so I'm hoping I can simplify it all
Yeah I assume most of my use cases will be related to what in my storage that Unraid is controlling, and with the new features in the v7 beta I think they covered most of the reasons I'd bother layering on top of Proxmox.
with my vaultwarden instance gone now I think for now my server is running literally nothing except smb and other filesyncing related things
I think VMs can be nice if you get into platform deployment code, and if you want to stage changes to a non-prod environment before going to prod
That's my plan at least
While it would be nice to separate my services, I feel like giving them access to shared actual files would be a pain, unless proxmox has some way around that
containers are good enough for me tbh
VM Manager Improvements and Improved Hardware Support
Take your virtualization to the next level with VM Clone and Snapshot capabilities, improved VM performance, and a whole host of improvements and optimizations. Additionally, thanks to the updated Linux kernel, Unraid 7 brings Intel ARC GPU support, ensuring you're ready for the latest hardware advancements.
I'm looking forward to unraid 7, but I'll probably wait for a .01 to let everyone else weed out the initial issues
It'll be nice to run on a zpool without an array though
I'm assuming by the time I figure out and get the rest of my hardware that it'll be in a decent state. And not having to learn both Proxmox and Unraid at once will be nice since I don't mess with server/systems stuff much
Right now I have an unused external drive as the Array since pre-7 it's still required
That's a weird requirement
I'm planning to use their ZFS since I have all new same-sized HDDs
ZFS support is still relatively new
Yeah I think they added it like a year or so ago
(Which is probably why there are remnants like assuming you'd be running your Array as randomly sized BTRFS drives)
maybe someday it'll be in the kernel...
I'm kinda considering writing code to make making knitting patterns easier
In particular instead of grading the pattern to set sizes, having users enter their measurements and gauge and just calculating the right pattern on the fly
Anyways did you know there is a latex library for this?
there's always a library for everything
||https://en.wikipedia.org › wiki › Throbber
Throbber - Wikipedia
A typical throbber animation like that seen on many websites when a blocking action is being performed in the background. A throbber, also known as a loading icon, is an animated graphical control element used to show that a computer program is performing an action in the background (such as downloading content, conducting intensive calculations or communicating with an external device).||
Have fun with this funny but cursed knowledge(censored for comedic effect)
exciting times: https://github.com/advisories/GHSA-hh2w-p6rv-4g7w
a security vulnerability in system.text.json
half our applications have suddenly stopped building due to it
uh oh
🎉
chaotic day
ironically the whole team responsible for handling security issues like this was away due to a security audit meeting
conspiracy...
@marble jewel what do you use for a reverse proxy in your lab? I’ve been using nginx for years but I’m thinking of switching to traefik
Traefik 100%. I too switched from nginx and now I could never go back.
It’s just getting to be a lot when I want to quickly set up some new stuff
Great for my web hosting needs but it just feels clunky with my services
whispers caddy
lmao
I have no idea what the differences between any of em are. I use caddy rn for my 1 hosted service cause it had a convenient guide for it
i used nginx proxy manager for a while, which was a GUI. It was nice to have a GUI, but also a pain to have a GUI lol. Caddy is much more streamlined IMO, but i havent tried traefik
I'm sure once I host more than 1 thing it'll be more important
im not even sure you need it if youre only hosting one thing
Caddy is great, I just want something a little more powerful
caddy2
Caddy is incredible. So easy to write config files
What about pdfs makes good programmers program like they are drunk
Please don't answer that question while I try to get a valid exit code from pdflatex
Never had I ever had so many issues with supposedly command line scripts dying if they didn't have someone on their stdout
Any of the lost-rustaceans thinking about the Bevy jam this year? Been kind of looking for an excuse to get back into it.
I'm waiting until bevy is more fleshed out to try it
I said the same thing 5 years ago
The main thing with bevy for me is its lack of tooling, particularly for 2d
@lethal walrus the normal within map pathfinders in stardew are a*
There is like four separate implementations all slightly different
ah
What lack of tooling? You mean the lack of an editor UI? (Personally, that's one of the things I like about it, for tile editing and such I can just use LDtk or Tiled).
