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!

«‹›1503 »

Bongo Bill: Jan 17, 2012

Also, smash the state.

No, that's wrong. Smash state.

# ? Sep 8, 2017 05:46

Adbot: ADBOT LOVES YOU

# ? May 21, 2024 04:25

vOv: Feb 8, 2014

Foxfire_ posted:

code:
int foo()
{
  int x = 0;
  return (x = 1) + (x = 2);
}
code:
def bar():
  x = []
  return x.append(1) + x.append(2)
Only real difference is that C doesn't specify an order to its + operator and Python does.

The difference is that it'd be legal for the C program to format your hard drive because it's undefined behavior, not just implementation-specified, but not for the Python program to do so

# ? Sep 8, 2017 05:52

Jeb Bush 2012: Apr 4, 2007; A mathematician, like a painter or poet, is a maker of patterns. If his patterns are more permanent than theirs, it is because they are made with ideas.

any project as large as CPython probably has undefined behaviour lurking somewhere, so the interpreter might just decide to format your hard drive anyway

# ? Sep 8, 2017 06:06

Gazpacho: Jun 18, 2004; by Fluffdaddy; Slippery Tilde

Foxfire_ posted:

Only real difference is that C doesn't specify an order to its + operator and Python does.

That's what happens when you standardize a language on paper for which multiple implementations already exist, rather than one holy reference implementation.

# ? Sep 8, 2017 07:32

KernelSlanders: May 27, 2013; Rogue operating systems on occasion spread lies and rumors about me.

If you made your lists immutable none of this would be an issue.

Python code:

def bar():
  x = []
  return x.append(1) + x.append(2)

x.append(1) can return [1] and not modify x. If so, this resolves very simply to [1] + [2] or [1,2].

I don't want to be religious about the no side effects thing. Imperative programming is fine too, but constructs like this that both modify and return are trying to have it both ways and just asking for trouble. Why should x.append(1) return anything?

KernelSlanders fucked around with this message at 03:44 on Sep 9, 2017

# ? Sep 9, 2017 03:33

Suspicious Dish: Sep 24, 2011; 2020 is the year of linux on the desktop, bro; Fun Shoe

KernelSlanders posted:

Why should x.append(1) return anything?

It doesn't.

# ? Sep 9, 2017 03:35

KernelSlanders: May 27, 2013; Rogue operating systems on occasion spread lies and rumors about me.

Suspicious Dish posted:

It doesn't.

heh, that's what i get for replying to a post without evaluating it critically I guess

# ? Sep 9, 2017 03:45

Foxfire_: Nov 8, 2010

and that's what I get for writing examples while tired without running them

everybody pretend I wrote this instead please

code:

def foo():
  x = [1,2]
  return x.pop() - x.pop()

# ? Sep 9, 2017 03:59

FoiledAgain: May 6, 2007

KernelSlanders posted:

Why should x.append(1) return anything?

Suspicious Dish posted:
It doesn't.

It does return something. It returns None.

# ? Sep 9, 2017 05:39

Zemyla: Aug 6, 2008; I'll take her off your hands. Pleasure doing business with you!

This is where Haskell shows its superiority. The operations with potential side effects live in the IO monad, so they have to be explicitly sequenced, and thus cannot have an undefined or implementation-dependent evaluation order.

# ? Sep 9, 2017 07:46

Linear Zoetrope: Nov 28, 2011; A hero must cook

Zemyla posted:

This is where Haskell shows its superiority. The operations with potential side effects live in the IO monad, so they have to be explicitly sequenced, and thus cannot have an undefined or implementation-dependent evaluation order.

Comedy response: until you call FFI C code

# ? Sep 9, 2017 07:49

PhantomOfTheCopier: Aug 13, 2008; Pikabooze!

FoiledAgain posted:

It returns None.

Python None is indeed a coding horror.

# ? Sep 9, 2017 08:54

Athas: Aug 6, 2007; fuck that joker

Zemyla posted:

This is where Haskell shows its superiority. The operations with potential side effects live in the IO monad, so they have to be explicitly sequenced, and thus cannot have an undefined or implementation-dependent evaluation order.

What is the result of

code:

undefined + error "farts"

? The Haskell trick only works if you state that different ways of diverging should not be distinguished.

# ? Sep 9, 2017 11:16

VikingofRock: Aug 24, 2008

Zemyla posted:

This is where Haskell shows its superiority. The operations with potential side effects live in the IO monad, so they have to be explicitly sequenced, and thus cannot have an undefined or implementation-dependent evaluation order.

Wait, but isn't the evaluation order still really important in Haskell because of how laziness works? For example it's important that && evaluates its first argument first, so that if you evaluate false && some_expensive_predicate then some_expensive_predicate isn't evaluated.

# ? Sep 9, 2017 17:52

Athas: Aug 6, 2007; fuck that joker

VikingofRock posted:

Wait, but isn't the evaluation order still really important in Haskell because of how laziness works? For example it's important that && evaluates its first argument first, so that if you evaluate false && some_expensive_predicate then some_expensive_predicate isn't evaluated.

