(#np2awqq) @prologic@twtxt.net uBlock Origin on Firefox, yes. I also wipe my browser profile on every shutdown – not sure if that makes a difference.

(#7jmjuwa) @xuu@txt.sour.is If only I could see them. All I get is Twitter’s login screen. 🫤

(#np2awqq) @prologic@twtxt.net I still haven’t seen this and I sure hope it stays that way. 🤔

(#mniqw3q) @jason@jasonsanta.xyz Do you have mutt’s threading view turned on?

set sort = threads
set sort_aux = date-sent
set strict_threads = yes

I find this to be very non-confusing. 🤔

(#zkjd2ya) @xuu@txt.sour.is Nice! 😊 I like it when those puzzles are solvable using pen and paper. Some of the other puzzles require you to write lots and lots of code, that can be really tedious.

(Yes, I know that it’s called “Advent of *Code*”. 😂)

(#y7edboa) @xuu@txt.sour.is Yes, exactly that. But: I do have more than 32 bits after all, when I use double or long double. 😲 I went with long double which has 80 bits even on DOS (63 bits mantissa, I think).

(I only wrote a brute-force thingy this morning, so I didn’t even think about doing anything related to polynomials or double. 😅 Even part 2 only took 55 ms. But for DOS, this wasn’t an option anyway, so I did the proper math while waiting for something at work. 🤣)

It’s still a bit surprising to me. Usually, AoC doesn’t require such tricks and most things fit neatly into 32 bits. Maybe there’s a better solution after all. 🤔 I might have a look at reddit.

Today’s AoC puzzle is a very simple problem on modern machines, but quite tricky for me: It involves a number that doesn’t fit into 32 bits. 🤔 I wonder if/how I can manage to port this beast to DOS. (I once wrote a “big int” library myself, but that was ages ago and I hardly remember it anymore.)

(#zktowyq) @xuu@txt.sour.is I see, interesting. 🤔 Now I’m curious if I could optimize my brute-force solution as well. 😅 (I wrote a “proper” solution in the meantime after watching a visualization by someone else.)

(#ebrldga) @lyse@lyse.isobeef.org What a great scenery. 😊 I hope the snow comes back soon. It’s all gone at the moment and well above 0°C.

(#zktowyq) @xuu@txt.sour.is And that is still a brute-force attempt? That’d be really fast. 😳 Are you doing anything clever or is your CPU just that powerful? 😅

(#irslaoq) @prologic@twtxt.net Doing AoC competitively is super hard. I don’t enjoy this at all. I just happened to be quick these first few days. 😅

I rather set my own goals. Like the DOS thing this year. That is a lot of fun. 😊 (I just have to remember not to check the leaderboards. 🤣)

(#irslaoq) … it just finished and brute-force worked. 18 minutes of computing time on my 11 year old machine, single-threaded.

(#irslaoq) @xuu@txt.sour.is I ask myself that question every year. 🤣

To be fair, the first part wasn’t really that difficult. If you have A LOT of experience with these kind of problems/puzzles and if you have a proper framework, I imagine it’s doable. (I, on the other hand, spent about 40 minutes just writing my C code to parse the input.)

Some of these people record themselves and then post it on YouTube. It’s pretty crazy to watch. 🥴

The second part of the puzzle, was/is pretty hard, though. At least for me, because I haven’t found “the trick” yet. I’m currently trying to brute-force it while having breakfast. 😅 (But given that it took ~8 minutes for the first person to get both stars, maybe they brute-forced it as well. With a faster machine and multithreading, ~8 minutes sounds about right. Brute-force is rarely the answer in AoC, though.)

(#txfbuaq) @eapl.me@eapl.me It’s been a while since I used QBasic/QuickBasic, but yep, I remember the experience being quite similar. 😊

(#ghidsjq) @jason@jasonsanta.xyz I got it. I was already asleep yesterday. 😅

(#ghidsjq) @jason@jasonsanta.xyz Sure thing. The address is on my website. 👌

It is a pleasure to work with the help system of Borland’s Turbo C++ 3.0 on DOS. The descriptions are clear and concise. There are short and simple examples. Pretty much every help page is cross-refenced and those links can be clicked.

(#s7b6peq) @lyse@lyse.isobeef.org

Holidays up until next year, hm?

Exactly that. 🛌

(#zkjovca) @prologic@twtxt.net Here you go: https://movq.de/v/dfb04df5c7/aoc2023-day4-part2.txt

(#s7b6peq) That second photo is from our local pond. See all those tracks in the snow? Those are from the ducks and the other birds that stayed here.

This video is from a couple of days ago – just frozen, no snow yet:

https://movq.de/v/527a76bd19/VID_20231203_104847.mp4.mp4

(#ghidsjq) @jason@jasonsanta.xyz If you can describe the problem, maybe I can help.

(#zkjovca) @prologic@twtxt.net I can try to explain, if you want to know. 😅 Or maybe this guy’s visualization will help (if you can see it, not sure): https://old.reddit.com/r/adventofcode/comments/18af47j/2023_day_4_scratchcard_counting/

A couple centimeters of snow today. 😊 And I’ve got a loooooooong vacation coming up. Time to relax. 😃

This is my favorite time of the year – by far. Even if it means the snow will freeze in my beard. 😂

(#zkjovca) @prologic@twtxt.net (Mild spoilers ahead.)