Is there a tiled lib for bevy/rust? That would go a long way
Yes.
Tiled and LDtk, both. Two of each, actually.
There's a newer project called EntiTiles that I've wanted to try, way faster than the others, supposed to work with both LDtk and Tiled.
(I prefer LDtk, personally, but it's largely just personal preference there I guess... it does have many weaknesses compared to Tiled)
The other thing I really want is at least a basic scene editor of some flavor. Animation being limited to gltf and sprite sheet is also somewhat irritating. I've been spoiled by godot.
Yeah, there's no concept of an "animated sprite" unless you're an Aseprite user. I got fed up back in 0.11 and built my own.
Depending on exactly what kind of animation you're talking about, some of that appears to have improved since then.
I don't mind code first but having to write your own editor really cuts time out of actually making a game
The main aspects of godots's system I like are that you can embed method calls in animations, and animate an object within a scene, rather than just the object itself
Makes some things waaaay easier
This is really more of a 3D thing, is it not? What do you use scenes for in 2D? I don't think I ever found a use for them.
Idk if this is the same in bevy's typical usage, but in godot I use scenes for making larger conglomerate objects by arranging lots of smaller, simpler ones. And that seems like an ideal use case in an ecs-driven system like bevy
Like in one of my recent projects there's enemies made of multiple pieces, each of which has its own hitbox, health pool, and behavior. Obviously they need to be in a specific arrangement to look right, and be easily reusable, so I use scenes
What does the "scene" part do that parent-child/transforms don't, though? Just an easier way of instantiating them all at once, like a prefab?
Right, I was figuring it was like multipart enemies or weak points or something.
Different approaches, I guess; I wanted to have everything be externalizable (i.e. moddable) so I didn't go the scenes route.
They are, sort of, just not in the right way for what I wanted to do.
They're an asset like any other asset, so as long as you can achieve what you want by merely editing the asset, then it's fine.
Not sure if there's a word for the design but I think of it as "fixed assets" - that is, you can mod the details of an asset (like a scene) but the nature and purpose of that asset still have to be defined ahead of time.
Now I'm just wondering what a tscn gets melted down into in a Godot build...
I poked in that code once on my fork, it was fun 😛
...do we not want to know the answer?
Which I compare to dynamic assets (again, not an official term at all) where the asset is just identified by something like an ID and can have entirely unique attributes or behavior. I guess sort of like Stardew's item system, if Object wasn't one giant monolithic class.
I mean, it's basically just a binary version of the tscn if I remember correctly.
Though that was in 3
Think it's this file, if it's the same as in 3 https://github.com/godotengine/godot/blob/master/scene/resources/packed_scene.cpp
I thought I clicked on the bevy server by accident
scenes are coming in the next verson of bevy most likely
cart has been working on BSN for a long while now which is essentially the scene format for eventually almost anything scene related
both writable as separate files and eventually possibly generated through an interface or as a bsn! macro
that'll have to wait 3 to 4 months tho
it didn't make it in this release cause of observers taking the focus near the end
Huh, the TeX hyphenated algorithm is less complex than I thought it would be
(The 1977 knuth algorithm, not the current one)
fuckin knuth
I stared at the dancing links algorithm he thought of for way more time than I'd like to admit when writing a sudoku solver
Man is a genius
Unfortunately he's on my shit list today for "dear lord please exit codes were a thing when you wrote latex"
Also I'm amused that even my currently feminine-hobby-associated username doesn't mean people don't assume I'm a guy
It has scenes already.... but yeah, BSN in particular is the newer spec.
I'd barely call it scenes unless you count like... loading in something from blender as a big ass gltf
I guess I'm the odd one out, I'm just allergic to the idea of putting the actual ECS structure in game data. Like dev outside of games, the structure you want for serialized data is rarely the same structure you want for in-memory representation.
There's already the .scn format, that's not gltf (is it?)
I mean that's still just as possible then as it is now right
right, it does have that
Sure, I mean you can read the scene and transform it into something else, but if you're doing that anyway then what's the point of the scene in the first place?
