The Daily WTF: Curious Perversions in Information Technology

PJH · 2015-02-12 Reply Admin

Wake up @PaulaBean!

http://thedailywtf.com/articles/accurate-comments

Kevin L saw the program crash with a null pointer exception. This lead to digging through stack traces, logs, and eventually the commit history to figure out who was responsible....

Paging @mark_bowytz, @apapadimoulis

Remy · 2015-02-12 Reply Admin

We are looking into the Paula Bean issue. We just need someone more brillant.

VinDuv · 2015-02-12 Reply Admin

Maybe @PaulaBean is trying to do a string padding on the article title?

aliceif · 2015-02-12 Reply Admin

                                                                                                                                                                                                                                                                                    fourth!

Is there a limit on spaces?!?!

PJH · 2015-02-12 Reply Admin

Apparently not...

[image]

Perhaps they need to be rate limited...

JBert · 2015-02-12 Reply Admin

About that final note:

Note: it'n

Didn't you mean it's?

Filed under: Spellar'd that for you

accalia · 2015-02-12 Reply Admin

JBert:
Filed under: ~~Spellar~~@accalia'd that for you

FTFY

Filed under: or would that more correctly be "anti-@accalia'd"...?

NedFodder · 2015-02-12 Reply Admin

un-@accalia'd?

accalia · 2015-02-12 Reply Admin

that works too.

:laughing:

RogerC · 2015-02-12 Reply Admin

Picks up the phone, calls the IT help desk: "Hello, I need a new keyboard. This one has run out of spaces."

Onyx · 2015-02-12 Reply Admin

Accurate Comments

[image]

Well... that's one way to ensure none of them are wrong...

powerlord · 2015-02-12 Reply Admin

Those who cannot remember printf are condemned to reimplement it... poorly.

Edit: I actually meant printf and its variants, so if this were C it'd be sprintf. Since this looks like Java, it'd be String.format.

Dlareg · 2015-02-12 Reply Admin

Maybe I am missing the point but why would this result in a null pointer exception? My C is very rusty since I found Python.

HardwareGeek · 2015-02-12 Reply Admin

Dlareg:
Maybe I am missing the point but why would this result in a null pointer exception? My C is very rusty since I found Python.

It's not C.
It would result in a null pointer exception if registrationNumber were null.

DocMonster · 2015-02-12 Reply Admin

I've seen these type of comments before, I don't get it. Why add a stupid "This isn't really good" comment instead of you know actually refactoring the damn code to be improved.

mruhlin · 2015-02-12 Reply Admin

If you've never been in a situation where a boss was afraid to touch anything and had developed an instinctive flight reaction against the word "refactor", consider yourself lucky. "It's a 5 minute fix, chief" "I don't care. QA says it'll take them 3 days to test it" "Um, shouldn't they just have to enter a couple strings?" "Probably, but I'm not going to tell them how to do their jobs" "Apparently they're telling us how to do ours...."

mruhlin · 2015-02-12 Reply Admin

// this is a much better way to do this.... 
    public static String padRegistrationNumber(String registrationNumber) {
        if (registrationNumber.length() == 11) {
            return " " + registrationNumber;
        }
        //
        // SNIP ....
        //
        if (registrationNumber.length() == 1) {
            return "           " + registrationNumber;
        }
        else if (registrationNumber.length() == 0 || registrationNumber == null){
            return "            ";
        }
        return registrationNumber;
    }

FIXED

HardwareGeek · 2015-02-12 Reply Admin

mruhlin:
FIXED

:laughing: "You keep using that word. I do not think it means what you think it means."

kupfernigk · 2015-02-12 Reply Admin

mruhlin:
FIXED

I assume that this is a joke. At least, I hope so.

The thing that interests me about string padding is that it is an extremely simple problem to state, and there are many solutions, but there is no obviously optimal solution. (Feel free to prove me wrong).

PleegWat · 2015-02-12 Reply Admin

There are several obviously-better-than-this solutions though.

machtyn · 2015-02-12 Reply Admin

That's generally because QA doesn't know what goes into programming. Nor do they have a full understanding of how a fix changes the system or affects surrounding code. The developer should be able to help QA understand that and that the test shouldn't take too long.

