A buffer is temporary storage for data. In Zig, functions often accept a caller-provided buffer as a mutable byte slice, []u8. For example, a stdout writer can collect output in a buffer, reducing the number of blocking writes. section 2.2.4
An array declared with undefined can supply buffer space without initializing its contents. Write data before reading it: reading an uninitialized value produces unpredictable results and can create security problems. section 2.2.4
Listing 2.12 Hello world, with buffers explained
const std = @import("std");
pub fn main(init: std.process.Init) !void {
const io = init.io;
const stdout_file = std.Io.File.stdout();
var stdout_buffer: [1024]u8 = undefined; #1
var stdout_writer = stdout_file.writer(io, &stdout_buffer); #2
const stdout = &stdout_writer.interface;
try stdout.print("All your codebase still belong to us!\n", .{}); #3
try stdout.flush(); #4
}
This stdout writer receives a 1,024-byte array, prints into the buffered writer, and flushes the output. section 2.2.4
Buffers can also provide short-lived memory across a C interface. Pass a pointer and length, turn them into a Zig slice, and initialize std.heap.FixedBufferAllocator with that slice. Its allocator interface lets code build an ingredient array in caller-provided memory; an append that exceeds the buffer returns would_overflow. section 5.4.2
Listing 5.19 Retrieving recipes with Benito
export fn recipe_get(
book: *const RecipeBook,
recipe_id: RecipeId,
buf_ptr: [*]u8, #1
buf_len: usize, #1
result: *RetrievedRecipe,
) Success {
const buffer = buf_ptr[0..buf_len]; #2
var fba: std.heap.FixedBufferAllocator = .init(buffer); #3
const buf_ally = fba.allocator(); #3
const recipe = book.getRecipe(recipe_id) catch |e| {#4
return translateErr(e);
};
var ingredients_copy: std.ArrayList(RetrievedIngredient) = .empty; #5
var ingredients = book.iterateIngredients(recipe_id);
while (ingredients.next()) |ing| {
ingredients_copy.append(buf_ally, .{ #6
.quantity = ing.quantity.ptr,
.name = ing.name.ptr,
}) catch return .would_overflow; #7
}
result.* = .{ #8
.name = recipe.name.ptr,
.author = recipe.author.ptr,
.instructions = recipe.instructions.ptr,
.ingredients = ingredients_copy.items.ptr,
.ingredient_count = ingredients_copy.items.len,
};
return .ok; #9
}
The recipe retrieval function converts a pointer and length into a buffer, then uses a fixed buffer allocator to build the ingredient array. section 5.4.2
Buffer size is a practical trade-off. Reading an entire file into an allocated buffer before writing it to stdout requires an allocation large enough for the file. Avoiding that allocation removes a potential out-of-memory failure at that stage. section 3.4.4