To me that makes customization and modding more difficult, not easier; you as the dev have less control over how the data is "deserialized" and therefore there's more requirements for the scene to be an actual complete bundle as opposed to data with lots of friendly names, IDs, cross-references and so on.
I guess you could treat a scene as just data, but then you might as well use plain old RON for it.
Probably I just don't understand. Maybe I'd grok it if I'd spent more time in the professional game dev world.
idk enough to know what the exact limitations are of the current scene format but I don't understand how a better way to define your game assets/scenes impedes in any way on anything that currently exists
why would it affect modding in anything but a positive way
it allows more to be defined in a loadable file format
if you prefer code only you can still just use either the old way or the bsn macro
and having a standard format that plays nicely with all the features of bevy allows tooling to be written that can write out to that format, for example a UI editor
I just don't think scenes are actually a better way.
They're a quicker way, maybe, when you're just starting out, and there are projects like bevy_proto that do similar things. But they don't scale well.
what do you use instead? level creator tool like ldtk that you then load in and associate components with?
Pretty much.
Level entities reference template IDs and I wrote a little template loader based on generics and RONs.
I just never ran into a scenario where I thought, "I wish I could have this load the entire component bundle for me", because I always needed more control over the spawning process, not less.
I'm not really sure in what way that has a benefit over a custom format specifically for all bevy features beyond already having existing tooling currently
ldtk is a nice tool but it doesn't offer any benefit to making a game behind already existing
it doesn't even support collision maps in its format
I don't think ldtk or tiled are in the same genre as scene loaders/editors anyway.
They're complementary tools.
I guess we just don't really understand each other here
Apparently not. Not surprising since I never understood the point of scenes anyway except for some 3D games and maybe point-and-click adventures.
You're saying there is a point, and it's better for X, and maybe you're right, but I just perceive it as being worse for the same X.
Even back when I played around in Unity I just never felt I found the sweet spot for prefabs, which I interpret as pretty much the same concept.
I don't really understand what you consider to be the alternative to some kind of "scene" beyond completely code only which would to be feasible for anything beyond either small projects or completely randomly generated ones
and can't work in general for a game project project with non programmers
which is almost all of them
What, exactly, is a scene supposed to represent? I mean, what physical comprehensible in-game structure does it correspond to? Levels are almost always better solved with a dedicated level editor. And major game engines almost always have a non-scene "data" format, like Gamebryo's plugin structures.
I consider things like ldtk to also create scenes, they just need an extra step to become understandable to bevy
I guess so, but a level editor is way friendlier than a generic scene editor for that kind of thing.
a scene is absolutely anything from entire levels to a single enemy going by the meanings used in Godot or bevy
Especially if we are talking about a project with non-programmers, as you said.
I mean look at the modders here, they have enough trouble with tiled, can you imagine foisting a bunch of property sheets on them and asking them to figure it out?
If a single enemy is a scene then what is the benefit of defining it as a "scene" vs. just defining a plain old rust object and putting a RON file in the game assets? What do scenes add to this process?
(Note that a lot of modders do this as their first real entry into computer science, many are kids.)
They'll learn. It takes time
Sure... didn't mean it as an insult, I'm just saying there's a learning curve, and learning to use a map editor is a million times easier than editing a bunch of data in scenes.
That said: I don't know what the base minimum computer science skills for like video game devs is
These kids are able to figure out tiled, even if it's difficult; if we look at the Bethesda side and people trying to figure out ESP data and file overrides, they just can't grok it.
Slash I don't interact with nearly anyone, professionally, who doesn't know how to program a little
I mean a RON file can also be a scene format just the same though you'd struggle to have that play nice with compile time correctness validation and autocomplete
Eh, bevy asset loader handles that pretty well, if it fails to deserialize a RON it'll tell you exactly where the error is.
the main issue rn for bevy and modding is that the dynamic query creation isn't fully done yet so it'll be difficult to develop some kind of modloader that allows for scripting new systems and the likes
I just want to be clear, I'm not trying to say you're wrong, I just haven't been able to find a good example of scenes demonstrating why scenes are unambiguously the best solution. I like being able to just edit a RON or YAML file to tweak, say, the hit points of an NPC.
