Don wrote:
bearophile wrote:
Thank you for your answers Simen kjaeraas.
No need for them when we have this: S s = S( 4 );<
If you have to initialize an array of many structs you have to repeat
the name of the struct many times, this is redundant, and takes more
space (and if you don't use an IDE it needs more time to type):
struct Vector2 { double x, y; }
Vector2[] data1 = [{1,2}, {3,4}, {0,1}, {1,2}, {2,3}, {3,4}, {4,5},
{5,6}, {6,7},
{7,8}, {8,9}, {9,10}, {10,11}, {11,12}, {12,13},
{13,14},
{14,15}, {15,16}, {16,17}, {17,18}, {18,19}, {19,20}];
Vector2[] data2 = [Vector2(0,1), Vector2(1,2), Vector2(2,3),
Vector2(3,4),
Vector2(4,5), Vector2(5,6), Vector2(6,7),
Vector2(7,8),
Vector2(8,9), Vector2(9,10), Vector2(10,11),
Vector2(11,12),
Vector2(12,13), Vector2(13,14), Vector2(14,15),
Vector2(15,16),
Vector2(16,17), Vector2(17,18), Vector2(18,19),
Vector2(19,20)];
void main() {}
That's interesting. I would tend to use a function to initialize such
things, though, so I suspect such use cases are quite rare. Note that
C-style struct initializers add a lot of complexity to the language.
Literals are allocated in the static data segment, and thus should be
immutable.<
In Python I am used to modify the contents of array (list) literals,
for example I can initialize them to some values and then change them.
So I'd like array literals to be mutable, but I can live with them
being immutable and dup them where necessary.
Currently, any use of array literals in D is a performance disaster.
Two days ago, I found that by changing one line of code in my app from:
immutable int [] setbits = [ 1, 2, 4, 8, 16, 32, 64, 128];
into
static immutable int [] setbits = [ 1, 2, 4, 8, 16, 32, 64, 128];
my entire app became about 5 times faster. That is ridiculous.
And such a clumsy syntax for something so simple.
Is this is bugzilla yet? FWIW I also submitted a related performance bug.
Andrei