p-lang thread: (now (have you (problems two)))

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > YOSPOS > p-lang thread: (now (have you (problems two)))

«‹›1784 »

FlapYoJacks: Feb 12, 2009

Python 3.10 introduced case statements. It�s the best version.

# ? Feb 22, 2022 08:38

Adbot: ADBOT LOVES YOU

# ? May 22, 2024 09:07

Cybernetic Vermin: Apr 18, 2005

for/else seems plenty useful, but in exactly the same way goto is plenty useful. overly arbitrary control flow is what i'm saying.

# ? Feb 22, 2022 10:21

cinci zoo sniper: Mar 15, 2013

DoomTrainPhD posted:

Python 3.10 introduced case statements. It�s the best version.

# ? Feb 22, 2022 11:19

eschaton: Mar 7, 2007; Don't you just hate when you wind up in a store with people who are in a socioeconomic class that is pretty obviously about two levels lower than your own?

but does it have for..case?

# ? Feb 22, 2022 11:20

cinci zoo sniper: Mar 15, 2013

eschaton posted:

but does it have for..case?

the what now

# ? Feb 22, 2022 11:21

ynohtna: Feb 16, 2007; backwoods compatible; Illegal Hen

python needs for...candles

# ? Feb 22, 2022 11:35

BobHoward: Feb 13, 2012; The only thing white people deserve is a bullet to their empty skull

oh you want a what now, do you?

so in verilog you get case statements like this:

code:

int x;
case(x)
    1: // do something
    2: // do something else
    default: // complain
endcase

i assume everyone can work out what's going on there. but guess what motherfuckers??? this here's 100% legal verilog too:

code:

int x, y;
case(1)
    x: // do something
    y: // do something else
    default: // complain
endcase

# ? Feb 22, 2022 11:42

Sweeper: Nov 29, 2007; The Joe Buck of Posting; Dinosaur Gum

BobHoward posted:

oh you want a what now, do you?

so in verilog you get case statements like this:
code:
int x;
case(x)
    1: // do something
    2: // do something else
    default: // complain
endcase
i assume everyone can work out what's going on there. but guess what motherfuckers??? this here's 100% legal verilog too:
code:
int x, y;
case(1)
    x: // do something
    y: // do something else
    default: // complain
endcase

if more than one is true does it invoke both?

# ? Feb 22, 2022 14:16

leper khan: Dec 28, 2010; Honest to god thinks Half Life 2 is a bad game. But at least he likes Monster Hunter.

BobHoward posted:

oh you want a what now, do you?

so in verilog you get case statements like this:
code:
int x;
case(x)
    1: // do something
    2: // do something else
    default: // complain
endcase
i assume everyone can work out what's going on there. but guess what motherfuckers??? this here's 100% legal verilog too:
code:
int x, y;
case(1)
    x: // do something
    y: // do something else
    default: // complain
endcase

Will it do just x or x and y if they're both 1? Lmao just please don't tell me it will only do y

# ? Feb 22, 2022 14:17

Phobeste: Apr 9, 2006; never, like, count out Touchdown Tom, man

is that synthesizable?

# ? Feb 22, 2022 14:22

TheFluff: Dec 13, 2006; FRIENDS, LISTEN TO ME
I AM A SEAGULL
OF WEALTH AND TASTE

BobHoward posted:

oh you want a what now, do you?

so in verilog you get case statements like this:
code:
int x;
case(x)
    1: // do something
    2: // do something else
    default: // complain
endcase
i assume everyone can work out what's going on there. but guess what motherfuckers??? this here's 100% legal verilog too:
code:
int x, y;
case(1)
    x: // do something
    y: // do something else
    default: // complain
endcase

maybe i'm missing something here but this doesn't seem that wacky to me? it's weird if you're used to c/c++, where the switch statement takes any expression but the case statement only takes constant expressions, but it's not that uncommon in other languages to allow any expression in either place. it's just a fancy way of writing a bunch of "if a == b" clauses so it doesn't matter which order you put the operands in.

# ? Feb 22, 2022 16:04

HappyHippo: Nov 19, 2003; Do you have an Air Miles Card?

leper khan posted:

Will it do just x or x and y if they're both 1? Lmao just please don't tell me it will only do y

Just x

Don't see why it wouldn't be synthesizable, like TheFluff points out, it's basically just a bunch of else ifs chained together.

HappyHippo fucked around with this message at 16:19 on Feb 22, 2022

# ? Feb 22, 2022 16:17

eschaton: Mar 7, 2007; Don't you just hate when you wind up in a store with people who are in a socioeconomic class that is pretty obviously about two levels lower than your own?

cinci zoo sniper posted:

the what now

the for case paradigm

# ? Feb 22, 2022 20:21

cinci zoo sniper: Mar 15, 2013

eschaton posted:

the for case paradigm

good god (i definitely can use that)

# ? Feb 22, 2022 20:23

toiletbrush: May 17, 2010

leper khan posted:

Will it do just x or x and y if they're both 1? Lmao just please don't tell me it will only do y

you can do this in Swift, and it chooses the first match

seems like the right thing to do, tbh

# ? Feb 22, 2022 20:25

eschaton: Mar 7, 2007; Don't you just hate when you wind up in a store with people who are in a socioeconomic class that is pretty obviously about two levels lower than your own?

see also Duff�s device, which is basically while�case

if you look at oldschool unix kernel, device driver, etc. code you might encounter it as a micro-optimization

# ? Feb 22, 2022 20:26

Cybernetic Vermin: Apr 18, 2005

toiletbrush posted:

you can do this in Swift, and it chooses the first match

seems like the right thing to do, tbh

in swift sure, but for a logic description language it arguably would make at least as much sense to have all matching cases "happen" at once, but i guess that is not the way verilog does things.

