Coding Horrors: You can gather all your technical debt into one easy framework!

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > The Cavern of COBOL > Coding Horrors: You can gather all your technical debt into one easy framework!

«‹›1501 »

Volmarias: Dec 31, 2002; EMAIL... THE INTERNET... SEARCH ENGINES...

dc3k posted:

everything in linux is a file
everything in javascript is a god drat nightmare

Removed the spoilers, it's not a surprise to anyone at this point

# ? Apr 4, 2024 08:34

Adbot: ADBOT LOVES YOU

# ? Apr 19, 2024 04:33

Breetai: Nov 6, 2005; 🥄Mah spoon is too big!🍌

Apologies for not contributing earlier.

code:


Select
num
, case when num/2 = int(num/2) then 'true' 
  else 'false'  
  end
  as iseven

from lnd.prd.db.allnums

Order by num

# ? Apr 4, 2024 14:58

ultrafilter: Aug 23, 2007; It's okay if you have any questions.

https://twitter.com/the_moisrex/status/1776661987855941776

# ? Apr 6, 2024 23:34

Nolgthorn: Jan 30, 2001; The pendulum of the mind alternates between sense and nonsense

GABA ghoul posted:

I think there is a very clever solution for this that could save a lot of work
code:
if (number < 0)
 number = int.Parse(number.ToString().Replace('-', ''))

I want to know if I can get this through code review, don't tempt me Satan.

# ? Apr 8, 2024 16:02

RPATDO_LAMD: Mar 22, 2013; 🐘🪠🍆

what yall really need is some solid mathematical background for your IsEvens

Python code:

def isEven(x):
    return not isRelativelyCoprime(x, 2)
    
def isRelativelyCoprime(a, b):
    if b > a:
        a, b = b, a
    # if two integers a and b are coprime to each other, then
    # for all integers i and j,
    # b*i % a == b*j % a implies i % a == j % a.
    for i in range(0, a):
        for j in range(i + 1, a):
            if (i * b) % a == (j * b) % a:
                if i % a != j % a:
                    return False # found a counterxample!
    return True

it uses modulos, and it only takes quadratic time to run!

# ? Apr 8, 2024 21:47

Volmarias: Dec 31, 2002; EMAIL... THE INTERNET... SEARCH ENGINES...

RPATDO_LAMD posted:

what yall really need is some solid mathematical background for your IsEvens

Python code:

def isEven(x):
    return not isRelativelyCoprime(x, 2)
    
def isRelativelyCoprime(a, b):
    if b > a:
        a, b = b, a
    # if two integers a and b are coprime to each other, then
    # for all integers i and j,
    # b*i % a == b*j % a implies i % a == j % a.
    for i in range(0, a):
        for j in range(i + 1, a):
            if (i * b) % a == (j * b) % a:
                if i % a != j % a:
                    return False # found a counterxample!
    return True

it uses modulos, and it only takes quadratic time to run!

My eyes just rolled into the back of my head

# ? Apr 9, 2024 07:38

Foxfire_: Nov 8, 2010

ultrafilter posted:

I don't think I hate this that much? it's icky, but it's also plain what it does and how it works.

(I do hate that resume() returns 0 on both the first iteration and after it's finished)

# ? Apr 9, 2024 20:43

Itaipava: Jun 24, 2007

In my head canon Duff's device was named after the Simpsons beer

# ? Apr 9, 2024 22:06

QuarkJets: Sep 8, 2008

Foxfire_ posted:

I don't think I hate this that much? it's icky, but it's also plain what it does and how it works.

(I do hate that resume() returns 0 on both the first iteration and after it's finished)

The post of a broken man

# ? Apr 10, 2024 02:40

Bruegels Fuckbooks: Sep 14, 2004; Now, listen - I know the two of you are very different from each other in a lot of ways, but you have to understand that as far as Grandpa's concerned, you're both pieces of shit! Yeah. I can prove it mathematically.

