The Nineteenth Byte - 2017-04-24 (page 1 of 5)

« first day (2277 days earlier) ← previous day next day → last day (2571 days later) »

12:00 AM

Hello

What does everyone think about the challenge?

@Downgoat what are you compiling

@HelkaHomba Hello

Interesting task, surprising to find a simple idea that uses a web request.

@ATaco It was a lot of fun to write the original program. :)

Never forget that Dennis will Outgolf you.

@ckjbgames libgit2

12:12 AM

@Downgoat Cool

@ATaco Of course. I wasn't trying to golf the original; I was just doing it as a project.

not really, all the building has raised my CPU temp to 78ºC

@Downgoat So, literally not cool.

yeah :P

@Downgoat :)

12:30 AM

Question: I have a problem I want to ask that has a cool math trick I know that can help people solve it. Should I reveal it in the question or let people discover it on their own?

I had a lot of fun discovering it myself so I thought it would be fun to leave it for other people to find, but it feels a little underhanded because the trick is pretty powerful.

you could put it in a spoiler

If you are going to make it a challenge, I would recommend posting it to the sandbox and people will judge if the trick should be part of the question if they figure it out

that way, people who want to figure it out themselves can, but they also can look to see what it is

And if it should be put in the problem put it into a spoiler

I think spoilers are >! text

>! test

maybe not in chat

12:34 AM

@CartManagerXD I think you should leave it to others to solve

I think it's really fun and gratifying when you discover cool tricks for an answer

Here is a straw poll so I can get some data.

I'm tempted to pick option 5...

@CartManagerXD So long as you can properly spec the challenge without revealing the trick I'd say you should leave it out

I voted "Yes"

I think I made the wording confusing here is a clarified version, feel free to vote on either one but this one is clearer

Ok looks like I wont include it thansk guys!

12:42 AM

Misunderstood; re-voted "No"

Thanks!

@CartManagerXD How are yes, but in a spoiler and no, but in a spoiler different?

Yeah that was super confusing

I don't think either of them makes much sense

see the second one where I fixed it

Oh, didn't notice the was another link. Four blue letters in the middle of a sentence are easy to miss. >_>

Bye

12:52 AM

o/

@Dennis, the !!superprime answer on the primality checking question isn't a valid answer because it isn't a programming language according to our standards.

We don't require submissions to be in a programming language

this challenge will also serve as a directory of proven programming languages

We require two possible programs for our definition

0

Q: Reverse and add degeneracy

Intro Reverse and add is as simple as it sounds, take n and add it to its digits in reverse order. (e.g. 234 + 432 = 666). If you apply this process repeatedly some numbers will eventually hit a prime number, and some will never reach a prime. Example I currently have 11431 rep. 11431 is n...

code-golf decision-problem

^ to be completely clear thats the problem we were discussing. Have fun!

1:07 AM

@CartManagerXD Wait, what? We have struct rules about what a valid programming language is, and submissions in other 'concepts' are invalid.

Yeo

welcome to bureaucracy!

?

Yes that is true. This is the result of a bureaucratic system.

I actually agree with the current rules but it is a bit strange.

We are less strict about programming languages than turing complete is. Turing complete is pretty darn lax about it too.

@Pavel If you meant, "Strict", I disagree. They're guidelines.

1:13 AM

No, they're pretty specific. If your language can determine primality and add two numbers, it counts, otherwise, it doesn't count.

As far as challenge creation goes, no-ones forced to use these guidelines as rules.

I don't think that's the best way it could have been defined (*cough* bubblegum)

Most importantly, Primacy Testing doesn't require Turing Completeness.

Sure, regex can do it.

@Pavel not exactly, though that's one of the main ones

1:17 AM

I made a language specifically not turing complete that could solve it.

I read that backwards my bad

Ah, all good.

.|(.+)\1+ matches composite numbers in unary.

Actually not quite, I messed that up

Oh well

@ATaco there is also Bubblegum by Dennis

Bubblegum is technically turing complete.

Through shinanigains.

1:19 AM

@ATaco We don't know that

How...

I'm going by the assumption that it is because it's easier for me.

@fəˈnɛtɪk It executes a program as python code if it matches a certain SHA-1 hash

I'm not a mathematician, sue me.

1:36 AM

SHA-256. And it's overwhelmingly probable that it's Turing complete.

Anyone have Notepad++?

LegionMammal978

Who defines the official names for Martian landforms? I'm pretty sure the IAU does it, but I'm not sure

@feersum On my Win10 disk, why?

@feersum Yes

I'm hoping someone can try to reproudce a bug.

I have notepad++.

1:38 AM

Shoot away, already open

Create a new file and paste the word expand_product_arg into it.

I'm afraid, because I'm rapidly growing more comfortable with VIM...

Then do replace (Ctrl+H). Choose the regex option.

@feersum Behavior seems normal

Oh there's more

Enter \B_\B as the pattern and - as the replacement

Then click replace.

Does it replace an underscore with a hyphen or not?

1:39 AM

expand-product-arg

Yep

TMK, most flavours see _ as a word boundary.

