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 »

TasteMyHouse: Dec 21, 2006

No Safe Word posted:

Stupid/made up/misspelled units
...
dekameters

This is the standard American spelling of "Decametres" as stated here and as taught to me in the 2nd grade.

# ? Aug 8, 2011 22:43

Adbot: ADBOT LOVES YOU

# ? Apr 19, 2024 20:46

No Safe Word: Feb 26, 2005

TasteMyHouse posted:

This is the standard American spelling of "Decametres" as stated here and as taught to me in the 2nd grade.

I always learned it as "decameter", and all units were prefaced with "deca-" and not "deka-". Weird.

# ? Aug 8, 2011 22:48

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

poemdexter posted:

What the hell software needs to go from nanometers to parsecs?

AutoCAD does! Some googling has indicated that this enum is probably retarded-up version of http://docs.autodesk.com/ACD/2010/ENU/AutoCAD%202010%20User%20Documentation/files/WS1a9193826455f5ffa23ce210c4a30acaf-4f8f.htm with some values rearranged and Hectometers replaced with Dynes for some inexplicable reason.

# ? Aug 8, 2011 22:52

Scaramouche: Mar 26, 2001; SPACE FACE! SPACE FACE!

No Safe Word posted:

I always learned it as "decameter", and all units were prefaced with "deca-" and not "deka-". Weird.

It's not super important, I think it might even just be a Greek/Roman thing. I know it's worthless but deka loses the GoogleFight: http://www.googlefight.com/index.php?lang=en_GB&word1=decameter&word2=dekameter

# ? Aug 8, 2011 23:31

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

I'm using a library for exact computation that supposed to be pretty close to drop-in (just include the headers and set up some typedefs and you're done). This is kind of neat*, but what if your library has some warnings to emit during use? You wouldn't want to throw an exception if it's just a warning - that's far too drastic. Instead, maybe you do something like this:

code:

// Usage: core_error(message, file_with_error, line_number, err_type)
//   where err_type=0 means WARNING, error_type=0 means ERROR
void core_error(std::string msg, std::string file, int lineno, bool err) {
  std::ofstream outFile(CORE_DIAGFILE, std::ios::app);  // open to append
  if (!outFile) {
     // perror("CORE ERROR: cannot open Core Diagnostics file");
     std::cerr << "CORE ERROR: can't open Core Diagnostics file"<<std::endl;
     std::exit(1); //Note: do not call abort()
  }
  outFile << "CORE " << (err? "ERROR" : "WARNING")
     << " (at " << file << ": " << lineno << "): "
     << msg << std::endl;
  outFile.close();

Since there couldn't possibly be any problem with just arbitrarily creating and writing to a file wherever the executable happens to be running, let's just do that. If it doesn't work the underlying file system must be unstable or something so obviously we have to just bail. No need to add capability to disable this "feature" - if they need to turn it off they can just dig around, comment this out, and recompile.

_{*It's also a horror on its own - the default behavior is to just redefine double.}

# ? Aug 9, 2011 00:46

GROVER CURES HOUSE: Aug 26, 2007; Go on...

No Safe Word posted:

Out of place units
Dyne

Stupid/made up/misspelled units
microinches
mils
dekameters

I'm fairly sure that the mil is a unit of angle and doesn't exactly belong there.

# ? Aug 9, 2011 00:55

Jerry Bindle: May 16, 2003

Broken Knees Club posted:

I'm fairly sure that the mil is a unit of angle and doesn't exactly belong there.

mil can also be 1/1000th of an inch.

# ? Aug 9, 2011 01:02

Zhentar: Sep 28, 2003; Brilliant Master Genius

Scaramouche posted:

It's not super important, I think it might even just be a Greek/Roman thing.

It also might be done to avoid confusion that could arise between decimeters and decameters.

# ? Aug 9, 2011 01:52

McGlockenshire: Dec 16, 2005; GOLLOCKS!

poemdexter posted:

What the hell software needs to go from nanometers to parsecs?

Han Solo.

# ? Aug 9, 2011 04:55

Malloc Voidstar: May 7, 2007; Fuck the cowboys. Unf. Fuck em hard.

McGlockenshire posted:

Han Solo.

Han Solo is firmware.

# ? Aug 9, 2011 05:51

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

poemdexter posted:

I had to double check if Parsecs was a thing.

What the hell software needs to go from nanometers to parsecs?

http://futureboy.us/frinkdocs/

# ? Aug 9, 2011 12:52

TasteMyHouse: Dec 21, 2006

Zhentar posted:

It also might be done to avoid confusion that could arise between decimeters and decameters.

This is the reason my teacher gave when I asked this question in 2nd grade! For whatever that's worth...

# ? Aug 9, 2011 17:53

Toady: Jan 12, 2009

Dooey posted:

Am I the only one who finds reading code in mixed styles not a big deal? I always use the same style for my own code, but at work we all have different styles and when I read other peoples code I have no problems at all.

I went through a period where I tried each style to see which one I truly preferred, and after the initial novelty of each one wore off, I didn't find any one more readable than the other, so I just chose one to be consistent.

Mustach posted:

I think that monospace fonts are an eyesore. I sometimes think that people stick to them just so they can pretend their code is some kind of WYSIWYG and waste time with valueless formatting like this:

Fixed-width fonts in Objective-C code let Xcode automatically line up colons in long messages, which is really nice:

code:

BOOL result = [someObject doA:firstArg
                        withB:secondArg
                       usingC:thirdArg
                        error:&outError];

Toady fucked around with this message at 21:35 on Aug 9, 2011

# ? Aug 9, 2011 21:21

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

Toady posted:

I went through a period where I tried each style to see which one I truly preferred, and after the initial novelty of each one wore off, I didn't find any one more readable than the other, so I just chose one to be consistent.

I think of written code as written prose: unless someone punches one of my pet peeves' buttons, I get used to different styles in a hurry.

Then the problem is reduced to limiting the number of pet peeves you have, which is much easier.

# ? Aug 9, 2011 21:24

Cocoa Crispies: Jul 20, 2001; Vehicular Manslaughter!; Pillbug

tef posted:

http://futureboy.us/frinkdocs/

I met the creator of that on Saturday during "Crash & Compile" at defcon; they won because one of the teams was a PDP-11 gimmick team that picked their language from a d12, and another team used the other gimmick team's d20 from last year.

# ? Aug 10, 2011 00:52

PrBacterio: Jul 19, 2000

BonzoESC posted:

I met the creator of that on Saturday during "Crash & Compile" at defcon; they won because one of the teams was a PDP-11 gimmick team that picked their language from a d12, and another team used the other gimmick team's d20 from last year.

Hahaha holy poo poo, that thing is the *definition* of the "kitchen sink" language. Honestly, that's quite impressive even though I don't think I'm ever going to use it

# ? Aug 10, 2011 01:46

Dren: Jan 5, 2001; Pillbug

In C++. Just found out that the base class for a ton of our objects is a "Utilities" class where all the methods are static methods.

The same codebase has namespace pollution all over the place, granted it's mostly using namespace std but it results in function prototypes like this:

code:

void doStuff( const string &the_message, const std::string &data );

Now those probably aren't two different types but I'd like to be able to tell for sure without running the preprocessor.

*sigh*

Dren fucked around with this message at 16:49 on Aug 10, 2011

# ? Aug 10, 2011 16:45

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

Dren posted:

I'd like to be able to tell for sure without running the preprocessor.

The true horror spotted.

# ? Aug 10, 2011 17:10

TasteMyHouse: Dec 21, 2006

It's cool when people declare virtual methods in abstract base classes that need to be implemented in subclasses as:

code:

protected virtual void cool_method() {}

instead of

code:

protected void abstract cool_method();

of course if I actually had documentation for this poo poo it wouldn't've been as big of a deal, but as is I had no idea why my subclass wasn't working.

# ? Aug 10, 2011 17:13

w00tz0r: Aug 10, 2006; I'm just so god damn happy.

Serializing a double (sanitized a bit for clarity):

code:

double d = 1.0;

volatile union TypeConverter {
	TypeConverter(double d) : _d(d) {}
	double _d;
	char _bytes[sizeof(double)];
} converter(d);

vector<char> buffer(sizeof(double));
for (int i = 0; i < sizeof(double); i++)
	buffer[i] = converter._bytes[i];

This was shown to me with the explanation of "this is the safest way I could think of to do this."

# ? Aug 10, 2011 19:55

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

Huh, I didn't actually realize you could give a union a constructor. I wonder if there's ever a good reason to.

The volatile there is extra-weird.

# ? Aug 10, 2011 20:11

1337JiveTurkey: Feb 17, 2005

Plorkyeran posted:

The volatile there is extra-weird.

Just a wild guess, but the user might have ripped it off from somewhere else where it makes sense and didn't bother to remove it. The status registers for a UART in an embedded system would be some horrible combination of a volatile union and structs.

# ? Aug 10, 2011 22:08

Brecht: Nov 7, 2009

TasteMyHouse posted:

It's cool when people declare virtual methods in abstract base classes that need to be implemented in subclasses as:
code:
protected virtual void cool_method() {}
instead of
code:
protected void abstract cool_method();
of course if I actually had documentation for this poo poo it wouldn't've been as big of a deal, but as is I had no idea why my subclass wasn't working.

'abstract' is not a C++ keyword? What is this?

Anyway, abstract methods should be virtual void cool_method() = 0; obv.

# ? Aug 10, 2011 22:10

nielsm: Jun 1, 2009

Brecht posted:

'abstract' is not a C++ keyword? What is this?

Probably C#.

# ? Aug 10, 2011 22:16

w00tz0r: Aug 10, 2006; I'm just so god damn happy.

1337JiveTurkey posted:

Just a wild guess, but the user might have ripped it off from somewhere else where it makes sense and didn't bother to remove it. The status registers for a UART in an embedded system would be some horrible combination of a volatile union and structs.

Not so much that he didn't bother to remove it, he wrote it that way. You're half-right on the embedded systems guess; he's an embedded C guy who's lead on our C++ desktop applications by virtue of being in the company the longest.

# ? Aug 10, 2011 22:18

Dren: Jan 5, 2001; Pillbug

I guess it's "safe" because it went in a vector? Considering that he's an embedded guy at least he didn't use any macros.

# ? Aug 10, 2011 22:51

TasteMyHouse: Dec 21, 2006

nielsm posted:

Probably C#.

Yeah, it's C#. I'm in the middle of writing a bunch of C++/CLI and C# classes to wrap a C library to write a plugin for a (completely undocumented) internal tool at work. It's fun kind of but also terrible.

# ? Aug 11, 2011 00:06

Scaevolus: Apr 16, 2007

Plorkyeran posted:

The volatile there is extra-weird.

I think the volatile is meant as an extra-strong hint for buggy compilers that don't properly handle the union's double/char array aliasing.

He's probably been bitten by bad compilers for embedded systems before, making him learn a more defensive style of programming.

Scaevolus fucked around with this message at 00:55 on Aug 11, 2011

# ? Aug 11, 2011 00:21

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

i guess disabling all optimizations is a way to avoid correctness problems due to aliasing violations

not a very good way

# ? Aug 11, 2011 01:58

mjau: Aug 8, 2008

Plorkyeran posted:

i guess disabling all optimizations is a way to avoid correctness problems due to aliasing violations

Union members are allowed to alias even with strict aliasing optimization switched on. That's kind of what they're for.

# ? Aug 11, 2011 02:14

Janitor Prime: Jan 22, 2004; PC LOAD LETTER

What da fuck does that mean; Fun Shoe

Plorkyeran posted:

i guess disabling all optimizations is a way to avoid correctness problems due to aliasing violations

not a very good way

Seriously what the gently caress is an aliasing violation? I tried looking it up, but the examples I saw didn't make any sense. Is that something that you C guys have to deal with constantly?

# ? Aug 11, 2011 02:25

New Yorp New Yorp: Jul 18, 2003; Only in Kenya.; Pillbug

TasteMyHouse posted:

It's cool when people declare virtual methods in abstract base classes that need to be implemented in subclasses as:
code:
protected virtual void cool_method() {}
instead of
code:
protected void abstract cool_method();
of course if I actually had documentation for this poo poo it wouldn't've been as big of a deal, but as is I had no idea why my subclass wasn't working.

Maybe they didn't know that abstract methods exist.

I like it when people make every method in a class virtual, because "I might need to override it some day!" It doesn't hurt anything other than readability, of course.

# ? Aug 11, 2011 02:40

First Time Caller: Nov 1, 2004

This is pretty miniscule compared to some of the things I've seen in this thread but I thought it was pretty retarded.

code:

<?php if("ying" == "yang"): ?>
//code block
<?php endif; ?>

# ? Aug 11, 2011 02:41

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

mjau posted:

Union members are allowed to alias even with strict aliasing optimization switched on. That's kind of what they're for.

This is in fact a very exciting and lively debate on the C committee. Neither C nor C++ actually permits aliasing through a union. Unions have an "active member", determined by which member was last stored through, and it is impermissible to read from any member other than the active member, with one exception: if the member accessed and the active member are both structs, and they share a common prefix, it is okay to access a member of that common prefix. However, some people have argued that unions ought to influence aliasing somehow, except that nobody seems quite certain what the language rule should be.

Anyway, char can alias anything, so if the code instead just casted the double* to char* and did that final loop, it would be perfectly legal by both standards.

# ? Aug 11, 2011 02:46

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

mjau posted:

Union members are allowed to alias even with strict aliasing optimization switched on. That's kind of what they're for.

writing to one union field then reading from another one is technically undefined behavior, and we're assuming a buggy and dumb compiler that might do something completely insane

Ithaqua posted:

I like it when people make every method in a class virtual, because "I might need to override it some day!" It doesn't hurt anything other than readability, of course.

and performance

# ? Aug 11, 2011 02:47

Ensign Expendable: Nov 11, 2008; Lager beer is proof that god loves us; Pillbug

Maybe they miss Java.

# ? Aug 11, 2011 02:49

New Yorp New Yorp: Jul 18, 2003; Only in Kenya.; Pillbug

Plorkyeran posted:

and performance

I'd never thought of that, actually. I always assumed that was something the compiler figured out, and at runtime it was already squared away.

Turns out I'm wrong! http://stackoverflow.com/questions/530799/what-are-the-performance-implications-of-marking-methods-properties-as-virtual

# ? Aug 11, 2011 03:10

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

MEAT TREAT posted:

Seriously what the gently caress is an aliasing violation? I tried looking it up, but the examples I saw didn't make any sense. Is that something that you C guys have to deal with constantly?

Basically "aliasing" is when you have one location that's referred to by multiple names.

If you assume that anything at all can be aliased to anything else in the current scope, then that really limits what optimizations you can do - you have to actually write everything to memory before you read through a pointer, for example. So for performance, the compiler is allowed to assume that in most cases there is no aliasing.

If there actually is aliasing when the compiler assumes there isn't, then you can run into some problems with your reads getting stale values and the like.

# ? Aug 11, 2011 03:21

Zamujasa: Oct 27, 2010; Bread Liar

First Time Caller posted:

This is pretty miniscule compared to some of the things I've seen in this thread but I thought it was pretty retarded.
php:
<?php if("ying" == "yang"): ?>
//code block
<?php endif; ?>

That's actually a pretty decent way of doing it, compared to a "{", HTML block, then a closing "}" possibly many lines away.

A bit weird and rarely used, but understandable.

# ? Aug 11, 2011 05:24

Adbot: ADBOT LOVES YOU

# ? Apr 19, 2024 20:46

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

Zamujasa posted:

That's actually a pretty decent way of doing it, compared to a "{", HTML block, then a closing "}" possibly many lines away.

A bit weird and rarely used, but understandable.

Surely the horror is the comparison that cannot possibly* be true, and/or the misspelling of "yin".

* I know nearly no PHP and would not be surprised to hear that I am wrong here.

# ? Aug 11, 2011 05:42

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 »