Of course, that would be in a sane, communicative environment.

machtyn · 2015-02-12 Reply Admin

Oh, no, no, no... This is a perfect chance to use a for loop! :anguished:

Yamikuronue · 2015-02-12 Reply Admin

machtyn:
Nor do they have a full understanding of how a fix changes the system or affects surrounding code

The number of times I've had a programmer tell me "But I didn't touch that part, why did it break!???!?!?" indicates to me that that sword cuts both ways

PleegWat · 2015-02-12 Reply Admin

machtyn:
Oh, no, no, no... This is a perfect chance to use a for loop!

ITYM recursion

public static String padRegistrationNumber(String registrationNumber) {
    if (registrationNumber.length() < 12) {
        return " " + padRegistrationNumber( registrationNumber );
    } else {
        return registrationNumber
    }
}

Yamikuronue · 2015-02-12 Reply Admin

machtyn:
QA doesn't know what goes into programming

Also, I know what you meant, but STFU :angry: The assumption that nobody in QA understands code is part of why many in QA never bother to learn code and why many coders utterly disregard QA

Onyx · 2015-02-12 Reply Admin

9/10, needs more factories.

EatenByAGrue · 2015-02-12 Reply Admin

there are many solutions, but there is no obviously optimal solution.

This one pretty much does the job, with some variation based on how you're storing your strings:

pad(original, targetLength, padChar=" "):
     padded = malloc(targetLength)
     orig_length = original.length  -- sometimes this is an O(n) traversal of the entire original
     split = max(targetLength - orig_length, 0)
     for i from 0 to split:
          padded[i] = padChar
     for i from 0 to min(orig_length, targetLength):
          padded[i + split] = original[i]
     return padded

That's guaranteed to be O(targetLength).

PleegWat · 2015-02-12 Reply Admin

That'll do horrible stuff if length(original) > targetLength

cconroy · 2015-02-12 Reply Admin

machtyn:
Oh, no, no, no... This is a perfect chance to use ~~a for loop! :anguished:~~ Duff's Device

FTFY

aliceif · 2015-02-12 Reply Admin

    public static String padRegistrationNumber(String registrationNumber) {
        try {
            switch (registrationNumber.length()) {
            case 0:
                registrationNumber += " ";
            case 1:
                registrationNumber += " ";
            case 2:
                registrationNumber += " ";
            case 3:
                registrationNumber += " ";
            case 4:
                registrationNumber += " ";
            case 5:
                registrationNumber += " ";
            case 6:
                registrationNumber += " ";
            case 7:
                registrationNumber += " ";
            case 8:
                registrationNumber += " ";
            case 9:
                registrationNumber += " ";
            case 10:
                registrationNumber += " ";
            case 11:
                registrationNumber += " ";
            case 12:
            default:
                return registrationNumber;
            }
        } catch (NullPointerException e) {
            return padRegistrationNumber("");
        }
    }

CoyneTheDup · 2015-02-12 Reply Admin

HardwareGeek:
It would result in a null pointer exception if registrationNumber were null.

If a function gives a null pointer exception when you pass a null input, then don't do that.

So in all "1,234" places where this function is called, all team members are required to do this:

String result = null;
if (regnum != null) result = padRegistrationNumber(regnum);

Fixed.

dcon · 2015-02-12 Reply Admin

machtyn:
Of course, that would be in a sane, communicative environment.

What's that?

CoyneTheDup · 2015-02-12 Reply Admin

What about this strategy?

public static String padRegistrationNumber(String registrationNumber) {
   if (registrationNumber == null || registrationNumber.length > 12) return registrationNumber;
   return ("            " + registrationNumber).substring(registrationNumber.length);
}

(It's a bit easier in languages that have a "right" function.)

Edit: Still fails for null. Fixed.

antiquarian · 2015-02-12 Reply Admin

CoyneTheDup:
If a function gives a null pointer exception when you pass a null input, then don't do that.
So in all "1,234" places where this function is called, all team members are required to do this:

String result = null; if (regnum != null) result = padRegistrationNumber(regnum);

Fixed.

+1 Funny

Filed under: inb4 slashdot is leaking again

PleegWat · 2015-02-12 Reply Admin

CoyneTheDup:

return ("            " + registrationNumber).substring(registrationNumber.length);

I don't think so

kupfernigk · 2015-02-12 Reply Admin

PleegWat:
There are several obviously-better-than-this solutions though.