Since && is a library function, this is entirely up to how you do the pattern matching. For example, this definition would not be short-circuiting:

code:

True && True = True
True && False = False
False && False = False
False && True = False

But this one is, because it does not always inspect its right argument:

code:

True && True = True
False && _ = False

The issue occurs for the few primitives that are built into the language, like addition of Ints. I have been programming Haskell for eight years and I don't really know the answer to that one.

# ? Sep 9, 2017 18:34

Zemyla: Aug 6, 2008; I'll take her off your hands. Pleasure doing business with you!

Linear Zoetrope posted:

Comedy response: until you call FFI C code

Non-comedy response: FFI C code should have its signature return an IO value if it is impure.

# ? Sep 9, 2017 18:44

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

Hey, it's that time again!

Nothing quite like sprinkling an existing data interchange format with a generous helping of new features. It's irresistable!

# ? Sep 10, 2017 05:46

TooMuchAbstraction: Oct 14, 2012; I spent four years making
Waves of Steel
Hell yes I'm going to turn my avatar into an ad for it.; Fun Shoe

The lack of trailing commas in JSON is a legit complaint. Comments, unquoted keys, and the other stuff aren't nearly as irritating, and IIRC there are good reasons to not support comments (though I forget what they are).

Given that, I'm still not dumb enough to try to make a new version of a standard, especially a non-backwards-compatible one.

# ? Sep 10, 2017 05:55

KernelSlanders: May 27, 2013; Rogue operating systems on occasion spread lies and rumors about me.

Within a particular schema comments can be added trivially by simply ignoring a particular key. I'm fond of "//".

code:

{
  "my_list": [1, 2, 3],
  "my_text": "Hello, world!",
  "//": "This is a comment and will be ignored by the application, because I wrote the application to ignore it"
}

# ? Sep 10, 2017 06:10

Steve French: Sep 8, 2003

KernelSlanders posted:

Within a particular schema comments can be added trivially by simply ignoring a particular key. I'm fond of "//".
code:
{
  "my_list": [1, 2, 3],
  "my_text": "Hello, world!",
  "//": "This is a comment and will be ignored by the application, because I wrote the application to ignore it"
}

What's the use case there, though? Comments only seem helpful to me in say config files, and there I'd rather change my application to use a parser that handles some JSON variant that allows comments (we use HOCON), than have to write custom logic to handle special keys in any arbitrary object, while severely limiting the number and location of comments. Perhaps there's something else I'm missing?

# ? Sep 10, 2017 06:15

KernelSlanders: May 27, 2013; Rogue operating systems on occasion spread lies and rumors about me.

My two principle use cases are config files (although I too will use HOCON when I can) and annotating JSON unit test fixtures.

# ? Sep 10, 2017 06:47

Sedro: Dec 31, 2008

JSON is fine for what it is, and HOCON is really good for config files that people will edit. Just not YAML, anything but YAML

# ? Sep 10, 2017 07:11

NihilCredo: Jun 6, 2011; iram omni possibili modo preme:
plus una illa te diffamabit, quam multæ virtutes commendabunt

Sedro posted:

JSON is fine for what it is, and HOCON is really good for config files that people will edit. Just not YAML, anything but YAML

lol. literally the only thing that matters to me as a dev is that I'm way more likely to have a Yaml deserialization library available than a hocon one

meanwhile as a user, hocon still has braces and commas and poo poo, so I'd much rather edit yaml, tyvm. if you're worried i might be an rear end in a top hat and do the weird corner case stuff you can find in the yaml spec, then toml is an acceptable substitute

# ? Sep 10, 2017 09:20

PhantomOfTheCopier: Aug 13, 2008; Pikabooze!

Durr

code:

{
  "key that matters A":  "value",
  "key that matters B":  "value",
  "comment with historical relevance":  "yay I'm ignored because why would anyone use this key?",
  "LOAD ME!":  "oh darn no one ever will"
}

Yay, comments are free. But the trailing comma thing, yeah, that's beyond ridiculous. Any language that can't parse ,} as a non-element on input, this is why we see leading commas.

And YAML... I've had to work on an application where someone decided to use it as a configuration format. It's a complete pain. The spacing is just about as ridiculous as Python's inability to understand program grouping, and it's really cute when the YAML parses beautifully and successfully until it's passed to Python where there's a bunch of parse errors because PyYAML gets confused (at least I seem to recall that's the module that was being used). Then one day I was using abbreviations only to find out that the app designer choose YAML despite the transport mechanism being SOAP (there's another horror story, I suppose), and it doesn't pass data unmodified because it's XML. Woohoo, let's blast all the ampersands! (on the way in but not on the way out)

Myself, I'm still partial to ASCII 0x1c--0x1f.