Or even have it reference an entirely different sprite.
you'd still be able to do that just the same in bsn, it's not like it's going to be anything that isn't human readable
idk I don't develop for bevy nor am I perfectly aware of what RON reasonably supports
I know RON has been considered in the past many times
RON is just the rust-friendlier version of JSON or YAML. Slap a serde on any rust type and you can probably deserialize it from a RON asset.
As far as I understand the scene proposal, the main difference between deserializing a RON file and loading a scene is that loading the scene actually loads components directly into the ECS - which is what I'm saying I don't want, because the ECS representation is far more complex than the "config data".
i haven't used it enough to really know. I'm from a web developer world where templating languages rule
whether I have to define an entity. spawn in a code file somewhere or have it be a bsn doesn't matter too much. You could still just use that for anything that is very dynamic
I understand wanting a difference between the internal representation and the external one. It's why APIs tend to have contracts that differ from the internal way they're used usually with some kind of mapping layer
That's my position, yes. Very obvious example being referencing other assets like textures/sprites; you want the mod data to just mention an asset path or name and have the loader figure out what that resolves to. But there can also be cases for referencing other entities or attaching to other world data.
Maybe the BSN loader is going to magically do all that, I don't know. Guessing that Pathos had to put in a lot of work to get CP off the ground and it required full control over the mod loader.
Evidently I'm in the minority since there's clearly a lot of demand for the feature and the Bevy team is putting a lot of resources into it. Doesn't change the fact that... I just don't get it.
I'm not remotely familiar with how modding works for any videogame so no comments there
Haha, wait, you're on this server and you never made any mods?
better yet
I mean
I've never used one
Crumble is purple
So what's your connection, were you a challenge player or something?
This is just the stardew server, it's not all about mods
I've been here since before the game even launched
I know the server's not all about mods; but this channel is under the "game modding" group.
I haven't looked inside a sdv related channel in like 7 years
Yeah so this is for lost programmers clearly
No one mentkon test I'm not a real programmer
Your code is better than mine 
No it is notdo
I basically joined this server to speculate on the features sdv might have when it would eventually launch and the potential release dates
then I made friends
stuck around in the off topic channels
I consider this to be an off topic channel, it's even in the name. I am a programmer and I spend my time in the off topic channels
this place was basically made for me 🙏
Alright then. In the future I will not assume any knowledge of mods or game dev.
this place was basically made for me
I was thinking of saying something like "so you're the reason this channel exists" but it sounded sort of rude in my head.
I do have a degree in game technologies fwiw but that's more stuff like computer graphics
I decided to not pursue game development further pretty early on
Didn't enjoy the vagaries of DirectX textures?
I actually learned a lot from my amateur forays into game development that helped in other parts of my career. But I do tend to hear the same story from most actual pro game devs (that it's a soul-sucking profession).
the game development market is just pretty much the least reliable, lowest paying and least available path to follow in software development. Like here in the Netherlands you either gotta be very experienced and be able to snag a job at guerilla games or something or you're working for some small indie team who probably can't even afford to pay you much despite likely having to work a lot
you need a very significant passion for the craft to stick with that industry
which I knew I didn't have
One of the interesting things I'm amused by is that my field like, working 6 days a week is kinda normal
(Electrical engineering)
But we get paid pretty damn well on average for it
I try to avoid places where working beyond the regular 40 hours is normalized
you need a very significant passion for the craft to stick with that industry
Funny how we're (well, some of us are) willing to do for free what you couldn't pay us to do.
not that Dutch law allows for much beyond it usually
modding as a fan is working on a game project you like
Yes, but in the distant past, all game dev was kind of like that.
if u r work in actual industry your chances of doing that is a lot lower
Modding seems very unfun I agree!
