The Daily WTF: Curious Perversions in Information Technology

2006-01-27 Reply Admin

Not the first time I've seen someone mystified by Perl's spaceship operator. And as far as I know every single Perl distribution comes with complete and clear documentation.

As someone who started with perl long before it had strict, I gotta agree with the last one. If I wanted to write clean and polished code I wouldn't be using perl in the first place.

chrismcb · 2006-01-27 Reply Admin

Anonymous:
Alex Papadimoulis:
--SNIP--

Jeremy Lew found this little comment in some C++ source file. Judging from the comment, the developer at least had an inkling of his questionable sanity ...
void sysDraw::close()
{
  // Is this necessary????
  // probably not but what the hell.
  // better safe than sorry
  if (!this) return;                 
  
  // ... snip ...
}
--SNIP--
I have to comment on this one, depending on what other hackery he was engaged in, this may have been a valid concern. I've developed C++ applications for small systems that used pointers to C++ member functions as call-backs and it is conceivable that
the actual call site might get bad/null data. Granted, the whole system was fugly, but that is a different story.

--The BitShifter, certified language abuse specialist.

Its pretty easy for it to happen, but it is kind of lame to write code to protect against it.

sysDraw *psd = NULL;

psd->close();

This will call sysDraw::close() function and this will be NULL. You should be tracking down why you are calling the function with a null pointer. Adding a if this equals null is just a bandaid on the problem.

2006-01-27 Reply Admin

WTF Batman fails because first, his name is WTF Batman, so it's impossible to take him seriously, and secondly, he used the signature feature, Gene actually SIGNS his post each time.

Also, this makes you the troll, since your only purpose was to provoke a response from me or someone like me.

Try harder!

2006-01-28 Reply Admin

Anonymous:
Alex Papadimoulis:
void sysDraw::close()
{
  // Is this necessary????
  // probably not but what the hell.
  // better safe than sorry
  if (!this) return;                 
  
  // ... snip ...
}
I have to comment on this one, depending on what other hackery he was engaged in, this may have been a valid concern. I've developed C++ applications for small systems that used pointers to C++ member functions as call-backs and it is conceivable that
the actual call site might get bad/null data. Granted, the whole system was fugly, but that is a different story.

Yes... In theory it's too late to test for that, because in order to get to that return you have already technically dereferenced a null pointer and invoked undefined behavior. In practice, the behavior that you'll actually get (assuming it's not something like a virtual function) is that the call will succeed with a null this pointer.

A better solution than just returning would be to crash hard so it's easier to find the bad call so you can add the null check there or whatever. Then again, if you're about to ship the code somewhere undebugable, I suppose this is a sensible panic action for a "close" function.

2006-01-28 Reply Admin

Anonymous:
Yes... In theory it's too late to test for that, because in order to get to that return you have already technically dereferenced a null pointer and invoked undefined behavior. In practice, the behavior that you'll actually get (assuming it's not something like a virtual function) is that the call will succeed with a null this pointer.

Oh yeah, not necessarily a null this pointer. If inheritance is involved, you may end up with a pointer that has been adjusted so that it won't compare equal to null. Basically, checking for null inside the function is fragile.

2006-01-28 Reply Admin

Alex Papadimoulis:
Ok, I realize I'm deviating from comments a bit, so before I go back, here's a function that Steve found while developing a plugin for a 3D Studio Max ...
MaybeAutoDelete();

I don't know what MaybeAutoDelete() do, but in the open source Gallery 2 there's a function begin with "maybe" which do the following:

    /**
     * Compact the access list map, if we deem that it's a good time to do so.
     *
     * @return object GalleryStatus a status code
     */
    function maybeCompactAccessLists() {
        /* We use a high tech genetic algorithm to make our decision */
        if (rand(1, 100) <= 50) {
            $ret = GalleryCoreApi::compactAccessLists();
            if ($ret->isError()) {
                return $ret->wrap(__FILE__, __LINE__);
            }
        }
        return GalleryStatus::success();
    }

2006-01-28 Reply Admin

Anonymous:

    /**
     * Compact the access list map, if we deem that it's a good time to do so.
     *
     * @return object GalleryStatus a status code
     */
    function maybeCompactAccessLists() {
        /* We use a high tech genetic algorithm to make our decision */
        if (rand(1, 100) <= 50) {
            $ret = GalleryCoreApi::compactAccessLists();
            if ($ret->isError()) {
                return $ret->wrap(__FILE__, __LINE__);
            }
        }
        return GalleryStatus::success();
    }

2006-01-28 Reply Admin

This is why I don't put comments in my programs.

