Timeline for Floating Point XOR
Current License: CC BY-SA 4.0
25 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Sep 17, 2019 at 23:31 | comment | added | Nick Kennedy | @PeterCordes I think I’ve modified my answer in accordance with this - thanks! | |
| Sep 17, 2019 at 23:30 | history | edited | Nick Kennedy | CC BY-SA 4.0 |
added 59 characters in body
|
| Sep 17, 2019 at 21:11 | comment | added | S.S. Anne | I have a better solution (but it's C-only). | |
| Sep 17, 2019 at 7:43 | comment | added | Peter Cordes |
You don't need a comma or GNU C statement-expression here; the golf rules allow returning a value by modifying an input operand by reference. Also, since this is golf, this version doesn't strictly need the () safety parens. The "caller" just has to make sure not to use it as part of a larger expression. It already has lots of other weirdness (like taking args by reference and destroying one of them).
|
|
| Sep 17, 2019 at 7:26 | comment | added | Peter Cordes |
Note that this is only "safe" in debug mode, or if you use gcc -fno-strict-aliasing. Maybe you should have specified C++ (MSVC) if you wanted to use type punning via pointer casting. Although it does work on G++ with optimization disabled. BTW, this also requires a C++ implementation where sizeof(int) == sizeof(float), so it won't work on G++ for AVR or MSP430 for example. (TIO runs on x86, which uses 32-bit int and float like all "normal" modern CPUs).
|
|
| Sep 16, 2019 at 18:36 | comment | added | Nick Kennedy | @PurpleP no because it takes pointers to floats as its arguments. | |
| Sep 16, 2019 at 18:22 | comment | added | Purple P |
Does this work on expressions like f(3.0+1.2,0)?
|
|
| Sep 16, 2019 at 7:12 | history | edited | Nick Kennedy | CC BY-SA 4.0 |
added 33 characters in body
|
| Sep 15, 2019 at 19:35 | history | edited | Nick Kennedy | CC BY-SA 4.0 |
added 17 characters in body
|
| Sep 15, 2019 at 19:33 | comment | added | Nishioka | 37 bytes after merging with @AZTECCO's solution. | |
| Sep 15, 2019 at 19:23 | comment | added | AZTECCO | Saves 2 more | |
| Sep 15, 2019 at 19:11 | comment | added | Nishioka |
Actually I've missed the obvious, so another -8 bytes: #define f(x,y)({*(int*)x^=*(int*)y;*x;})
|
|
| Sep 15, 2019 at 19:03 | history | edited | Nick Kennedy | CC BY-SA 4.0 |
added 10 characters in body
|
| Sep 15, 2019 at 19:00 | comment | added | Nishioka |
Taking into consideration @Neil's suggestion it would get to 48 bytes: #define f(x,y)({*(int*)x^=*(int*)y;*(float*)x;})
|
|
| Sep 15, 2019 at 18:54 | comment | added | AZTECCO | 62 | |
| Sep 15, 2019 at 18:47 | comment | added | Neil |
Since you're using pointers, is it legal to use one of the inputs as the output? If so, you could write (*(int*)x^=*(int*)y).
|
|
| Sep 15, 2019 at 18:41 | history | edited | Nick Kennedy | CC BY-SA 4.0 |
deleted 26 characters in body; deleted 3 characters in body
|
| Sep 15, 2019 at 18:34 | comment | added | Nishioka |
With gcc extensions, 54 bytes: #define f(x,y)({int z=*(int*)x^*(int*)y;*(float*)&z;})
|
|
| Sep 15, 2019 at 18:33 | comment | added | Nishioka |
Function-like macro, 63 bytes: #define f(x,y)[&]{int z=*(int*)x^*(int*)y;return*(float*)&z;}()
|
|
| Sep 15, 2019 at 18:25 | history | edited | Nick Kennedy | CC BY-SA 4.0 |
added 54 characters in body
|
| Sep 15, 2019 at 17:07 | comment | added | Nishioka |
-2 bytes: using F=float;F f(F*x,F*y){int z=*(int*)x^*(int*)y;return*(F*)&z;}
|
|
| Sep 15, 2019 at 17:00 | history | edited | Nick Kennedy | CC BY-SA 4.0 |
added 53 characters in body
|
| Sep 15, 2019 at 16:46 | comment | added | Arnauld |
You can save 4 more bytes with z=*(int*)x^*(int*)y;.
|
|
| Sep 15, 2019 at 16:40 | comment | added | 12Me21 | You can remove the linebreaks to save 2 chars, and this also works in C | |
| Sep 15, 2019 at 16:08 | history | answered | Nick Kennedy | CC BY-SA 4.0 |