1 | \chapter{\CFA{} Existing Features}
|
---|
2 | \label{c:existing}
|
---|
3 |
|
---|
4 | \CFA is an open-source project extending ISO C with
|
---|
5 | modern safety and productivity features, while still ensuring backwards
|
---|
6 | compatibility with C and its programmers. \CFA is designed to have an
|
---|
7 | orthogonal feature-set based closely on the C programming paradigm
|
---|
8 | (non-object-oriented), and these features can be added incrementally to an
|
---|
9 | existing C code-base,
|
---|
10 | allowing programmers to learn \CFA on an as-needed basis.
|
---|
11 |
|
---|
12 | Only those \CFA features pertaining to this thesis are discussed.
|
---|
13 | A familiarity with
|
---|
14 | C or C-like languages is assumed.
|
---|
15 |
|
---|
16 | \section{Overloading and \lstinline{extern}}
|
---|
17 | \CFA has extensive overloading, allowing multiple definitions of the same name
|
---|
18 | to be defined~\cite{Moss18}.
|
---|
19 | \begin{cfa}
|
---|
20 | char i; int i; double i;
|
---|
21 | int f(); double f();
|
---|
22 | void g( int ); void g( double );
|
---|
23 | \end{cfa}
|
---|
24 | This feature requires name mangling so the assembly symbols are unique for
|
---|
25 | different overloads. For compatibility with names in C, there is also a syntax
|
---|
26 | to disable name mangling. These unmangled names cannot be overloaded but act as
|
---|
27 | the interface between C and \CFA code. The syntax for disabling/enabling
|
---|
28 | mangling is:
|
---|
29 | \begin{cfa}
|
---|
30 | // name mangling on by default
|
---|
31 | int i; // _X1ii_1
|
---|
32 | extern "C" { // disables name mangling
|
---|
33 | int j; // j
|
---|
34 | extern "Cforall" { // enables name mangling
|
---|
35 | int k; // _X1ki_1
|
---|
36 | }
|
---|
37 | // revert to no name mangling
|
---|
38 | }
|
---|
39 | // revert to name mangling
|
---|
40 | \end{cfa}
|
---|
41 | Both forms of @extern@ affect all the declarations within their nested lexical
|
---|
42 | scope and transition back to the previous mangling state when the lexical scope
|
---|
43 | ends.
|
---|
44 |
|
---|
45 | \section{Reference Type}
|
---|
46 | \CFA adds a reference type to C as an auto-dereferencing pointer.
|
---|
47 | They work very similarly to pointers.
|
---|
48 | Reference-types are written the same way as pointer-types, but each
|
---|
49 | asterisk (@*@) is replaced with a ampersand (@&@);
|
---|
50 | this includes cv-qualifiers (\snake{const} and \snake{volatile})
|
---|
51 | and multiple levels of reference.
|
---|
52 |
|
---|
53 | Generally, references act like pointers with an implicit dereferencing
|
---|
54 | operation added to each use of the variable.
|
---|
55 | These automatic dereferences may be disabled with the address-of operator
|
---|
56 | (@&@).
|
---|
57 |
|
---|
58 | % Check to see if these are generating errors.
|
---|
59 | \begin{minipage}{0,5\textwidth}
|
---|
60 | With references:
|
---|
61 | \begin{cfa}
|
---|
62 | int i, j;
|
---|
63 | int & ri = i;
|
---|
64 | int && rri = ri;
|
---|
65 | rri = 3;
|
---|
66 | &ri = &j;
|
---|
67 | ri = 5;
|
---|
68 | \end{cfa}
|
---|
69 | \end{minipage}
|
---|
70 | \begin{minipage}{0,5\textwidth}
|
---|
71 | With pointers:
|
---|
72 | \begin{cfa}
|
---|
73 | int i, j;
|
---|
74 | int * pi = &i
|
---|
75 | int ** ppi = π
|
---|
76 | **ppi = 3;
|
---|
77 | pi = &j;
|
---|
78 | *pi = 5;
|
---|
79 | \end{cfa}
|
---|
80 | \end{minipage}
|
---|
81 |
|
---|
82 | References are intended to be used when the indirection of a pointer is
|
---|
83 | required, but the address is not as important as the value and dereferencing
|
---|
84 | is the common usage.
|
---|
85 | Mutable references may be assigned to by converting them to a pointer
|
---|
86 | with a @&@ and then assigning a pointer to them, as in @&ri = &j;@ above.
|
---|
87 |
|
---|
88 | \section{Operators}
|
---|
89 |
|
---|
90 | \CFA implements operator overloading by providing special names, where
|
---|
91 | operator expressions are translated into function calls using these names.
