#programmers-off-topic
1 messages · Page 41 of 1
seems reasonable to me that they might have their own ideas about how to format your text, so they want to guarantee that they'll be able to easily copy-paste it into their target document and not have to deal with How Pdfs Are Sometimes
(Eh, appearance rules are pretty normal.)
I have the strong opinion that, for research papers, the submitter should not be in charge of thr formatting
Please provide word template and latex template
i submit that this should be true in almost every text submission context, not just research
one of the principal advantages of text is it's easy to format and reflow as needed (hence TeX and friends)
yeah, that would be nice to have a TeX template for that report.
Special shoutout to this health care portal, whose programmer couldn't figure out how to do this themselves
doesn't seem like this could be a concern at all, I have no connection issues, and this issue is 100% consistent
here's a demo
it's supposed to be a gif, maybe that was a bad export choice on my part
when I select text, then type a letter to replace it, the cursor is placed before the new letter instead of after it
even databases have a built in function for this
Bold of you to assume they're not using plain ol' text files.
speaking of
I discovered the absolute worst import format I've ever seen
it's a text file without any kind of delimiters. Instead there's some hardcoded assumption built into the importer about the number of characters in each field and just goes off that
they can't increase the size of field because it would fundamentally break all their existing imports/exports
there's workarounds in place where the same field was introduced again at the end of the row data but then with the new character count, now ignoring the original spot there were in(but it still needs that space filled with the exact number of desired characters anyway)
I have no idea
it's some ERP tool
another shit stain developed by industrial engineers
I can say that because it's my major
man why web development course are the most boring shit in the second year 
what's the material?
fixed-width formats get more common with how "enterprise" or "just old" software is, in my experience
We deal with a company where half (the more important half, too) is in a "database"/application whose "tables" are all fixed-width fields, the other half is in MySQL (because we put it there). In this case it's because the database software is literally from the 1980s. They usually had the foresight to sprinkle in "open_1", "open_2", etc fields of various sizes into most tables, at least, so you could shoehorn new fields into existing tables later without forcing a rebuild of the table, but the software does technically support rebuilding itself with an expanded format, too, luckily. We've been slowly migrating tables over to MySQL over the years; eventually we'll hit some critical point where all the rest have to come over.
What's the name of the database software?
Yup, I can definitely imagine how younger developers might have never seen them before, but fixed-width formats were very common in the 1990s and even somewhat into the 2000s. In fact, databases themselves used to be like this, we had char(n) before varchar(n).
they kind of make sense in extremely constrained environments
FilePro
It's both a database and has it's own "programming" you could do in it. It's literally from the 80s so it's not really fair to compare to modern languages, but it sure made me cross-eyed trying to decipher what it was doing so I could port it to the web app
It's interface is in the terminal, so when they used to use it directly it was a blue-screen-of-fields type interface where you'd tab through to the field you needed to write into
Have to switch some Putty settings when we have to interact with it so it will both display properly and understand our keyboard input properly
definitely a step worse than my current process of porting a microsoft access program
We're doing that for a different client right now lmao
though that one is also 30 years old by now
If they don't keep compressing their Access DB it will hit 4GB and their entire system will cease functioning
So we've been porting tables and stored procs over to modern SQL Server
Before their inevitable implosion
they don't even use the db part of microsoft access. They use it to create forms that then get all their data from an oracle database
well, they do use its database in a way of course
at least it means I don't have any porting to do for the data itself
Yeah there's not really a good modern form-builder to take Crystal Reports' place (since they refuse to update past .NET Framework)
If you have MS's ecosystem SSRS is alright from what I hear, but nothing more universal
as far as I know any kinds of reports are slowly being moved into power bi dashboards
though that's not happening till I have the rest api done for gathering that data from the oracle db in a format a sane human being can actually understand
since the database itself works with a giant generic shared items table with column names that are essentially just property16 , property48
Yeah another client stores their documents using an app that uses database tables with columns like string1, string2, .. string10, int1, int2, etc
are the things in the strings columns at least actually strings
When did people invent csv
Not sure off the top of my head, wouldn't be surprised if all the fields are actually varchar though
CSV itself(as in the spec we know it as now) is from 2005
but I'm sure they didn't come up with the concept of delimited text files
Crystal Reports, man that takes me back. And not in a good way.
Well, it is my CPU. New motherboard didn't help either.
That sucks I'm so sorry
Definitely bites to go through all the trouble of replacing a mobo only to have it not fix anything. You definitely already tried underclocking the memory specifically (not CPU clock)? The Intel IMCs for the last 3-4 generations are flaky as hell.
It's an AMD system, so Intel anything doesn't really apply. 😛
But yeah I've messed around with a lot of stuff in software. As far as I can tell without being able to actually test anything in Windows, it's probably a CPU voltage issue.
Yeah, with AMD I've had entire CPUs basically die after a few years, at least with Intel it's just the underpowered IMCs and generally dumb silicon lottery.
At least the AMDs are cheaper to replace.
This is the first AMD chip I've ever had die on me, so it's rather frustrating.
(I mean it'd be rather frustrating either way, but the lack of expectation is also frustrating?)
It's frustrating when you aren't prepared and don't have a backup machine, yeah.
hey im sorry to interrupt the current convo -- i'm having trouble enabling chat cheats and was wondering if someone could help? tysm!
this is a non-Stardew channel
#stardew-spoilers perhaps
ty!
It's amazing how much time variance there is between two lines in python code
My Steam Deck has been doing its best as a backup machine. I'm very thankful that I had that.
Ugh. Windows has this annoying issue for me where if I use my USB switch too many times, it eventually fails to detect and the only way I can use my keyboard/mouse again is to restart.
did you get a dirt cheap usb switch
It's pretty well-reviewed and it wasn't exactly cheap
ah, okay, so not a "this usb switch is failing to identify itself correctly" issue then
I've come to accept it's just windows being windows
Like it'll work flawlessly for days at a time, and I'll constantly be switching back and forth between devices throughout the day, and then just randomly it decides nope not gonna work anymore
What's weird is that I can plug another mouse into the front usb and it'll detect correctly, but whatever usb port my devices were connected through is just completely broken for the remainder of the session
It could be a faulty USB controller, or even a short. On one of my older builds I'd often have a few days of USB weirdness eventually followed by a hard crash and then failure to reboot with that enraging "USB over current" error even with all the peripherals unplugged.
Sometimes it's really quite shocking the absolute trash tier components they'll put on motherboards just to save a few literal pennies on manufacturing. Or just terrible quality control. I honestly don't think it's a Windows thing, I've used a lot of cheapo USB switches and never had a problem precisely like that.
My guess would be a stupid USB controller / driver.
Drivers can also be pretty terrible, yeah.
It's probably exhausting its pool of device IDs or something lol
I have seen how hardware manufacturers develop software (drivers). It's not pretty.
Is your front USB on the same controller? Because your PC probably has at least two USB controllers.
Also on the topic of USB, the motherboard I ordered for my updated build today has this for ports and I am pretty pumped. Every single port is USB 3.2 half of them are better than gen 1.
But on another topic holy heck do I hate the USB versioning scheme
Isn't that pretty normal these days?
Look my old motherboard is like 4 years old I can be excited over USB lol
Haha, sure. Wasn't sure if there was something specific I was supposed to be paying attention to.
2.5G ethernet would be so much more exciting if any of my home network wasn't limited to gigabit.
I am not good at flooring and drywall and stuff yet, so I never tried to install a wired backhaul... with wireless backhaul I'm limited to about 400 Mbps on my theoretically 1 gig connection.
Kinda sad the onboard audio doesn't have toslink, but I haven't actually used that in a bit since I switched dacs
I have cat 6 run everywhere that's relevant but my switches and stuff are just gigabit.
Haven't used Toslink in forever. These days if you want low-latency audio, you normally get a USB3 device (or Firewire if you've got it, but recent evidence is that it's become largely inferior to USB unless you've got heavy bus traffic).
It's a shame, really. Toslink has electrical isolation and nothing else does.
Not that I tend to get that audiophile-y about it, but... computers can really generate electrical noise.
Ehhh, I know what you're talking about but balanced audio cables make that go away really fast.
Seriously, I had horrible noise over the USB until I realized I was using an unbalanced cable that looked like a balanced one.
Oh, but you know what does generate noise on the cable? Freaking cell phones, if they get too close.
I haven't had issues with my current dac but my old sound setup had a pretty noticeable hum in some situations. And my speakers are 13"x18"x45" monstrosities so it kinda matters lol. They get loud
Man now I'm thinking about networking again and how shiny it could be to upgrade my Ubiquiti stuff
I am so jealous of the rgb ports on their new stuff.
RGB on network peripherals sounds... distracting.
No it's actually amazing. Picture this: you're working on a network rack. You need to replace a patch. So you just open the management software and you set it to flash the port or color it bright red or something.