Foxfire_ posted:

I don't think I hate this that much? it's icky, but it's also plain what it does and how it works.

(I do hate that resume() returns 0 on both the first iteration and after it's finished)

One of the reasons I like the switch statement is how it breaks down into assembly pretty easily...

e.g.

code:

int x = // some value;

switch (x) {
    case 1:
        // Code for case 1
        break;
    case 2:
        // Code for case 2
        break;
    case 3:
        // Code for case 3
        break;
    default:
        // Code for default case
        break;
}

turns into something like

code:

; Assume 'x' is stored in the register eax
    mov eax, [x]      ; Move the value of x into eax
    cmp eax, 1        ; Compare eax with 1
    je case1          ; Jump to case1 if eax is 1
    cmp eax, 2        ; Compare eax with 2
    je case2          ; Jump to case2 if eax is 2
    cmp eax, 3        ; Compare eax with 3
    je case3          ; Jump to case3 if eax is 3
    jmp default_case  ; If none of the above, jump to default case

case1:
    ; Code for case 1
    jmp end_switch

case2:
    ; Code for case 2
    jmp end_switch

case3:
    ; Code for case 3
    jmp end_switch

default_case:
    ; Code for default case

end_switch:
    ; Continuation of code after switch

Pretty logical.

Using that construct in the screenshot is going to result in assembly language that looks like this:

code:

; Pseudo-assembly representation of the 'resume' function in the range_struct

; Assume 'this' pointer is passed in register ecx

resume:
    mov eax, [ecx + state]   ; Load the current state into eax
    cmp eax, At_start
    je  at_start
    cmp eax, In_loop
    je  in_loop
    cmp eax, Done
    je  done

at_start:
    mov dword [ecx + i], 0   ; i = 0
    mov eax, [ecx + stop]    ; Load stop value into eax
    cmp [ecx + i], eax       ; Compare i and stop
    jge done_label           ; If i >= stop, jump to done
    mov [ecx + state], In_loop   ; state = In_loop
    mov eax, [ecx + i]       ; Return i
    ret

in_loop:
    add [ecx + i], [ecx + step]  ; i += step
    mov eax, [ecx + stop]        ; Load stop value into eax
    cmp [ecx + i], eax           ; Compare i and stop
    jge done_label               ; If i >= stop, jump to done
    mov eax, [ecx + i]           ; Return i
    ret

done_label:
    mov [ecx + state], Done   ; state = Done

done:
    xor eax, eax              ; Return 0
    ret

So really gross for no good reason.

# ? Apr 10, 2024 16:40

Volte: Oct 4, 2004; woosh woosh

I don't think there's anything that gross about it that's not also gross about any switch statement that uses fallthrough logic. The fact that it slices up the for-loop construct is kind of viscerally offputting but starting partway into the for loop isn't really conceptually any different than a do-while loop. It's what I would mildly classify as spaghetti code and not something I would ever do in general, but it's pretty clear what it does and how it works.

# ? Apr 10, 2024 17:47

Soricidus: Oct 21, 2010; freedom-hating statist shill

The problem with it is that having a case statement in the middle of a loop is the kind of thing that might plausibly be undefined behavior or a compiler-specific extension. It�s obvious what the code intends to do, but it�s not a common idiom, so unless you�re familiar with the trick, it�s not obvious whether it will actually do that reliably, even if you run it and it seems to work on your machine.

That�s why it�s bad code.

# ? Apr 10, 2024 19:22

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

Soricidus posted:

The problem with it is that having a case statement in the middle of a loop is the kind of thing that might plausibly be undefined behavior or a compiler-specific extension. It’s obvious what the code intends to do, but it’s not a common idiom, so unless you’re familiar with the trick, it’s not obvious whether it will actually do that reliably, even if you run it and it seems to work on your machine.