|
---|
92 | An operator name is created by taking the operator symbols and joining them with
|
---|
93 | @?@s to show where the arguments go.
|
---|
94 | For example,
|
---|
95 | infixed multiplication is @?*?@, while prefix dereference is @*?@.
|
---|
96 | This syntax makes it easy to tell the difference between prefix operations
|
---|
97 | (such as @++?@) and postfix operations (@?++@).
|
---|
98 |
|
---|
99 | As an example, here are the addition and equality operators for a point type.
|
---|
100 | \begin{cfa}
|
---|
101 | point ?+?(point a, point b) { return point{a.x + b.x, a.y + b.y}; }
|
---|
102 | int ?==?(point a, point b) { return a.x == b.x && a.y == b.y; }
|
---|
103 | {
|
---|
104 | assert(point{1, 2} + point{3, 4} == point{4, 6});
|
---|
105 | }
|
---|
106 | \end{cfa}
|
---|
107 | Note that this syntax works effectively as a textual transformation;
|
---|
108 | the compiler converts all operators into functions and then resolves them
|
---|
109 | normally. This means any combination of types may be used,
|
---|
110 | although nonsensical ones (like @double ?==?(point, int);@) are discouraged.
|
---|
111 | This feature is also used for all builtin operators as well,
|
---|
112 | although those are implicitly provided by the language.
|
---|
113 |
|
---|
114 | %\subsection{Constructors and Destructors}
|
---|
115 | In \CFA, constructors and destructors are operators, which means they are
|
---|
116 | functions with special operator names, rather than type names as in \Cpp.
|
---|
117 | Both constructors and destructors can be implicity called by the compiler,
|
---|
118 | however the operator names allow explicit calls.
|
---|
119 | % Placement new means that this is actually equivant to C++.
|
---|
120 |
|
---|
121 | The special name for a constructor is @?{}@, which comes from the
|
---|
122 | initialization syntax in C, \eg @Example e = { ... }@.
|
---|
123 | \CFA generates a constructor call each time a variable is declared,
|
---|
124 | passing the initialization arguments to the constructor.
|
---|
125 | \begin{cfa}
|
---|
126 | struct Example { ... };
|
---|
127 | void ?{}(Example & this) { ... }
|
---|
128 | {
|
---|
129 | Example a;
|
---|
130 | Example b = {};
|
---|
131 | }
|
---|
132 | void ?{}(Example & this, char first, int num) { ... }
|
---|
133 | {
|
---|
134 | Example c = {'a', 2};
|
---|
135 | }
|
---|
136 | \end{cfa}
|
---|
137 | Both @a@ and @b@ will be initalized with the first constructor,
|
---|
138 | @b@ because of the explicit call and @a@ implicitly.
|
---|
139 | @c@ will be initalized with the second constructor.
|
---|
140 | Currently, there is no general way to skip initialization.
|
---|
141 | % I don't use @= anywhere in the thesis.
|
---|
142 |
|
---|
143 | % I don't like the \^{} symbol but $^\wedge$ isn't better.
|
---|
144 | Similarly, destructors use the special name @^?{}@ (the @^@ has no special
|
---|
145 | meaning).
|
---|
146 | \begin{cfa}
|
---|
147 | void ^?{}(Example & this) { ... }
|
---|
148 | {
|
---|
149 | Example d;
|
---|
150 | ^?{}(d);
|
---|
151 |
|
---|
152 | Example e;
|
---|
153 | } // Implicit call of ^?{}(e);
|
---|
154 | \end{cfa}
|
---|
155 |
|
---|
156 | Whenever a type is defined, \CFA creates a default zero-argument
|
---|
157 | constructor, a copy constructor, a series of argument-per-field constructors
|
---|
158 | and a destructor. All user constructors are defined after this.
|
---|
159 |
|
---|
160 | \section{Polymorphism}
|
---|
161 | \CFA uses parametric polymorphism to create functions and types that are
|
---|
162 | defined over multiple types. \CFA polymorphic declarations serve the same role
|
---|
163 | as \Cpp templates or Java generics. The ``parametric'' means the polymorphism is
|
---|
164 | accomplished by passing argument operations to associate \emph{parameters} at
|
---|
165 | the call site, and these parameters are used in the function to differentiate
|
---|
166 | among the types the function operates on.