And then you just look at it and easily see which port it is because it's literally lit up
Or you could, say, color all the ports on a specific vlan a specific color.
I don't think I have any one device with more than 4 ports, and most of them don't have all the ports in use. But I can see why you'd find that useful if you're running something with a lot of complexity. Some smart-home setup or something.
"I'm tired, I'm going to sleep" in Atraese = "I'm considering the possibility of planning to go to sleep sometime in the next seven hours"
Technically I said I was tired and would not be devoting any more brain cells to understand spritebatch.draw
Which was technically true! I went off to understand dishwashers
Hmm, so you did. You are technically correct.
The best kind of correct?
I don't have a "bureaucrat" emoji on hand so you'll have to imagine your own.
I'm also doing not bad on sleep
See?
That's like legit better than I usually do ngl
God, anyone who tries to wake me up at 5:30 am had better be wearing protective gear.
i'm not sure the folks behind USB could have made it more confusing if they'd tried
That's almost seven full hours! Pretty good
Ahh, to be young again, and also a robot.
Getting up before the sun has even thought about coming out is a big nope from me. But props if you can pull it off consistently and still function in the morning.
I’m younger than Atra and still hate mornings
I've learned to love them. No one wants to try to talk to me at 5:30am
What if theyre from a different timezone
timezones aren't real
huh now Swift is getting Spans? https://github.com/swiftlang/swift-evolution/blob/main/proposals/0447-span-access-shared-contiguous-storage.md
Today I learned: Google has routers
like, their nest wifi things?
My new parts got here today. I live again!
And then I immediately shut down Windows to clone my boot drive onto a newer, faster NVMe.
Attempting to build consensus at work over making a homemade sodastream
Boss's boss is unconvinced
Must try harder
what does that do?
I experienced a loss of my primary drive a couple years ago, and when I reconfigured Windows 11, I used their Storage Pool feature which is just their own branding for raid essentially. Just FYI.
Gives me some peace of mind
Oh, I don't need backups. I have a NAS with automatic nightly backups of all my PCs.
I have backups as well, but the failover for uptime is what this affords me
Like if a drive fails, everything will still work as I try to replace it
Whereas my backup would take more time to restore while I have downtime
Yeah, I would have some downtime if my drive failed, but I'm not really concerned about it. As things that kill my PC go, drive failure is almost never it lol
Carbonated water!
I wasn't so lucky with what happened to me. A brownout caused my NVME SSD to fail. PNW power is so flaky when it gets windy out.
I definitely think my response was an overcorrection to the issue, but it got me into a whole new hobby
can u carbonate coffee
yeah it was a sort of craze for a bit. i forget what coffee shops used to call it. nitro or something
maybe they still do. i do not typically go to coffee shops because i'm too precious
Ugh, power. I have UPSes for all my electronics and my house has an automatic backup generator.
Alright. NVMe is cloned. Now I'm making a bootable Windows 10 installation USB so I can run repair if the boot sectors don't feel like cooperating
After that whole bomb cyclone thing, I need a backup generator. Lost power for a week and had to throw out all of my food.
Same concept, different implementation. N2O cartridges are used for foams, whipped toppings, etc. CO2 cartridges are used for soda.
I don't remember seeing either in a coffee shop, but I also make 99% of my own coffee at home. In theory you could use the whipper for capuccino or whatever, but the fancy espresso machines do a much better job.
A whipped cream charger (colloquially called a whippet, nos or nang when used recreationally) is a steel cylinder or cartridge filled with nitrous oxide (N2O) that is used as a whipping agent in whipped cream. The narrow end of a charger has a foil covering that is broken to release the gas. This is usually done by a sharp pin inside the whippin...
Nitrous oxide is used because it dissolves easily into the cream, and does not cause the cream to oxidize while it is in the can.
It is; that was my second sentence. Usually the capuccinos, mochas, etc. that you get at coffee shops just whip milk (not cream) with air (not nitrous). The nitrous whippers are used for other purposes, generally.
There was one place I used to go to a long time ago that made mochas with real whipped cream, and might very well have used a charger for it. Used to go there a lot when I was way younger and still consumed sugar.
You probably wouldn't be surprised to know I live off caffeine
Sometimes with a little sugar
I like how Wikipedia spends multiple paragraphs covering N2O chargers' use in recreational drugs and precisely zero on its use in molecular gastronomy. Friggin' Wikipedia.
For the first time ever toml spec has surprised me
What do you think this deserializes as
[Blah.Header]
1.5 = 1
3.0 = 7
4.5 = 20
5.5 = 20
Hint: not a dict[float, float]
Isn't the left side always strings?
what would a number on the left side of a toml expression even mean
I'd assume they're strings yeah
Yeah, I guess a foo.bar under [Blah.Header] would be the setting Blah.Header.foo.bar. So with these "floats" you'd have Blah.Header.foo.bar = 1
I guess it makes sense
It's more consistent than JavaScript would be, anyway.
arr['1'] vs. arr[1] vs. obj['1'] vs. obj[1] and....
being more consistent than Javascript is such a low bar
Yeah, it's fully consistent, just not what I expected
Also, the laziest meal I've made in a while: kale and beef
Literally just kale and beef
I mean, throw in a potato and that sounds like a regular Friday night dinner to me
Lose the kale and you've got a deal.
I have never understood why people like kale.
I love kale
I do Kale in a way that even never Kalers seem to enjoy. It ends up crispy with some spicy seasoning.
Tbh I love kale because it is a very good veggies for meal prep
One of a few veggies you can just blance3
Kale, eggs, and chickpeas 💖
oh oh if you've never tried it, fried kale with mustard is delicious
I went grocery shopping and glared at the kale because of this discussion
i wouldn't say i "like" kale but it is Fine if you need a sturdy green
there's a different variety that looks less like a fractal which i usually choose if the choice is available
Speaking of kale and Swiss chard, I should try growing some
This is how I make dinner with Kale.
Damn, that looks delicious. How'd you season that meat?
Salt+Pepper+Garlic
Cinnamon and nutmeg?
You put cinnamon on meat?
I've seen cinnamon and even unsweetened chocolate go into some stews and sauces, but as a seasoning or rub... pass.
You know, given I've thoroughly enjoyed a five spice rub on a steak... I think I get it.
Fresh broccoli, stir fried with a huge amount of garlic 
Apparently this is the stardew cooking channel now
Broccoli is my favorite vegetable, and by a huge margin.
That reminds me, my mom is going to be cooking up beef and broccoli for one of the like four Christmas dinners we're doing so I get to have that soon. 
Oh that sounds awesome!
Boss's boss got me pocky!!!!
Christmas pocky?
I mean it's for the full office but
What flavor?
pocky flavor
so that's the first time I got a javascript out of memory error (aoc 23)
that's a lot of memory to store a little graph
I'd have expected the out of memory exception to occur on day 21 instead
soon advent of code will be over once again
Anyone here familiar with Lua and using it in wikis?
Lua in wikis…?
Yeah, I'm trying to create a module to calculate the time in hours or minutes, so for example 600 would be 10 hours.
Then I'd invoke that module in a template
I just didn’t know wikis used Lua at all
I've written a basic code and it doesn't seem to throw any errors in the sandbox, but you know how that goes, lol
Just because it says there isn't errors doesn't mean it works, lol
I suppose the biggest hel pwe could get it... which wiki? Is it some particular... extension or something like that Scribunto that's enabling it? Gotta have something to look into!
Scribunto is enabling it, yeah. An example is here: https://eastscarp.wiki.gg/wiki/Module:CalculateEdibility
Weird, I feel like the time conversion should be a LOT simpler than that...
Oh it is - this is more complicated because it has more variables
This is what I've got so far, but I'm almost absolutely certain I'm returning the results incorrectly
I think I should be returning (result) not time
But I'm over my head, lol
The return in line 7 is maybe wrong - it’s numeric, the others are strings
You do abs twice in line 9
Lines 13 and 15 return the thing that was passed in with no math involved
Yeah, looks like you would be changing out time on lines 13 and 15 for result?
Yeah, I don't know why I have the math listed twice
Not sure what exactly you mean about line 7, Shockah - I thought I defined "time" as a number
Basically, what I'm trying to do is if someone put 0 in then it would stay 0
(though why they would use the template is silly - just leave it blank, I guess)
you're returning a number though
(0)
idk if that will or will not cause issues
I'm not really familiar with Lua, but I... think line 7 is fine with it being dynamically typed and this probably going through to HTML in the end?
it depends on how that Lua code's result is then transformed into HTML, yeah.
So I'd have to implement it to see if it actually works
but i'd probably still try to keep the same type across the board
and return '0' instead
Ah, I see, yeah
The other results are to concatenate a noncalculated string with the calculated results
that really made more sense in my head
Ah yeah, I'd definitely do what Shockah said!