@ATaco If you got that, you didn't follow the directions because you should only click replace once :P

(I hit replace all)

betternikebot.com This bot automatically parses sites that sell Nike products and auto purchases items when they become available

1:40 AM

(It's basically instinct at this point)

like, what the heck

The bug doesn't occur for me with replace all.

Only with single replace.

So try it again.

How quaint, it doesn't replace.

It found both occurrences, but did nothing with them.

However, it works if you don't use \B's

1:42 AM

Of course it does. If find and replace never worked in genereal, someone wolud have noticed.

\b is word boundary, what's \B?

No word boundary.

Essentially matches the center of AA but not A.

So to find 'ing' at the end of a word, one can use \Bing\b

So it's a zero-width group (?=[^\s_])

Yeah.

1:45 AM

That's... actually really useful.

@ATaco Considering the _ is matched by \w (word), I'd be surprised if that were true.

I was mistaken, disregard that...

@feersum I've been testing, and it has anything to do to \B and replace one

([aeiouy]\w*)\Bing\b/$1in''s a favourite regex for me, because it replaces words like flying and running with flyin' and runnin', but preserves thing bling sing and such.

the underscore, and dash are irrelevant

1:58 AM

No clue. It was how to raise your i.q. by eating gifted children for a while, so I'd consider this an improvement. — Dennis ♦ 2 hours ago

3

I want to write a language that like its MathJax

this is almost scarier than yahoo answers' "am i prangnent"

@ATaco Isn't that called Mathematica?

As funny as that may be, it doesn't exactly support \sum^5_n=3{n}

@Riker To be fair, it's the title of a book.

2:00 AM

yea, I noticed

amazon.com/Raise-Your-Eating-Gifted-Children/dp/0595002366

dafuq

> Explore the benefits of aerobic typing.

I think I'm going to buy this book.

@ATaco Yeah, if you convert that to <name of form I've forgotten> it will format and be able to evaluate it.

Question: is allowing user to execute an arbitrary grep pattern on a directory in anyway possible to be unsafe?

(so not an arbitrary grep command)

@Riker amazon.com/Practical-Guide-Racism-C-Dalton/dp/1592404308/…

@Downgoat Grep can do find and replace iirc

2:05 AM

No, it cannot.

what about an entire grep command?

Hmm, thinking of something else.

@Downgoat grep .* && rm -rf /*

@Pavel rm command =/= grep command

@Downgoat That can go easily in an infinite loop. E.g., grep verylongstring /dev/random.

assuming the directory was already added (so only flags & options)

2:07 AM

What prevents the user from adding more?

@Downgoat grep . /dev/

grep can take a directory as an argument

Actually no I was right the first time grep [OPTIONS] PATTERN [FILE...]

Not without the -r or -R flag, no.

@Dennis hm :/ I want user to be able to run grep on cloned repo for gitgoat but obviously has to be secure

grep . /dev/zero

That terminates instantly. That's why I changed my previous message.

2:10 AM

The point is, yes you can.

Depends on how you sanitize your input. One could easily close the argument.

@Downgoat If they're just running it locally then the worst they can do is crash GitGoat

@Downgoat I'm not sure I understand. The user would run this on their own device?

@Dennis yes all local

but obviously apple will reject if unsafe

So there are no terminal emulators for Apple devices?

2:12 AM

Well is there a shell calling this?

grep $(:(){ : $@$@;};: :) is hostile enough for me.

What does that do?

Eat memory.

It looks kinda like a fork bomb but I'm not sure.

@Dennis ive tried but didn't go very well

OK, thanks for reminding me why I should stick with Android.

2:13 AM

I wish I had one ;-;

@Pavel It's single-thread, it just appends it's own arguments recursively.

grep $(:(){ : <(:);};:) is also fun. Even when run in cygwin, it eats memory so fast my Windows install is reaping apps left and right.

And it doesn't kill Cygwin..?

I killed it first. Since cygwin isn't that memory-heavy, it'll probably be killed last.

First Chrome.

No one has answered my challenge should I reveal the trick or wait until someone finds it? Perhaps this is harder than I had thought it to be.

But it sees the command you're running as memory being consumed by Cygwin, no?

2:18 AM

Eh, this kept running and running on my VPS until I remembered I'm using zram.

zram?

So? Windows still sees Chrome as the best target because killing a huge memory footprint frees up memory faster than trying to keep up with the forks.

@Pavel Compressed in-RAM swapfiles.

The closest thing to downloading more RAM there is.

@Dennis That's the same reason Windows will tolerate this for quite some time. It compresses RAM segments and page segments if memory is low.

@mınxomaτ Try it online!

Seems to run out of memory entirely

Doesn't get killed at all

2:21 AM

It hits the child process limit. That has nothing to do with running out of memory.

oic

Though Windows takes a really long time to uncompress memory >.>

.code.tio: fork: retry: Resource temporarily unavailable made me think it was a memory issue

zram seems to be really fast. It uses LZO iirc.

.code.tio: fork: retry: No child processes
.code.tio: fork: retry: No child processes
.code.tio: fork: retry: No child processes
.code.tio: fork: retry: No child processes
.code.tio: fork: retry: No child processes
Usage: grep [OPTION]... PATTERN [FILE]...
Try 'grep --help' for more information.