That’s why it’s bad code.

what.

duff's device is a little silly, but it's not against the spec

# ? Apr 10, 2024 19:47

Hammerite: Mar 9, 2007; And you don't remember what I said here, either, but it was pompous and stupid.; Jade Ear Joe

Soricidus posted:

The problem with it is that having a case statement in the middle of a loop is the kind of thing that might plausibly be undefined behavior or a compiler-specific extension. It�s obvious what the code intends to do, but it�s not a common idiom, so unless you�re familiar with the trick, it�s not obvious whether it will actually do that reliably, even if you run it and it seems to work on your machine.

That�s why it�s bad code.

I was going to say that it should be ok because no initialization of variables was being skipped, but actually it potentially is if I understand correctly.

The initialization of i to 0 occurs in the for loop header and it ought to run because the state is initialized to "at start" - but having said that, all the properties are public, so the user could construct one of these, then change the state and then call resume() to get it to run without i being initialized? But you could fix that by just initializing i to 0 along with all the other members so I don't know why the author didn't just do that.

Even if the enum member was (say) made private so that you couldn't do that, it seems like it's asking a lot of the compiler to prove that i isn't accessed before it's initialized? You would think the compiler would warn about it.

I guess in C++ both the "switch" and "for loop" language features translate pretty directly to labels and gotos, and it's clear how you convert that code into labels and gotos and it shouldn't cause any problems, so other than the uninitialized variable thing, I don't know how plausible it really is that it would be undefined.

It's a while since I touched C++.

# ? Apr 10, 2024 20:14

Zopotantor: Feb 24, 2013; ...und ist er drin dann lassen wir ihn niemals wieder raus...

leper khan posted:

what.

duff's device is a little silly, but it's not against the spec

This is not really Duff's device though; it�s an implementation of a state machine or coroutine. In Duff's device the loop actually has a function, but here it would not really be necessary if the author had properly initialised "i" in the constructor.

efb

# ? Apr 10, 2024 20:17

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

Zopotantor posted:

This is not really Duff's device though; it’s an implementation of a state machine or coroutine. In Duff's device the loop actually has a function, but here it would not really be necessary if the author had properly initialised "i" in the constructor.

efb

case statements in a loop have no issues

# ? Apr 10, 2024 20:55

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

switch in C is unstructured control flow. The nested statement is just an ordinary statement block, and the associated cases can be recursively anywhere inside that block except (of course) within another switch; the switch jumps to the case, entering any scopes along the way and bypassing any earlier code in that scope. (In C++, bypassing initialized declarations is illegal, but otherwise it�s the same.) The only thing special about the switch statement block is that it�s the only statement block in the language that�s never possible to enter from the top.

for is structured control flow; there are no branches into arbitrary scopes. You just keep entering the loop body from the top as long as the condition expression evaluates to true.

You can desugar a for loop into labels and goto, but at some level that�s true of any local non-conditional control flow. You could desugar if statements into a more primitive �if � goto� instead of using nested blocks, and that would indeed look a lot more like assembly language; that doesn�t mean if isn�t a structured feature, it just means it has definable semantics.

rjmccall fucked around with this message at 21:02 on Apr 10, 2024

# ? Apr 10, 2024 20:59

RPATDO_LAMD: Mar 22, 2013; 🐘🪠🍆

Hammerite posted:

I was going to say that it should be ok because no initialization of variables was being skipped, but actually it potentially is if I understand correctly.

The initialization of i to 0 occurs in the for loop header and it ought to run because the state is initialized to "at start" - but having said that, all the properties are public, so the user could construct one of these, then change the state and then call resume() to get it to run without i being initialized? But you could fix that by just initializing i to 0 along with all the other members so I don't know why the author didn't just do that.

Even if the enum member was (say) made private so that you couldn't do that, it seems like it's asking a lot of the compiler to prove that i isn't accessed before it's initialized? You would think the compiler would warn about it.

I guess in C++ both the "switch" and "for loop" language features translate pretty directly to labels and gotos, and it's clear how you convert that code into labels and gotos and it shouldn't cause any problems, so other than the uninitialized variable thing, I don't know how plausible it really is that it would be undefined.

