source: doc/rob_thesis/ctordtor.tex @ 122aecd

Last change on this file since 122aecd was 93afb96, checked in by Rob Schluntz <rschlunt@…>, 7 years ago

fix typos in thesis

  • Property mode set to 100644
File size: 77.1 KB
2\chapter{Constructors and Destructors}
5% TODO: discuss move semantics; they haven't been implemented, but could be. Currently looking at alternative models. (future work)
7% TODO: as an experiment, implement Andrei Alexandrescu's ScopeGuard
8% doesn't seem possible to do this without allowing ttype on generic structs?
10% If a Cforall constructor is in scope, C style initialization is
11% disabled by default.
12% * initialization rule: if any constructor is in scope for type T, try
13%   to find a matching constructor for the call. If there are no
14%   constructors in scope for type T, then attempt to fall back on
15%   C-style initialization.
16% + if this rule was not in place, it would be easy to accidentally
17%   use C-style initialization in certain cases, which could lead to
18%   subtle errors [2]
19% - this means we need special syntax if we want to allow users to force
20%   a C-style initialization (to give users more control)
21% - two different declarations in the same scope can be implicitly
22%   initialized differently. That is, there may be two objects of type
23%   T that are initialized differently because there is a constructor
24%   definition between them. This is not technically specific to
25%   constructors.
27% C-style initializers can be accessed with @= syntax
28% + provides a way to get around the requirement of using a constructor
29%   (for advanced programmers only)
30% - can break invariants in the type => unsafe
31% * provides a way of asserting that a variable is an instance of a
32%   C struct (i.e. a POD struct), and so will not be implicitly
33%   destructed (this can be useful at times, maybe mitigates the need
34%   for move semantics?) [3]
35% + can modernize a code base one step at a time
37% Cforall constructors can be used in expressions to initialize any
38% piece of memory.
39% + malloc() { ... } calls the appropriate constructor on the newly
40%   allocated space; the argument is moved into the constructor call
41%   without taking its address [4]
42% - with the above form, there is still no way to ensure that
43%   dynamically allocated objects are constructed. To resolve this,
44%   we might want a stronger "new" function which always calls the
45%   constructor, although how we accomplish that is currently still
46%   unresolved (compiler magic vs. better variadic functions?)
47% + This can be used as a placement syntax [5]
48% - can call the constructor on an object more than once, which could
49%   cause resource leaks and reinitialize const fields (can try to
50%   detect and prevent this in some cases)
51%   * compiler always tries to implicitly insert a ctor/dtor pair for
52%     non-@= objects.
53%     * For POD objects, this will resolve to an autogenerated or
54%       intrinsic function.
55%     * Intrinsic functions are not automatically called. Autogenerated
56%       are, because they may call a non-autogenerated function.
57%     * destructors are automatically inserted at appropriate branches
58%       (e.g. return, break, continue, goto) and at the end of the block
59%       in which they are declared.
60%   * For @= objects, the compiler never tries to interfere and insert
61%     constructor and destructor calls for that object. Copy constructor
62%     calls do not count, because the object is not the target of the copy
63%     constructor.
65% A constructor is declared with the name ?{}
66% + combines the look of C initializers with the precedent of ?() being
67%   the name for the function call operator
68% + it is possible to easily search for all constructors in a project
69%   and immediately know that a function is a constructor by seeing the
70%   name "?{}"
72% A destructor is declared with the name ^?{}
73% + name mirrors a constructor's name, with an extra symbol to
74%   distinguish it
75% - the symbol '~' cannot be used due to parsing conflicts with the
76%   unary '~' (bitwise negation) operator - this conflict exists because
77%   we want to allow users to write ^x{}; to destruct x, rather than
78%   ^?{}(&x);
80% The first argument of a constructor must be a pointer. The constructed
81% type is the base type of the pointer. E.g. void ?{}(T *) is a default
82% constructor for a T.
83% + can name the argument whatever you like, so not constrained by
84%   language keyword "this" or "self", etc.
85% - have to explicitly qualify all object members to initialize them
86%   (e.g. this->x = 0, rather than just x = 0)
88% Destructors can take arguments other than just the destructed pointer
89% * open research problem: not sure how useful this is
91% Pointer constructors
92% + can construct separately compiled objects (opaque types) [6]
93% + orthogonal design, follows directly from the definition of the first
94%   argument of a constructor
95% - may require copy constructor or move constructor (or equivalent)
96%   for correct implementation, which may not be obvious to everyone
97% + required feature for the prelude to specify as much behavior as possible
98%   (similar to pointer assignment operators in this respect)
100% Designations can only be used for C-style initialization
101% * designation for constructors is equivalent to designation for any
102%   general function call. Since a function prototype can be redeclared
103%   many times, with arguments named differently each time (or not at
104%   all!), this is considered to be an undesirable feature. We could
105%   construct some set of rules to allow this behaviour, but it is
106%   probably more trouble than it's worth, and no matter what we choose,
107%   it is not likely to be obvious to most people.
109% Constructing an anonymous member [7]
110% + same as with calling any other function on an anonymous member
111%   (implicit conversion by the compiler)
112% - may be some cases where this is ambiguous => clarify with a cast
113%   (try to design APIs to avoid sharing function signatures between
114%   composed types to avoid this)
116% Default Constructors and Destructors are called implicitly
117% + cannot forget to construct or destruct an object
118% - requires special syntax to specify that an object is not to be
119%   constructed (@=)
120% * an object will not be implicitly constructed OR destructed if
121%   explicitly initialized like a C object (@= syntax)
122% * an object will be destructed if there are no constructors in scope
123%   (even though it is initialized like a C object) [8]
125% An object which changes from POD type to non POD type will not change
126% the semantics of a type containing it by composition
127% * That is, constructors will not be regenerated at the point where
128%   an object changes from POD type to non POD type, because this could
129%   cause a cascade of constructors being regenerated for many other
130%   types. Further, there is precedence for this behaviour in other
131%   facets of Cforall's design, such as how nested functions interact.
132% * This behaviour can be simplified in a language without declaration
133%   before use, because a type can be classified as POD or non POD
134%   (rather than potentially changing between the two at some point) at
135%   at the global scope (which is likely the most common case)
136% * [9]
138% Changes to polymorphic type classes
139% * dtype and ftype remain the same
140% * forall(otype T) is currently essentially the same as
141%   forall(dtype T | { @size(T); void ?=?(T *, T); }).
142%   The big addition is that you can declare an object of type T, rather
143%   than just a pointer to an object of type T since you know the size,
144%   and you can assign into a T.
145%   * this definition is changed to add default constructor and
146%     destructor declarations, to remain consistent with what type meant
147%     before the introduction of constructors and destructors.
148%     * that is, forall(type T) is now essentially the same as
149%       forall(dtype T | { @size(T); void ?=?(T *, T);
150%                          void ?{}(T *); void ^?{}(T *); })
151%     + this is required to make generic types work correctly in
152%       polymorphic functions
153%     ? since declaring a constructor invalidates the autogenerated
154%       routines, it is possible for a type to have constructors, but
155%       not default constructors. That is, it might be the case that
156%       you want to write a polymorphic function for a type which has
157%       a size, but non-default constructors? Some options:
158%       * declaring a constructor as a part of the assertions list for
159%         a type declaration invalidates the default, so
160%         forall(otype T | { void ?{}(T *, int); })
161%         really means
162%         forall(dtype T | { @size(T); void ?=?(T *, T);
163%                            void ?{}(T *, int); void ^?{}(T *); })
164%       * force users to fully declare the assertions list like the
165%         above in this case (this seems very undesirable)
166%       * add another type class with the current desugaring of type
167%         (just size and assignment)
168%       * provide some way of subtracting from an existing assertions
169%         list (this might be useful to have in general)
171% Implementation issues:
172% Changes to prelude/autogen or built in defaults?
173% * pointer ctors/dtors [prelude]
174%   * other pointer type routines are declared in the prelude, and this
175%     doesn't seem like it should be any different
176% * basic type ctors/dtors [prelude]
177%   * other basic type routines are declared in the prelude, and this
178%     doesn't seem like it should be any different
179% ? aggregate types [undecided, but leaning towards autogenerate]
180%   * prelude
181%     * routines specific to aggregate types cannot be predeclared in
182%       the prelude because we don't know the name of every
183%       aggregate type in the entire program
184%   * autogenerate
185%     + default assignment operator is already autogenerated for
186%       aggregate types
187%       * this seems to lead us in the direction of autogenerating,
188%         because we may have a struct which contains other objects
189%         that require construction [10]. If we choose not to
190%         autogenerate in this case, then objects which are part of
191%         other objects by composition will not be constructed unless
192%         a constructor for the outer type is explicitly defined
193%       * in this case, we would always autogenerate the appropriate
194%         constructor(s) for an aggregate type, but just like with
195%         basic types, pointer types, and enum types, the constructor
196%         call can be elided when when it is not necessary.