I was once responsible for converting a C++ desktop program into a web based app (don't ask).

Anyhoo, just about every damned comment in the program was incorrect. I had to ignore every comment I ran across. The original programmers had made so many changes to the program but had never bothered to update the comments.

I see this everywhere. Open source programs are notorious for this (as we've seen).

I figure there's only one way to ensure a new programmer understands what's going on in an app: go over the whole damned program line by line.

hash · 2006-01-28 Reply Admin

I don't care whether his sig was generated by a modified toaster, it's still f'ing annoying

CornedBee · 2006-01-28 Reply Admin

The Koders search also turned up this nice bit from the Moz source:

         // ATEV: I don't understand what this does
        // stabilize the component manager, etc.
        nsCOMPtr<nsIComponentManager> kungFuDeathGrip= mCompMgr;
        if (mModules) delete mModules;
        kungFuDeathGrip= 0;

2006-01-28 Reply Admin

Alex Papadimoulis:

  // Is this necessary????
  // probably not but what the hell.
  // better safe than sorry

Now I'm scared. This morning I wrote:

# Is this necessary?
# probably not but what the hell
# Better safe than sory
$script = 0 unless ($script);

Is there anything scarier than seeing something which looks like your own code on the daily WTF?

(In my case, it was code written out of the office when I didn't have a live database to check how it deals with empty strings)

Corwinoid · 2006-01-28 Reply Admin

josh:

Yes... In theory it's too late to test for that, because in order to get to that return you have already technically dereferenced a null pointer and invoked undefined behavior. In practice, the behavior that you'll actually get (assuming it's not something like a virtual function) is that the call will succeed with a null this pointer.

A better solution than just returning would be to crash hard so it's easier to find the bad call so you can add the null check there or whatever. Then again, if you're about to ship the code somewhere undebugable, I suppose this is a sensible panic action for a "close" function.

That's not necessarily true, depending on when the code was written. VC 5 has a buggy code generator that would strip the this pointer in some cases, during optimization; especially during callbacks. After spending a few days trying to hunt down this particular bug, I wrote a pretty nifty exception catch that would flag the debugger locally, and blow up all over the place if it errored.

The key code boiled down to if (!this) __asm int 3; at the beginning of a particular message handler. (Ok, it doesn't seem that nifty when I strip it down, admittedly... but that was effectively what got built into the function)

The result? With identical builds I never got the interrupt locally. I shipped with a note to the company I developed it for that there was a "Known bug, and if it errored to give me a call with debugging information." This was built, primarily, for a software development shop, so I figured "You know, this guy's going to know when it crashed and dumps to record the dump, start up his debugger, give me a call and ask WTF?"

6 years later I've not been able to duplicate it again with that tool; the client never called, and over lunch with some of the guys at the company I developed it for, they said the solution fixed a runtime error in what they believed was also a compiler bug -- I checked that code, and it looked fine to me also.

It makes perfect sense that this code is reasonable if originally written on buggy implementations. VC5 was notorious for this and other numerous optimization errors, generally some slight code change (such as if (!this)) would be enough to straighten things out again.

2006-01-28 Reply Admin

hash:
Can I just say that ending every post with

Sincerely,

Gene Wirchenko

is not only annoying as hell, but redundant. Your username is Gene Wirchenko, you don't need to reinforce the fact that it was you who posted the message. Maybe i'm the only one that cringes everytime I read

Sincerely,

Gene Wirchenko

but oh well.