It's a while since I touched C++.

"this is bad because if you gently caress with internal object state you can put it in an invalid state" is a very "you're holding it wrong" kind of problem to have

the real problem is that it's using the switch as a fancy goto which is just fucky and unintuitive to read.
you could easily refactor this to use separate start(), loop(), done() functions instead of jumping into the middle of a block of shared code.

# ? Apr 10, 2024 21:40

Soricidus: Oct 21, 2010; freedom-hating statist shill

leper khan posted:

what.

duff's device is a little silly, but it's not against the spec

I didn�t say it was against the spec. I know it�s not against the spec. I said it was the kind of thing that looks like it might be against the spec, and is bad code for that reason.

leper khan posted:

case statements in a loop have no issues

Coding Horrors: case statements in a loop have no issues

# ? Apr 10, 2024 22:21

Hammerite: Mar 9, 2007; And you don't remember what I said here, either, but it was pompous and stupid.; Jade Ear Joe

RPATDO_LAMD posted:

"this is bad because if you gently caress with internal object state you can put it in an invalid state" is a very "you're holding it wrong" kind of problem to have

It's not internal object state though, it's public. The language gives you the tools to make it internal, but the author of this code neglected to do that.

It's not like Python where everything is based on conventions and promises. If something is left public in C++ then the user is entitled to suppose that that means you intended for them to be able to modify it.

# ? Apr 10, 2024 22:31

JawnV6: Jul 4, 2004; So hot ...

Bruegels Fuckbooks posted:

So really gross for no good reason.

well yeah but you chose x86 asm yourself

# ? Apr 10, 2024 22:34

pokeyman: Nov 26, 2006; That elephant ate my entire platoon.

Soricidus posted:

it was the kind of thing that looks like it might be against the spec, and is bad code for that reason.

This is me writing any line of C++. "But does this really work the way I think it does???"

# ? Apr 10, 2024 22:56

Jen heir rick: Aug 4, 2004; when a woman says something's not funny, you better not laugh your ass off

pokeyman posted:

This is me writing any line of C++. "But does this really work the way I think it does???"

# ? Apr 10, 2024 23:14

Breetai: Nov 6, 2005; 🥄Mah spoon is too big!🍌

Our org uses SAS programs for regular scheduled reporting, and our data warehouse is a Teradata SQL environment that we can connect to using SAS EG. Generally speaking our programs are set up to to a whole bunch of data manipulation in Teradata, then bring refined datasets into the SAS grid in order to perform a few final transformations and then create the report outputs.

However, occasionally I will get a 'can you see why this program is running poorly' maintenance request and well...

So someone creates a SAS view via a connection to Teradata. Something like this:

code:

Proc sql;
<Teradata connection credentials>
Create view SASVIEW1 as select * from connection to teradata
(
Select <variables>
from <teradata table 1>

left join

	(select <variables>
	from <teradata table 2>

		left join

		(select <variables>
		from <teradata table 3>


			left join

			(select <variables>
			from <teradata table4>

on.....

)

And then they make HALF A DOZEN SAS views just like this. And then this happens:

code:

Proc sql;
<Teradata connection credentials>
crate table work.FINAL as
select * from connection to Teradata(

select * from SASVIEW1

union all

select * from SASVIEW2

union all

select * from SASVIEW3

union all

select * from SASVIEW4

union all

select * from SASVIEW5

union all

select * from SASVIEW6);

"Whycome my code run slow?" :gonk:

# ? Apr 10, 2024 23:35

Adbot: ADBOT LOVES YOU

# ? Apr 19, 2024 04:33

ExcessBLarg!: Sep 1, 2001

I wrote a for-case loop the other day for a project. At first I felt pretty bad about it but it's way better than these case-for shenanigans.

# ? Apr 11, 2024 00:56

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > The Cavern of COBOL > Coding Horrors: You can gather all your technical debt into one easy framework!

«‹›1501 »