Real time: 15.140 s
User time: 0.012 s
Sys. time: 0.027 s
CPU share: 0.26 %
.code.tio: fork: retry: Resource temporarily unavailable
..code.tio: fork: retry: No child processes
code.tio: fork: retry: No child processes

2:23 AM

user image

At least it used every list bit of HDD space before reaping.

It weirds out completely.

Why is there output after the timing information? I've never seen TIO do that before.

Is anyone actually interested in solving my challenge? Because if its not as interesting as I think I can provide a sample algorithm to make it easier.

I thought it was fun but it might not be fun for other people.

@CartManagerXD Definitely interested, no idea how to start.

@Dennis I'm still working on it, though :P

I do still want to figure it out on my own

2:26 AM

Ok I'll leave it unspoilt.

If you think you need to, put it in a spoiler block.

@Downgoat If running user-supplied commands is deemed a security risk, every terminal emulator they let you make will inevitably be useless.

@Dennis terminal emulator is easy. I am using a lightweight linux VM

giving a weird error

Apple is annoyingly strict about running code in apps

You could use something like the TIO sandboxing.

2:27 AM

Apr 1 at 2:20, by Downgoat

> >>PANIC<< add param 'general' to bx_list_c 'bochs': list capacity exceeded

@Downgoat VM? I can use TIO is the terminal won't let me do anything.

@ATaco really? I mean usually it has to be sandboxed but otherwise they're pretty lenient

So this is interesting. If I force my PC to the memory limit (i.e. compress all compressible memory), I gain more free memory because Windows is very reluctant to uncompress any of it.

@Pavel Idea: TIO iOS app

I would totally use that, if I had an iPhone

2:29 AM

I don't like writing code from a phone anyway.

I mean there is vihan.org/p/tio which is pretty mobile friendly /selfadvertising

(Universal Windows App, anyone? No? I'll just go back to my hole.)

Gasp! Not using Ace?

@Downgoat Mobile friendly? I have a 6" screen and can only see 24x2 chars of my code at once...

great, more incentive to golf! :P

2:36 AM

Also, Denis' website is absolutely great for mobile.

;_; making into collapsible side bar then...

user image

The only issue I find with TIO on my phone is that the CopyLink buttons don't work.

user image

Both sites have the issue of "poorly sized tap targets"

Do you have a link to the tool that's scoring these..?

2:40 AM

testmysite.thinkwithgoogle.com

user image

user image

That's cheating, because it's just the homepage.

@Pavel Huh? Size tap targets appropriately has a green checkmark.

I thought I would inform everyone that I have just upped the reward on the fastest miniflak Quine, there is now a 1100 rep indefinite bounty for beating me.

Am I blind?

@Pavel that is incorrect, I am using @media tags to rescale on mobile devices. All buttons have minimum 44x44 tap target

2:43 AM

user image

I am not a webpage designer.

I should not be held accountable for dog-slow apache web server >_>

It's a node.js server.

Mine is, atleast.

Um, I also use Apache.

The speed difference is most likely syntax highlighting, vim mode, and whatnot.

apache is not problem problem is I am using apache -> nodejs -> some other weird library thing I found -> another nodejs -> static html files

I should not be allowed to route servers

So you should be held accountable

2:46 AM

ok maybe

but not directly my fault ¯\_(ツ)_/¯

My school district's website gets 64/57/61

:O 100/100 on mobile friendlyness

@Downgoat What is?

I mean I'm serving a static html page with minimal dependencies, and I'm getting 64 so likely my web server

@Pavel stories

user image

2:48 AM

:/ of course google.com is all 100s

It's either Google Tooting their own horn, or their legitimately applying the tools they've learnt.

I'm guessing it's the latter.

yeah, I think they have like a team dedicated to minifying JS to decrease load time

not sure if myth or fact but seems like a google thing to do

XD yahoo bombs the test

How to apply for Paid Codegolf?

That, and Google.com is almost a static webpage, with almost no images, powered by the most powerful server farms on the planet.

user image

That's in the Div that is the Google Logo on the homepage

2:51 AM

they know that window.lol&&lol is redundant right

Maybe not for IE4 or something silly like that.

though google even works with JS disabled so not too surprised if they support netscape

I remember writing a library that would bridge ES5 functions to work in ES3-pre netscape JS

@Downgoat Why? The first check for existence, the second executes.

user image

ouch

TIO > GMail

Yes!

2:55 AM

Not in friendliness.

Haha!

I guess the only logical next step is to launch TIO Mail.

2

I wouldn't mind the email ATaco@tio.run

Flat-out fails to analyze TNB

user165474

user image

user165474

2:58 AM

Yay my firebase-hosted webpage is ok

user165474

But it's firebase

user165474

Which is a google service now

"Wait, can I set up a node.js MAIL server?"

« first day (2277 days earlier) ← previous day next day → last day (2571 days later) »

Transcript for

Apr '1724

The Nineteenth Byte

The Nineteenth Byte: General discussion for codegolf.stackexc...

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile