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 »

Vanadium: Jan 8, 2005

Plorkyeran posted:

It's not very hard to write a generic RAII container for c api things:

C++ code:

template<class T, class Del = void(*)(T)>
class scoped_holder {
	T value;
	Del destructor;

	scoped_holder(scoped_holder const&);
	scoped_holder& operator=(scoped_holder const&);
public:
	operator T() const { return value; }
	T operator->() const { return value; }

	scoped_holder(T value, Del destructor)
	: value(value)
	, destructor(destructor)
	{
	}

	~scoped_holder() { if (value) destructor(value); }
};

...

{
  scoped_holder<api_handle_type> handle(api_alloc_a_thing(), api_delete_a_thing);
  if (!handle)
    return;

  api_do_stuff_with_a_thing(handle);
}

I never get the type conversion operators right :saddowns:

std::unique_ptr seems fairly reasonable for most such things, though. I'd be happier if I could squeeze the api_delete_a_thing in there at the type level so the call would be, like, inlined and everything, but that's probably just an imaginary concern.

# ? Jun 20, 2012 18:37

Adbot: ADBOT LOVES YOU

# ? May 21, 2024 00:18

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

That was meant to be C code. Unfortunately, there's no syntax highlighting for plain C. And yes, I know about the multiple labels trick, but I chose to keep it simple on purpose. And at least people aren't linking Djikstra on my rear end and calling me human scum this time.

# ? Jun 20, 2012 18:41

Johnny Cache Hit: Oct 17, 2011

Suspicious Dish posted:

I'm curious how people feel about the goto statement as an approach to this.

I'm pretty sure this is the technique Rochkind uses in Advanced UNIX Programming. I think he wraps a macro or two around it, though...

It really is the only decent approach I've ever seen.

# ? Jun 20, 2012 19:33

raminasi: Jan 25, 2005; a last drink with no ice

Lumpy posted:

So I had to poke around in a homegrown CMS system that a client's old web dev shop wrote ~6 years ago. It is full of horrors, but the README file takes the cake:
code:
THIS IS OUT OF DATE: UPDATE WITH INFO ABOUT HOW THIS WORKS!

This directory holds files that will eventually become actual documents on
a website: the editor scripts, etc.

That README is still leagues better than a straight-up incorrect one.

# ? Jun 20, 2012 19:36

tef: May 30, 2004; -> some l-system crap ->

Suspicious Dish posted:

I'm curious how people feel about the goto statement as an approach to this.
C++ code:
  request = asprintf ("HTTP 1.1\nGET %s", url->path);
_{Yes, I know, it's an "HTTP client".}

HTTP/1.1 sends \r\n as line separators, and has a Host: header :argh:

# ? Jun 20, 2012 19:45

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

Since you won't take "it was dumb example code" for an answer, how about "it was wrong on purpose so that a copy/paste programmer wouldn't have any chance of making it work, promoting proper software engineering practices instead".

Yeah, that's it.

# ? Jun 20, 2012 21:00

pseudorandom name: May 6, 2007

tef posted:

HTTP/1.1 sends \r\n as line separators, and has a Host: header

If it makes you feel any better, it won't get that far because the check after the call to my_socket_open() is inverted.

# ? Jun 20, 2012 21:44

McGlockenshire: Dec 16, 2005; GOLLOCKS!

No Safe Word posted:

Yeah though I mean 99% of Stack Overflow is a horror anyway, those tags just ensure the deck is fully stacked.

At least those of us that hang out in the PHP chat room are competent. We've formed a little mob of sanity enforcers and copy/paste "DON'T DO THAT" commenters to help spread best practice.

It's sad that such things are necessary.

# ? Jun 20, 2012 22:37

Lysidas: Jul 26, 2002; John Diefenbaker is a madman who thinks he's John Diefenbaker.; Pillbug

Suspicious Dish posted:

And at least people aren't linking Djikstra on my rear end and calling me human scum this time.

The core Linux kernel developers are extremely competent C programmers, and their views on this subject are pretty informative.