Sincerely,

hash.

i agree, i cringe as well

2006-01-28 Reply Admin

Yup! THAT's why I use Gentoo. ;)

2006-01-28 Reply Admin

Anonymous:
Alex Papadimoulis:
--SNIP--

Jeremy Lew found this little comment in some C++ source file. Judging from the comment, the developer at least had an inkling of his questionable sanity ...
void sysDraw::close()
{
  // Is this necessary????
  // probably not but what the hell.
  // better safe than sorry
  if (!this) return;                 
  
  // ... snip ...
}
--SNIP--
I have to comment on this one, depending on what other hackery he was engaged in, this may have been a valid concern. I've developed C++ applications for small systems that used pointers to C++ member functions as call-backs and it is conceivable that
the actual call site might get bad/null data. Granted, the whole system was fugly, but that is a different story.

--The BitShifter, certified language abuse specialist.

Another possiblity is that the function might get called while the object is in the destructor - in that case, things tend to go splat really badly.

2006-01-28 Reply Admin

The Real WTF is (TM)......

That he is using multiple instances of the single line comment // instead of putting all comments within a block of

/*
... .... */

masklinn · 2006-01-29 Reply Admin

Anonymous:
The Real WTF is (TM)......
That he is using multiple instances of the single line comment // instead of putting all comments within a block of

/*
... .... */

Not that much, some people don't like block comments (they're annoying for formatting purposes for example), and modern editors can comment/uncomment whole blocks with line-comments with a single keystroke.

Mikademus · 2006-01-29 Reply Admin

Anonymous:
The Real WTF is (TM)......
That he is using multiple instances of the single line comment // instead of putting all comments within a block of

/*
... .... /

You use line-comments within blocks you might want to comment out with / ... */ later. Also, you might want to write several lines of comments behind the code, and then you can't use //.

Super-Galactically Mega Co-Starring,

MIKADEMUS

sao · 2006-01-29 Reply Admin

hash:
Can I just say that ending every post with

Sincerely,

Gene Wirchenko

is not only annoying as hell, but redundant. Your username is Gene Wirchenko, you don't need to reinforce the fact that it was you who posted the message. Maybe i'm the only one that cringes everytime I read

Sincerely,

Gene Wirchenko

but oh well.

Sincerely,

hash.

you know what i think is reduntant. some dickhead constantly going on at someone for doing this.

personally, i prefer Gene's input. He adds comments that actually refer to the topic.
They are generally intelligent responses.

If his preference is to write that, whos to stop him? you?

Gene, i say keep it up. If it makes him cringe, do it multiple times in one post -
we might end up getting him riggling on the floor in some sort of a semi-epileptic fit.

2006-01-29 Reply Admin

I you could do it with one hand, you could play Chopin

2006-01-29 Reply Admin

Rank Amateur:
Anonymous:

KungFuDeathGrip is the semi-official name for CTRL-ALT-DELETE equivalent on SGI systems. I think it was CTRL-ALT-F7-BACKSPACE. Try this out and you can get a good idea of where the name came from.

With one hand?

Try that again. . .

If you could do it with one hand, you could play Chopin

2006-01-30 Reply Admin

hash:
Can I just say that ending every post with

Sincerely,

Gene Wirchenko

is not only annoying as hell, but redundant. Your username is Gene Wirchenko, you don't need to reinforce the fact that it was you who posted the message. Maybe i'm the only one that cringes everytime I read

Sincerely,

Gene Wirchenko

but oh well.

Sincerely,

hash.

It's not just you. It makes me want to sign my posts:

Seruptitiously,

Rich

2006-01-30 Reply Admin

I thought OTP was One Time Perl, which commenting out 'use strict;' is the surest way to achieve.

Ulvhamne · 2006-01-30 Reply Admin

Anonymous:
Rank Amateur:
Anonymous:

KungFuDeathGrip is the semi-official name for CTRL-ALT-DELETE equivalent on SGI systems. I think it was CTRL-ALT-F7-BACKSPACE. Try this out and you can get a good idea of where the name came from.

With one hand?

Try that again. . .

If you could do it with one hand, you could play Chopin

I can. But it feels really unnatural. ;)

2006-01-30 Reply Admin

"Its pretty easy for it to happen, but it is kind of lame to write code to protect against it. sysDraw *psd = NULL; psd->close(); This will call sysDraw::close() function and this will be NULL. You should be tracking down why you are calling the function with a null pointer. Adding a if this equals null is just a bandaid on the problem." I've done something like that. Was passing a member function to a C library as a callback (to get messages from it, nothing critical) but I set this up in the constructor, so it was possible to get a callback occuring before the constructor finished (said library being multithreaded). It was horrible code though.

2006-01-30 Reply Admin

re: GetGlobalTime( const TCHAR* Filename )

umm, based on the "nature" of this function, shouldn't it have
been called GetUniversalTime(...)?

Scott

codermeta · 2006-01-30 Reply Admin

hahahah no comments!!

Arancaytar · 2006-01-30 Reply Admin

Anonymous:
Funny -- after seeing the Commentator for the first time yesterday, I put together a post on some of the best comments that I've come across in various chunks of code.

My all-time favorite? A little one-liner dumped into a couple thousand lines of byzantine code. I like to picture the expression on the coder's face as they finally figured out what it was all actually supposed to do:

<font face="courier, courier new">;;; oh sh!t, it tries to write to oracle, too...</font>

Aw, and I almost thought this "Commentator" was actually available.

Seems like I'll have to emulate the verbosity=10,relevance=0,selfimportance=10 myself.

/* Oh, by the well, did I tell you about that one

 * time

turbothy · 2006-01-30 Reply Admin

hash:
Can I just say that ending every post with

Sincerely,

Gene Wirchenko

is not only annoying as hell, but redundant.

I love it.

GalacticCowboy · 2006-01-30 Reply Admin

Anonymous:
...the universe was created by God Almighty in 4004 BC.

Monday, October 14th, if I recall correctly. At around 8:45AM, GMT.

masklinn · 2006-01-30 Reply Admin

GalacticCowboy:
Anonymous:
...the universe was created by God Almighty in 4004 BC.

Monday, October 14th, if I recall correctly. At around 8:45AM, GMT.

T'was Sunday the 21th of October actually, but around 8:45AM indeed, for God liked to get work done early in the morning while he was feeling fresh.

OneFactor · 2006-01-30 Reply Admin

Anonymous:
OneFactor:
Isn't the time of the creation of the universe undecided between 9 and 20 billion years because of the imprecision of the Cepheid yardstick and therefore the Hubble "constant"?

No, no, no... everyone knows that in accordance with today's religious nutjobbery, the universe was created by God Almighty in 4004 BC.

FYI, the 4004 BC thing is religious nutjobbery from about one hundred years ago. Today's religious nutjobs are saying that bloodclotting is unlikely to evolve without intelligent help because it is irreducibly complex. The scaffolding counterargument claims otherwise. As far as I know, there is no longer a debate about the overall empirical dating and it has shifted into the realm of "the dice are loaded", "are not", "are too", "are not", "are too", "dee two".

2006-01-30 Reply Admin

Isn't the time of the creation of the universe undecided between 9 and 20 billion years because of the imprecision of the Cepheid yardstick and therefore the Hubble "constant"?
Why express time in metres when time and the speed of light can be measured more precisely than distances?
To transform between frames, can't you just apply a Lorentz transformation - I thought the (Feynman?) Path Integral was for bringing in quantum mechanical effects.

All of you are forgetting one thing. Your Flux Capacitor.

Bustaz Kool · 2006-01-30 Reply Admin

Anonymous:

No, no, no... everyone knows that in accordance with today's religious nutjobbery, the universe was created by God Almighty in 4004 BC.

I believe that it was on a Monday. Six days later, he rested.

2006-01-30 Reply Admin

Anonymous:
$script = 0 unless ($script);

Try

$script ||= 0;

in the future.

WWWWolf · 2006-01-30 Reply Admin

Someone explained the kung fu death grip thing a while ago in Slashdot. Apparently, it's part of the garbage collection mechanism in event handlers - some operations mess with the reference counts, and if you want to make sure something isn't deallocated while you aren't looking and you need to keep the object around, you apply the kung fu death grip on the thing.

And that's all I know about Mozilla source code. =)

And, bleh, #use strict; is just asking for trouble. Definitely saying "Please don't whine about my sloppiness." It is every Perl coder's duty to Read the fine Camel Book if you get some mysterious errors from your newbieish, messy code. The book is very informative indeed. use strict; is the first step towards making the code good.

hank miller · 2006-01-30 Reply Admin

masklinn:
GalacticCowboy:
Anonymous:
...the universe was created by God Almighty in 4004 BC.

Monday, October 14th, if I recall correctly. At around 8:45AM, GMT.

T'was Sunday the 21th of October actually, but around 8:45AM indeed, for God liked to get work done early in the morning while he was feeling fresh.

Assuming we are referring to the Jewish/christian/Muslim God. (which are essentially the same, though none will admit the others know anything about him)

The universe was created on a Sunday (the sabbath is Saturday, not Sunday!), and God started in the Evening. (What time zone is not specified, though usually Jerusalem time is assumed. Some would guess Garden of Eden time but they don't know where that is. Might be other theories as well)

Reference (KJV): Genesis 1:5: And God called the light Day, and the darkness he called Night. And the evening and the morning were the first day. Genesis 2:2: And on the seventh day God ended his work which he had made; and he rested on the seventh day from all his work which he had made.

2006-01-30 Reply Admin

Are Perl and the Necronomicon isomorphic?

They can be. It depends heavily on whether you 'use strict;'.

Is the alternative 'use eldritch;', 'use cyclopean;' or 'use rugose;'?

2006-01-30 Reply Admin

Anonymous:
Anonymous:
$script = 0 unless ($script);

Try

$script ||= 0;

in the future.

People like you are what turned me into a super-villain.

I don't even write shit like that in "C" any more.

for(p=buffer,q=p;*q&&*(p+=(*q++!=TOK_END))!=TOK_END;prepare(&p,&q))
continue;

masklinn · 2006-01-30 Reply Admin

OneFactor:
FYI, the 4004 BC thing is religious nutjobbery from about one hundred years ago. Today's religious nutjobs are saying that bloodclotting is unlikely to evolve without intelligent help because it is irreducibly complex. The scaffolding counterargument claims otherwise. As far as I know, there is no longer a debate about the overall empirical dating and it has shifted into the realm of "the dice are loaded", "are not", "are too", "are not", "are too", "dee two".

hank miller:
Assuming we are referring to the Jewish/christian/Muslim God. (which are essentially the same, though none will admit the others know anything about him)
The universe was created on a Sunday (the sabbath is Saturday, not Sunday!), and God started in the Evening. (What time zone is not specified, though usually Jerusalem time is assumed. Some would guess Garden of Eden time but they don't know where that is. Might be other theories as well)

Reference (KJV): Genesis 1:5: And God called the light Day, and the darkness he called Night. And the evening and the morning were the first day. Genesis 2:2: And on the seventh day God ended his work which he had made; and he rested on the seventh day from all his work which he had made.

Some people here are seriously lacking in the Pratchett/Gaiman department...

OneFactor · 2006-01-30 Reply Admin

masklinn:
OneFactor:

FYI, the 4004 BC thing is religious nutjobbery from about one hundred years ago. Today's religious nutjobs are saying that bloodclotting is unlikely to evolve without intelligent help because it is irreducibly complex. The scaffolding counterargument claims otherwise. As far as I know, there is no longer a debate about the overall empirical dating and it has shifted into the realm of "the dice are loaded", "are not", "are too", "are not", "are too", "dee two".

hank miller:
Assuming we are referring to the Jewish/christian/Muslim God. (which are essentially the same, though none will admit the others know anything about him) The universe was created on a Sunday (the sabbath is Saturday, not Sunday!), and God started in the Evening. (What time zone is not specified, though usually Jerusalem time is assumed. Some would guess Garden of Eden time but they don't know where that is. Might be other theories as well) Reference (KJV): Genesis 1:5: And God called the light Day, and the darkness he called Night. And the evening and the morning were the first day. Genesis 2:2: And on the seventh day God ended his work which he had made; and he rested on the seventh day from all his work which he had made.

Some people here are seriously lacking in the Pratchett/Gaiman department...

I have read and own every Pratchett up to and including Night Watch. Who is Gaiman?

2006-01-30 Reply Admin

OneFactor:
I have read and own every Pratchett up to and including Night Watch. Who is Gaiman?

