You, I heard you have some memory to waste

Doesn't even save time on computation, cache abuse since it destroys locality and what else... Horrible code

Are you fucking kidding me... This code is a disgrace...

@filthyranter because they are dumb. They thought that if you give a long long, you should get a long long, even though it maxes out in value at fucking 18.

looks like ascii art

* code porn!

Technically correct... B+

Return num.length()-1;

Except for first use case roughly, treat num as string

@hoggchan it's more like code rape and homicide.

I use a one liner for that; see pseudocode:
int n
nLength = floor(log10(n)) + 1

Ha! What prat wrote that code? Obviously he should have used a for loop instead...

@ryanmhoffman lol

There may be good reasons for writing such code, although I believe there are none in this case, as proven by the return type and the redundant >= tests

@bkwilliams this is floating point math, you don't need it here

Come on guys, at least it's got documentation 👍😎👌

Looking at is again, thanks to notifs; it's also wrong for negative numbers there is no default return.

@bkwilliams +1 for thinking log10 :)

This hurts me, i am cry

I'm thinking about how even if someone intentionally wanted to write that, they'd run a loop to print the code.

I like how there's a staircase of small travesties leading up to the big one. It helps the reader mentally prepare.

I predict, that you haven't ever heared about logarithms =) a guy above posted a good inline form of counting nunber of digits in the number ( floor(log10(x)) + 1) )

@OrestH my guess actually is the code in the image is more efficient (discounting the long long and unnecessary comparisons) because there would be one instruction for each case.

A log is probably a few instructions, one more for add, one comparison...

I'll confess I probably would have used sprintf and strlen because I'm lazy and that's the worst option..

@alebianco the cpu will likely execute the instructions out of order - i.e. in parallel in the same core, since there are no real interdependencies there. So it may very well be faster than any floating-point operation (or a conversion to string, or an unrolled loop)

I think convert to string and count length is slightly faster than log? I guess it depends on the language

What really bugs me is that on line 10 the semicolon is suddenly displaced by 3 to the right instead of 2 on the other lines

This is being reposted too much

@mzpro10 this was posted 3 weeks ago.

The patience it took to write this....

@soulkicks it was dead, no necromancy!!

@iam13islucky what is dead may never die

return num.toString().length;

Is he paid by the number of lines of code that he writes?