|
---|
167 |
|
---|
168 | Polymorphic declarations start with a universal @forall@ clause that goes
|
---|
169 | before the standard (monomorphic) declaration. These declarations have the same
|
---|
170 | syntax except they may use the universal type names introduced by the @forall@
|
---|
171 | clause. For example, the following is a polymorphic identity function that
|
---|
172 | works on any type @T@:
|
---|
173 | \begin{cfa}
|
---|
174 | forall( T ) T identity( T val ) { return val; }
|
---|
175 | int forty_two = identity( 42 );
|
---|
176 | char capital_a = identity( 'A' );
|
---|
177 | \end{cfa}
|
---|
178 | Each use of a polymorphic declaration resolves its polymorphic parameters
|
---|
179 | (in this case, just @T@) to concrete types (@int@ in the first use and @char@
|
---|
180 | in the second).
|
---|
181 |
|
---|
182 | To allow a polymorphic function to be separately compiled, the type @T@ must be
|
---|
183 | constrained by the operations used on @T@ in the function body. The @forall@
|
---|
184 | clause is augmented with a list of polymorphic variables (local type names)
|
---|
185 | and assertions (constraints), which represent the required operations on those
|
---|
186 | types used in a function, \eg:
|
---|
187 | \begin{cfa}
|
---|
188 | forall( T | { void do_once(T); } )
|
---|
189 | void do_twice(T value) {
|
---|
190 | do_once(value);
|
---|
191 | do_once(value);
|
---|
192 | }
|
---|
193 | \end{cfa}
|
---|
194 |
|
---|
195 | A polymorphic function can be used in the same way as a normal function. The
|
---|
196 | polymorphic variables are filled in with concrete types and the assertions are
|
---|
197 | checked. An assertion is checked by verifying each assertion operation (with
|
---|
198 | all the variables replaced with the concrete types from the arguments) is
|
---|
199 | defined at a call site.
|
---|
200 | \begin{cfa}
|
---|
201 | void do_once(int i) { ... }
|
---|
202 | int i;
|
---|
203 | do_twice(i);
|
---|
204 | \end{cfa}
|
---|
205 | Any value with a type fulfilling the assertion may be passed as an argument to
|
---|
206 | a @do_twice@ call.
|
---|
207 |
|
---|
208 | Note, a function named @do_once@ is not required in the scope of @do_twice@ to
|
---|
209 | compile it, unlike \Cpp template expansion. Furthermore, call-site inferencing
|
---|
210 | allows local replacement of the specific parametric functions needs for a
|
---|
211 | call.
|
---|
212 | \begin{cfa}
|
---|
213 | void do_once(double y) { ... }
|
---|
214 | int quadruple(int x) {
|
---|
215 | void do_once(int & y) { y = y * 2; }
|
---|
216 | do_twice(x);
|
---|
217 | return x;
|
---|
218 | }
|
---|
219 | \end{cfa}
|
---|
220 | Specifically, the complier deduces that @do_twice@'s T is an integer from the
|
---|
221 | argument @x@. It then looks for the most specific definition matching the
|
---|
222 | assertion, which is the nested integral @do_once@ defined within the
|
---|
223 | function. The matched assertion function is then passed as a function pointer
|
---|
224 | to @do_twice@ and called within it.
|
---|
225 | The global definition of @do_once@ is ignored, however if @quadruple@ took a
|
---|
226 | @double@ argument, then the global definition would be used instead as it
|
---|
227 | would then be a better match.\cite{Moss19}
|
---|
228 |
|
---|
229 | To avoid typing long lists of assertions, constraints can be collected into
|
---|
230 | a convenient package called a @trait@, which can then be used in an assertion
|
---|
231 | instead of the individual constraints.
|
---|
232 | \begin{cfa}
|
---|
233 | trait done_once(T) {
|
---|
234 | void do_once(T);
|
---|
235 | }
|
---|
236 | \end{cfa}
|
---|
237 | and the @forall@ list in the previous example is replaced with the trait.
|
---|
238 | \begin{cfa}
|
---|
239 | forall(dtype T | done_once(T))
|
---|
240 | \end{cfa}
|
---|
241 | In general, a trait can contain an arbitrary number of assertions, both
|
---|
242 | functions and variables, and are usually used to create a shorthand for, and
|
---|
243 | give descriptive names to, common groupings of assertions describing a certain
|
---|
244 | functionality, like @summable@, @listable@, \etc.
|
---|
245 |
|
---|
246 | Polymorphic structures and unions are defined by qualifying an aggregate type
|
---|
247 | with @forall@. The type variables work the same except they are used in field
|
---|
248 | declarations instead of parameters, returns and local variable declarations.