197%     + constructors will have to be explicitly autogenerated
198%       in the case where they are required for a polymorphic function,
199%       when no user defined constructor is in scope, which may make it
200%       easiest to always autogenerate all appropriate constructors
201%     - n+2 constructors would have to be generated for a POD type
202%       * one constructor for each number of valid arguments [0, n],
203%         plus the copy constructor
204%         * this is taking a simplified approach: in C, it is possible
205%           to omit the enclosing braces in a declaration, which would
206%           lead to a combinatorial explosion of generated constructors.
207%           In the interest of keeping things tractable, Cforall may be
208%           incompatible with C in this case. [11]
209%       * for non-POD types, only autogenerate the default and copy
210%         constructors
211%       * alternative: generate only the default constructor and
212%         special case initialization for any other constructor when
213%         only the autogenerated one exists
214%         - this is not very sensible, as by the previous point, these
215%           constructors may be needed for polymorphic functions
216%           anyway.
217%     - must somehow distinguish in resolver between autogenerated and
218%       user defined constructors (autogenerated should never be chosen
219%       when a user defined option exists [check first parameter], even
220%       if full signature differs) (this may also have applications
221%       to other autogenerated routines?)
222%     - this scheme does not naturally support designation (i.e. general
223%       functions calls do not support designation), thus these cases
224%       will have to be treated specially in either case
225%   * defaults
226%     * i.e. hardcode a new set of rules for some "appropriate" default
227%       behaviour for
228%     + when resolving an initialization expression, explicitly check to
229%       see if any constructors are in scope. If yes, attempt to resolve
230%       to a constructor, and produce an error message if a match is not
231%       found. If there are no constructors in scope, resolve to
232%       initializing each field individually (C-style)
233%     + does not attempt to autogenerate constructors for POD types,
234%       which can be seen as a space optimization for the program
235%       binary
236%     - as stated previously, a polymorphic routine may require these
237%       autogenerated constructors, so this doesn't seem like a big win,
238%       because this leads to more complicated logic and tracking of
239%       which constructors have already been generated
240%     - even though a constructor is not explicitly declared or used
241%       polymorphically, we might still need one for all uses of a
242%       struct (e.g. in the case of composition).
243%   * the biggest tradeoff in autogenerating vs. defaulting appears to
244%     be in where and how the special code to check if constructors are
245%     present is handled. It appears that there are more reasons to
246%     autogenerate than not.
248% --- examples
249% [1] As an example of using constructors polymorphically, consider a
250% slight modification on the foldl example I put on the mailing list a
251% few months ago:
253% context iterable(type collection, type element, type iterator) {
254%   void ?{}(iterator *, collection); // used to be makeIterator, but can
255%                             // idiomatically use constructor
256%   int hasNext(iterator);
257%   iterator ++?(iterator *);
258%   lvalue element *?(iterator);
259% };
262% forall(type collection, type element, type result, type iterator
263%   | iterable(collection, element, iterator))
264% result foldl(collection c, result acc,
265%     result (*reduce)(result, element)) {
266%   iterator it = { c };
267%   while (hasNext(it)) {
268%     acc = reduce(acc, *it);
269%     ++it;
270%   }
271%   return acc;
272% }
274% Now foldl makes use of the knowledge that the iterator type has a
275% single argument constructor which takes the collection to iterate
276% over. This pattern allows polymorphic code to look more natural
277% (constructors are generally preferred to named initializer/creation
278% routines, e.g. "makeIterator")
280% [2] An example of some potentially dangerous code that we don't want
281% to let easily slip through the cracks - if this is really what you
282% want, then use @= syntax for the second declaration to quiet the
283% compiler.
285% struct A { int x, y, z; }
286% ?{}(A *, int);
287% ?{}(A *, int, int, int);
289% A a1 = { 1 };         // uses ?{}(A *, int);
290% A a2 = { 2, 3 };      // C-style initialization -> no invariants!
291% A a3 = { 4, 5, 6 };   // uses ?{}(A *, int, int, int);
293% [3] Since @= syntax creates a C object (essentially a POD, as far as
294% the compiler is concerned), the object will not be destructed
295% implicitly when it leaves scope, nor will it be copy constructed when
296% it is returned. In this case, a memcpy should be equivalent to a move.
298% // Box.h
299% struct Box;
300% void ?{}(Box **, int};
301% void ^?{}(Box **);
302% Box * make_fortytwo();
304% //
305% Box * make_fortytwo() {
306%   Box *b @= {};
307%   (&b){ 42 }; // construct explicitly
308%   return b; // no destruction, essentially a move?
309% }
311% [4] Cforall's typesafe malloc can be composed with constructor
312% expressions. It is possible for a user to define their own functions
313% similar to malloc and achieve the same effects (e.g. Aaron's example
314% of an arena allocator)
316% // CFA malloc
317% forall(type T)
318% T * malloc() { return (T *)malloc(sizeof(T)); }
320% struct A { int x, y, z; };
321% void ?{}(A *, int);
323% int foo(){
324%   ...
325%   // desugars to:
326%   // A * a = ?{}(malloc(), 123);
327%   A * a = malloc() { 123 };
328%   ...
329% }
331% [5] Aaron's example of combining function calls with constructor
332% syntax to perform an operation similar to C++'s std::vector::emplace
333% (i.e. to construct a new element in place, without the need to
334% copy)
336% forall(type T)
337% struct vector {
338%   T * elem;
339%   int len;
340%   ...
341% };
343% ...
344% forall(type T)
345% T * vector_new(vector(T) * v) {
346%   // reallocate if needed
347%   return &v->elem[len++];
348% }
350% int main() {
351%   vector(int) * v = ...
352%   vector_new(v){ 42 };  // add element to the end of vector
353% }
355% [6] Pointer Constructors. It could be useful to use the existing
356% constructor syntax even more uniformly for ADTs. With this, ADTs can
357% be initialized in the same manor as any other object in a polymorphic
358% function.
360% // vector.h
361% forall(type T) struct vector;
362% forall(type T) void ?{}(vector(T) **);
363% // adds an element to the end
364% forall(type T) vector(T) * ?+?(vector(T) *, T);
366% //
367% // don't want to expose the implementation to the user and/or don't
368% // want to recompile the entire program if the struct definition
369% // changes
371% forall(type T) struct vector {
372%   T * elem;
373%   int len;
374%   int capacity;
375% };
377% forall(type T) void resize(vector(T) ** v) { ... }
379% forall(type T) void ?{}(vector(T) ** v) {
380%   vector(T) * vect = *v = malloc();
381%   vect->capacity = 10;
382%   vect->len = 0;
383%   vect->elem = malloc(vect->capacity);
384% }
386% forall(type T) vector(T) * ?+?(vector(T) *v, T elem) {
387%   if (v->len == v->capacity) resize(&v);
388%   v->elem[v->len++] = elem;
389% }
391% //
392% #include "adt.h"
393% forall(type T | { T ?+?(T, int); }
394% T sumRange(int lower, int upper) {
395%   T x;    // default construct
396%   for (int i = lower; i <= upper; i++) {
397%     x = x + i;
398%   }
399%   return x;
400% }
402% int main() {
403%   vector(int) * numbers = sumRange(1, 10);
404%   // numbers is now a vector containing [1..10]
406%   int sum = sumRange(1, 10);
407%   // sum is now an int containing the value 55
408% }
410% [7] The current proposal is to use the plan 9 model of inheritance.
411% Under this model, all of the members of an unnamed struct instance
412% become members of the containing struct. In addition, an object
413% can be passed as an argument to a function expecting one of its
414% base structs.
416% struct Point {
417%   double x;
418%   double y;
419% };
421% struct ColoredPoint {
422%   Point;        // anonymous member (no identifier)
423%                 // => a ColoredPoint has an x and y of type double
424%   int color;
425% };
427% ColoredPoint cp = ...;
428% cp.x = 10.3;    // x from Point is accessed directly
429% cp.color = 0x33aaff; // color is accessed normally
430% foo(cp);        // cp can be used directly as a Point
432% void ?{}(Point *p, double x, double y) {
433%   p->x = x;
434%   p->y = y;
435% }
437% void ?{}(ColoredPoint *cp, double x, double y, int color) {
438%   (&cp){ x, y };  // unambiguous, no ?{}(ColoredPoint*,double,double)
439%   cp->color = color;
440% }
442% struct Size {
443%   double width;
444%   double height;
445% };
447% void ?{}(Size *s, double w, double h) {
448%   p->width = w;
449%   p->height = h;
450% }
452% struct Foo {
453%   Point;
454%   Size;
455% }
457% ?{}(Foo &f, double x, double y, double w, double h) {
458%   // (&F,x,y) is ambiguous => is it ?{}(Point*,double,double) or
459%   // ?{}(Size*,double,double)? Solve with a cast:
460%   ((Point*)&F){ x, y };
461%   ((Size*)&F){ w, h };
462% }
464% [8] Destructors will be called on objects that were not constructed.