# ? Jun 20, 2012 22:38

geonetix: Mar 6, 2011

Lysidas posted:

The core Linux kernel developers are extremely competent C programmers, and their views on this subject are pretty informative.

I saw this excerpt in that conversation and felt it needed some attention:

quote:

> I thought Edsger Dijkstra coined the "gotos are evil" bit in his
> structured programming push?

Yeah, he did, but he's dead, and we shouldn't talk ill of the dead. So
these days I can only rant about Niklaus Wirth,

Oh Linus, never stop being a horrible dick.

# ? Jun 20, 2012 22:44

csammis: Aug 26, 2003; Mental Institution

geonetix posted:

I saw this excerpt in that conversation and felt it needed some attention:

Oh Linus, never stop being a horrible dick.

My new favorite image

# ? Jun 20, 2012 22:48

ohgodwhat: Aug 6, 2005

"Hey guys, I'm not very good at C, but you're doing it all wrong, could you change it to be more like what I learned about in my introductory programming class?"

These people :psyduck:

# ? Jun 20, 2012 22:49

Plorkyeran: Mar 22, 2007; To Escape The Shackles Of The Old Forums, We Must Reject The Tribal Negativity He Endorsed

basically gently caress people who talk about something despite not knowing anything about it beyond the name of a famous paper

# ? Jun 20, 2012 23:08

Zamujasa: Oct 27, 2010; Bread Liar

Haha. Not a coding horror in itself, but I just realized that the UPS we have here at work is apparently powering 4 outlet strips, three of which make up a huge daisy-chain. It's really starting to reflect the way coding is done at this place.

Basically:

code:

UPS
  |
  +-Power strip 1
  |  |-3 unknown items
  |  `-Power strip 3
  |     |-2 unknown items
  |     `-Power strip 4
  |        `-Several screens and two computers
  |
  +-Power strip 2
  |  `-6 unknown items
  |
  `-A cord that goes behind some shelves

Somehow I do not think this UPS will be very useful in the event of a power outage.

E: Oh, they managed to plug the microwave into the UPS at one point a few months ago too. That was all sorts of fun!

Zamujasa fucked around with this message at 23:22 on Jun 20, 2012

# ? Jun 20, 2012 23:14

Safe and Secure!: Jun 14, 2008; OFFICIAL SA THREAD RUINER
SPRING 2013

That was actually really neat. I had just accepted that gotos are bad, but after reading that, I've changed my mind. The arguments against goto are pretty similar to the arguments I've seen against including pass-by-reference and operator-overloading in Java.

Safe and Secure! fucked around with this message at 23:36 on Jun 20, 2012

# ? Jun 20, 2012 23:30

PrBacterio: Jul 19, 2000

Safe and Secure! posted:

That was actually really neat. I had just accepted that gotos are bad, but after reading that, I've changed my mind. The arguments against goto are pretty similar to the arguments I've seen against including pass-by-reference and operator-overloading in Java.

Please don't do that. Yes, goto's aren't in and of themselves evil, and sometimes it's better to use a well-placed goto in place of some more complicated circumscription where it makes for cleaner and more readable code, but nobody should take that as a license to pepper their code with gotos all over the place. In practice, using a goto statement is a bad idea *most of the time*, which is why the rule exists and people are usually taught never to use them, because that is *generally speaking* correct. If you come across one of the exceptions to this rule, where your code would be cleaner when using a goto, you'll know and should probably just go ahead and do it anyway. But knowing when it is OK to break such a rule, which is so commonly taught for a very good reason, first requires you to understand why it is in place in the first place and thus, adhere to it in most cases anyway.

# ? Jun 20, 2012 23:58

N.Z.'s Champion: Jun 8, 2003; Yam Slacker

code:

.nowrap { display: inline-block; }

# ? Jun 21, 2012 00:19

No Safe Word: Feb 26, 2005

PrBacterio posted:

Please don't do that. Yes, goto's aren't in and of themselves evil, and sometimes it's better to use a well-placed goto in place of some more complicated circumscription where it makes for cleaner and more readable code, but nobody should take that as a license to pepper their code with gotos all over the place. In practice, using a goto statement is a bad idea *most of the time*, which is why the rule exists and people are usually taught never to use them, because that is *generally speaking* correct. If you come across one of the exceptions to this rule, where your code would be cleaner when using a goto, you'll know and should probably just go ahead and do it anyway. But knowing when it is OK to break such a rule, which is so commonly taught for a very good reason, first requires you to understand why it is in place in the first place and thus, adhere to it in most cases anyway.

Pretty much. Really when it's appropriate to use a goto you should be at the point of already asking yourself, "Really? Really?? Is this one of those 0.01% of cases? Well goddamn..."

# ? Jun 21, 2012 00:24

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

LOOK I AM A TURTLE posted:

God drat do I hate "single point of exit" as a design philosophy. It's the cause of conditional blocks the size of skyscrapers that are impossible to parse when all you need is if (gently caress == null) return false;

Granted, this only happens if the code is also full of massive functions. It's probably okay to use single point of exit if you manage to keep functions small, and it's probably (somewhat) okay to make a lot of long functions if you can make them readable, but combine single point of exit with behemoth functions and you have planted the seeds of despair.

But of course when the functions are small single-point-of-exit is pointless because you can easily scan the whole function for returns.

I have no use for single-point-of-exit as a design philosophy. If I'm going to use it for clean-up or whatever it's because that's the best way to write that particular function not because it's a good approach to programming in general.

# ? Jun 21, 2012 00:36

Dicky B: Mar 23, 2004

"Blanket rules" in programming are the worst thing.

# ? Jun 21, 2012 01:16

teamdest: Jul 1, 2007

Dicky B posted:

"Blanket rules" in programming are the worst thing.

You have a funny way of spelling "PHP".

# ? Jun 21, 2012 01:19

Zhentar: Sep 28, 2003; Brilliant Master Genius

Dicky B posted:

"Blanket rules" in programming are the worst thing.

Spoken like someone who hasn't had a co-worker try to sneak a for loop implemented with goto through a code review.

# ? Jun 21, 2012 01:27

Doctor w-rw-rw-: Jun 24, 2008

teamdest posted:

You have a funny way of spelling "PHP".

This joke is old and tired. Please stop.

PHP doesn't cause bad code (though it certainly helps), people do. PHP is a very practical and productive tool, and in the hands of a capable programmer, can avoid a lot of badness. I dislike almost everything about PHP, but since I have the privilege of working amongst good programmers, the PHP code I see at work is not crap. Good PHP is still going to be better than bad Python or bad Scala or what-have-you.

# ? Jun 21, 2012 01:42

trex eaterofcadrs: Jun 17, 2005; My lack of understanding is only exceeded by my lack of concern.

Zhentar posted:

Spoken like someone who hasn't had a co-worker try to sneak a for loop implemented with goto through a code review.

Idiot coworkers aren't a reason to abandon useful tools wholesale.

# ? Jun 21, 2012 01:45

xf86enodev: Mar 27, 2010; dis catte!

Suspicious Dish posted:

I'm curious how people feel about the goto statement as an approach to this.
C++ code:
char *
grab_http_response(char *string_url)
{
  char *response = NULL;
  URL *url = NULL;
  my_socket *sock = NULL;
  char *request;

  url = parse_url (string_url, url);
  if (url == NULL)
    goto out;

  sock = my_socket_open (url->host, url->port);
  if (sock != NULL)
    goto out;

  request = asprintf ("HTTP 1.1\nGET %s", url->path);
  my_socket_send (sock, request);

  response = my_socket_recvall (sock);

 out:
  if (sock != NULL)
    my_socket_close (sock);
  if (url != NULL)
    free (url);

  return response;
}
_{Yes, I know, it's an "HTTP client". The code is probably incorrect, too. It was the simplest example that I think everybody would understand that required multiple resources to be closed in an error path.}

If this were C I'd completely agree with the goto usage. OTOH, I don't know C++ that well, but wouldn't it be possible to wrap the response in an object and have the clean-up done be its destructor? ~*patterns*~