Other than that, it seems to work: https://onecompiler.com/lua/433yfspkt
Just with a quick tweak to directly accept a string instead of this frame thing.
In your case, you would definitely want the frame, since I assume that's just whatever NPC schedule or whatever's being displayed.
I don't know what the frame actually does except I believe it "frames" the object resulting from the code
Yep
I was going to post the same thing
Holy macaroni, I kinda can't believe that I had a somewhat working bit of code
It's a bit of a 
Thank you both for the help! 
You're making the next even bot now, though. /j

It's alive https://eastscarp.wiki.gg/wiki/Module:CalculateDuration results in
We'll get you that blacksmith role eventually. 

Computerphile is supported by Jane Street. Learn more about them (and exciting career opportunities) at: https://jane-st.co/computerphile
This video was filmed and edited by Sean Riley.
Computerphile is a sister project to Brady Haran's Numberphile. More at https://www.bradyharanblog.com
Holy crap, I knew I recognise the voice behind the camera! I think it's the same guy who did all of these? https://www.youtube.com/watch?v=zvcaXM2hMgA
Martyn Poliakoff, CBE, is an acclaimed chemistry professor at the University of Nottingham. He is the brother of writer and director Stephen Poliakoff. More about Martyn and other Nottingham scientists at http://www.test-tube.org.uk/
no more advent of code
freedom at last
merry christmas
my goal of less than 1 second of runtime I think I ended up getting a little too comfortably
anybody remember this project hahaha
i think it’s booting but idk? no output on the screen though
how do i know which GPIO pins to connect it to? it is the “with headers” version so it’s just a matter of slotting it in i assume but idk which ones it goes in
there's already the headers on the screen so it just consumes all the gpio
nah cos there’s pins left over
idk i’ve ran the script on my laptop in ubuntu to flash the drivers onto the microsd but it’s not showing any display on the screen
well i think i’ve got the right pins now cos i crossreferenced the pinout and stuff, still no display of anything though
can you ssh in
remind me, what is the exact model numbers of everything you bought?
That... looks like it might be in the wrong position on the GPIO?
Looks like it call for being connected to pins 1-26, and I think you have it on the opposite end?
Going by this pinout image and this photo? Unless this is the underside...
yeah, going by the pictures (hard to tell) the second position may be correct
(it looks to me like the standoff positions should also line up.)
It's hard to tell, because it looks like there are differences between revisions, so it's kinda tough to see...
looks like there's only one spi on the zero
one second
what's the clock speed you need for the display?
it's gotta be pretty high
(also, I drew in that "spi" marker but I'm looking for csb still, that's three-out-of-four for required spi lines.)
Okay, I've determined I... think I'm right? There is no real underside on the Pi Zero, most everything is on the one side. So going by that pinout me and Atra posted, I think it wants to be shifted all the way to the left?
Going by:
(Absolutely trust Atra more than me on this, though!
)
ah, gpio 8 and 7 are chip selects
okay, all lines accounted for
do you have an oscilloscope
Am I way off there though, Atra? 
you're not way off! I think we're in agreement
like, I didnt' find that pinout, I was using their graphical version and just matching up pins with the pinout of the zero I found on some guy's blog
Yeah, like pins 1, 2, 4, the power pins, etc. are on the left there unconnected.
harvz did reverse it later,
OH! Okay, definitely missed that then.
I think the boards line up on each over, stand off over stand off
the holes I marked out are the standoffs
wait so it’s in the wrong spot?
No, this variant seems correct
uhh i wasn’t at a proper pc but i put the ssh file on the boot partition to give it the SSID and password of my wifi network so i will see if i can ssh into it or not when i’m next at my pc
Yeah, I missed that one photo! That one looks right to me too, yeah.
thank you all for the support though appreciate the research 🙏
okay nice
so it means there’s some kind of driver issue that’s making the display blank then
Although I suppose it's possible that the board needs to be flipped? Like it might be the right position on the GPIO, but the board itself the wrong orientation?
I haven't seen a pinout diagram of the display board itself...
i tried nearly every possible orientation you could think of
i find it very hard to believe that it’s the position of the GPIO stuff that’s wrong and not the drivers
but idk it could be 🤷♂️
Yeah, if you've tried all of the orientation, then...
So, stuff like this may not be idiot proofed
I'd expect not to be, honestly!
And I don't have the actual hardware in front of me but it's possible plugging it in wrong may permanently break one or more items
SPI is a pretty generic protocol
I feel like that's only super likely if the power pins were connected but in the wrong position, and that would put them on... let's see...
I'm super not used to the world of raspi but I thought raspi was a 3.3 system with 5 as an option
Now, given that raspi is a teaching system and overdrive protection isn't that hard, I would kinda expect it
3V3 on 26 TP_CS Touch Panel chip selection, low active, and... 5V +/- on unused pins?
so i’ve blown my rpi 😭
More likely the display, I think!
fiya
It's also possibly a weird driver thing...
Yeah, my biggest worry would be if you somehow managed a power to ground short
But I don't think so
I would verify it is sending a valid spi command but tbh that's an oscilloscope thing
Which. Uh
My cheap one was $250
Yeah, looking at the pinouts I don't think that's been possible at least, which is good.
I want an oscilloscope...
I hardly use mine lol, we just abuse the fact that work has nice ones
well i’ll try to ssh into it tomorrow and then do the driver install natively on the pi instead of through ubuntu and see if that works
if not i’ll return the display and try again with another
Keep us updated, though!
will do! thank you for all the help. very confusing all this for a noob like me, but it was fun whilst i (thought i) was doing it!
also my first time using linux which was exciting!
Does anyone use Code instead of Visual Studio for modding SDV?
I got a new PC and I'm being very cautious of not bloating it with software, and Visual Studio is a bloated nightmare.
Is Rider reasonably lightweight?
me i use code it sucks
to be more specific
- the LSP works alright, sometimes it dies and i have to restart vsc
- need to hand write csproj which doesnt seem to be the norm
- build from solution explorer is awkward for build configs, so terminal
dotnet build -c <config> - need some fiddling to have debug and hotreload
- u can get ilspy for code if u want
thats a great summary, thanks
Visual Studio is a smoother experience than Rider for most of us, though Rider is alright. VSCode really does suck, I'd call it far more than "fiddly" for debugging and hot reload. Maybe it's because I only really tried to do it on a Mac, but it was honestly a nightmare whereas Rider just worked off the shelf.
I straight up gave up on my attempt to get VSCode doing anything useful for C#. It's a great editor-on-steroids and a terrible IDE.
got it
Visual StudioVSCode- Rider (maybe)
Haha... you really find VS that bad in terms of bloat? It takes up a lot of space, sure, but I've never found it doing anything untoward, or outside its "scope", or when it's not running, etc.
it's not that bad honestly
but I don't like having to install an installer for it
Yeah, I guess the complexity of the install is very high because Visual Studio supports so many languages and frameworks. Rider is just C#/.NET.
The corollary with JetBrains is that you literally have to install a different product for every language, or at least every category of languages.
All I need right now is C# so I'm up to try something new and focused
I'll apply for an open source license and try Rider
You don't actually need the open source license anymore, they made it free for personal use.
I cant code .NET without Resharper lol
Funny... I can't code with resharper, from the very beginning it always drove me crazy. But everyone else swears by it so I guess it's just me.
I liked VS fine but they killed VS for Mac
And I REFUSE to believe VS code is an actual ide
It's not. I mean, Microsoft does not call it an IDE at any point. They call it an editor. They are at least honest about that.
They do bill it as a “totally just as good” replacement for VS Mac though
Like their retirement notices sounded really delusional to me
"We've come up with a very sweet solution"
i have to use vscode for C++ at work rn it sucks^10
C# LSP at least works C++ is me fiddling with cpp_configurations.json for days 
C# on Mac+VSCode was me fiddling with launch.json for days, same kind of thing.
The amount of AI bullshit on the website for VSCode these days is so gross
At least VSCodium exists.
And ugh, this one really does bug me. IntelliJ Ultimate, for instance, can do everything that... PyCharm, RubyMine, PhpStorm, DataGrip, GoLand, and WebStorm can do via official plugins only available for the ultimate version.
But you want .NET or C/C++? Nope, sorry, you're screwed. Separate IDE for you!
just noticed the Leah bot doesn't have access to this channel. probably doesn't matter too much though
anyway, Harmony woes strike again
i was implementing something in my mod loader, used SomeStructType as one of the params, but in reality i needed SomeStructType?
no errors, other than my stuff not working
....... thanks #programmers-off-topic for being a great rubber duck, i realized that was in a transpiler, so it's actually my fault
weird that it didn't cause any exceptions though
it just reinterpreted the memory of the struct as a different struct
you’re welcome
To be fair, VSCode probably felt like it needed to respond to Cursor and Windsurf
That's almost definitely what they were doing and to me it's part of the problem - companies with a long legacy and tons of technical talent that should be leading the space are following "responding to" dumb ideas from startups chasing trends and investor money.