466% struct A { ... };
467% ^?{}(A *);
468% {
469%   A x;
470%   A y @= {};
471% } // x is destructed, even though it wasn't constructed
472%   // y is not destructed, because it is explicitly a C object
475% [9] A type's constructor is generated at declaration time using
476% current information about an object's members. This is analogous to
477% the treatment of other operators. For example, an object's assignment
478% operator will not change to call the override of a member's assignment
479% operator unless the object's assignment is also explicitly overridden.
480% This problem can potentially be treated differently in Do, since each
481% compilation unit is passed over at least twice (once to gather
482% symbol information, once to generate code - this is necessary to
483% achieve the "No declarations" goal)
485% struct A { ... };
486% struct B { A x; };
487% ...
488% void ?{}(A *);  // from this point on, A objects will be constructed
489% B b1;           // b1 and b1.x are both NOT constructed, because B
490%                 // objects are not constructed
491% void ?{}(B *);  // from this point on, B objects will be constructed
492% B b2;           // b2 and b2.x are both constructed
494% struct C { A x; };
495% // implicit definition of ?{}(C*), because C is not a POD type since
496% // it contains a non-POD type by composition
497% C c;            // c and c.x are both constructed
499% [10] Requiring construction by composition
501% struct A {
502%   ...
503% };
505% // declared ctor disables default c-style initialization of
506% // A objects; A is no longer a POD type
507% void ?{}(A *);
509% struct B {
510%   A x;
511% };
513% // B objects can not be C-style initialized, because A objects
514% // must be constructed => B objects are transitively not POD types
515% B b; // b.x must be constructed, but B is not constructible
516%      // => must autogenerate ?{}(B *) after struct B definition,
517%      // which calls ?{}(&b.x)
519% [11] Explosion in the number of generated constructors, due to strange
520% C semantics.
522% struct A { int x, y; };
523% struct B { A u, v, w; };
525% A a = { 0, 0 };
527% // in C, you are allowed to do this
528% B b1 = { 1, 2, 3, 4, 5, 6 };
529% B b2 = { 1, 2, 3 };
530% B b3 = { a, a, a };
531% B b4 = { a, 5, 4, a };
532% B b5 = { 1, 2, a, 3 };
534% // we want to disallow b1, b2, b4, and b5 in Cforall.
535% // In particular, we will autogenerate these constructors:
536% void ?{}(A *);             // default/0 parameters
537% void ?{}(A *, int);        // 1 parameter
538% void ?{}(A *, int, int);   // 2 parameters
539% void ?{}(A *, const A *);  // copy constructor
541% void ?{}(B *);             // default/0 parameters
542% void ?{}(B *, A);          // 1 parameter
543% void ?{}(B *, A, A);       // 2 parameters
544% void ?{}(B *, A, A, A);    // 3 parameters
545% void ?{}(B *, const B *);  // copy constructor
547% // we will not generate constructors for every valid combination
548% // of members in C. For example, we will not generate
549% void ?{}(B *, int, int, int, int, int, int);   // b1 would need this
550% void ?{}(B *, int, int, int);                  // b2 would need this
551% void ?{}(B *, A, int, int, A);                 // b4 would need this
552% void ?{}(B *, int, int, A, int);               // b5 would need this
553% // and so on
557% TODO: talk somewhere about compound literals?
559Since \CFA is a true systems language, it does not provide a garbage collector.
560As well, \CFA is not an object-oriented programming language, i.e. structures cannot have routine members.
561Nevertheless, one important goal is to reduce programming complexity and increase safety.
562To that end, \CFA provides support for implicit pre/post-execution of routines for objects, via constructors and destructors.
564% TODO: this is old. remove or refactor
565% Manual resource management is difficult.
566% Part of the difficulty results from not having any guarantees about the current state of an object.
567% Objects can be internally composed of pointers that may reference resources which may or may not need to be manually released, and keeping track of that state for each object can be difficult for the end user.
569% Constructors and destructors provide a mechanism to bookend the lifetime of an object, allowing the designer of a type to establish invariants for objects of that type.
570% Constructors guarantee that object initialization code is run before the object can be used, while destructors provide a mechanism that is guaranteed to be run immediately before an object's lifetime ends.
571% Constructors and destructors can help to simplify resource management when used in a disciplined way.
572% In particular, when all resources are acquired in a constructor, and all resources are released in a destructor, no resource leaks are possible.
573% This pattern is a popular idiom in several languages, such as \CC, known as RAII (Resource Acquisition Is Initialization).
575This chapter details the design of constructors and destructors in \CFA, along with their current implementation in the translator.
576Generated code samples have been edited to provide comments for clarity and to save on space.
578\section{Design Criteria}
580In designing constructors and destructors for \CFA, the primary goals were ease of use and maintaining backwards compatibility.
582In C, when a variable is defined, its value is initially undefined unless it is explicitly initialized or allocated in the static area.
584int main() {
585  int x;        // uninitialized
586  int y = 5;    // initialized to 5
587  x = y;        // assigned 5
588  static int z; // initialized to 0
591In the example above, @x@ is defined and left uninitialized, while @y@ is defined and initialized to 5.
592Next, @x@ is assigned the value of @y@.
593In the last line, @z@ is implicitly initialized to 0 since it is marked @static@.
594The key difference between assignment and initialization being that assignment occurs on a live object (i.e. an object that contains data).
595It is important to note that this means @x@ could have been used uninitialized prior to being assigned, while @y@ could not be used uninitialized.
596Use of uninitialized variables yields undefined behaviour, which is a common source of errors in C programs. % TODO: *citation*
598Declaration initialization is insufficient, because it permits uninitialized variables to exist and because it does not allow for the insertion of arbitrary code before the variable is live.
599Many C compilers give good warnings most of the time, but they cannot in all cases.
601int f(int *);  // never reads the parameter, only writes
602int g(int *);  // reads the parameter - expects an initialized variable
604int x, y;
605f(&x);  // okay - only writes to x
606g(&y);  // will use y uninitialized
608Other languages are able to give errors in the case of uninitialized variable use, but due to backwards compatibility concerns, this cannot be the case in \CFA.
610In C, constructors and destructors are often mimicked by providing routines that create and teardown objects, where the teardown function is typically only necessary if the type modifies the execution environment.
612struct array_int {
613  int * x;
615struct array_int create_array(int sz) {
616  return (struct array_int) { malloc(sizeof(int)*sz) };
618void destroy_rh(struct resource_holder * rh) {
619  free(rh->x);
622This idiom does not provide any guarantees unless the structure is opaque, which then requires that all objects are heap allocated.
624struct opqaue_array_int;
625struct opqaue_array_int * create_opqaue_array(int sz);
626void destroy_opaque_array(opaque_array_int *);
627int opaque_get(opaque_array_int *);  // subscript
629opaque_array_int * x = create_opaque_array(10);
630int x2 = opaque_get(x, 2);
632This pattern is cumbersome to use since every access becomes a function call.
633While useful in some situations, this compromise is too restrictive.
634Furthermore, even with this idiom it is easy to make mistakes, such as forgetting to destroy an object or destroying it multiple times.
636A constructor provides a way of ensuring that the necessary aspects of object initialization is performed, from setting up invariants to providing compile-time checks for appropriate initialization parameters.
637This goal is achieved through a guarantee that a constructor is called implicitly after every object is allocated from a type with associated constructors, as part of an object's definition.
638Since a constructor is called on every object of a managed type, it is impossible to forget to initialize such objects, as long as all constructors perform some sensible form of initialization.
640In \CFA, a constructor is a function with the name @?{}@.
641Every constructor must have a return type of @void@ and at least one parameter, the first of which is colloquially referred to as the \emph{this} parameter, as in many object-oriented programming-languages (however, a programmer can give it an arbitrary name).
642The @this@ parameter must have a pointer type, whose base type is the type of object that the function constructs.
643There is precedence for enforcing the first parameter to be the @this@ parameter in other operators, such as the assignment operator, where in both cases, the left-hand side of the equals is the first parameter.
644There is currently a proposal to add reference types to \CFA.
645Once this proposal has been implemented, the @this@ parameter will become a reference type with the same restrictions.
647Consider the definition of a simple type encapsulating a dynamic array of @int@s.
650struct Array {
651  int * data;
652  int len;
656In C, if the user creates an @Array@ object, the fields @data@ and @len@ are uninitialized, unless an explicit initializer list is present.
657It is the user's responsibility to remember to initialize both of the fields to sensible values.
658In \CFA, the user can define a constructor to handle initialization of @Array@ objects.
661void ?{}(Array * arr){
662  arr->len = 10;    // default size
663  arr->data = malloc(sizeof(int)*arr->len);
664  for (int i = 0; i < arr->len; ++i) {
665    arr->data[i] = 0;
666  }
668Array x;  // allocates storage for Array and calls ?{}(&x)
671This constructor initializes @x@ so that its @length@ field has the value 10, and its @data@ field holds a pointer to a block of memory large enough to hold 10 @int@s, and sets the value of each element of the array to 0.
672This particular form of constructor is called the \emph{default constructor}, because it is called on an object defined without an initializer.