I might get into modding someday though probably not sdv cause (and I'm sure this is a strange thing to say) I'm not actually all that into sdv
I've never reached year 2
Before the AAA "era", people worked on games because they liked games and wanted the games to get played.
yeah passion is what dominates much of the industry
I have little to no creative drive in that regard
Hardware is similar to me, I'd rather just fool around modding devices or trying to make my own gadgets.
I never play past year 2, if that's any consolation. I do get amused hearing some of the players talking about what they accomplished in year 10.
to be clear. I also have zero passion for what I develop for for my job, it just pays reliably and I enjoy the act of programming itself quite a lot even if I don't particularly care about the end product it's for
programming for me loses the appeal when there's no puzzle to solve
Don't care about cash registers, huh?
Tbh I find programming inherently fun to a point
The point is when users show up expecting things to work
having spent the last 3 weeks in the absolute hellish depths of bookkeeping law and invoices I can't say it's my favorite thing...
Meanwhile i have a lot of sorrow when it comes to weirdly shaped sockets and bad uses of lvds
this is why nothing I do is ever released
ok I'm working on a genuine longterm project that I also expect to actually use as well
I'm pretty enthused at the knitting pattern thing I mentioned earlier
But it also comparatively simple
It's a delicate dance, for both game dev and conventional programming... if you completely divorce the users from it then usually the result is trash. Games by people who are just pursuing their "artistic vision" and don't care about what the players want are usually bad games. On the other hand, products driven by UX research and customer surveys are also usually trash.
Seems you have to listen to the users.... but not too closely.
I mean, I have no desire to be a game dev and at this point couldn't care less
It was just an example. Like I said, applies to all fields of software.
My interaction with programming is more as applied math
I have no intention for anything I make to have users beyond myself tbh
other than the longterm project which is extends to also being used by my sister
of course if I made customer facing projects I'd want them to be something the user's enjoy using
but that's not happening any time soon
I mean, artistic vision doesn't describe software to me either
More applied math problems?
I gotta invent applications to make to have an excuse to program
Meanwhile I'm just checking out hypenation algorithms for no real reason than "spent half the day fighting latex"
My college had a game dev track for the comp sci degree, but I decided against it since I didn't want to work at a giant studio and indie games weren't really a "thing" yet
I've had a hobby of making games for years (or... tinkering with them, at least) but I would never do game dev at a major studio
Indie games sound like a good time but that's also a moonshot. Either you get lucky and the thing you wanted to make is also something lots of people want to play, or you grind your own soul to dust churning out trend-chasing games
(assuming you don't just go broke)
same here. i can't imagine shipping or even working on a game in any capacity besides "individual passion project"
I can imagine it, but game devs don't have unions and studios have consistently shown that they will crunch you until you die
and actually indie-ing a game to completion seems similar to opening a restaurant: fun to dream about but nightmarish to do (which is why i haven't tried it)
I would definitely try if I had some other means of supporting myself, but like. I don't.
If I win the lottery I'll pick up game dev lol
Or maybe if I ever retire
Dont have to try and make money off it, it can just be a fun hobby
Users seem tiring
I mean, that's what it is now. But I still have to pay rent and buy groceries and I need an income to do that, and that means spending time on things that are no indie game dev
Tbh I voluntarily signed up for six day work weeks and I still prefer my little corner of the world
At least we get job stability out of this
(Also because actually users sound tiring and my idea of "fun code" looks more like scientific papers than useful code.)
I definitely get lost in the weeds sometimes
Right now I'm sufficiently exhausted that my creativity goes to knitwear...
