* Why can't you return a raw C-style array :cpp:arrays:elaborative-why: :PROPERTIES: :ANKI_NOTE_TYPE: Basic :END: ** Front Why is returning a raw C-style array illegal in C++? ** Back Arrays /decay to pointers/ when passed around. Returning one would return a pointer to a local variable that's destroyed when the function returns — undefined behavior. The compiler forbids it. #+begin_src c++ int arr[26]; return arr; // ❌ returns pointer to dead stack memory #+end_src * std::array vs int[]: key differences :cpp:arrays:comparative: :PROPERTIES: :ANKI_NOTE_TYPE: Basic :END: ** Front How does ~std::array~ differ from ~int[26]~? ** Back | Feature | ~int[26]~ | ~std::array~ | |----------------------+-----------+---------------------| | Return by value | No | Yes | | Copy/assign | No | Yes | | Knows its own size | No | ~.size()~ | | Works with STL algos | No | Yes | | Zero-init with ~{}~ | No | Yes | Same machine code — zero overhead. * Task: return character frequency count from string :cpp:arrays:production: :PROPERTIES: :ANKI_NOTE_TYPE: Basic :END: ** Front Write a function that counts character frequencies in a string and returns the result. The string has only lowercase letters. ** Back #+begin_src c++ std::array calc(const std::string& str) { std::array arr{}; for (char c : str) { arr[c - 'a']++; } return arr; } #+end_src * Stack vs heap: std::array vs std::vector :cpp:arrays:comparative: :PROPERTIES: :ANKI_NOTE_TYPE: Basic :END: ** Front Where is the data stored for ~std::array~ vs ~std::vector(26)~? ** Back - ~std::array~ → *stack*. Data stored inline inside the struct, no dynamic allocation. - ~std::vector~ → *heap*. Internally calls ~new[]~, pointer indirection. Rule: use ~std::array~ when size is known at compile time (fast, no alloc). Use ~std::vector~ when size is dynamic. * Why const in const T& parameter :cpp:references:elaborative-why: :PROPERTIES: :ANKI_NOTE_TYPE: Basic :END: ** Front Why write ~const std::string&~ instead of ~std::string&~ for a read-only parameter? ** Back Two reasons: 1. *Enforces intent* — compiler catches accidental mutation inside the function. 2. *Allows more callers* — without ~const~, you can't pass temporaries or const variables: #+begin_src c++ calc("hello"); // ❌ without const const std::string s = "hi"; calc(s); // ❌ without const #+end_src * Why & instead of pass-by-value for string param :cpp:references:elaborative-why: :PROPERTIES: :ANKI_NOTE_TYPE: Basic :END: ** Front Why pass ~std::string~ by reference (~const std::string&~) instead of by value? ** Back ~std::string~ contains a heap allocation. Passing by value copies it — allocating memory and copying every character. #+begin_src c++ calc(std::string str) // copies entire string — heap allocation calc(const std::string& str) // passes address only — no copy #+end_src For a function that only reads, the copy is pure waste. * Reference vs pointer for function parameters :cpp:references:comparative: :PROPERTIES: :ANKI_NOTE_TYPE: Basic :END: ** Front When do you use ~T&~ vs ~T*~ for a function parameter? ** Back | Feature | Reference ~T&~ | Pointer ~T*~ | |-------------------+----------------+--------------------| | Can be null | No | Yes | | Dereference | ~str[i]~ | ~(*str)[i]~ | | Can be reassigned | No | Yes | Use ~T*~ when null is a valid input. Use ~T&~ when the value is *guaranteed to exist* — which is almost always the case for function parameters. * Rule of thumb: C++ function parameter types :cpp:references:factual: :PROPERTIES: :ANKI_NOTE_TYPE: Basic :END: ** Front What's the general rule for choosing parameter types in C++ functions? ** Back - *Non-trivial types* (~string~, ~vector~, structs): use ~const T&~ for read-only, ~T&~ for mutating - *Cheap types* (~int~, ~char~, ~bool~, pointers): pass by value - *Nullable input*: use ~T*~ - *Guaranteed non-null*: use ~T&~