673In other words, a default constructor is a constructor that takes a single argument, the @this@ parameter.
675In \CFA, a destructor is a function much like a constructor, except that its name is \lstinline!^?{}!.
676A destructor for the @Array@ type can be defined as such.
678void ^?{}(Array * arr) {
679  free(arr->data);
682Since the destructor is automatically called at deallocation for all objects of type @Array@, the memory associated with an @Array@ is automatically freed when the object's lifetime ends.
683The exact guarantees made by \CFA with respect to the calling of destructors are discussed in section \ref{sub:implicit_dtor}.
685As discussed previously, the distinction between initialization and assignment is important.
686Consider the following example.
688Array x, y;
689Array z = x;  // initialization
690y = x;        // assignment
692By the previous definition of the default constructor for @Array@, @x@ and @y@ are initialized to valid arrays of length 10 after their respective definitions.
693On line 3, @z@ is initialized with the value of @x@, while on line @4@, @y@ is assigned the value of @x@.
694The key distinction between initialization and assignment is that a value to be initialized does not hold any meaningful values, whereas an object to be assigned might.
695In particular, these cases cannot be handled the same way because in the former case @z@ does not currently own an array, while @y@ does.
697\begin{cfacode}[emph={other}, emphstyle=\color{red}]
698void ?{}(Array * arr, Array other) {  // copy constructor
699  arr->len = other.len;               // initialization
700  arr->data = malloc(sizeof(int)*arr->len)
701  for (int i = 0; i < arr->len; ++i) {
702    arr->data[i] =[i];     // copy from other object
703  }
705Array ?=?(Array * arr, Array other) { // assignment
706  ^?{}(arr);                          // explicitly call destructor
707  ?{}(arr, other);                    // explicitly call constructor
708  return *arr;
711The two functions above handle these cases.
712The first function is called a \emph{copy constructor}, because it constructs its argument by copying the values from another object of the same type.
713The second function is the standard copy-assignment operator.
714These four functions are special in that they control the state of most objects.
716It is possible to define a constructor that takes any combination of parameters to provide additional initialization options.
717For example, a reasonable extension to the array type would be a constructor that allocates the array to a given initial capacity and initializes the array to a given @fill@ value.
719void ?{}(Array * arr, int capacity, int fill) {
720  arr->len = capacity;
721  arr->data = malloc(sizeof(int)*arr->len);
722  for (int i = 0; i < arr->len; ++i) {
723    arr->data[i] = fill;
724  }
727In \CFA, constructors are called implicitly in initialization contexts.
729Array x, y = { 20, 0xdeadbeef }, z = y;
731In \CFA, constructor calls look just like C initializers, which allows them to be inserted into legacy C code with minimal code changes, and also provides a very simple syntax that veteran C programmers are familiar with.
732One downside of reusing C initialization syntax is that it isn't possible to determine whether an object is constructed just by looking at its declaration, since that requires knowledge of whether the type is managed at that point.
734This example generates the following code
736Array x;
737?{}(&x);                  // implicit default construct
738Array y;
739?{}(&y, 20, 0xdeadbeef);  // explicit fill construct
740Array z;
741?{}(&z, y);               // copy construct
742^?{}(&z);                 // implicit destruct
743^?{}(&y);                 // implicit destruct
744^?{}(&x);                 // implicit destruct
746Due to the way that constructor calls are interleaved, it is impossible for @y@ to be referenced before it is initialized, except in its own constructor.
747This loophole is minor and exists in \CC as well.
748Destructors are implicitly called in reverse declaration-order so that objects with dependencies are destructed before the objects they are dependent on.
751\label{sub:syntax} % TODO: finish this section
752There are several ways to construct an object in \CFA.
753As previously introduced, every variable is automatically constructed at its definition, which is the most natural way to construct an object.
755struct A { ... };
756void ?{}(A *);
757void ?{}(A *, A);
758void ?{}(A *, int, int);
760A a1;             // default constructed
761A a2 = { 0, 0 };  // constructed with 2 ints
762A a3 = a1;        // copy constructed
763// implicitly destruct a3, a2, a1, in that order
765Since constructors and destructors are just functions, the second way is to call the function directly.
767struct A { int a; };
768void ?{}(A *);
769void ?{}(A *, A);
770void ^?{}(A *);
772A x;               // implicitly default constructed: ?{}(&x)
773A * y = malloc();  // copy construct: ?{}(&y, malloc())
775?{}(&x);    // explicit construct x
776?{}(y, x);  // explit construct y from x
777^?{}(&x);   // explicit destroy x
778^?{}(y);    // explicit destroy y
780// implicit ^?{}(&y);
781// implicit ^?{}(&x);
783Calling a constructor or destructor directly is a flexible feature that allows complete control over the management of a piece of storage.
784In particular, constructors double as a placement syntax.
786struct A { ... };
787struct memory_pool { ... };
788void ?{}(memory_pool *, size_t);
790memory_pool pool = { 1024 };  // create an arena of size 1024
792A * a = allocate(&pool);      // allocate from memory pool
793?{}(a);                       // construct an A in place
795for (int i = 0; i < 10; i++) {
796  // reuse storage rather than reallocating
797  ^?{}(a);
798  ?{}(a);
799  // use a ...
802deallocate(&pool, a);         // return to memory pool
804Finally, constructors and destructors support \emph{operator syntax}.
805Like other operators in \CFA, the function name mirrors the use-case, in that the first $N$ arguments fill in the place of the question mark.
807struct A { ... };
808struct B { A a; };
810A x, y, * z = &x;
811(&x){}          // default construct
812(&x){ y }       // copy construct
813(&x){ 1, 2, 3 } // construct with 3 arguments
814z{ y };         // copy construct x through a pointer
815^(&x){}         // destruct
817void ?{}(B * b) {
818  (&b->a){ 11, 17, 13 };  // construct a member
821Constructor operator syntax has relatively high precedence, requiring parentheses around an address-of expression.
822Destructor operator syntax is actually an statement, and requires parentheses for symmetry with constructor syntax.
824\subsection{Function Generation}
825In \CFA, every type is defined to have the core set of four functions described previously.
826Having these functions exist for every type greatly simplifies the semantics of the language, since most operations can simply be defined directly in terms of function calls.
827In addition to simplifying the definition of the language, it also simplifies the analysis that the translator must perform.
828If the translator can expect these functions to exist, then it can unconditionally attempt to resolve them.
829Moreover, the existence of a standard interface allows polymorphic code to interoperate with new types seamlessly.
831To mimic the behaviour of standard C, the default constructor and destructor for all of the basic types and for all pointer types are defined to do nothing, while the copy constructor and assignment operator perform a bitwise copy of the source parameter (as in \CC).
833There are several options for user-defined types: structures, unions, and enumerations.
834To aid in ease of use, the standard set of four functions is automatically generated for a user-defined type after its definition is completed.
835By auto-generating these functions, it is ensured that legacy C code will continue to work correctly in every context where \CFA expects these functions to exist, since they are generated for every complete type.
837The generated functions for enumerations are the simplest.
838Since enumerations in C are essentially just another integral type, the generated functions behave in the same way that the builtin functions for the basic types work.
839% TODO: examples for enums
840For example, given the enumeration
842enum Colour {
843  R, G, B
846The following functions are automatically generated.
848void ?{}(enum Colour *_dst){
849  // default constructor does nothing
851void ?{}(enum Colour *_dst, enum Colour _src){
852  (*_dst)=_src;  // bitwise copy
854void ^?{}(enum Colour *_dst){
855  // destructor does nothing
857enum Colour ?=?(enum Colour *_dst, enum Colour _src){
858  return (*_dst)=_src; // bitwise copy
861In the future, \CFA will introduce strongly-typed enumerations, like those in \CC.
862The existing generated routines will be sufficient to express this restriction, since they are currently set up to take in values of that enumeration type.
863Changes related to this feature only need to affect the expression resolution phase, where more strict rules will be applied to prevent implicit conversions from integral types to enumeration types, but should continue to permit conversions from enumeration types to @int@.
864In this way, it will still be possible to add an @int@ to an enumeration, but the resulting value will be an @int@, meaning that it won't be possible to reassign the value into an enumeration without a cast.
866For structures, the situation is more complicated.
867For a structure @S@ with members @M$_0$@, @M$_1$@, ... @M$_{N-1}$@, each function @f@ in the standard set calls \lstinline{f(s->M$_i$, ...)} for each @$i$@.
868That is, a default constructor for @S@ default constructs the members of @S@, the copy constructor with copy construct them, and so on.
869For example given the struct definition
871struct A {
872  B b;
873  C c;
876The following functions are implicitly generated.