I'm being lazy and working out of gitlab vscode and I keep on trying to :wq lol
...for whatever reason, using a new AssemblyBuilder/ModuleBuilder for each type proxied in Pintail makes the whole process SO MUCH FASTER
rather than using just one
That's very strange. I imagine there could be some global caching that reduces the overhead for new instances, but hard to understand how it could outperform a single instance. I hope you're double- and triple-checking your benchmark/profiling setup to make 100% sure it's not observation error.
i'm actually only doing this, because i've seen someone saying that on StackOverflow
and it does seem correct
The only analogous situation I can think of is how Visual Studio seems to grind to a halt when you have 1 file with 5000 lines as opposed to 50 files with 100 lines, but I always assumed that to be an issue with the IDE in particular (don't think I've noticed it in other IDEs).
Interesting. Though one thing that goes unexplored is the cost of having 100 dynamic assemblies at runtime as opposed to 1 dynamic assembly.
Not the up-front cost of building them but what the consequences are of having them all loaded.
i'm profiling my mod loader. the current stable version of Pintail was taking 67% of total load time, because we have a lot of complex APIs used across the mods, and we proxy non-API instances to add new "events" too. after making some changes to Pintail to create a new AssemblyBuilder/ModuleBuilder per context pair (unordered pair of mod unique IDs), it went down to 12% of total load time
so yeah. i'll be doing some more testing whether anything gets broken, and if not, it'll go as the next Pintail version. the default behavior will still be the old one, since you had to pass in a ModuleBuilder. but there's gonna be an alternative constructor with a Func<IdkWhatExactlyYet, ModuleBuilder>, so you can opt in
I did not expect that post to be seven years old. That seems like the sort of thing the .NET team should've addressed in that amount of time, lol
I don't think they've touched the old-school IL emit stuff in even longer than that. The new hotness was expression trees, then it was dynamic, then it was source generators, etc. I wonder how often the low-level APIs are used outside of game mods.
I didn’t even know this stuff existed before y’all but it’s interesting to see
I already feel like I’m using some deep internal C# stuff just using shit like GetGenericDefinition
It's used heavily by folks like xmlser and newtonsoft
...at least I know xmlser uses it lol
no idea what that is but I assume it’s an xml serializer
Correct, lol
I’m very familiar with newtonsoft though we’ve been gradually removing it in favor of system.text.json
Tbh I care less about like
Time to make a dynamic method
And more about runtime. In theory you don't make that many dynamic methods
I hope I never will
Pintail makes enough of these that it matters
Yeah, but who still uses XmlSerializer?
some 70 year old dude at my previous job absolutely loved XML and would figure out horrible ways to integrate it anywhere he could
we suffered the consequences of his actions
Xml is only really useful for schemas and xslt
So far at work today, both the graphics and wifi drivers on my work laptop are mad at me

sufficiently tired to fail to count to 20 multiple times
it happens
still debating whether someone in a programming server I’m in is an extremely dedicated elaborate troll account or someone genuinely extremely misguided on just about every fundamental aspect of programming with an unwillingness to change their beliefs
it’s been months
that was my mini rant as I recover from witnessing another one of those discussions I apologize
I generally assume incompetence over malice. However, I mostly encounter that attitude in people trying to make or install mods, so it’s fairly easy to tell them that this just isn’t going to work and they should find a hobby better suited to their talents.
For example, people who don’t know what a Downloads folder is or where to find it
Or don’t know how to unzip files
Honestly i feel a willingness to Google is huge
If unzipping files is a challenge, installing SMAPI is a moonshot
Sometimes I try to send them Microsoft help pages, but honestly those are only medium helpful. Usually I ask if they have anyone they trust who’s better at computers, in the hopes of keeping them from accidentally installing viruses in the process
I don’t wanna slander this random person too much but their biggest issue actually has nothing to do with not googling or not finding information. It’s that they make two basic fundamental assumptions that seem to cause most of their ignorance.
- There is always an objectively correct approach to any programming problem
- Any software that does anything that isn’t exactly what they expected is simply the result of shit programming and the person who built it was simply incompetent and couldn’t have had a reason for why they did it that way
Today’s discussion was about their belief that a smaller binary is always better and more performant and that anything that would cause there to be “more” instructions would make it slower.
This was due to them complaining that generics in Rust make the binary much larger and that rust would be much faster if they “didn’t need so many instructions” to do generics. (While the actual reason generics in rust cause the binary to become larger is due to monomorphization which actually makes the program faster since it avoids dynamic dispatch)
they’re a big rust fanboy though so this one was a change of pace
it’s very different type of discussion than someone who has no background in programming trying to figure out how to mod
I see, that makes sense
Honestly I know I’m not good enough at programming to know that kind of stuff for sure
I just listen to whatever Atra says
rust discords have funky conversations
the off topic channel can be anything from twitter drama to discussing different threadpool implementation strategies
Got it. Making a binary that is
0 goto 0
perfect
Aside, one thing I've always been told (when picking my career) is that I was going into an industry that just...did not believe in the 40 hour work week.
like. At all.
like, it's pretty normal to work 60-80
and like, I'm perfectly okay with it, but I'm starting to hazard a guess that it's not like....a thing most people would accept
Yeah I mean ok considering my choices…dunno if I have room to talk, but that would not be a plus for me and might be a dealbreaker if it was a) actually physically at work and b) started early in the morning at fixed hours
It's at least partly an age/experience thing. Everyone starts out bright-eyed and bushy-tailed, but eventually makes their way to "fuck you, pay me".
The process happens a lot faster with exposure to "enterprise", but it happens invariably.
Exceptions might be made if the founder/CEO/whatever is truly, phenomenally inspiring, but there are 100x as many who just think they are than those who actually are.
I had an internship where they indicated that working on Saturdays wasn’t expected of interns but it was of full timers
And I was like “fuck that, im not moving to a state where i don’t have any friends to work 6 days a week for decent but not amazing pay”
It was a very educational experience
I learned that businesses that don’t have HR can be a massive red flag
Ah, the good old Death March.
"We're behind schedule [unstated: because of stupid management decisions and bureaucracy] so we'll need you to start coming in Saturdays. And maybe Sundays."
I was working 40-50 hours a week as an intern and wasn’t mad about that, it was very interesting work, but I was so tired after working a full 8 hrs/day
My older friends had to gently tell me that hardly anyone actually works for 8 hrs/day
If someone is genuinely a workaholic then ok, I get it. Sometimes you just get obsessed with a particular problem and have to solve it. But that's different from the business/industry expecting it from everyone
Also, if someone is genuinely a workaholic, they should ask themselves what is truly most worthy of their time; if you're able and willing to work 60-80 hour weeks then you should seriously consider starting your own business.
I've known the true workaholics, and was one for maybe a year or two of my career. Also known a lot of folks who were just rationalizing.
I mean empirically I can, it just also drives me crazy at 80hrs/week
And I’ve never wanted to start my own business
Wasn't suggesting that you personally should. Was more referring to Atra saying he's fine with it. If one enjoys or is at least ambivalent about working that many hours, why waste those hours on wage work?
Entrepreneurship scales extremely non-linearly with hours worked, the difference between 80 hours a week and 40 hours a week can be 10x the revenue rather than double. Whereas putting in 80 hours a week in a salaried job gets you... nothing. Literally.
you'd be hard pressed to find anyone at my job working more than 40 hours consistently
the office is essentially abandoned by like 4:30 pm
I'm relatively certain that consistently working 60+ hours in my country is considered illegal business practices
the absolute most that is legally allowed is 55 hours a week for a maximum of 4 weeks and it needs to be averaged out to a maximum of 48 hours a week over 16 weeks
for the most part this is actually adhered to... unless you work in Healthcare
so sad
a tragedy
"This repository's code has not been indexed yet. Try again later"
RAAAAAHH you're owned by one of the biggest tech companies in the world that runs a multitrillion dollars cloud service, the hell you mean it's not indexed yet after 2 fucking weeks
it's Microsoft man
they haven't even figured out indexing yet for their file explorer
I guess on the flip side if you are gonna be at work that much you wont' have time for friends but I think you made a good choice
I've found that it really loves to do this with forks in particular.
(Though I've had it happen on masters too)
Honestly, I can understand why indexing might just fail once in a while when you've got ten million repositories; those 0.01% bugs become certainties. The crime is not having a button to click on that says "ok fine, nobody's perfect, but seriously, hurry up and index that shit right now".
hi
hi
nothing quite like the minor annoyance that is a spelling error in a configuration file that you cannot fix cause it'd break shit
someone typed "responsed" instead of "responded" 10 years ago and now I suffer
stardew version (sorry):
public class IslandActivityAssigments
{```

If I ever find myself at crumble company I wil dedicate myself to minor typos
As in intentionally making them
cruel
A client of ours has a "Premissions" table for user permissions
the consequences of a typo
one thing that happens a lot here is people mixing in Dutch words in their programming which can drive me mad

there’s suddenly random dutch or it’s even worse and they combined dutch and english into the same variable name
the “GebruikerRoles” (user roles) table
that one is made up but it’s the same kind of thing
at my old job we had an even worse category
half dutch half english but the dutch is also wrong
-# Dutch is always wrong
we had a “WichtQuantity” aka a “weight quantity” but the dutch word for weight isn’t “wicht” it’s “gewicht” so now it’s some weird middleground. Wicht does have a dutch meaning but it’s like… a not so kind word for a girl/woman
I wonder if they are trying to write English but somehow getting it mixed up with an unrelated Dutch word (that sounds similar) in their head.
Real talk: is an electrical engineer "supposed to" be able to replace a ceiling light
Mom and I disagree
I think your average electrical engineer probably could. A second year uni student probably could
She says no, she knows plenty that probably couldn't
I can do that and I barely know how to hold a screwdriver
I feel like this might be some discrepancy between what I'm (and I guess Crumble) are used to by "ceiling light"... being a super basic twist it in and out kind of thing, versus a fully mounted thing you'd need to rewire?
well are we replacing the light or the light fixture
That's the question! I assumed the light itself like I think you did...
most of the trouble installing the actual ceiling light fixture has nothing to do with the actual wires in my experience... It's figuring out how tf to hang it
often this funky ass setup with a little hook
Light fixture
Then I'm pretty sure I would agree with you...
It might not be the kind of thing an EE would be expected to do for a job, but if you can't do that, then...
shoutout to my 2 dangling ceiling lights for which I never installed a proper thing
Yup. If you can recognize colors you can wire a light. White to white, black to black, green or copper to earth, done
Or alternatively in the UK:
you have 3 wires for your ceiling lights?
maybe it’s me not having lived in a building built within the last 70 years before but I’ve never seen that
don’t remember if the wires have colors
I’ve done it for a fluorescent light, it involved a lot of cursing but it was fine
And I’m not an EE
An EE student is surely going to understand the theory behind it, how the circuit works, etc. On the other hand, they've probably never heard of a non-contact voltage tester, have no idea how to trace which breaker is responsible, or any of the other tools and skills for doing it safely. Also, unless they've done other handy-work, they might not know how to work with solid wire, or any of the mounting hardware, and certainly will struggle the first time around with juggling their tools and wire nuts and other pieces around while trying to hold the damn thing up there.
I'd say your average EE could figure it out fairly quickly if given a couple of tools and a 5-minute crash course, but otherwise have to agree that it's mostly a different skill set. Knowing what the circuit diagram looks like is precisely zero help.
Hmmm
I feel like a lot of that falls into figure it out though
Sure
It's awkward af
It's just another theory vs. practice thing. EE students and the majority of all college students have never done any actual trade work, and to the extent they've worked on "physical" projects it's usually in some heavily sandboxed lab setting.
Sure, they'd be able to figure out in 5 seconds which wires to connect, but so would a lot of eight-year-olds. It's all the tasks other than identifying the wires that make the difference.
(Also, sometimes the colors don't precisely match. You might have a fixture with, for example, both red and white wires.)
Yeah, but I think most minor home repair is within the abilities of your average homeowner
Not even remotely. If you'd said it's within their capacity then yes, I'd agree. But "abilities" implies they can perform those tasks now, without a guide or external aid.
Most homeowners can probably learn. Most homeowners also never do.
I include with a guide in ability
Anyways
YouTube has convinced me I can replace a toilet
The hardest part is probably dealing with the bulk. Getting a one-piece up a flight of stairs...
Going to the gym more is on my new years resolution anyways
My grandmas house had black, white, and yellow iirc?? Also half of the wires weren’t color coded, they were just a color
I’m pretty sure you can, just shut off the water carefully!
Hot Reload isn't working for me in Rider.
Anyone seen this error?
[ENC2014] Changes made in project 'DaLion.Harmonics' require restarting the application: Could not found output assembly for mvid: 3187fef9-e01d-43e8-b7a1-4fe9bba9f850
It's... also not working for me, for what it's worth.
Has it ever worked?
Absolutely!
Quite when it broke, I have no idea, because I was purely on Linux for a good while. Sometime between the start of that period and me needing to go back to Windows for a while.
In Windows did you go back to Visual Studio?
I've just been suffering, mainly.
But on the occasion I really need to hot reload to test something, yeah.
I see, that sucks
I even tried rolling all the way back to Rider 2022, and it was still broken. So it would seem to be something in Windows/.NET/SMAPI.
I was never able to get it working on Rider Mac either. Though I didn't get any error, it just... didn't do anything.
On Windows I just stick to VS.
Can't quite see how it would be .NET's fault, though, since this is all still on the .NET 6 framework which isn't getting any updates. (Unless hot reload is tied to the IDE's SDK version and not the runtime version... which is definitely possible)
I still generally run through the later .NET tooling/msbuild, but that's fair. I even tried using .NET 6's entire tooling to no avail.
This is the same issue I've had with Rider for a long time, and so I eventually abandoned it for VS2022. I've mostly gotten used to the differences, and have found Resharper to be sufficient for most of the things I missed.
hot reload worked flawlessly for AoC for me on linux
from what I remember
I used to use resharper but for the large projects at work it just wasn’t feasible to keep using
it would make visual studio almost completely unresponsive to the point of triggering windows’ “this application seems to have frozen” popup
How long ago was that? I've found the new 64-bit 2022 is pretty responsive even with Resharper
You also have a nasa level machine
Well even with my beefy machine, it was much snappier when compared to the non-preview, non-64-bit version
I have a long history of the feature simply not working correctly in Rider for me, and have found I'm not alone in that, and it's something that it works for some and not for others. The same code with all of the recommended configuration supports Hot Reload in VS and not in Rider (for me).
I've even tested it in a Windows VM that I installed from scratch to find that the issue could be easily replicated
It's something they seem to have no intention of fixing anytime soon since it's been at least 2 years of me waiting for a solution
like 5 months ago
I'm not sure if I ever noticed the "as the target framework" bit, that would explain an awful lot.
Well I suppose relatively better can still be bad depending on your starting point. VS used to be borderline unusable for me which is one of the reasons why I found Rider such a joy to work with.
For some reason I had it in my head as "need .NET 8 installed"
I had to learn pretty quick that opening a webform’s aspx file would basically just render visual studio completely unusable if resharper was enabled
same for the 40k+ line “main” (vba) code behind file from the cash register’s main form
I might have just experienced more disastrous codebases with resharper
Hot reload in Rider works flawlessly for me on anything, but it's specifically Stardew it's broken with.
rough
to be clear, these things aren’t particularly snappy without resharper either. They just were kinda workable
this was on some extremely beefed up 64 GB ram i9 laptop
I think it's time for my annual "post the terrible thing that I'd done a while back" comment: https://github.com/ameisen/SV-SpriteMaster/blob/master/SpriteMaster/Extensions/ArrayExt.cs
It isn't with Unreal, at least. It was bad with 4.27, it's even worse with 5.0+.
One thing I've been wondering but haven't taken the time to look up: are wire nuts or lever nuts preferred if they're being used somewhere that isn't fully environmentally-controlled, like an attic?
My instinct says lever-nuts, since the springs should compensate for expansion/contraction, but I'd also expect them to fatigue over time. But twist nuts also experience differential expansion...
I am 90% positive I've seen wire nuts in my garage. But I am definitely not any kind of expert on home wiring or electrical codes, just a barely-competent DIYer.
I do know that for electronics that might be stored outside, the answer is "neither" - that shit needs to be soldered, heat-wrapped and put in a proper enclosure.
Well, in the US, lever nuts are very rare overall - electricians don't seem to trust them for some reason.
However, I don't believe that the NEC allows you to solder electrical wires.
Probably not, no.
so, what electricians in the US do isn't necessarily indicative of the best choice. Probably true elsewhere.
Wire nuts are fine
Home wiring is always on breakers, though.
You've got short protection at the source.
WAGO lever nuts are also relatively recently introduced.
If you put them outdoors though
The issue isn't shorts. The issue is that the contacts come loose due to expansion/contraction, leading to an increase in impedence, and thus temperatures, spiraling the impedence, and then a fire.
or arcing.
the latter can be protected against with AFCIs, but only if an arc actually occurs.
You twist the wires together first
That's more than enough surface area for a gd light bulb
That actually goes against what the wire nut instructions always say.
I've never actually seen contacts come apart in a wire nut. Doesn't mean that it doesn't happen, but can't be common.
you're not supposed to twist them together first.
Eh? I've never seen instructions saying that. Always twisted 'em.