# ? Feb 22, 2022 20:30

BobHoward: Feb 13, 2012; The only thing white people deserve is a bullet to their empty skull

Sweeper posted:

if more than one is true does it invoke both?

IEEE 1800-2017 posted:

The case_expression shall be evaluated exactly once and before any of the case_item_expressions. The case_item_expressions shall be evaluated and then compared in the exact order in which they appear. If there is a default case_item, it is ignored during this linear search. During the linear search, if one of the case_item_expressions matches the case_expression, then the statement associated with that case_item shall be executed, and the linear search shall terminate. If all comparisons fail and the default item is given, then the default item statement shall be executed. If the default statement is not given and all of the comparisons fail, then none of the case_item statements shall be executed.

except... if you prefix the case statement with the "unique" keyword, this is an assertion that you think exactly one case item (not 0, not 2 or more) will match. this makes it safe to evaluate all branches of the case statement in parallel. this is quite useful for certain synthesizable constructs where you want to make the logic generated for each branch of the case fully independent from the rest, but it comes with some footguns. the synthesized HW is generated by assuming that you are correct to assert uniqueness, and won't do anything to protect you if multiple branches of the case statement match.

in a simulation, the simulator will throw an error if multiple matches occur when the unique case statement is evaluated, so you better make sure your sim has adequate coverage and never throws that error.

# ? Feb 22, 2022 21:35

animist: Aug 28, 2018

if i were designing a programming language i would simply evaluate all the arguments at the same time

# ? Feb 22, 2022 22:16

Destroyenator: Dec 27, 2004; Don't ask me lady, I live in beer

i feel like there's one of those formal proof languages that has the rule that if multiple branches match one is selected at random to be evaluated

my mind keeps telling me it's related to z, but i can't find an example of control flow in that

# ? Feb 22, 2022 22:21

tinaun: Jun 9, 2011; tell me...

here's a rust quiz that's related

Rust code:

fn id(n: i32) {
    const X: i32 = 1;
    let y: i32 = 2;
    
    match n {
        X => println!("that's one"),
        y => println!("that's two"),
        _ => println!("something else"),
    }
}

fn main() {
    id(1);
    id(2);
    id(3);
}

# ? Feb 22, 2022 22:43

gonadic io: Feb 16, 2011; >>=

Destroyenator posted:

i feel like there's one of those formal proof languages that has the rule that if multiple branches match one is selected at random to be evaluated

my mind keeps telling me it's related to z, but i can't find an example of control flow in that

go does this with channels

# ? Feb 22, 2022 22:51

gonadic io: Feb 16, 2011; >>=

tinaun posted:

here's a rust quiz that's related

I got this correct lol, but I'm cheating because it comes up in haskell too. It is also a rustc warning!

code:

warning: unreachable pattern
 --> src/main.rs:8:9
  |
7 |         y => println!("that's two"),
  |         - matches any value
8 |         _ => println!("something else"),
  |         ^ unreachable pattern
  |
  = note: `#[warn(unreachable_patterns)]` on by default

warning: unused variable: `y`
 --> src/main.rs:3:9
  |
3 |     let y: i32 = 2;
  |         ^ help: if this is intentional, prefix it with an underscore: `_y`
  |
  = note: `#[warn(unused_variables)]` on by default

gonadic io fucked around with this message at 22:57 on Feb 22, 2022

# ? Feb 22, 2022 22:53

rjmccall: Sep 7, 2007; no worries friend; Fun Shoe

swift doesn�t have that problem but i�m honestly not sure it was worth it because it�s sooo wordy having to write let in patterns

# ? Feb 23, 2022 01:58

Internet Janitor: May 17, 2008; "That isn't the appropriate trash receptacle."

Destroyenator posted:

i feel like there's one of those formal proof languages that has the rule that if multiple branches match one is selected at random to be evaluated

my mind keeps telling me it's related to z, but i can't find an example of control flow in that

dijkstra's "guarded command language" in a discipline of programming featured loops and conditionals which were each a set of clauses with "guard" preconditions, and they were intended to select a matching guard nondeterministically. the only difference between a conditional and a loop is that conditionals must have one matching clause, whereas loops iterate as long as there is at least one matching clause

Internet Janitor fucked around with this message at 03:38 on Feb 23, 2022

# ? Feb 23, 2022 03:35

eschaton: Mar 7, 2007; Don't you just hate when you wind up in a store with people who are in a socioeconomic class that is pretty obviously about two levels lower than your own?

rjmccall posted:

swift doesn�t have that problem but i�m honestly not sure it was worth it because it�s sooo wordy having to write let in patterns

oh no the horror you might have to use more characters to express something that someone might have to read, understand, and change 30 years from now

# ? Feb 23, 2022 05:27

eschaton: Mar 7, 2007; Don't you just hate when you wind up in a store with people who are in a socioeconomic class that is pretty obviously about two levels lower than your own?

whjust do what most MUMPS implementations do and what DCL does, and support using the minimum spelling that�s unambiguous at any point in the program!