878void ?{}(A * this) {
879  ?{}(&this->b);  // default construct each field
880  ?{}(&this->c);
882void ?{}(A * this, A other) {
883  ?{}(&this->b, other.b);  // copy construct each field
884  ?{}(&this->c, other.c);
886A ?=?(A * this, A other) {
887  ?=?(&this->b, other.b);  // assign each field
888  ?=?(&this->c, other.c);
890void ^?{}(A * this) {
891  ^?{}(&this->c);  // destruct each field
892  ^?{}(&this->b);
895It is important to note that the destructors are called in reverse declaration order to resolve conflicts in the event there are dependencies among members.
897In addition to the standard set, a set of \emph{field constructors} is also generated for structures.
898The field constructors are constructors that consume a prefix of the struct's member list.
899That is, $N$ constructors are built of the form @void ?{}(S *, T$_{\text{M}_0}$)@, @void ?{}(S *, T$_{\text{M}_0}$, T$_{\text{M}_1}$)@, ..., @void ?{}(S *, T$_{\text{M}_0}$, T$_{\text{M}_1}$, ..., T$_{\text{M}_{N-1}}$)@, where members are copy constructed if they have a corresponding positional argument and are default constructed otherwise.
900The addition of field constructors allows structs in \CFA to be used naturally in the same ways that they could be used in C (i.e. to initialize any prefix of the struct), e.g., @A a0 = { b }, a1 = { b, c }@.
901Extending the previous example, the following constructors are implicitly generated for @A@.
903void ?{}(A * this, B b) {
904  ?{}(&this->b, b);
905  ?{}(&this->c);
907void ?{}(A * this, B b, C c) {
908  ?{}(&this->b, b);
909  ?{}(&this->c, c);
913For unions, the default constructor and destructor do nothing, as it is not obvious which member if any should be constructed.
914For copy constructor and assignment operations, a bitwise @memcpy@ is applied.
915In standard C, a union can also be initialized using a value of the same type as its first member, and so a corresponding field constructor is generated to perform a bitwise @memcpy@ of the object.
916An alterantive to this design is to always construct and destruct the first member of a union, to match with the C semantics of initializing the first member of the union.
917This approach ultimately feels subtle and unsafe.
918Another option is to, like \CC, disallow unions from containing members that are themselves managed types.
919This restriction is a reasonable approach from a safety standpoint, but is not very C-like.
920Since the primary purpose of a union is to provide low-level memory optimization, it is assumed that the user has a certain level of maturity.
921It is therefore the responsibility of the user to define the special functions explicitly if they are appropriate, since it is impossible to accurately predict the ways that a union is intended to be used at compile-time.
923For example, given the union
925union X {
926  Y y;
927  Z z;
930The following functions are automatically generated.
932void ?{}(union X *_dst){  // default constructor
934void ?{}(union X *_dst, union X _src){  // copy constructor
935  __builtin_memcpy(_dst, &_src, sizeof(union X ));
937void ^?{}(union X *_dst){  // destructor
939union X ?=?(union X *_dst, union X _src){  // assignment
940  __builtin_memcpy(_dst, &_src, sizeof(union X));
941  return _src;
943void ?{}(union X *_dst, struct Y src){  // construct first field
944  __builtin_memcpy(_dst, &src, sizeof(struct Y));
948% This feature works in the \CFA model, since constructors are simply special functions and can be called explicitly, unlike in \CC. % this sentence isn't really true => placement new
949In \CCeleven, this restriction has been loosened to allow unions with managed members, with the caveat that any if there are any members with a user-defined operation, then that operation is not implicitly defined, forcing the user to define the operation if necessary.
950This restriction could easily be added into \CFA once \emph{deleted} functions are added.
952\subsection{Using Constructors and Destructors}
953Implicitly generated constructor and destructor calls ignore the outermost type qualifiers, e.g. @const@ and @volatile@, on a type by way of a cast on the first argument to the function.
954For example,
956struct S { int i; };
957void ?{}(S *, int);
958void ?{}(S *, S);
960const S s = { 11 };
961volatile S s2 = s;
963Generates the following code
965const struct S s;
966?{}((struct S *)&s, 11);
967volatile struct S s2;
968?{}((struct S *)&s2, s);
970Here, @&s@ and @&s2@ are cast to unqualified pointer types.
971This mechanism allows the same constructors and destructors to be used for qualified objects as for unqualified objects.
972Since this applies only to implicitly generated constructor calls, the language does not allow qualified objects to be re-initialized with a constructor without an explicit cast.
974Unlike \CC, \CFA provides an escape hatch that allows a user to decide at an object's definition whether it should be managed or not.
975An object initialized with \ateq is guaranteed to be initialized like a C object, and has no implicit destructor call.
976This feature provides all of the freedom that C programmers are used to having to optimize a program, while maintaining safety as a sensible default.
978struct A { int * x; };
979// RAII
980void ?{}(A * a) { a->x = malloc(sizeof(int)); }
981void ^?{}(A * a) { free(a->x); }
983A a1;           // managed
984A a2 @= { 0 };  // unmanaged
986In this example, @a1@ is a managed object, and thus is default constructed and destructed at the end of @a1@'s lifetime, while @a2@ is an unmanaged object and is not implicitly constructed or destructed.
987Instead, @a2->x@ is initialized to @0@ as if it were a C object, due to the explicit initializer.
988Existing constructors are ignored when \ateq is used, so that any valid C initializer is able to initialize the object.
990In addition to freedom, \ateq provides a simple path to migrating legacy C code to Cforall, in that objects can be moved from C-style initialization to \CFA gradually and individually.
991It is worth noting that the use of unmanaged objects can be tricky to get right, since there is no guarantee that the proper invariants are established on an unmanaged object.
992It is recommended that most objects be managed by sensible constructors and destructors, except where absolutely necessary.
994When the user declares any constructor or destructor, the corresponding intrinsic/generated function and all field constructors for that type are hidden, so that they will not be found during expression resolution unless the user-defined function goes out of scope.
995Furthermore, if the user declares any constructor, then the intrinsic/generated default constructor is also hidden, making it so that objects of a type may not be default constructable.
996This closely mirrors the rule for implicit declaration of constructors in \CC, wherein the default constructor is implicitly declared if there is no user-declared constructor. % TODO: cite C++98 page 186??
998struct S { int x, y; };
1000void f() {
1001  S s0, s1 = { 0 }, s2 = { 0, 2 }, s3 = s2;  // okay
1002  {
1003    void ?{}(S * s, int i) { s->x = i*2; }
1004    S s4;  // error
1005    S s5 = { 3 };  // okay
1006    S s6 = { 4, 5 };  // error
1007    S s7 = s5; // okay
1008  }
1009  S s8, s9 = { 6 }, s10 = { 7, 8 }, s11 = s10;  // okay
1012In this example, the inner scope declares a constructor from @int@ to @S@, which hides the default constructor and field constructors until the end of the scope.
1014When defining a constructor or destructor for a struct @S@, any members that are not explicitly constructed or destructed are implicitly constructed or destructed automatically.
1015If an explicit call is present, then that call is taken in preference to any implicitly generated call.
1016A consequence of this rule is that it is possible, unlike \CC, to precisely control the order of construction and destruction of subobjects on a per-constructor basis, whereas in \CC subobject initialization and destruction is always performed based on the declaration order.
1018struct A {
1019  B w, x, y, z;
1021void ?{}(A * a, int i) {
1022  (&a->x){ i };
1023  (&a->z){ a->y };
1026Generates the following
1028void ?{}(A * a, int i) {
1029  (&a->w){};   // implicit default ctor
1030  (&a->y){};   // implicit default ctor
1031  (&a->x){ i };
1032  (&a->z){ a->y };
1035Finally, it is illegal for a subobject to be explicitly constructed for the first time after it is used for the first time.
1036If the translator cannot be reasonably sure that an object is constructed prior to its first use, but is constructed afterward, an error is emitted.
1037More specifically, the translator searches the body of a constructor to ensure that every subobject is initialized.
1039void ?{}(A * a, double x) {
1040  f(a->x);
1041  (&a->x){ (int)x }; // error, used uninitialized on previous line
1044However, if the translator sees a subobject used within the body of a constructor, but does not see a constructor call that uses the subobject as the target of a constructor, then the translator assumes the object is to be implicitly constructed (copy constructed in a copy constructor and default constructed in any other constructor).
1046void ?{}(A * a) {
1047  // default constructs all members
1048  f(a->x);
1051void ?{}(A * a, A other) {
1052  // copy constructs all members
1053  f(a->y);
1056void ^?{}(A * a) {
1057  ^(&a->x){}; // explicit destructor call
1058} // z, y, w implicitly destructed, in this order
1060If at any point, the @this@ parameter is passed directly as the target of another constructor, then it is assumed that constructor handles the initialization of all of the object's members and no implicit constructor calls are added. % TODO: confirm that this is correct. It might be possible to get subtle errors if you initialize some members then call another constructor... -- in fact, this is basically always wrong. if anything, I should check that such a constructor does not initialize any members, otherwise it'll always initialize the member twice (once locally, once by the called constructor).
1061To override this rule, \ateq can be used to force the translator to trust the programmer's discretion.
1062This form of \ateq is not yet implemented.
1064Despite great effort, some forms of C syntax do not work well with constructors in \CFA.
