## digitalmars.D.learn - integer max, min?

- Paolo Invernizzi <arathorn NO_SPAMfastwebnet.it> Oct 02 2007
- BLS <nanali nospam-wanadoo.fr> Oct 02 2007
- downs <default_357-line yahoo.de> Oct 02 2007
- Paolo Invernizzi <arathorn NO_SPAMfastwebnet.it> Oct 02 2007
- Nathan Reed <nathaniel.reed gmail.com> Oct 02 2007
- downs <default_357-line yahoo.de> Oct 02 2007
- Paolo Invernizzi <arathorn NO_SPAMfastwebnet.it> Oct 02 2007
- Bill Baxter <dnewsgroup billbaxter.com> Oct 02 2007
- "Steven Schveighoffer" <schveiguy yahoo.com> Oct 02 2007
- Paolo Invernizzi <arathorn NO_SPAMfastwebnet.it> Oct 02 2007

Silly question... Where are in Phobos the max and min for integers? Thanks, Paolo

Oct 02 2007

Paolo Invernizzi schrieb:Silly question... Where are in Phobos the max and min for integers? Thanks, Paolo

http://www.digitalmars.com/d/property.html HTH Bjoern

Oct 02 2007

BLS wrote:Paolo Invernizzi schrieb:Silly question... Where are in Phobos the max and min for integers? Thanks, Paolo

http://www.digitalmars.com/d/property.html HTH Bjoern

template supertype(T...) { static assert (T.length); static if (T.length==1) alias T[0] supertype; else static if (T.length==2) alias typeof(T[0].init+T[1].init) supertype; else alias supertype!( typeof(T[0].init+T[1].init), T[2..$] ) supertype; } supertype!(T) max(T...)(T t) { static assert(T.length); supertype!(T) res=t[0]; foreach (v; t[1..$]) if (v>res) res=v; return res; } supertype!(T) min(T...)(T t) { static assert(T.length); supertype!(T) res=t[0]; foreach (v; t[1..$]) if (v<res) res=v; return res; } That should work. Have fun! --downs

Oct 02 2007

God! It's an overkill! There's not in Phobos something like fmin fmax but for integers? Or must I always convert the result back from real? Cheers, Paolo downs wrote:BLS wrote:Paolo Invernizzi schrieb:Silly question... Where are in Phobos the max and min for integers? Thanks, Paolo

http://www.digitalmars.com/d/property.html HTH Bjoern

template supertype(T...) { static assert (T.length); static if (T.length==1) alias T[0] supertype; else static if (T.length==2) alias typeof(T[0].init+T[1].init) supertype; else alias supertype!( typeof(T[0].init+T[1].init), T[2..$] ) supertype; } supertype!(T) max(T...)(T t) { static assert(T.length); supertype!(T) res=t[0]; foreach (v; t[1..$]) if (v>res) res=v; return res; } supertype!(T) min(T...)(T t) { static assert(T.length); supertype!(T) res=t[0]; foreach (v; t[1..$]) if (v<res) res=v; return res; } That should work. Have fun! --downs

Oct 02 2007

Paolo Invernizzi wrote:God! It's an overkill! There's not in Phobos something like fmin fmax but for integers? Or must I always convert the result back from real?

(a) If you're talking about the maximum and minimum values an int can take on, they are named int.max and int.min. (b) If you're talking about functions max(a,b) and min(a,b) that return the max and min of their arguments, see Downs' post, which is a bit overkill if you just want it on numeric types, but will (I think) do lexicographic comparison for string and array types. If you just want to use it with numeric types (or types that have opCmp), this will do: T min(T) (T a, T b) { return (a < b) ? a : b; } T max(T) (T a, T b) { return (a > b) ? a : b; } Thanks, Nathan

Oct 02 2007

Nathan Reed wrote:(b) If you're talking about functions max(a,b) and min(a,b) that return the max and min of their arguments, see Downs' post, which is a bit overkill if you just want it on numeric types, but will (I think) do lexicographic comparison for string and array types.

Actually, it only works with numbers. Most of the complicated-looking code just determines the type that can hold all possible results (that's what the supertype stuff is for). That's also the reason it won't work with strings and arrays - they don't define "+". Sorry. ^^ But yeah, it is a bit overkill. Gets handy when you want to min/max more than two values of different types. --downs

Oct 02 2007

Natan, I'm talking about the (b) case, sorry for the confusion. I know I can use a template, I'm just wondering if THAT template was in some Phobos module, as I cannot find it! I'm already including std.math, so I was only wondering why... Cheers, Paolo Nathan Reed wrote:Paolo Invernizzi wrote:God! It's an overkill! There's not in Phobos something like fmin fmax but for integers? Or must I always convert the result back from real?

(a) If you're talking about the maximum and minimum values an int can take on, they are named int.max and int.min. (b) If you're talking about functions max(a,b) and min(a,b) that return the max and min of their arguments, see Downs' post, which is a bit overkill if you just want it on numeric types, but will (I think) do lexicographic comparison for string and array types. If you just want to use it with numeric types (or types that have opCmp), this will do: T min(T) (T a, T b) { return (a < b) ? a : b; } T max(T) (T a, T b) { return (a > b) ? a : b; } Thanks, Nathan

Oct 02 2007

Paolo Invernizzi wrote:Natan, I'm talking about the (b) case, sorry for the confusion. I know I can use a template, I'm just wondering if THAT template was in some Phobos module, as I cannot find it! I'm already including std.math, so I was only wondering why...

Nope, crazy as it may seem it's not there. There's also no standard swap() function like: void swap(T)(ref T a, ref T b) { T tmp = a; a = b; b = tmp; } --bb

Oct 02 2007

"Paolo Invernizzi" <arathorn NO_SPAMfastwebnet.it> wrote in message news:fdt65v$1306$1 digitalmars.com...Silly question... Where are in Phobos the max and min for integers? Thanks, Paolo

try x < y ? x : y for min(x, y) and x > y ? x : y for max(x, y) -Steve

Oct 02 2007

Thanks Steven, I know the ternary operator, and that's good for simple variable comparison, but if you use it in complex expressions, say also more than one times, sometimes it's not so readable as plain old min/max. Paolo. Steven Schveighoffer wrote:"Paolo Invernizzi" <arathorn NO_SPAMfastwebnet.it> wrote in message news:fdt65v$1306$1 digitalmars.com...Silly question... Where are in Phobos the max and min for integers? Thanks, Paolo

try x < y ? x : y for min(x, y) and x > y ? x : y for max(x, y) -Steve

Oct 02 2007