PhantomOfTheCopier fucked around with this message at 10:08 on Sep 10, 2017

# ? Sep 10, 2017 10:04

SupSuper: Apr 8, 2009; At the Heart of the city is an Alien horror, so vile and so powerful that not even death can claim it.

We should start classifying data formats as "human-writable" instead of "human-readable". Because every plain-text data format is human-readable, but users will gently caress up anything that's not as basic as an INI file.

# ? Sep 10, 2017 15:11

Vanadium: Jan 8, 2005

I wish all my config files were .nix!

Stop the .toml madness!

# ? Sep 10, 2017 15:22

pseudorandom name: May 6, 2007

XML isn't human readable. :colbert:

# ? Sep 10, 2017 16:09

Pollyanna: Mar 5, 2005; Milk's on them.

Ranzear posted:

There was probably some debug logging stuffed in there at one point.

You could just do this in that case:

code:

            if (responseJSON == null)  // assuming you're checking specifically for the null case and not for just plain falsiness
            {
                console.log("gently caress")
            }

            return responseJSON;

# ? Sep 10, 2017 16:44

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

pseudorandom name posted:

XML isn't human readable.

<butt size="fat" scent="foul"/> ? Unreadable garbage fit only for robots. But {"butt":{"size":"fat","scent":"foul"}} is good and very human friendly.

# ? Sep 10, 2017 17:03

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.

Internet Janitor posted:

Hey, it's that time again!

Nothing quite like sprinkling an existing data interchange format with a generous helping of new features. It's irresistable!

Man, I saw that first sentence and immediately thought "oh gently caress," but I actually did cook up something that works exactly like the handlers parameter in JSON plus for a project a few years ago, I guess I should've made it a standalone library.

# ? Sep 10, 2017 17:18

Ellie Crabcakes: Feb 1, 2008; Stop emailing my boyfriend Gay Crungus

TooMuchAbstraction posted:

The lack of trailing commas in JSON is a legit complaint. Comments, unquoted keys, and the other stuff aren't nearly as irritating, and IIRC there are good reasons to not support comments (though I forget what they are)

Because it's an interchange format for relatively simple data structures. Using it as a primary config format or concatenating it is silly and will bite you in the rear end. There are even better interchange formats.

Except YAML. YAML sucks balls.

# ? Sep 10, 2017 22:22

Doc Hawkins: Jun 15, 2010; Dashing? But I'm not even moving!

Vanadium posted:

I wish all my config files were .nix!

Stop the .toml madness!

You may want to look into dhall, a typed, functional, total language with a toNix converter. Actually, I think I've heard it grew from a project of adding types to nix?

I'm still not sure 100% if this is a horror, but weirdly, it can also import expressions by url, and its prelude is hosted on ipfs.

code:

$ dhall
let not = [url]https://ipfs.io/ipfs/QmQ8w5PLcsNz56dMvRtq54vbuPe9cNnCCUXAQp6xLc6Ccx/Prelude/Bool/not[/url]
in  not True
<Ctrl-D>
Bool

False

e: pretend those auto-added [url] tags are not there

# ? Sep 10, 2017 22:54

QuarkJets: Sep 8, 2008

always define your own human-readable data containers and don't tell anyone else how they work

# ? Sep 10, 2017 22:55

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

QuarkJets posted:

always define your own human-readable data containers and don't tell anyone else how they work

Adjust the value of "human" as appropriate

# ? Sep 10, 2017 23:06

QuarkJets: Sep 8, 2008

I usually encode data so that it's readable by dogs

# ? Sep 10, 2017 23:10

PhantomOfTheCopier: Aug 13, 2008; Pikabooze!

QuarkJets posted:

always define your own human-readable data containers and don't tell anyone else how they work

There is nothing better than Data:: Dumper. If it's not quite clean enough for you, I recommend Acme::Bleach.

# ? Sep 10, 2017 23:20

Ellie Crabcakes: Feb 1, 2008; Stop emailing my boyfriend Gay Crungus

PhantomOfTheCopier posted:

There is nothing better than Data:: Dumper.

quote:

If it's not quite clean enough for you, I recommend Acme::Bleach.

Now we're talking.

# ? Sep 10, 2017 23:46

KernelSlanders: May 27, 2013; Rogue operating systems on occasion spread lies and rumors about me.

Maybe the constraints on data interchange formats and config file formats are sufficiently different that they deserve two different file formats?

:shrug:

# ? Sep 10, 2017 23:55

Absurd Alhazred: Mar 27, 2010; by Athanatos

KernelSlanders posted:

Maybe the constraints on data interchange formats and config file formats are sufficiently different that they deserve two different file formats?

Config files are just serialized behavior, though. :shrug:

# ? Sep 11, 2017 01:05

Adbot: ADBOT LOVES YOU

# ? May 21, 2024 04:25

Steve French: Sep 8, 2003

Absurd Alhazred posted:

Config files are just serialized behavior, though.

So are files containing code, both as written by humans and compiled code.

So let's just use JSON for everything!

# ? Sep 11, 2017 01:10

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!

«‹›1503 »