# ? Feb 23, 2022 05:29

Kazinsal: Dec 13, 2011

eschaton posted:

whjust do what most MUMPS implementations do and what DCL does, and support using the minimum spelling that�s unambiguous at any point in the program!

also applicable to cisco boxes

sw t p v e 1,2,3

# ? Feb 23, 2022 06:25

Destroyenator: Dec 27, 2004; Don't ask me lady, I live in beer

Internet Janitor posted:

dijkstra's "guarded command language" in a discipline of programming featured loops and conditionals which were each a set of clauses with "guard" preconditions, and they were intended to select a matching guard nondeterministically. the only difference between a conditional and a loop is that conditionals must have one matching clause, whereas loops iterate as long as there is at least one matching clause

that was it, thank you!

# ? Feb 23, 2022 08:17

Cybernetic Vermin: Apr 18, 2005

eschaton posted:

oh no the horror you might have to use more characters to express something that someone might have to read, understand, and change 30 years from now

i for one add a "/* let */" in every place it is syntactically allowed in my programs, to help the people in 30 years.

# ? Feb 23, 2022 10:55

leper khan: Dec 28, 2010; Honest to god thinks Half Life 2 is a bad game. But at least he likes Monster Hunter.

Cybernetic Vermin posted:

i for one add a "/* let */" in every place it is syntactically allowed in my programs, to help the people in 30 years.

Knuth over here trying to solve illiteracy through writing.

# ? Feb 23, 2022 14:44

Captain Foo: May 11, 2004; we vibin'
we slidin'
we breathin'
we dyin'

casema balls

# ? Feb 23, 2022 15:28

Subjunctive: Sep 12, 2006; ✨sparkle and shine✨

https://github.com/Kindelia/HVM

quote:

High-order Virtual Machine (HVM) is a pure functional compile target that is lazy, non-garbage-collected and massively parallel. It is also beta-optimal, meaning that, in several cases, it can be exponentially faster than most functional runtimes, including Haskell's GHC.

That is possible due to a new model of computation, the Interaction Net, which combines the Turing Machine with the Lambda Calculus. Previous implementations of this model have been inefficient in practice, however, a recent breakthrough has drastically improved its efficiency, giving birth to the HVM. Despite being a prototype, it already beats mature compilers in many cases, and is set to scale towards uncharted levels of performance.

Welcome to the inevitable parallel, functional future of computers!

I don�t know enough about Haskell or beta optimality or higher-kinded anything to really evaluate it, but I do know that cherry-picked benchmarks can even make a compiler I write look good

is this

so restricted in its computing model that there aren�t meaningful non-toy applications
some other flavour of kookery
actually promising

# ? Feb 25, 2022 20:19

Carthag Tuek: Oct 15, 2005; Tider skal komme,
tider skal henrulle,
sl�gt skal f�lge sl�gters gang

idk that depends.... is it highly opinionated?

# ? Feb 25, 2022 20:25

CPColin: Sep 9, 2003; Big ol' smile.

I'm sigma-optimal

# ? Feb 25, 2022 20:27

rjmccall: Sep 7, 2007; no worries friend; Fun Shoe

i mean you can do a lot if you just completely assume purity. even haskell compilers have to model effects

# ? Feb 25, 2022 21:10

gonadic io: Feb 16, 2011; >>=

This came up with STM too, with haskell (and co) the only ones to be able to implement it without insane footguns of: literally any side effect in code.

# ? Feb 25, 2022 21:46

Poopernickel: Oct 28, 2005; electricity bad; Fun Shoe

sigma nutts

# ? Feb 26, 2022 00:35

Nomnom Cookie: Aug 30, 2009

Subjunctive posted:

https://github.com/Kindelia/HVM

I don�t know enough about Haskell or beta optimality or higher-kinded anything to really evaluate it, but I do know that cherry-picked benchmarks can even make a compiler I write look good

is this

so restricted in its computing model that there aren�t meaningful non-toy applications

some other flavour of kookery

actually promising

?

the author is excited about something or other that enables even higher levels of typewank than have previously been achieved in efficient computations. there will probably be a shitload of publications citing this but aside from that who knows

# ? Feb 26, 2022 03:09

Adbot: ADBOT LOVES YOU

# ? May 22, 2024 09:07

akadajet: Sep 14, 2003

Poopernickel posted:

sigma nutts

new ux people at work have decided to go from zepplin (which is good) to figma (which sucks)

so figma balls

# ? Feb 26, 2022 03:10

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > YOSPOS > p-lang thread: (now (have you (problems two)))

«‹›1784 »