1065In particular, constructor calls cannot contain designations (see \ref{sub:c_background}), since this is equivalent to allowing designations on the arguments to arbitrary function calls.
1066In C, function prototypes are permitted to have arbitrary parameter names, including no names at all, which may have no connection to the actual names used at function definition.
1067Furthermore, a function prototype can be repeated an arbitrary number of times, each time using different names.
1069// all legal forward declarations in C
1070void f(int, int, int);
1071void f(int a, int b, int c);
1072void f(int b, int c, int a);
1073void f(int c, int a, int b);
1074void f(int x, int y, int z);
1076f(b:10, a:20, c:30);  // which parameter is which?
1078As a result, it was decided that any attempt to resolve designated function calls with C's function prototype rules would be brittle, and thus it is not sensible to allow designations in constructor calls.
1079% Many other languages do allow named arguments, such as Python and Scala, but they do not allow multiple arbitrarily named forward declarations of a function.
1081In addition, constructor calls cannot have a nesting depth greater than the number of array components in the type of the initialized object, plus one.
1082For example,
1084struct A;
1085void ?{}(A *, int);
1086void ?{}(A *, A, A);
1088A a1[3] = { { 3 }, { 4 }, { 5 } };
1089A a2[2][2] = {
1090  { { 9 }, { 10 } },  // a2[0]
1091  { {14 }, { 15 } }   // a2[1]
1093A a3[4] = {
1094  { { 11 }, { 12 } },  // error
1095  { 80 }, { 90 }, { 100 }
1098% TODO: in CFA if the array dimension is empty, no object constructors are added -- need to fix this.
1099The body of @A@ has been omitted, since only the constructor interfaces are important.
1100In C, having a greater nesting depth means that the programmer intends to initialize subobjects with the nested initializer.
1101The reason for this omission is to both simplify the mental model for using constructors, and to make initialization simpler for the expression resolver.
1102If this were allowed, it would be necessary for the expression resolver to decide whether each argument to the constructor call could initialize to some argument in one of the available constructors, making the problem highly recursive and potentially much more expensive.
1103That is, in the previous example the line marked as an error could mean construct using @?{}(A *, A, A)@, since the inner initializer @{ 11 }@ could be taken as an intermediate object of type @A@ constructed with @?{}(A *, int)@.
1104In practice, however, there could be many objects that can be constructed from a given @int@ (or, indeed, any arbitrary parameter list), and thus a complete solution to this problem would require fully exploring all possibilities.
1105It should be noted that unmanaged objects can still make use of designations and nested initializers in \CFA.
1107\subsection{Implicit Destructors}
1109Destructors are automatically called at the end of the block in which the object is declared.
1110In addition to this, destructors are automatically called when statements manipulate control flow to leave a block in which the object is declared, e.g., with return, break, continue, and goto statements.
1111The example below demonstrates a simple routine with multiple return statements.
1113struct A;
1114void ^?{}(A *);
1116void f(int i) {
1117  A x;  // construct x
1118  {
1119    A y; // construct y
1120    {
1121      A z; // construct z
1122      {
1123        if (i == 0) return; // destruct x, y, z
1124      }
1125      if (i == 1) return; // destruct x, y, z
1126    } // destruct z
1127    if (i == 2) return; // destruct x, y
1128  } // destruct y
1132%% having this feels excessive, but it's here if necessary
1133% This procedure generates the following code.
1134% \begin{cfacode}
1135% void f(int i){
1136%   struct A x;
1137%   ?{}(&x);
1138%   {
1139%     struct A y;
1140%     ?{}(&y);
1141%     {
1142%       struct A z;
1143%       ?{}(&z);
1144%       {
1145%         if ((i==0)!=0) {
1146%           ^?{}(&z);
1147%           ^?{}(&y);
1148%           ^?{}(&x);
1149%           return;
1150%         }
1151%       }
1152%       if (((i==1)!=0) {
1153%           ^?{}(&z);
1154%           ^?{}(&y);
1155%           ^?{}(&x);
1156%           return ;
1157%       }
1158%       ^?{}(&z);
1159%     }
1161%     if ((i==2)!=0) {
1162%       ^?{}(&y);
1163%       ^?{}(&x);
1164%       return;
1165%     }
1166%     ^?{}(&y);
1167%   }
1169%   ^?{}(&x);
1170% }
1171% \end{cfacode}
1173The next example illustrates the use of simple continue and break statements and the manner that they interact with implicit destructors.
1175for (int i = 0; i < 10; i++) {
1176  A x;
1177  if (i == 2) {
1178    continue;  // destruct x
1179  } else if (i == 3) {
1180    break;     // destruct x
1181  }
1182} // destruct x
1184Since a destructor call is automatically inserted at the end of the block, nothing special needs to happen to destruct @x@ in the case where control reaches the end of the loop.
1185In the case where @i@ is @2@, the continue statement runs the loop update expression and attemps to begin the next iteration of the loop.
1186Since continue is a C statement, which does not understand destructors, a destructor call is added just before the continue statement to ensure that @x@ is destructed.
1187When @i@ is @3@, the break statement moves control to just past the end of the loop.
1188Like the previous case, a destructor call for @x@ is inserted just before the break statement.
1190\CFA also supports labelled break and continue statements, which allow more precise manipulation of control flow.
1191Labelled break and continue allow the programmer to specify which control structure to target by using a label attached to a control structure.
1192\begin{cfacode}[emph={L1,L2}, emphstyle=\color{red}]
1193L1: for (int i = 0; i < 10; i++) {
1194  A x;
1195  L2: for (int j = 0; j < 10; j++) {
1196    A y;
1197    if (j == 0) {
1198      continue;    // destruct y
1199    } else if (j == 1) {
1200      break;       // destruct y
1201    } else if (i == 1) {
1202      continue L1; // destruct y
1203    } else if (i == 2) {
1204      break L1;    // destruct x,y
1205    }
1206  } // destruct y
1207} // destruct X
1209The statement @continue L1@ begins the next iteration of the outer for-loop.
1210Since the semantics of continue require the loop update expression to execute, control branches to the \emph{end} of the outer for loop, meaning that the block destructor for @x@ can be reused, and it is only necessary to generate the destructor for @y@.
1211Break, on the other hand, requires jumping out of the loop, so the destructors for both @x@ and @y@ are generated and inserted before the @break L1@ statement.
1213Finally, an example which demonstrates goto.
1214Since goto is a general mechanism for jumping to different locations in the program, a more comprehensive approach is required.
1215For each goto statement $G$ and each target label $L$, let $S_G$ be the set of all managed variables alive at $G$, and let $S_L$ be the set of all managed variables alive at $L$.
1216If at any $G$, $S_L \setminus S_G = \emptyset$, then the translator emits an error, because control flow branches from a point where the object is not yet live to a point where it is live, skipping the object's constructor.
1217Then, for every $G$, the destructors for each variable in the set $S_G \setminus S_L$ is inserted directly before $G$, which ensures each object that is currently live at $G$, but not at $L$, is destructed before control branches.
1219int i = 0;
1221  L0: ;     // S_L0 = { x }
1222    A y;
1223  L1: ;     // S_L1 = { x }
1224    A x;
1225  L2: ;     // S_L2 = { y, x }
1226    if (i == 0) {
1227      ++i;
1228      goto L1;    // S_G = { y, x }
1229      // S_G-S_L1 = { x } => destruct x
1230    } else if (i == 1) {
1231      ++i;
1232      goto L2;    // S_G = { y, x }
1233      // S_G-S_L2 = {} => destruct nothing
1234    } else if (i == 2) {
1235      ++i;
1236      goto L3;    // S_G = { y, x }
1237      // S_G-S_L3 = {}
1238    } else if (false) {
1239      ++i;
1240      A z;
1241      goto L3;    // S_G = { z, y, x }
1242      // S_G-S_L3 = { z } => destruct z
1243    } else {
1244      ++i;
1245      goto L4;    // S_G = { y, x }
1246      // S_G-S_L4 = { y, x } => destruct y, x
1247    }
1248  L3: ;    // S_L3 = { y, x }
1249    goto L2;      // S_G = { y, x }
1250    // S_G-S_L2 = {}
1252L4: ;  // S_L4 = {}
1253if (i == 4) {
1254  goto L0;        // S_G = {}
1255  // S_G-S_L0 = {}
1258Labelled break and continue are implemented in \CFA in terms of goto statements, so the more constrained forms are precisely goverened by these rules.
1260The next example demonstrates the error case.
1263    goto L1; // S_G = {}
1264    // S_L1-S_G = { y } => error
1265    A y;
1266  L1: ; // S_L1 = { y }
1267    A x;
1268  L2: ; // S_L2 = { y, x }
1270goto L2; // S_G = {}
1271// S_L2-S_G = { y, x } => error
1274\subsection{Implicit Copy Construction}
1275When a function is called, the arguments supplied to the call are subject to implicit copy construction (and destruction of the generated temporary), and the return value is subject to destruction.