The twisting of the wire nut twists them together and compresses them.
wirenuts are not UL tested pre-twisted.
the instruction sheet will very much tell you to not pre-twist.
That doesn't mean that you shouldn't, just that they say not to.
I prefer using Ideal's nuts, your mileage may vary with other manufacturers
I Googled it and... wow. Controversial topic, apparently.
A thermal expansion coefficient of 29 ppm per degree C, a maximum swing of (and I'm assuming you live in Minnesota) no more than 150C
some makes probably handle it better than others.
it's more of an issue when it's repeated instances.
metal fatigues, even on small movements.
and it also has this habit of working its way loose
(because of entropy)
Sorting through all the emotions, the general message I'm getting is that wire nut manufacturers don't say not to twist, just that it isn't required because the nuts are supposed to do it themselves. And in a lot of people's experiences (including mine), the nuts just... don't.
I suspect that if the nut isn't twisting it sufficiently on its own, you either haven't applied the nut fully, or the nut isn't very good.
Maybe it works better for stranded wire.
I hate using twist nuts on stranded.
unlike solid, there's just so much more give and it never seems secure.
I just use WAGOs there.
if it's not full electrical, crimps.
or the nut isn't very good.
Would hardly surprise me that cheap parts are cheaply made.
As said, I use Ideal stuff. I work on stuff in my home
I think this is the most electrical engineering I’ve ever seen this channel get
and if I'm doing it myself, I want it done well, and with the least trouble to me.
so, ideal wingnuts, and WAGO lever nuts
WAGOs are preferred in most studies; they're just uncommon here.
This is more like electrical wiring than engineering... electricianing.
I'm just unsure of certain environments like the attic.
this is electrical... programming.
I don't do the stuff professionally, so I don't buy anything cheap. No benefit to that.
it’s ok. it’s enough for me to have no idea what it’s about and that’s what matters
I have little to gain by cheaping out on nuts and either burning my house down or making my hand hurt.
or, worst-case, both
I think the wire gauge may have something to do with the effectiveness of nuts as well. It's hard to conceive how a regular wire nut could apply enough force to a bundle of 12-gauge solid wires to actually get them properly compressed and twisted. You have to use pliers.
well, ideal wingnuts are pretty thick and sturdy
and the wings let you apply a lot of torque (and if you use the tool it's even easier)
And this tool does something other than the equivalent of pre-twisting?
lever nuts do have the advantage of, well, just working
It stops my fingers from hurting after twisting 40 nuts
:D
Lever nuts, at least according to Home Depot, are about $1 per nut. So there's that.
You can certainly argue "don't cheap out on safety" but still, that's more than 10 times the price.
that's a bit much. They're much cheaper elsewhere.
on amazon, 2-conductor are about 60c each for 10, 28c each for 100.
but, as said, I'm working in my house. The cost difference here is negligible.
the advantage (from an electrician's viewpoint) is that WAGOs are still significantly faster and easier to install.
NASA uses crimping.
(I've already read that document in the past)
Yup. They would know how to deal with fast, repeated temperature swings
I could learn to love crimping I guess
crimping is objectively the best, it's just a pain sometimes
have to make sure you use the right tools and do it right
NASA also used wire-wrapping in the past, IIRC
Wire wrap, now that is a pain
we should go back to knob-and-tube
I can't use ROMEX here anyways
the electrical code in Chicagoland requires all electrical conductors to be within metal casing - EMT or RIGID.
The electrical code in... wherever I live, is: "Building code? What's that?"
if anyone wants to participate in my experiment "How well can AI mimic your friends", please DM me for more info!
(i desperately need more people, as my science fair deadline is in 13 days)
heyy does anyone here have any knowledge of switch hardware ?
I swear Phone Link is the dumbest bit of software MS has made
My goal:
- Copy one LONG link from my computer to
sorry wrong chat
I pretty much message myself on Discord to accomplish that
hi discord I'm just buying shit from grainger again, lolz
I have a channel on my server called #clipboard where I just paste things between my phone and PC
it's so silly
I tried higher tech options, but it hard to overcome the simplicity of something that just works, and is easy to support
anyways
this is what phone link is supposed to do
let me copy and paste from my laptop to my phone
it does not do that
it can mimic my phone screen and I guess show me what would have been displayed on my phone (no idea why I would want that.)
but using kb+mouse to type in text boxes? Nope. Using copy and paste? nope
why did you do this, microsoft
Oddly enough, it works fine for me to copypaste between my pc and phone and vice versa
ah yes, my favorite type
i'm actually about to use the above in Pintail
that's very nullable
also i can't figure out try/finally blocks in IL
and i could really use one right now
couldn't you write one and then use an IL inspector
i did, and i get an invalid program then
oh that's exciting
i just wanted to pool some objects
apparently i was doing so much Pintail magic in my latest Cobalt Core mods, that i've managed to make GC pauses take half of the total runtime, all due to Pintail creating dumb Dictionary<string, Type> objects on every single proxy
i've managed to greatly reduce these allocations by just... checking if i even need that dictionary to begin with - they're only needed for generic types
but i'd love to pool the remaining instances
Hey y'all. I'm tired of dealing with the political drama of big corporations. Is the grass greener at smaller companies or startups?
Instructions also have block metadata for try/catches. I have no idea how you create those with Emit though, only how to copy them via harmony. I'm sure there's documention somewhere...
We're playing "what's that mod?" at work again, except "mod" in this context is an edit to the pcb
I haven't a clue why this is done, what this does, or even "what structure is this soldered to because it sure isn't on the schematic."
Fun stuff
Not inherently so. With a startup you could presumably meet the founders and know what kinds of people they are and what kinds of politics they're into (or not). You'll also probably be getting less pay, less benefits, less job security, often be asked to work long hours or deal with crunches, etc. Definitely not grass is greener, more like a choice between Fescue and Kentucky Blue.
The real grass-is-greener is starting your own, but that has its own set of major difficulties.
In the corporate world, I feel like I'm dealing with land grabs every year. Leaders fighting over who should be in charge of X, and meanwhile I just want to have my head down and do work without having to deal with all of these negotiation meetings.
If that's the only political drama you have to deal with, you're honestly kind of fortunate.
The grass is brown everywhere.
It's not the only drama, but it has been the most taxing lately
Literally had to be in a room for two full work days fighting over prioritization of my team's capacity
Why do you personally have to be at the meetings? I've been through a dozen reorgs in my day, just ignored them all.
I deal with it so that my team doesn't have to. They probably don't appreciate what goes on behind the scenes because I try to shield them from it.
Sure, I had to deal with the planning and prioritization of work streams when I was playing lead, but that's different from wholesale changes in project ownership.
It's just silly what's happening imo. No one is thinking about what's going to lead to more efficiency/productivity. All they care about is that they get to be in charge.
🫂
i.e. land grabs. People trying to be in charge just because they want to be in charge, even if they're incapable of providing any value to it.
It's like congratulations, you're in charge just so that my team can do all the work and you can take all the credit.
I think TPM has just lost its appeal to me, and I miss doing dev work
That happens to a lot of people. You can just say you're done with that, back to IC-ing, if they change all the priorities tomorrow then oh well, not your problem.
I'm guessing your company has some sort of formalized and possibly peer-review-based promotion system which essentially requires being a project owner in order to move up. So, yeah, everyone fights for ownership.
I learned how to play those games but didn't enjoy it.
I've been at a place with less than 20 people for ~10 years and there's approaching zero office politics, but they've been around for ~30 years and the majority of the company has still been there longer than me. I think the trick is to find a smallish, established business that isn't vying to IPO in 2 years and pays you the going rate for your area and expertise. The main thing would be that the work isn't going to be the same as what you're doing at larger places; everything will probably move slower and not be as shiny and new, so it depends what you're looking for.
That honestly sounds like a dream. The one thing I'd be missing is being able to use shiny new technology.
I actually was part of a different company that operated more like a startup, which got acquired by a bigger company
There was something special about finding out how to accomplish things with more limited options for tech/tools
Now there is no shortage of tools, which is fun, but they have jobs/roles that oversee every little thing
The owner prefers proven over shiny and new, and I can hardly blame him with how stable things have been for the company. We obviously upgrade things and try new stuff, but usually after it's been around a couple years. We also deal with a lot of smaller clients who only come back to us for feature changes, so their software tends to stagnate until its time to just rewrite it. Sometimes it just sticks around on life support, though, but a lot of it is internal webapps and things so security is less of a concern.