For that building mod I was working on the other day I wrote this thing that used reflection to find subclasses that overrode a virtual method and combine them all into a switch statement based on class name. So I could be supremely lazy if I wanted to add more
Overengineered as fuck but it was fun to write and worked first try
That said as someone who did not grow up with video games, I don't have any attachment to the genre personally
Or to put it another way, writing a minimally allocating json parser sounds fun
That does sound fun
I wanted to be a game maker back in college, even joined a couple clubs. I bounced out of it, 5% due to my (at the time) lack of artistic skill and 95% due to the absolute horror stories coming from either inside big studios or indie developers chasing an impossible dream
I now have a cushy, stable good ol' software dev job and couldn't be more content. Once in a while my mother sends me an article of an indie/mobile dev striking it rich and says "This could have been you!" and I just reply "No, it couldn't" lmao
I mean it could've been you, the same way I could've been rich if I bought a lottery ticket last time I went to the drug store. We'll never know for sure, but I'm comfortable in the fact that I didn't waste $20 on the lottery
A client recently sent in this jacket after getting caught in the rain. We tested a variety of solutions to remove this dye bleed however it is not correctable.
Very disappointing to see this occur on such a unique and expensive piece. A simple test should have taken place before this garment was sent for manufacturing. Also the FTC requires a...
Is it weird that I kinda like the look of this
I mean would never buy it but maybe would try to make it
Or you... don't chase trends. Personally, I like trying to remake older games. But I don't honestly care if I ever make a single dollar from it... I get the sense that some of you folks are talking about jobs.
yeah I mean. that was literally the point of the conversation. why I'm not doing games as a career and keeping them as a hobby instead. I said as much in my first sentence
I still do mod stuff all the time for free because I want to make things
Command line program, I am piping your stdout to a file. Why is it still on my console
Grump
Grumps at badly behaved programs
Also smapi has spoiled me i need my warnings in yellow now
console logs that don't have color coding are evil
Its probably stderr youre seeing
If youre piping it, that would be on the shell command if the output isn't redirecting properly
....it was stderr
I didn't expect that since it was basically stuff like "compression level = 74%"
(This is actually a nightmarish mess of python orchestrating a bunch of subprocesses)
So I had to tell stderr to go to.
So that Zed editor isn't bad!
oh it has a stable linux release now
Yup!
Just tried it out because of that, and it's... pretty snazzy.
I hate their install method, though.
curl https://zed.dev/install.sh | sh
basically the same as how people install rust
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
and rust has no flaws as we all know
The search results page is pretty, too.
might try it out tomorrow or over the weekend
I wouldn't mind having a replacement for vscode
performance has never been an issue for me with for vscode but vscode has just been incredibly unreliable with my usage
tbh I'm used to visual studio. I've probably become numb to editor performance
in the webforms page I've been working in lately it takes roughly 20 seconds to save the file as it generates the new designer file
Huh
VS to me is unstable but more powerful
Vsc I've also never had performance issues with
But VS will like
Occasionally go "you tried to delete a file? fatal error sowwy"
it's really bad too cause while it's saving the entirety of visual studio just ceases to react to anything at all. If you click the window while it's going windows pops open the "this application seems to have frozen" warning
I sometimes have to restart VS because it locks the DLL I compile even though the program stopped running (usually happens with my mods)
vsc is like "asked me to rename v to voltage in this method? Okay I'll do it for the whole damn file 😸 "
smh v is velocity
also fun
tbh every editor I've used has had issues other than jetbrains rider
but that's probably cause I haven't used jetbrains rider enough
I only use it when I have to C# on my mac, since VS died
I only use it on linux for hobby stuff and I practically never use C# for hobby stuff
since I am not a modder
C++ used to be my favorite language but that was mainly because it was where I learned most of my programming skills.
I'd much prefer sticking with C# over C++ at my job 
C# is definitely the language I'm the best at and it's been my job for as long as I've had a job
I'll take a lot of things over Unreal-style C++!
shivers
(That's what I deal with, kinda, though I can't really elaborate further because NDA)

That and Perforce, bane of my existance
I'm learning more and more that the scripting languages for game engines are all kind of cursed in their own ways that differ from the language normally
VCS I've touched:
Git
Mercurial
That is all.
I used SVN back on google code
vcs I've touched:
Git
and google drive zip files
my source control of choice back in uni
Intelligence on tap
Save time and keystrokes by generating code with AI.