1276When a value is returned from a function, the copy constructor is called to pass the value back to the call site.
1277Exempt from these rules are intrinsic and builtin functions.
1278It should be noted that unmanaged objects are subject to copy constructor calls when passed as arguments to a function or when returned from a function, since they are not the \emph{target} of the copy constructor call.
1279This is an important detail to bear in mind when using unmanaged objects, and could produce unexpected results when mixed with objects that are explicitly constructed.
1281struct A;
1282void ?{}(A *);
1283void ?{}(A *, A);
1284void ^?{}(A *);
1286A f(A x) {
1287  return x;
1290A y, z @= {};
1294Note that @z@ is copy constructed into a temporary variable to be passed as an argument, which is also destructed after the call.
1295A special syntactic form, such as a variant of \ateq, could be implemented to specify at the call site that an argument should not be copy constructed, to regain some control for the C programmer.
1297This generates the following
1299struct A f(struct A x){
1300  struct A _retval_f;
1301  ?{}((&_retval_f), x);
1302  return _retval_f;
1305struct A y;
1307struct A z = { 0 };
1309struct A _tmp_cp1;     // argument 1
1310struct A _tmp_cp_ret0; // return value
1311_tmp_cp_ret0=f((?{}(&_tmp_cp1, y) , _tmp_cp1)), _tmp_cp_ret0;
1312^?{}(&_tmp_cp_ret0);   // return value
1313^?{}(&_tmp_cp1);       // argument 1
1315struct A _tmp_cp2;     // argument 1
1316struct A _tmp_cp_ret1; // return value
1317_tmp_cp_ret1=f((?{}(&_tmp_cp2, z), _tmp_cp2)), _tmp_cp_ret1;
1318^?{}(&_tmp_cp_ret1);   // return value
1319^?{}(&_tmp_cp2);       // argument 1
1323A known issue with this implementation is that the return value of a function is not guaranteed to have the same address for its entire lifetime.
1324Specifically, since @_retval_f@ is allocated and constructed in @f@ then returned by value, the internal data is bitwise copied into the caller's stack frame.
1325This approach works out most of the time, because typically destructors need to only access the fields of the object and recursively destroy.
1326It is currently the case that constructors and destructors which use the @this@ pointer as a unique identifier to store data externally will not work correctly for return value objects.
1327Thus is it not safe to rely on an object's @this@ pointer to remain constant throughout execution of the program.
1329A * external_data[32];
1330int ext_count;
1331struct A;
1332void ?{}(A * a) {
1333  // ...
1334  external_data[ext_count++] = a;
1336void ^?{}(A * a) {
1337  for (int i = 0; i < ext_count) {
1338    if (a == external_data[i]) { // may never be true
1339      // ...
1340    }
1341  }
1344In the above example, a global array of pointers is used to keep track of all of the allocated @A@ objects.
1345Due to copying on return, the current object being destructed will not exist in the array if an @A@ object is ever returned by value from a function.
1347This problem could be solved in the translator by mutating the function signatures so that the return value is moved into the parameter list.
1348For example, the translator could restructure the code like so
1350void f(struct A x, struct A * _retval_f){
1351  ?{}(_retval_f, x);  // construct directly into caller's stack frame
1354struct A y;
1356struct A z = { 0 };
1358struct A _tmp_cp1;     // argument 1
1359struct A _tmp_cp_ret0; // return value
1360f((?{}(&_tmp_cp1, y) , _tmp_cp1), &_tmp_cp_ret0), _tmp_cp_ret0;
1361^?{}(&_tmp_cp_ret0);   // return value
1362^?{}(&_tmp_cp1);       // argument 1
1364This transformation provides @f@ with the address of the return variable so that it can be constructed into directly.
1365It is worth pointing out that this kind of signature rewriting already occurs in polymorphic functions which return by value, as discussed in \cite{Bilson03}.
1366A key difference in this case is that every function would need to be rewritten like this, since types can switch between managed and unmanaged at different scope levels, e.g.
1368struct A { int v; };
1369A x; // unmanaged
1371  void ?{}(A * a) { ... }
1372  void ^?{}(A * a) { ... }
1373  A y; // managed
1375A z; // unmanaged
1377Hence there is not enough information to determine at function declaration to determine whether a type is managed or not, and thus it is the case that all signatures have to be rewritten to account for possible copy constructor and destructor calls.
1378Even with this change, it would still be possible to declare backwards compatible function prototypes with an @extern "C"@ block, which allows for the definition of C-compatible functions within \CFA code, however this would require actual changes to the way code inside of an @extern "C"@ function is generated as compared with normal code generation.
1379Furthermore, it isn't possible to overload C functions, so using @extern "C"@ to declare functions is of limited use.
1381It would be possible to regain some control by adding an attribute to structs which specifies whether they can be managed or not (perhaps \emph{manageable} or \emph{unmanageable}), and to emit an error in the case that a constructor or destructor is declared for an unmanageable type.
1382Ideally, structs should be manageable by default, since otherwise the default case becomes more verbose.
1383This means that in general, function signatures would have to be rewritten, and in a select few cases the signatures would not be rewritten.
1385__attribute__((manageable)) struct A { ... };   // can declare constructors
1386__attribute__((unmanageable)) struct B { ... }; // cannot declare constructors
1387struct C { ... };                               // can declare constructors
1389A f();  // rewritten void f(A *);
1390B g();  // not rewritten
1391C h();  // rewritten void h(C *);
1393An alternative is to instead make the attribute \emph{identifiable}, which states that objects of this type use the @this@ parameter as an identity.
1394This strikes more closely to the visibile problem, in that only types marked as identifiable would need to have the return value moved into the parameter list, and every other type could remain the same.
1395Furthermore, no restrictions would need to be placed on whether objects can be constructed.
1397__attribute__((identifiable)) struct A { ... };  // can declare constructors
1398struct B { ... };                                // can declare constructors
1400A f();  // rewritten void f(A *);
1401B g();  // not rewritten
1404Ultimately, this is the type of transformation that a real compiler would make when generating assembly code.
1405Since a compiler has full control over its calling conventions, it can seamlessly allow passing the return parameter without outwardly changing the signature of a routine.
1406As such, it has been decided that this issue is not currently a priority.
1409\subsection{Array Initialization}
1410Arrays are a special case in the C type system.
1411C arrays do not carry around their size, making it impossible to write a standalone \CFA function that constructs or destructs an array while maintaining the standard interface for constructors and destructors.
1412Instead, \CFA defines the initialization and destruction of an array recursively.
1413That is, when an array is defined, each of its elements is constructed in order from element 0 up to element $n-1$.
1414When an array is to be implicitly destructed, each of its elements is destructed in reverse order from element $n-1$ down to element 0.
1415As in C, it is possible to explicitly provide different initializers for each element of the array through array initialization syntax.
1416In this case, each of the initializers is taken in turn to construct a subsequent element of the array.
1417If too many initializers are provided, only the initializers up to N are actually used.
1418If too few initializers are provided, then the remaining elements are default constructed.
1420For example, given the following code.
1422struct X {
1423  int x, y, z;
1425void f() {
1426  X x[10] = { { 1, 2, 3 }, { 4 }, { 7, 8 } };
1429The following code is generated for @f@.
1431void f(){
1432  struct X x[((long unsigned int )10)];
1433  // construct x
1434  {
1435    int _index0 = 0;
1436    // construct with explicit initializers
1437    {
1438      if (_index0<10) ?{}(&x[_index0], 1, 2, 3);
1439      ++_index0;
1440      if (_index0<10) ?{}(&x[_index0], 4);
1441      ++_index0;
1442      if (_index0<10) ?{}(&x[_index0], 7, 8);
1443      ++_index0;
1444    }
1446    // default construct remaining elements
1447    for (;_index0<10;++_index0) {
1448      ?{}(&x[_index0]);
1449    }
1450  }
1451  // destruct x
1452  {
1453    int _index1 = 10-1;
1454    for (;_index1>=0;--_index1) {
1455      ^?{}(&x[_index1]);
1456    }
1457  }
1460Multidimensional arrays require more complexity.
1461For example, a two dimensional array
1463void g() {
1464  X x[10][10] = {
1465    { { 1, 2, 3 }, { 4 } }, // x[0]
1466    { { 7, 8 } }            // x[1]
1467  };
1469Generates the following
1471void g(){
1472  struct X x[10][10];
1473  // construct x
1474  {
1475    int _index0 = 0;
1476    for (;_index0<10;++_index0) {
1477      {
1478        int _index1 = 0;
1479        // construct with explicit initializers
1480        {
1481          switch ( _index0 ) {
1482            case 0:
1483              // construct first array
1484              if ( _index1<10 ) ?{}(&x[_index0][_index1], 1, 2, 3);
1485              ++_index1;
1486              if ( _index1<10 ) ?{}(&x[_index0][_index1], 4);
1487              ++_index1;
1488              break;
1489            case 1:
1490              // construct second array
1491              if ( _index1<10 ) ?{}(&x[_index0][_index1], 7, 8);
1492              ++_index1;
1493              break;
1494          }
1495        }
1496        // default construct remaining elements
1497        for (;_index1<10;++_index1) {
1498            ?{}(&x[_index0][_index1]);
1499        }
1500      }
1501    }
1502  }
1503  // destruct x
1504  {
1505    int _index2 = 10-1;
1506    for (;_index2>=0;--_index2) {
1507      {
1508        int _index3 = 10-1;
1509        for (;_index3>=0;--_index3) {
1510            ^?{}(&x[_index2][_index3]);
1511        }
1512      }
1513    }
1514  }
1517% It is possible to generate slightly simpler code for the switch cases, since the value of @_index1@ is known at compile-time within each case, however the procedure for generating constructor calls is complicated.