At the smaller company I got to see them upgrade from Excel spreadsheets, to Access databases, to SQL Server, and each of those was a game changer
Yeah we have systems people and then us app devs, so if its not hardware related or infrastructure like server, we are usually handling it, so we're all more jack-of-all-trades then experts in some specific stack, but that keeps things much more interesting
It's not uncommon for us to collect Excel files from new clients to translate them into a webapp with a database for them. And then they ask for exports so they can mess with the data in Excel and we have to push back real hard until they learn to ask for us to make the webapp do it for them
So many small businesses run on spreadsheets
I have nostalgia for those simpler times
I remember I added some automatic record creation to an existing site and a guy told me I freed up 3 days of his workweek from not having to hand-enter data from spreadsheets back into the site after he exported it and futzed with it
I once taught a guy how to do search and replace in Excel, and he said it saved him hours every week
I was a young analyst who was going through a bunch of inefficient processes and helped automate them
People will brute force Excel unless you teach them a few simple formulas
You mean compared to "can we get chatgpt to do it for us?"
Doomed
Anyway I think voltaek is basically on the money. Ask yourself if you are the "it's just a job" kind of guy or a "my job is my identity" kind of guy. If you don't care what you work on as long as you get paid and don't have to deal with too much BS, startups or rather small businesses can be great, especially non-tech startups that only employ a handful of tech people to begin with. But if you have to have the newest and shiniest, or if you need to feel like you're constantly learning new things, then you're going to be unhappy. Expect to see a lot of Java and Oracle and the like.
Ew Java. I’m glad I get to use Python and C# for work.
The python part of that is a toss-up for me, honestly.
I learn new stuff all the time and appreciate that I get to (mostly) set the pace of the my work, such that I can produce polished, maintainable results. When I do the estimate and then also do the work, it works like that
We just use proven frameworks/software stacks, is all
The flip side is that clients bring in busted, barely taped together software/sites that we have to revamp, and while that can be very interesting, it can also be frustrating, so there are ups and downs for sure.
I don't think that's entirely typical of small businesses, though. More often, being polished and maintainable is an afterthought. It's good that you personally can set your own pace but someone considering a new job shouldn't expect to find that easily.
You're correct, that's not typical and I can appreciate that it's difficult to find such a place. I think it will usually take people several jobs before they land on one that they like and are hopefully also compensated properly for.
Sometimes you have to go through a few places before you even know what it is you want in a job and a company, too, though
Microsoft: "researchers in France are turning to AI to assist with the non research parts of their jobs."
Me: "have we considered maybe not making them do those parts of their jobs."
no
You'd think tech companies would understand the importance of having humans in charge of accuracy
I think this is more positive than you make it out to be. At least if we take the headline at face value, with the caveat that those are usually a lie - they want to automate (which in practice means eliminating much of) the pointless bureaucratic make-work. What's your alternative, create a bunch more clerical jobs and make the bureaucracy even bigger?
Less bureaucratic work 
Tbh I'm complaining over stuff I've only rarely had to deal with
That is, in effect, what automating it will accomplish, because once something is automated there is always a concerted effort to minimize scope.
There was a lovely accountant at my last place who would handle stuff really well; often I just sent him the receipt + account number + project
Like. Over email
Several of the places I've worked at have had some online expense portal thing where I put in the same info. Doubt that's the kind of work they're trying to LLMify.
Why is the dwarf debug info doc 1 giant pdf 
Hey, I am new here and I dont really know where to put this message lol. I am a little comfused and i was wondering if the SMPI is only used for modding or if we can also use it to get information from the game as in levels that the player is at, day of the season, season, time, etc.
I'm not sure what you mean by "get information from the game" if not for modding.
All that information is stored in the save file though, which is just xml, so if what you're trying to make is an external tool, that's a good place to start.
I guess I should ask- what specifically are you trying to do or make happen
Yeah, so i am trying to make a game assistant kind of to have the stardew valley wiki accessible from the game, kind of like a little chat bot. But i dont want to be telling the chat like how much money I have, the day, the season and all that stuff if that makes sense. I was trying to figure out if i could get all that information from the API and what other information is available (as in is the greenhouse fixed, what level fishing are you, what did you choose fisher or trapper)
Yes you can, but that just sounds like making a mod, yes?
I guess i am a little confused of how mods work. the API mods are in C#? I want to use python if possible but i do feel that doing a mod would be easier. I guess i am a little confused on the restrictions of modding
To be clear- c# in general does not really have IPC capabilities out of the box, so if you want any real-time information into or out of the game, you WILL need to make a mod for that
Yes, you will have to use c#. Python has no way to send data to or from the game
(unless you make a mod to do it)
The game is written in C#. SMAPI allows your C# code to run with the game.
could you do a mod just to send real time information to another script?
Absolutely
You could. There have been mods that incorporate their own scripting language as part of their features. I even think an old mod called PyTK used to do stuff with Python Lua. The mod itself would be more complicated C# code to do the scripting stuff than writing it all in C#.
so what i want to make is by hitting some key strokes i want a window to pop up where you can ask a question like "what should I plant" and without any more context needed i would get anohter popup window with the answer and i dont really know a lot of c# but i know that python has many libraries for doing that and i dont know if it would be easier to do a mod to get the info to the other python script or just doing it all in the mod
Sounds like a good place to start doing research
if i ended up doing the mod i would probably do it in c# just to not do extra work lol
Stardew does have ui capabilities
that is true
Considering the whole game was written in C#, I'm pretty sure C# can handle UI
Since, you know, the game has UI
C# has libraries for nearly everything. It is extremely powerful, flexible, and concise.
Also, there's this which is a new mod that helps other mods implement UI more easily https://discord.com/channels/137344473976799233/1293051032904925255
Another detail that's probably worth noting. You theoretically could write a mod in Python. However, you would need to write a Python program that would run all the time alongside SDV, checking for various events to happen and reacting appropriately. This is a bit cumbersome, so the reason why C# is recommended is that the mechanisms are in place for the C# modding code to be initialized and run alongside the game, not as an external program
you can also just write mods in python directly https://github.com/SinZ163/StardewMods/blob/main/SinZationalSnakeShowcase/modentry.py
that is very true
omg you guys are so helpful, thank you very much!
:sigh:
Fuckin' passwords, how do they work?
the one most illogical here is the special characters
which makes me think those passwords are stored in plaintext without sanitisation or something
tbh the real annoying part is that I need to make a phone call to get a dental appointment
Does that mean special characters being a requirement for passwords is a sign of poor security?
limiting the special characters you can use is a sign of it
limiting your password in general is a bad sign. it just reduces the number of possible passwords there can be
they shouldn't matter
and in theory, length shouldn't even matter - most hashing algorithms have the same length output always
in theory I should be able to make a password that is 100% cat emojis and the chinese character for biangbiang noodles and it should be fine
I can imagine a length limit, but 15 is real low.
Duang (Mandarin pronunciation: [twáŋ]; pinyin: duāng; Zhuyin Fuhao: ㄉㄨㄤ; written as 動L in Hong Kong Cantonese with Jyutping dung6 eu6) is a Chinese neologism that has become a viral meme despite its meaning being unclear. It has become a popular hashtag on Sina Weibo with more than 8 million mentions by the start of March 2015.
very sad this isnt in unicode
(ie, I shouldn't be able to submit the entirety of War and Peace as my password, but 15 feels low.)
tbh it took me a few tries to realize that was why they were rejecting my autogenerated password
Can we also get a round of boos for sites that disable or try to disable pasting into the password field?



BOOOOOOOO
But if it's inconvenient for me to log into a website, it must be inconvenient to hackers too! /s
Making me want to make the least friendly login/sign up form of all time...
just make a login form where you simply always have to fill in a captcha to log in and have it forget your session every refresh
I'm thinking of including features such as limiting the amount of characters that can be inserted with a certain timeframe, too.
Too fast, it's obviously a brute forcing tool, and we don't want that.
anything below 20 wpm is considered being idle however which the website does not allow to avoid extra network traffic to idle clients
you will be forcefully removed from the website
reminding me of the password game
Forcefully removed by way of testing every possible known browser vulnerability on you
And since your machine is clearly idle, run a cryptominer so at least something productive is happening
🙏
If my password can't be the entire bee movie script what's even the point
alternative title: fun things to put into the production server on your last day of work
Gonna go out on a limb and say that's fake.
impossible!
Can a tech wizard explain why having a web browser open takes up space on my hard drive? Mozilla AND Google Chrome take up about 2GB when I'm on YouTube for some wild reason
Are you sure it’s hard drive space and not RAM?
Maybe if you've got a dynamically sized pagefile. Didn't think that was still a thing, but who knows.
Or it could be cache that is cleared on exit, depending on your settings.
There are a lot of possible reasons but really the only way to know for sure is to do a before/after snapshot and see what was actually deleted or changed.
I can imagine it shoving a bunch of stuff in a Temp file
$ md5sum /mnt/b2/test
c3c2969a32764125a5858c7f6bca5f72 /mnt/b2/test
$ md5sum /mnt/b2/test2
734e9abb7b05916c808fa109473884a0 /mnt/b2/test2
``` 
hmmm
I assume the -a flag is doing something to it
It shouldn't be, it just means it copies attributes and timestamps and stuff
Is /mnt/b2/test an alias for /dev/random or something?