🤮
seems like these are the same people that made/contributed to atom/electron
do they ever say why they stopped doing that and made a new text editor
there must be some motiviation
cant blame them for trying to grab that bag tbh
i can 🤮
ironically I know treesitter better than atom
Yeah, that bit I hope to just ignore entirely.
this is the important part of course
I just want something like VSCode that isn't Electron-based. So right now... Sublime and Zed seem to be my only real options.
I just like trying funky new things
I'm still astounded by how well VSC performs whenever I remember that it is an Electron app
It's just the startup time I dislike. Other than that, it's shockingly fast yeah.
does electron really have that much of an impact on performance beyond the base cost of running the chromium layer
it uses more resources when doing nothing than I'd like but I don't think it affects much for anything you're actually doing beyond that
it will do nothing sluggishly 
I've just made the switch from sublime text to vscode, it has been fun
I used sublime since high school I think
I really need to just buy the Sublime pack at some point. I'm truly evil.
I used sublime for like 15 minutes at one point thinking I needed it for some project only to realize I misremembered and it was atom I needed
zed looks neat, a bit too tech company for me, but it's not like vscode is any better
yeah!!! I bought it after around 10 years of use, but I did!
well worth it
yeah! it starts up faster than intellij anyway, so that is fine for me
the editor I'm most interested in from a nerdy perspective is Helix but the keybinds being different from vim completely fucks me up and it doesn't ultimately gain me anything over using a preconfigured neovim config rn
mouse gang 4 life
I do like the mouse
the built like a videogame thing for zed's marketing is funny to me
it's such a roundabout way of just saying it makes good use of hardware acceleration
bit too tech company startup marketing nonsense
Ok, going to rubber duck something, or maybe one of you has an answer. Is this a "go home C# you're drunk" situation or am I just blind?
I have T explicitly guarded as notnull, but VS insists that the value can be null when passed as an arg. Ironically, if I remove the notnull constraint, then it stops complaining. WTF?
You see how you have Name = tab?
because you added a ? you've basically told it to look for null
You said earlier that tab is not null, but then guarded against jt anyways
So the compiler is like "uhhhhh okay this is inconsistent"
Yeah, if you ever use ? on a variable, then the interpretter will think it can be null from that point. Likewise, if you use ! on a variable, then the interpretter will think it cannot be null from that point.
Yep, you're right. Not the first time I've had that happen either.
Kind of dumb. It should be squigglifying the tab? part and saying "this actually can't be null", not everything else.
tbh I think this comes more from when your project is set to nullable = false where it needs to rely on stuff like this to figure out nullability
C# rightly assumes that, if you are going out of your way to guard something as possibly null, then it might be null. The interpreter isn't perfect, and if you're calling into questionable libraries (like the game) then nullability might not hold up at runtime, so it trusts that you know what you're doing if you use ? or !
I never set nullable to false.
you could argue a better warning would be to point to the question mark in the message and be like "inconsistent null signature" or something
good
That's what I did argue, lol
I suffer on every project where it is
which is most
cause most projects are much older than that option
Most of my projects I set nullable warnings to errors.
It's actually only the SDV mods that don't go that extra step, mostly because that's what I found in the default project template. I never actually leave any warnings.
so it trusts that you know what you're doing
Well that's its problem right there.
a very foolish error I do not make myself
Look, we hope the programmer is competent
Oh itll take 5 years to start up then
This definition of "loseless" is not one I recognize
I thought the definition was "you can undoubtedly undo the compression to get the original back"
But their definition is "your human eyes won't notice the difference"
yeah thats not lossless lol
fuckin webp confused me to no end with their lossless
(Just as a quick heads up/warning to people I don't think will abuse it, Discord has new Markdown that absolutely nobody will abuse!)
-# This is a real message from Discord for real I promise honestly and seriously. dismiss
oh lord
wtf
(Just as a quick heads up/warning to people I don't think will abuse it, Discord has new Markdown that absolutely nobody will abuse!)
-# This is a real message from Discord for real I promise honestly and seriously. [dismiss](<https://www.stardewvalley.net>)
Yup!
since when has -# been markdown

-# test tests
Oh @safe dragon, the Markdown URLs trigger a timeout! 
I noticed