1518% It is simple to remove the increment statements for @_index1@, but it is not simple to remove the
1519%% technically, it's not hard either. I could easily downcast and change the second argument to ?[?], but is it really necessary/worth it??
1521\subsection{Global Initialization}
1522In standard C, global variables can only be initialized to compile-time constant expressions.
1523This places strict limitations on the programmer's ability to control the default values of objects.
1524In \CFA, constructors and destructors are guaranteed to be run on global objects, allowing arbitrary code to be run before and after the execution of the main routine.
1525By default, objects within a translation unit are constructed in declaration order, and destructed in the reverse order.
1526The default order of construction of objects amongst translation units is unspecified.
1527% TODO: not yet implemented, but g++ provides attribute init_priority, which allows specifying the order of global construction on a per object basis
1529% suggestion: implement this in CFA by picking objects with a specified priority and pulling them into their own init functions (could even group them by priority level -> map<int, list<ObjectDecl*>>) and pull init_priority forward into constructor and destructor attributes with the same priority level
1530It is, however, guaranteed that any global objects in the standard library are initialized prior to the initialization of any object in the user program.
1532This feature is implemented in the \CFA translator by grouping every global constructor call into a function with the GCC attribute \emph{constructor}, which performs most of the heavy lifting. % CITE:
1533A similar function is generated with the \emph{destructor} attribute, which handles all global destructor calls.
1534At the time of writing, initialization routines in the library are specified with priority \emph{101}, which is the highest priority level that GCC allows, whereas initialization routines in the user's code are implicitly given the default priority level, which ensures they have a lower priority than any code with a specified priority level.
1535This mechanism allows arbitrarily complicated initialization to occur before any user code runs, making it possible for library designers to initialize their modules without requiring the user to call specific startup or teardown routines.
1537For example, given the following global declarations.
1539struct X {
1540  int y, z;
1542void ?{}(X *);
1543void ?{}(X *, int, int);
1544void ^?{}(X *);
1546X a;
1547X b = { 10, 3 };
1549The following code is generated.
1551__attribute__ ((constructor)) static void _init_global_ctor(void){
1552  ?{}(&a);
1553  ?{}(&b, 10, 3);
1555__attribute__ ((destructor)) static void _destroy_global_ctor(void){
1556  ^?{}(&b);
1557  ^?{}(&a);
1561\subsection{Static Local Variables}
1562In standard C, it is possible to mark variables that are local to a function with the @static@ storage class.
1563Unlike normal local variables, a @static@ local variable is defined to live for the entire duration of the program, so that each call to the function has access to the same variable with the same address and value as it had in the previous call to the function. % TODO: mention dynamic loading caveat??
1564Much like global variables, in C @static@ variables must be initialized to a \emph{compile-time constant value} so that a compiler is able to create storage for the variable and initialize it before the program begins running.
1566Yet again, this rule is too restrictive for a language with constructors and destructors.
1567Instead, \CFA modifies the definition of a @static@ local variable so that objects are guaranteed to be live from the time control flow reaches their declaration, until the end of the program, since the initializer expression is not necessarily a compile-time constant, but can depend on the current execution state of the function.
1568Since standard C does not allow access to a @static@ local variable before the first time control flow reaches the declaration, this restriction does not preclude any valid C code.
1569Local objects with @static@ storage class are only implicitly constructed and destructed once for the duration of the program.
1570The object is constructed when its declaration is reached for the first time.
1571The object is destructed once at the end of the program.
1573Construction of @static@ local objects is implemented via an accompanying @static bool@ variable, which records whether the variable has already been constructed.
1574A conditional branch checks the value of the companion @bool@, and if the variable has not yet been constructed then the object is constructed.
1575The object's destructor is scheduled to be run when the program terminates using @atexit@, and the companion @bool@'s value is set so that subsequent invocations of the function will not reconstruct the object.
1576Since the parameter to @atexit@ is a parameter-less function, some additional tweaking is required.
1577First, the @static@ variable must be hoisted up to global scope and uniquely renamed to prevent name clashes with other global objects.
1578Second, a function is built which calls the destructor for the newly hoisted variable.
1579Finally, the newly generated function is registered with @atexit@, instead of registering the destructor directly.
1580Since @atexit@ calls functions in the reverse order in which they are registered, @static@ local variables are guaranteed to be destructed in the reverse order that they are constructed, which may differ between multiple executions of the same program.
1582Extending the previous example
1584int f(int x) {
1585  static X a;
1586  static X b = { x, x };  // depends on parameter value
1587  static X c = b;         // depends on local variable
1590Generates the following.
1592static struct X a_static_var0;
1593static void __a_dtor_atexit0(void){
1594  ((void)^?{}(((struct X *)(&a_static_var0))));
1596static struct X b_static_var1;
1597static void __b_dtor_atexit1(void){
1598  ((void)^?{}(((struct X *)(&b_static_var1))));
1600static struct X c_static_var2;
1601static void __c_dtor_atexit2(void){
1602  ((void)^?{}(((struct X *)(&c_static_var2))));
1604int f(int x){
1605  int _retval_f;
1606  __attribute__ ((unused)) static void *_dummy0;
1607  static _Bool __a_uninitialized = 1;
1608  if ( __a_uninitialized ) {
1609    ((void)?{}(((struct X *)(&a_static_var0))));
1610    ((void)(__a_uninitialized=0));
1611    ((void)atexit(__a_dtor_atexit0));
1612  }
1614  __attribute__ ((unused)) static void *_dummy1;
1615  static _Bool __b_uninitialized = 1;
1616  if ( __b_uninitialized ) {
1617    ((void)?{}(((struct X *)(&b_static_var1)), x, x));
1618    ((void)(__b_uninitialized=0));
1619    ((void)atexit(__b_dtor_atexit1));
1620  }
1622  __attribute__ ((unused)) static void *_dummy2;
1623  static _Bool __c_uninitialized = 1;
1624  if ( __c_uninitialized ) {
1625    ((void)?{}(((struct X *)(&c_static_var2)), b_static_var1));
1626    ((void)(__c_uninitialized=0));
1627    ((void)atexit(__c_dtor_atexit2));
1628  }
1632\subsection{Constructor Expressions}
1633In \CFA, it is possible to use a constructor as an expression.
1634Like other operators, the function name @?{}@ matches its operator syntax.
1635For example, @(&x){}@ calls the default constructor on the variable @x@, and produces @&x@ as a result.
1636The significance of constructors as expressions rather than as statements is that the result of a constructor expression can be used as part of a larger expression.
1637A key example is the use of constructor expressions to initialize the result of a call to standard C routine @malloc@.
1639struct X { ... };
1640void ?{}(X *, double);
1641X * x = malloc(sizeof(X)){ 1.5 };
1643In this example, @malloc@ dynamically allocates storage and initializes it using a constructor, all before assigning it into the variable @x@.
1644If this extension is not present, constructing dynamically allocated objects is much more cumbersome, requiring separate initialization of the pointer and initialization of the pointed-to memory.
1646X * x = malloc(sizeof(X));
1647x{ 1.5 };
1649Not only is this verbose, but it is also more error prone, since this form allows maintenance code to easily sneak in between the initialization of @x@ and the initialization of the memory that @x@ points to.
1650This feature is implemented via a transformation produceing the value of the first argument of the constructor, since constructors do not themslves have a return value.
1651Since this transformation results in two instances of the subexpression, care is taken to allocate a temporary variable to hold the result of the subexpression in the case where the subexpression may contain side effects.
1652The previous example generates the following code.
1654struct X *_tmp_ctor;
1655struct X *x = ?{}((_tmp_ctor=((_tmp_cp_ret0=
1656  malloc(sizeof(struct X))), _tmp_cp_ret0))), 1.5), _tmp_ctor);
1658It should be noted that this technique is not exclusive to @malloc@, and allows a user to write a custom allocator that can be idiomatically used in much the same way as a constructed @malloc@ call.
1660It is also possible to use operator syntax with destructors.
1661Unlike constructors, operator syntax with destructors is a statement and thus does not produce a value, since the destructed object is invalidated by the use of a destructor.
1662For example, \lstinline!^(&x){}! calls the destructor on the variable @x@.
Note: See TracBrowser for help on using the repository browser.