No, nothing sneaky like that
I want to know what test is ,whether multiple cp -a runs result in the same hash, and to peek at the before and afters in a hex editor...
A random file, head -c 50G /dev/urandom > /mnt/b2/test
...well I'm certainly not looking at those side by side in a hex editor.
Yeah
$ md5sum /mnt/b2/test3
f59488c604ae003955a56f38e021f71e /mnt/b2/test3
$ md5sum /mnt/b2/test4
c7fb26956a30b2cd1b157755d642e5bd /mnt/b2/test4
$ cp -a /mnt/b2/test4 /mnt/b2/test5
$ md5sum /mnt/b2/test5
ce3641709b282a769be4aa9372cc2f2f /mnt/b2/test5
Repeating the md5sum on the same file does give the same result though
Good thought! Is this going on a FAT file system, maybe?
If they were being truncated, wouldn't they be truncated at the same place each time?
I would assume so, but at this point...
idk if it's truncation or something about the MD5 calculation picking a non-standard chunk of data
But no, ext4
If you try a smaller filesize do you have the same issue
That's what I need to try
And you know... try it on a different mount point, different drive, etc. Worst case would be a failing drive or I/O controller.
It must be something, not sure what yet. Not controller or cable, may be drive, still investigating
Did a compare on a 1G pair of files that were doing the same. 1 bit difference
is it a random bit different or one at the start/end?
In the middle
is it at least at a consistent middle
And it's not every file that happens to exist on whatever drive/controller this is?
Not every file, no, smaller files seem less likely to have a problem
And now it seems to have stopped happening, but I didn't change anything 
Hi everyone,
Does anyone know how I can get the item sprites?
I've been downloading them from the wiki one by one but there are too many
wrong channel but if I had to guess they should be available in the game files, possibly as xnb files?
sorry, i unpacking thats files but the items are not there
Ah got it
can someone make a mod where your children can help on the farm?
!modideas
If you have a mod idea that you aren't planning to make yourself, you can put it in the mod ideas github: https://github.com/StardewModders/mod-ideas
However, this does not mean anyone is guaranteed to work on your idea—modders who are looking for ideas sometimes go through and work on what they find interesting off this list. If you want to pay someone to make your mod idea, there are a few people who do commissions (mostly art, sometimes code); you can ask around, search usernames for the word comms, or see !commissions.
spent the last few days experimenting with blazor vs svelte for my job and man blazor sucks
I've used blazor professionally for years now and I still can't get it to work right while I got it working exactly how I wanted within an afternoon in svelte
As much as I like c# I am skeptical of blazor
they tried to do too many things at the same time with it, making it shit for every one of those options
Do you think it's better or worse than php
I've never used it so can't judge
I can tell you it's in some ways actually worse than asp.net webforms was
though I'd still pick blazor between those two
Yikes
blazor has so many core issues that have been just moved to later releases over and over for years now
the framework is honestly only really usable at all if you
- turn off prerendering completely
- only use blazor wasm or blazor server and never auto
and for both blazor wasm and blazor server it's very far from ideal
Fat clients my beloathed
blazor wasm takes genuinely ages to load even if you try everything you can to reduce load times and the debug experience is sometimes actually worse than if it didn't exist
blazor server just has core issues with requiring permanent stable connectivity to function due to the websocket but on top of that the framework missing some extremely basic important features like debounce that you'd need to prevent some types of events to not completely overload the websocket queue with events (like onmouseover you simply cannot use without building some custom funky ass workaround through javascript bindings)
the amount of data sent over the wire by blazor server is also quite significant when compared to for example phoenix liveview(which uses the same concept of a permanent websocket connection)
Phoenix liveview uses absolutely tiny little update packets that very precisely specify which parts need to change, blazor will gladly blast over the entire grid's worth of data through the socket if you look at it wrong, sometimes several times within a second
I've been wanting to use htmx for the frontend of my next project, but I'm not sure what to use for the backend yet.
C# is great for api stuff, but its templating seems terrible.
Rust has some promising libraries but lacks higher-level utilities
Codeigniter seems pretty good but my php experience is very minimal so I can't judge it well
on a more fundamental issues level, blazor tries to be a "reactive" framework (aka updating the dom immediately based on changes in the data) without actually introducing any kind of mechanism to track updates. Which means it will just assume that all data changed every single time any event happens (or StateHasChanged is called)
...what.
Oh yikes
blazor doesn't even support observables/notifypropertychanged
But...
like basically every react inspired framework introduces some kind of system where if you try to update something, the framework can detect this and attempt a minimal dom patch if necessary
some like svelte go a step further using proxies + signals to detect the exact change you make even if it's only 1 field in 1 object in a list of 1000 items
but blazor? nope
nothing
So is it like... they just havenn't shipped a Blazor-native way of doing that, or with how it's made it'll actively fight you trying to do it yourself?
anything that is not a primitive type like an integer or something is always assumed to have changed unless you manually override the ShouldRender method and then manually implementing change detection
ShouldRender wasn't available either for most of my time using it
yeah...
there's also features that are on by default that don't work when used together which is wild
they have some feature to "persist state" from prerendering so that it doesn't have to be retrieved again once the client itself takes over... they also introduced "enhanced navigation" so that when you click a link that brings you to a different page on the website it won't actually do a full page load and just update the DOM/url accordingly...
these do not work together so if you go to a different page while enhanced navigation is on and it tries to prerender the page, the state is not persisted anyway so it does it twice regardless
yet both are on be default, making both features fundamentally broken out of the box
when they introduced built in authentication features they actually ran into this themselves and built some funky workaround manually. Then in that issue tracker someone of their own team commented like "this is clearly an issue we should give more priority to fix"... which was years ago at this point
so yeah
as I talk I keep thinking of more complaints I have
but I'll leave it at that
hello
hello
Mystery solved. Switched to Windows on the same machine, still had the same problem. Moved the drive to a different Windows machine, no problem there, so it wasn't the drive. While I was doing that, ran memtest on the old machine, failed. Removed one RAM stick, still failed. Switched to the other stick, pass. Put the drive back in the first machine and tried again, now no problem. 
That's quite a bad stick of RAM
1 bad nibble I think
latex is a bit like programming — today I was trying to figure out how to get \section{} to use header tagging for accessibility but it turns out the answer is ???? pain and suffering for the most part (there's apparently a package straight up named accessibility that does it, but also is kind of old and broken, so lots of conflicting advice)
TeX is a turing complete programming language
uh oh
C# drama
fluent assertions changed its license to require commercial users to pay
and it's over 100 dollars per seat
who would ever do that
fluent assertions does nothing except make test assertions easier to read
I luckily didn't use it in the first place though I considered it
nah, the difference is huge
huge
I never understood the appeal of those frameworks. Like it's really that hard to read Assert.Equal(a, b)?
There seems to be a certain type of programmer who that more abstraction and more complexity automtically means better design, even if the abstraction doesn't do anything useful.
Some people are just addicted to chaining method calls
I blame JavaScript devs
the only thing this does that I find somewhat "useful" is that I sometimes forget which of the two in an Assert.Equal is supposed to have the one you're testing and which one the value you're expecting
this is something that can be solved within like a quarter of a second and visual studio even tells you if you put them the wrong way around but still!
it's something
I've always just used the standard built in asserts regardless
I love fluent api's but for this it doesn't seem to actually get you much
I'd be more worried if this had been about FluentValidation instead of FluentAssertions
cause DataAnnotations ain't great and I don't wouldn't want to migrate back to that
Always check the license. Apache, MIT and even GPL can't be retroactively revoked, so the worst that happens is you get rev-locked. And then, probably, someone forks the open version and continues maintaining that as open, and the original company that pulled the bait-and-switch fades into oblivion.
If it's "free" but not open, well...
yeah fluent assertions was apache so you can just continue using the old release
and considering assertions don't change particularly often I don't think you'd miss out on much
I really don't get it
Tbh
I like fluent apis (perhaps too much)
But I don't find them more readable
And especially not for assertions
honestly my guess is that the one who developed this was kinda done with maintaining it and decided to cash out, not caring very much that this would definitely fail
That's usually not how it happens. Usually it's a change in ownership, with the new owners not being particularly bright re: how open source actually works.
apparently the company that purchased it (and now employ the guy who made it) have pulled the same trick before
with the wpf toolkit
It happens depressingly often. I'm trying to remember the specific example that I came across a few months ago, but for the life of me cannot remember which product it was. It was something I hadn't used in years, and had never used very much in the first place. The acquiring company blasted out a bunch of emails saying we were going to get charged monthly - no ifs, ands or buts. One wonders how they were going to do that with users who hadn't ever provided a credit card number or other FOP.