In Python you'd do it something similar. You'd have with response(butt) as poop: where response has its clean-up defined somewhat like a finally iirc.

edit: I guess what I'm getting at is why wouldn't you use exceptions if this is C++?

edit2: vvv

Suspicious Dish posted:

It only says "C++ code" because there's no C highlighter.

gently caress C++ and gently caress radium for his inferior syntax highlighting

xf86enodev fucked around with this message at 03:59 on Jun 21, 2012

# ? Jun 21, 2012 02:04

shrughes: Oct 11, 2008; (call/cc call/cc)

quote:

I guess what I'm getting at is why wouldn't you use exceptions if this is C++?

Because they're bad and dangerous and unclear.

Other reasons may include legacy systems or legacy compilers that don't support them.

A good reason is because your coworkers don't know how to write exception-safe code. Probably you don't either, assuming you're a randomly distributed C++ developer.

# ? Jun 21, 2012 02:34

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

xf86enodev posted:

If this were C

It is. It only says "C++ code" because there's no C highlighter.

# ? Jun 21, 2012 02:40

Vanadium: Jan 8, 2005

Suspicious Dish posted:

It is. It only says "C++ code" because there's no C highlighter.

We should get Radium to make it say C/C++ and drive all the C nerds and the C++ nerds into a rage.

# ? Jun 21, 2012 02:43

xf86enodev: Mar 27, 2010; dis catte!

Vanadium posted:

We should get Radium to make it say C/C++ and drive all the C nerds and the C++ nerds into a rage.

Posting C/C++ code should just crash your browser.
Posting Java would be fine though. But it will crash browsers trying to look at the post unless they have signed up to Oracle's Open Post Streaming Intellisense (Enterprise Edition).

# ? Jun 21, 2012 04:26

LOOK I AM A TURTLE: May 22, 2003; "I'm actually a tortoise."; Grimey Drawer

xf86enodev posted:

Posting C/C++ code should just crash your browser.
Posting Java would be fine though. But it will crash browsers trying to look at the post unless they have signed up to Oracle's Open Post Streaming Intellisense (Enterprise Edition).

In the future, probations will be automatically handed out for posts that don't follow PEP8.

# ? Jun 21, 2012 06:17

Zhentar: Sep 28, 2003; Brilliant Master Genius

Suspicious Dish posted:

I'm curious how people feel about the goto statement as an approach to this.

And what's wrong with doing this instead?

C++ code:

char *
grab_http_response(char *string_url)
{
  char *response = NULL;
  URL *url = NULL;
  my_socket *sock = NULL;
  char *request;

  url = parse_url (string_url, url);
  if (url != NULL) {    
  
    sock = my_socket_open (url->host, url->port);
    if (sock != NULL) {

      request = asprintf ("HTTP 1.1\nGET %s", url->path);
      my_socket_send (sock, request);

      response = my_socket_recvall (sock);

      my_socket_close (sock);
    }
    free (url);
  }
  return response;
}

I keep hearing smart people insist that goto is a useful tool, but they make it hard to believe them when practically every example I ever see is either a workaround for a language shortcoming or an allergy to nested ifs.

# ? Jun 21, 2012 06:18

Jabor: Jul 16, 2010; #1 Loser at SpaceChem

Because code that looks like

code:

///////////
  ///////////
    ///////////
      ///////////
        ///////////
          ///////////
            ///////////
            ///////////
          ///////////
        ///////////
      ///////////
    ///////////
  ///////////
///////////

Is actually pretty horrible. Two levels is fine, three or four is pushing it, once you get to 5 it's horrendous unless you want to waste an entire monitor for a single bit of code.

But that said, yes, goto is often a workaround for "missing" language features. Structured programming is important. When your programming language doesn't support the ideal structure for this situation, emulating the proper structure with gotos is often far better than trying to hack something up with the control-flow structures it does support, which will end up even worse.

As a quick example, nesting ifs like this for cleanup purposes is a lovely hack to compensate for the lack of deterministic destructors. Faking it with gotos is still a hack, but it's far less lovely than doing it with if statements.

