rapid spire
I have a struct
Struct {
Username: [32:0]u8,
Email: [65:0]u8
}
When I check the contents it's 'biom4st3r\0' followed by 0xaa's, but when I print it via client.stream.writer().print("{s}", .{data.username}) ,with client being a StreamServer.Connection, it comes out as biom4st3r followed by a bunch of Unicode junk.
I have a function that just returns a slice ending at the sentinel, but I thought I had it working earlier without the function
willow surge
does client.stream.writer().print("{s}", .{&data.username}) work?
rapid spire
I also tried that and got the same result
valid path
[32:0]u8 is 32 u8s followed by a guarateed 0, if you only want up to the first zero, use std.mem.sliceTo()
willow surge
whats funny is that this works: std.debug.print("{s}\n", .{@as([*:0]u8, &str)});, but thats just cuz it calls std.mem.span for you lol
sliceTo is def the way to go tho
rapid spire
It just seems strange that fmt wouldnt take the sentinel into account
valid path
the sentinel is only guaranteed to be at the last index
a sentinel terminated slice or array is allowed to contain the sentinel as part of it's content, because the length is independently known
rapid spire
It's not just printing the sentinel. It's printing the whole buffer including everything after the sentinel
valid path
it's printing the entire array, again, the array doesn't end at the first sentinel
tall grove
It's not just printing the sentinel. It's printing the whole buffer including ev...
The sentinel element of a [32:0]u8 is, by definition, at index 32
This is an important distinction between sentineled arrays and sentineled pointers like [*:0]u8. The latter type just has a semantic meaning of "I promise there will be a sentinel 0 at some point", whereas sentineled arrays are actually different in memory to non-sentineled arrays: they store an extra element
[10:0]u8 is like a [11]u8 but Zig makes sure that the last element (arr[10]) is initialized to 0