Today’s puzzle was one of those where I immediately knew that there must be a solution that does not require actually copying anything. AoC often has these kinds of problems that make it look like you have to create lots of duplicates, multiply objects, stuff like that. That’s often a sign that you can solve it by doing something simple. 😅 (If only that was true in real life as well. 😂)

(#qwm7hha) A little “visualization” / “progress bar” of day 3 part 2:

https://movq.de/v/6fc13f8ce1/MVI_6766.MOV.mp4

Running in a DOS box under OS/2 Warp 4, including some of OS/2’s silly sound effects. 🤣

(#apnoyca) @prologic@twtxt.net Yours is more clever and cleaner, I think. 🤔 You were able to reuse more code for the second part. I, instead, wrote a different search logic for part 2. 😅

(#apnoyca) @prologic@twtxt.net

Did I write too much code again? 😅

I don’t know. 😅 My solution is rather long as well:

https://www.uninformativ.de/git/advent-of-code/file/2023/day03/s.c.html

Still happily runs on DOS.

(#qwm7hha) @prologic@twtxt.net Oof, it’s already getting a bit tricky. 😬

(#vdtfw2a) @lyse@lyse.isobeef.org All I can say is: My client didn’t report any errors. :-) (Does that mean anything? 😅)

(#6oq4ywq) @xuu@txt.sour.is Ah, you went with the “scanning” approach as well. I did that, too.

It’s quite surprising to see (imho) how many people on reddit started substituting strings (one becomes 1 etc.). That makes the puzzle much harder by introducing nasty corner cases.

(Maybe I was just lucky this time to pick the correct approach right from the start. 🤣 Or maybe it’s a bit of experience from doing past AoC events …)

Here’s some more #AdventOfCode nostalgia:

https://movq.de/v/6441f2fe2a/MVI_6760.MOV.mp4

It shows the machine’s boot sequence and the copy process: I somehow have to grab the files from my normal PC and I do that using FTP under Windows 3.11, there’s a PCI Ethernet card in that machine. Then some glorious WinZip action to decompress the files. 😃 Finally the first two AoC 2023 puzzles are being run.

(Yes, there’s a GRUB on that machine. 🥴 It’s a left-over from some experiments with Linux ages ago and I didn’t bother to uninstall it.)

(#wg7xx5q) @prologic@twtxt.net That is a lot of code, yes. 😅 Mine is shorter, has more naive searching (it searches a word like red and then backtracks to the number before it; completely ignoring the semi-colons because they don’t matter), and – like any good C program – will crash horribly on malformed input. 👌😂🥴

https://www.uninformativ.de/git/advent-of-code/file/2023/day02/s.c.html

It really depends on what you’re after. I’d never write code like this in “real-world applications”. But for these puzzles and my DOS use case, it’s “good enough”. 😅

(#mxyoxrq) My immediate reaction to today’s puzzle was: “Ahhh, shit.” 😂 Parsing these input files in C is not a lot of fun. 🥴

(#kr3qhra) @prologic@twtxt.net It sure does! 😊 (Was the video big/large enough? I should probably capture in a higher resolution next time. 🤔)

(#kr3qhra) @prologic@twtxt.net The solutions will go here:

git clone https://uninformativ.de/git/advent-of-code.git

About demos … I made a video yesterday, but I’m not sure if it’s that interesting. 😅

https://movq.de/v/de32f1481f/MVI_6756.MOV.mp4

I’d probably have to do some cool visualizations for the more interesting puzzles. 🤔 Not sure if I can pull that off, though. 😅 It’s probably going to be hard enough anyway. (I wonder how long I can get away with just conventional memory, i.e. only ~500 kB.) We’ll see.

(#mldneoa) @prologic@twtxt.net Ah, right, kids and school. Yeah, oof, nothing you can do about that. 😐

(#mxyoxrq) @prologic@twtxt.net It’s the perfect time of day for Aussies, isn’t it? People in the EU would have to get up around 6am. 🤣 I just happened to be awake anyway today.

(#kr3qhra) @prologic@twtxt.net I joined as movq42rax. Don’t expect speedy results from me, because porting my solutions to DOS is my personal goal this year, intended to be runnable on that old Pentium. 😅

(#mldneoa) @prologic@twtxt.net Argh! 🤢 I wish you guys a speedy recovery. (Any idea where you got it?)

(#hzu4kfq) @eapl.me@eapl.me Great timing. 😅

Yeah, it can be quite exhausting to do one puzzle per day. I guess taking some vacation is mandatory. 😂

(#7hvbypa) @lyse@lyse.isobeef.org Now that’s what I’d call winter. 😊 Awesome to look at. Poor camera, though. 😮

(#spbxjha) @eapl.me@eapl.me Whaaaaaaaat, you’re doing a programming challenge at this time of the year and it ISN’T Advent of Code?! 😱 😅

(#fbgdevq) @lyse@lyse.isobeef.org Oh lovely, look at all that snow. 😍 Good luck to the restaurant guys.

(#hz2qwyq) @prologic@twtxt.net I walked. 😅 I mean, I walk rather fast and it’s not a relaxed stroll, but it certainly isn’t running. 😅 The goal isn’t to lose weight but to be outdoors, enjoy nature, and clear my mind.

Just to be clear, it was 100km over the course of a whole month. It was 23 tracks with 4.4km per track on average (4.4km is roughly 50 minutes). It’s actually not that much, it’s mostly time consuming. 😂

(#hz2qwyq) Made it to 100km today. 🥳 (It was harder than it should have been, because I was basically knocked out for about a week. 🤣)

(#ngkm4ya) @lyse@lyse.isobeef.org No need for snowplows here. All the snow is already gone. 😢

I posted this link about Windows 3’s architecture and VMs a while ago, but this topic continues to fascinate me. Raymond Chen brought it up again recently.

I’m aware that virtualization itself is much older than Windows 3 (IBM did it in the 1960ies, I believe?), but knowing that similar concepts existed in my tiny little machine that ran Windows 3.1 is just mindblowing. 🤯 (Alright, it wasn’t exactly “tiny”. It was an IBM PS/2 Model 80. 🤣)