# ? Jun 21, 2012 06:43

Wheany: Mar 17, 2006; Spinya^{ha^{ha^haha}ha}ha_{ha_{ha_haha}ha}ha!; Doctor Rope

Gotos and returns keep your code from looking like >

# ? Jun 21, 2012 06:43

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

I noticed that I never free(request); in my code sample, so I'm actually leaking memory. Oh well.

# ? Jun 21, 2012 07:05

The Gripper: Sep 14, 2004; i am winner

Zhentar posted:

And what's wrong with doing this instead?
C++ code:
snip
I keep hearing smart people insist that goto is a useful tool, but they make it hard to believe them when practically every example I ever see is either a workaround for a language shortcoming or an allergy to nested ifs.

I prefer that over goto any day, plus it adds the convenience of being able to fold code blocks properly and not have to gently caress around following gotos to skim program flow.

Maybe goto is a useful tool but if your only reason for using it is because you hate nesting, I hate you.

Edit;

redleader posted:

Today I saw some of the html output from our platform:

Haha it's like one of those chain emails with "SCROLL DOWN --------> boo" in it.

The Gripper fucked around with this message at 09:39 on Jun 21, 2012

# ? Jun 21, 2012 09:35

redleader: Aug 18, 2005; Engage according to operational parameters

Today I saw some of the html output from our platform:

code:

<script>
//
//
//
//
//
//
//
//
//
//
//
//
//
//
//
//
//
//
//
//
//
//
foo();
//
</script>

I actually laughed at this.

# ? Jun 21, 2012 09:35

bobthecheese: Jun 7, 2006; Although I've never met Martha Stewart, I'll probably never birth her child.

Doctor w-rw-rw- posted:

This joke is old and tired. Please stop.

PHP doesn't cause bad code (though it certainly helps), people do. PHP is a very practical and productive tool, and in the hands of a capable programmer, can avoid a lot of badness. I dislike almost everything about PHP, but since I have the privilege of working amongst good programmers, the PHP code I see at work is not crap. Good PHP is still going to be better than bad Python or bad Scala or what-have-you.

I'm of the personal opinion that if you can't write clean, clear, and consistent code in PHP, the language is not the problem.

Sure, it's missing a few features that you'd expect in, say, Java or python, or whatever else, but it works and it doesn't force you to write your code in one particular way.

# ? Jun 21, 2012 11:31

KaneTW: Dec 2, 2011

bobthecheese posted:

Sure, it's missing a few features that you'd expect in, say, Java or python, or whatever else, but it works and it doesn't force you to write your code in one particular way.

I don't quite think features are the problem in PHP.

Horrors. I used to work on a HL1 based mod and most weapons were created by copy-pasting and changing a few values. Hmm. Let's run simian against the server-side codebase:

code:

Found 22067 duplicate lines in 2005 blocks in 147 files
Processed a total of 105422 significant (206434 raw) lines in 249 files

Compare that with regular HL1 MP SDK:

code:

Found 1395 duplicate lines in 181 blocks in 37 files
Processed a total of 25789 significant (51060 raw) lines in 94 files

E: And here the singleplayer SDK:

code:

Found 4779 duplicate lines in 405 blocks in 74 files
Processed a total of 48011 significant (95291 raw) lines in 144 files

KaneTW fucked around with this message at 12:01 on Jun 21, 2012

# ? Jun 21, 2012 11:54

Zombywuf: Mar 29, 2008

tef posted:

HTTP/1.1 sends \r\n as line separators, and has a Host: header

It also requires the verb, then the url, then the version number.

Doctor w-rw-rw- posted:

This joke is old and tired. Please stop.

No joke.

quote:

PHP doesn't cause bad code

PHP is bad code.

# ? Jun 21, 2012 12:03

Adbot: ADBOT LOVES YOU

# ? May 21, 2024 00:18

Zombywuf: Mar 29, 2008

Zhentar posted:

And what's wrong with doing this instead?

A complete lack of RAII.

# ? Jun 21, 2012 12:09

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 »