Well, yes, especially all the ones that start with a test for null and behave appropriately.

The problem is to find one that is the most efficient in resources. I have in the past had the problem of reconstructing a space padded fixed field text file from a database for legacy reasons. When you are constructing a file that will exceed 100Mbytes and has, say, ten space padded fields, the efficiency of the padding algorithm makes an awful lot of difference to execution time.

Don't ask, it was over 10 years ago.

PleegWat · 2015-02-12 Reply Admin

That'll pad right instead of left.

CreatedToDislikeThis · 2015-02-12 Reply Admin

As awful as that function is, it would take a special set of circumstances for it to be the root cause of a null pointer exception. While it might be expected that it would either assert or throw a specialized exception on null argument, a string manipulation function shouldn't be required to silently ignore null string references.

aliceif · 2015-02-12 Reply Admin

It was supposed to be tongue-in-cheek. Sheesh.

aliceif · 2015-02-12 Reply Admin

Whoops. I'll do a quick fix:

    public static String padRegistrationNumber(String registrationNumber) {
        try {
            switch (registrationNumber.length()) {
            case 0:
                registrationNumber += " ";
            case 1:
                registrationNumber += " ";
            case 2:
                registrationNumber += " ";
            case 3:
                registrationNumber += " ";
            case 4:
                registrationNumber += " ";
            case 5:
                registrationNumber += " ";
            case 6:
                registrationNumber += " ";
            case 7:
                registrationNumber += " ";
            case 8:
                registrationNumber += " ";
            case 9:
                registrationNumber += " ";
            case 10:
                registrationNumber += " ";
            case 11:
                registrationNumber += " ";
            case 12:
            default:
                int lastSpace = registrationNumber.lastIndexOf(" ");
                return lastSpace == -1 ? registrationNumber
                        : registrationNumber.substring(lastSpace + 1)
                                + registrationNumber
                                        .substring(0, lastSpace + 1);
            }
        } catch (NullPointerException e) {
            return padRegistrationNumber("");
        }
    }

Much better!

kupfernigk · 2015-02-12 Reply Admin

aliceif:
It was supposed to be tongue-in-cheek.Sheesh.

Sorry. On teh interwebs, we have still not sorted out tags for humor and irony.

dkf · 2015-02-12 Reply Admin

I wonder what's wrong with:

public static String padRegistrationNumber(String registrationNumber) {
    assert registrationNumber != null;
    if (registrationNumber.length() >= 12)
        return registrationNumber;
    String s = "            " + registrationNumber;
    return s.substring(s.length() - 12);
}

dcon · 2015-02-12 Reply Admin

aliceif:
Much better!

C# isn't my native language, but I thought they didn't allow fall through. So you need lots of gotos.

dkf · 2015-02-12 Reply Admin

dcon:
C# isn't my native language, but I thought they didn't allow fall through. So you need lots of gotos.

The capitalized String should tip you off: it's Java…

dcon · 2015-02-12 Reply Admin

dkf:
The capitalized String should tip you off: it's Java…

It's not C/C++, so no.

Nprz · 2015-02-12 Reply Admin

Pretty sure I got a capitalized S as an option in c#.

Nprz · 2015-02-12 Reply Admin

PleegWat:

ITYM recursion

public static String padRegistrationNumber(String registrationNumber) { if (registrationNumber.length() < 12) { return " " + padRegistrationNumber( registrationNumber ); } else { return registrationNumber } }

Won't this cause a stack recursion fault? You need to put the " " inside the recursive call.

dkf · 2015-02-12 Reply Admin

dcon:
It's not C/C++, so no.

WAT?

Nprz:
Pretty sure I got a capitalized S as an option in c#.

Maybe, but no. The method capitalizations are wrong. The exception name is wrong. While yes, you could assemble this whole thing in C# if you wanted to and have it compile… No, you're just demonstrating your general ignorance.

PleegWat · 2015-02-12 Reply Admin

aliceif:
Much better!

Are you sure?

PleegWat · 2015-02-12 Reply Admin

Nprz:
Won't this cause a stack recursion fault? You need to put the " " inside the recursive call.

You're correct.

Mine has bugs too.

Accurate Comments

Leave a comment on “Accurate Comments”