Do you know your bitwise operators? 18 points by amenghra 1376 days ago | hide | past | web | 31 comments | favorite

 Validator seems broken.
 I keep getting various parse errors like Expected assignment, comment or return statement but " " found. When trying to get rid of parse errors, I did something like "return x" and it says Solution did not return correct output for x=1. Expected output: 1, got: 0 Which is clearly not right.Just tried doing "return 1" and still got Solution did not return correct output for x=1. Expected output: 1, got: 0
 I get the same thing as you, and I get the same thing when I have "return 0." The puzzle is annoying.For what it's worth, here's my answer: 1>>(x&~(~x+1))&~(1>>x) . (I didn't figure out that the tester was broken before I developed my answer.)I tested it with positive integers on Python. I haven't tried it with negatives in Javascript.The '(x&~(~x+1))' is a work-around to get 'x&(x-1)', which is 0 if x is a power of two. The (1>> value) is a workaround to get "value == 0". Thus, '1>>(x&~(~x+1))' means 'is x a power of two?'The ~(1>>x) means 'is x == 0'. Combine the two together gives 1 if a single bit is set, else 0.
 The work-around for 'x&(x-1)' I found is one-character longer: '(x&(~x+x+x))'
 Definitely. function one_bit(x) { return 1 } Solution did not return correct output for x=1. Expected output: 1, got: 0 umm... ok
 I certainly can't get it to work. For the first problem, I tried: function one_bit(x) { return x & (x-1) == 0 } Only to see this error: Expected "}" or comment but "&" found. WAT
 You can't use '-'.
 That returns true or false, which I believe strictly speaking aren't numeric in Javascript. By definition of problem, probably should be:function one_bit(x) { return x & (x-1) == 0 ? 1 : 0; }
 It appears to have a limited JS validator to prevent you from using anything other than the limited operators it allows.
 It really doesn't like having anything after "return" except a single value or variable.
 That's not it. For example, it accepts "return x|x&x;" . The restriction is that only a handful of Javascript operations are allowed. Anything else gives a hard-to-understand error message.
 Should be fixed now...
 :(Oops, i'll fix it.
 This is stated to be a puzzle. But, when does this actually get used? Are there use cases of bit manipulation in JS? I've never understood when using bit operators could be necessary or even useful. Of course, I'm speaking from my experience building web apps; ruby, python, js, php, etc. I don't have much other programming experience
 Embedded applications use bitwise ops all the time. Particularly for performing changes to registers.Example, a microcontroller has an 8 bit status register. Each bit controls some hardware function. You wish to change bit 5, but you can't change anything else because it would cause unwanted side effects. So this rules out simply saying myregister = 0b00001000.Common solution is to make a macro like:#define setbit(BYTE,BIT) BYTE |= (1 << BIT)Thus you can say I want to set bit 5 in my register by calling setbit(foo, 5) while keeping the rest of the byte intact. There are analogous macros for clearing and toggling bits.Or you might want to perform an operation based on a particular bit in a byte, e.g. byte 5 is a status flag that tells you when a timer has overflowed.#define getbit(BYTE, BIT) !!(BYTE & ~(1 << BIT))Then you can say:if( getbit(myregister, 5)){ // do something }Fun fun :)
 - Packing lots of small pieces of information into a single memory word - for example, 3-3-2 bit RGB [1]- Interacting with hardware that uses certain bits in a register to control behavior - for example, Arduino digital ports [2]- Doing arithmetic extra-fast with numbers that are powers of two - modulo becomes bitwise-and, division and multiplication become shifts
 The only time I find myself using bit manipulations in production code is when I am writing a library that uses an int to represent a set of option flags (where each bit is a boolean flag). It's pretty common in C libs to define some constants like int FLAG_1 = 1; int FLAG_2 = 2; int FLAG_3 = 4; //These would actually probably be formatted like (1 << N) ... And then pass in your options to a function like f(... FLAG_1+FLAG_2) so you can have an optional number of flags in a single argument.
 To do this exercise in javascript is pretty silly. But as a systems developer, I can tell you that bit manipulation happens all over the place in my line of work.
 The first time I used bit manipulation in web apps was for declaring user roles and permissions. A single integer field in your database can be used to declare any combination of up to 32 different capabilities/permissions for users by setting the individual bits in the value. Also, when dealing with colors as an integer value, you can isolate your ARGB values using bit shifting.
 This is mostly done in lower level code, so no wonder you have never bumped into it. One easy example, though is passing a set of binary options around, which is useful, for example, when you want to say "enable those features, and disable those other features". You can pack up to 32 options into a 4-byte integer.
 That will get fairly hairy, given that JavaScript doesn't have integers. x should be expected to be a IEEE 754 double (can one discriminate between NaNs in JavaScript?), but might also be a string (does "@" have one bit set?, or might it be EBCDIC?), an array, an object, null or undefined (do those have a bit pattern?)
 JavaScript supports various bit manipulation operators, which treat their operands as 32 bits (and return a numerical value).
 Because of that, I am not sure can use them to count the number of 1 bits in a JavaScript number. For example, if x&0xFFFFFFFF equals 356621, what do you know of x?
 The validator doesn't work at all.
 Sorry. Fixed.
 This is making me crazy, does someone have the solution?The best I could come up with is:function one_bit(x) { return 1 >>> (x & (x + ~0)) }("x + ~0" is equivalent to "x - 1")But this doesn't work when x is 0 All other case should be fine, afaik.
 That would work if the `>>>` operator worked as expected, but actually it just looks at the first 5 bits of the second variable so that `a>>>b` is effectively equivalent to `a>>>(b&31)`. His description of it is wrong.
 Here's an ugly solution.
 Nice. Find a shorter one now (afaik, the best score is 12 and it's like in golf, smaller scores are better).

Search: