From cbcb570248a5297a0644fc22357bf81aac54545a Mon Sep 17 00:00:00 2001 From: Chris Boesch Date: Sun, 21 Jun 2026 16:16:31 +0200 Subject: expanded explanation about self vs *self --- exercises/047_methods.zig | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) (limited to 'exercises') diff --git a/exercises/047_methods.zig b/exercises/047_methods.zig index 3221ca2..3e88719 100644 --- a/exercises/047_methods.zig +++ b/exercises/047_methods.zig @@ -37,6 +37,24 @@ // self, others use a lowercase version of the type name, but feel // free to use whatever is most appropriate. // +// "But hold on," you say, eyeing a() and b() suspiciously, "why +// does one take 'self' and another take '*self'?" Sharp eye! +// +// It all comes down to a single question: does the function need +// to CHANGE the struct? +// +// * Needs to change it? Take a pointer (*Bar). Without it you'd +// be scribbling on a COPY, and your changes would evaporate the +// instant the function returns. Poof. +// * Only reads it? Plain Bar is just fine. +// (For a big, bulky struct you might still write *const Bar to +// avoid copying it around, but for small ones a copy is cheap.) +// +// You'll see this below: zap() takes 'self: HeatRay' by value +// because it only reads the ray's damage, but it takes the alien +// as '*Alien' because zapping is supposed to HURT - and that means +// changing the alien's health for real, not on a throwaway copy. +// // Okay, you're armed. // // Now, please zap the alien structs until they're all gone or -- cgit v1.2.3