You've never heard of Good Omens?

sao · 2006-01-30 Reply Admin

hank miller:
masklinn:
GalacticCowboy:
Anonymous:
...the universe was created by God Almighty in 4004 BC.

Monday, October 14th, if I recall correctly. At around 8:45AM, GMT.

T'was Sunday the 21th of October actually, but around 8:45AM indeed, for God liked to get work done early in the morning while he was feeling fresh.

Assuming we are referring to the Jewish/christian/Muslim God. (which are essentially the same, though none will admit the others know anything about him)

The universe was created on a Sunday (the sabbath is Saturday, not Sunday!), and God started in the Evening. (What time zone is not specified, though usually Jerusalem time is assumed. Some would guess Garden of Eden time but they don't know where that is. Might be other theories as well)

Reference (KJV): Genesis 1:5: And God called the light Day, and the darkness he called Night. And the evening and the morning were the first day. Genesis 2:2: And on the seventh day God ended his work which he had made; and he rested on the seventh day from all his work which he had made.

Unto God, a Thousand Years is as a Day, a Day is as a Thousand Years.
So whos to say that things werent more laid back back then? The Earth took its time spinning about, as did everything else.

On the jewish/christian/muslim thing. Are Christians executing Muslims in the name of this 'common' god.
(I also understand this is not done by all muslims. i have nothing against them. dont misread my thoughts.)

But no more religious talk from me. You dont want to go down that path. Lets keep it Coding related people.

Regards,
Sao.

Regards,
Sao.

Regards,
Sao.

(Convulse all you signed-post-alergic people out there, Convulse!!!)

devdas · 2006-01-31 Reply Admin

Neil Gaiman, author of the Sandman series. He collaborated with Pratchett on the book 'Good Omens'.

http://www.amazon.com/exec/obidos/search-handle-url/ref=br_ss_hs/104-2808118-5695143?platform=gurupa&url=index%3Dstripbooks%3Arelevance-above&field-keywords=neil+gaiman&Go.x=0&Go.y=0&Go=Go
should help.

masklinn · 2006-01-31 Reply Admin

OneFactor:
I have read and own every Pratchett up to and including Night Watch.

You obviously haven't since the quotes and reflections on the creation of the earth in 4004 come from Good Omens, written in 1990 in collaboration between Terry Pratchett and Neil Gaiman

OneFactor:
Who is Gaiman?

An english author of science fiction books and comics, very well known in the US for his Sandman serie (well known enough that he's credited before Pratchett in the US edition of Good Omens, while Pratchett is usually credited first e.g. the UK and french editions), he's the author of the recent American Gods (2001) and Anansi Boys (2005), and -- even though I doubt you care the least about it -- a friend of singer Tori Amos.

OneFactor · 2006-01-31 Reply Admin

Anonymous:
OneFactor:

I have read and own every Pratchett up to and including Night Watch. Who is Gaiman?

You've never heard of Good Omens?

Oops. I meant to say every discworld up to NightWatch. I haven't read any Pratchett apart from discworld.

2006-01-31 Reply Admin

You don't want to lose the pointer, so you use the ku-fu death grip on it. It makes perfect sense.

2006-01-31 Reply Admin

This thread has some of the geekiest replies I've ever read! You guys know who you are...

Roelf_ · 2006-02-01 Reply Admin

sao:
quoted bla

On the jewish/christian/muslim thing. Are Christians executing Muslims in the name of this 'common' god.
(I also understand this is not done by all muslims. i have nothing against them. dont misread my thoughts.)

bla

they did a few centuries ago

masklinn · 2006-02-01 Reply Admin

OneFactor:
Anonymous:
OneFactor:

I have read and own every Pratchett up to and including Night Watch. Who is Gaiman?

You've never heard of Good Omens?

Oops. I meant to say every discworld up to NightWatch. I haven't read any Pratchett apart from discworld.

Then I'd really recommend reading Good Omens, it's good.

And the Tiffany Aching arc (The Wee Free Men and A Hat Full Of Sky, with at least 2 other upcoming books. It's like Harry Potter, but good.). Going Postal is also very good (and much funnier than Night Watch, which is kind of apart from the whole Discworld collection), and extremely geeks-oriented.

A Collection Of Comments

Leave a comment on “A Collection Of Comments”