Prefixes are one of the most important y_amx concepts used throughout YSI for very efficient filtering and finding of functions. For example, the core of y_hooks looks like:
#define hook%0(%1) @yH_%0(%1)
while ((idx = AMX_GetPublicEntryPrefix(idx, entry, _A<@yH_>)))
{
// Do something with the found hook.
The define adds @yH_ to the start of every hook function, and the loop then looks for all functions starting with those four bytes. The four part is important because that's the size of a cell, and in pawn @ first makes the function public, so this prefix serves double duty. The macros are also optimised for prefixes with @ and _ in for better support of the old compiler. Further, to support 64-bit cells the prefixes are extended as minimally as possible to eight bytes, which requires declaring the functions slightly differently:
#define hook%0(%1) _F<@yH_>%0(%1)_F<> will either return the prefix exactly, or with padding bytes for larger cell sizes.
_A<> converts the input to a number representing those four characters in C string order, which is not the same as pawn packed string order:
new prefx = _A<code>;Becomes:
new prefix = ('c') | ('o' << 8) | ('d' << 16) | ('e' << 24);There is also _C<> if you want the pawn equivalent:
new prefx = _C<code>;Becomes:
new prefix = ('c' << 24) | ('o' << 16) | ('d' << 8) | ('e');