|
---|
249 | \begin{cfa}
|
---|
250 | forall(dtype T)
|
---|
251 | struct node {
|
---|
252 | node(T) * next;
|
---|
253 | T * data;
|
---|
254 | };
|
---|
255 | node(int) inode;
|
---|
256 | \end{cfa}
|
---|
257 | The generic type @node(T)@ is an example of a polymorphic type usage. Like \Cpp
|
---|
258 | template usage, a polymorphic type usage must specify a type parameter.
|
---|
259 |
|
---|
260 | There are many other polymorphism features in \CFA but these are the ones used
|
---|
261 | by the exception system.
|
---|
262 |
|
---|
263 | \section{Control Flow}
|
---|
264 | \CFA has a number of advanced control-flow features: @generator@, @coroutine@,
|
---|
265 | @monitor@, @mutex@ parameters, and @thread@.
|
---|
266 | The two features that interact with
|
---|
267 | the exception system are @coroutine@ and @thread@; they and their supporting
|
---|
268 | constructs are described here.
|
---|
269 |
|
---|
270 | \subsection{Coroutine}
|
---|
271 | A coroutine is a type with associated functions, where the functions are not
|
---|
272 | required to finish execution when control is handed back to the caller.
|
---|
273 | Instead,
|
---|
274 | they may suspend execution at any time and be resumed later at the point of
|
---|
275 | last suspension.
|
---|
276 | Coroutine
|
---|
277 | types are not concurrent but share some similarities along with common
|
---|
278 | underpinnings, so they are combined with the \CFA threading library.
|
---|
279 | % I had mention of generators, but they don't actually matter here.
|
---|
280 |
|
---|
281 | In \CFA, a coroutine is created using the @coroutine@ keyword, which is an
|
---|
282 | aggregate type like @struct,@ except the structure is implicitly modified by
|
---|
283 | the compiler to satisfy the @is_coroutine@ trait; hence, a coroutine is
|
---|
284 | restricted by the type system to types that provide this special trait. The
|
---|
285 | coroutine structure acts as the interface between callers and the coroutine,
|
---|
286 | and its fields are used to pass information in and out of coroutine interface
|
---|
287 | functions.
|
---|
288 |
|
---|
289 | Here is a simple example where a single field is used to pass (communicate) the
|
---|
290 | next number in a sequence.
|
---|
291 | \begin{cfa}
|
---|
292 | coroutine CountUp {
|
---|
293 | unsigned int next;
|
---|
294 | };
|
---|
295 | CountUp countup;
|
---|
296 | \end{cfa}
|
---|
297 | Each coroutine has a @main@ function, which takes a reference to a coroutine
|
---|
298 | object and returns @void@.
|
---|
299 | %[numbers=left] Why numbers on this one?
|
---|
300 | \begin{cfa}
|
---|
301 | void main(CountUp & this) {
|
---|
302 | for (unsigned int next = 0 ; true ; ++next) {
|
---|
303 | this.next = next;
|
---|
304 | suspend;$\label{suspend}$
|
---|
305 | }
|
---|
306 | }
|
---|
307 | \end{cfa}
|
---|
308 | In this function, or functions called by this function (helper functions), the
|
---|
309 | @suspend@ statement is used to return execution to the coroutine's caller
|
---|
310 | without terminating the coroutine's function.
|
---|
311 |
|
---|
312 | A coroutine is resumed by calling the @resume@ function, \eg @resume(countup)@.
|
---|
313 | The first resume calls the @main@ function at the top. Thereafter, resume calls
|
---|
314 | continue a coroutine in the last suspended function after the @suspend@
|
---|
315 | statement. In this case there is only one and, hence, the difference between
|
---|
316 | subsequent calls is the state of variables inside the function and the
|
---|
317 | coroutine object.
|
---|
318 | The return value of @resume@ is a reference to the coroutine, to make it
|
---|
319 | convent to access fields of the coroutine in the same expression.
|
---|
320 | Here is a simple example in a helper function:
|
---|
321 | \begin{cfa}
|
---|
322 | unsigned int get_next(CountUp & this) {
|
---|
323 | return resume(this).next;
|
---|
324 | }
|
---|
325 | \end{cfa}
|
---|
326 |
|
---|
327 | When the main function returns, the coroutine halts and can no longer be
|
---|
328 | resumed.
|
---|
329 |
|
---|
330 | \subsection{Monitor and Mutex Parameter}
|
---|
331 | Concurrency does not guarantee ordering; without ordering, results are
|
---|
332 | non-deterministic. To claw back ordering, \CFA uses monitors and @mutex@
|
---|
333 | (mutual exclusion) parameters. A monitor is another kind of aggregate, where
|
---|
334 | the compiler implicitly inserts a lock and instances are compatible with
|
---|
335 | @mutex@ parameters.
|
---|
336 |
|
---|
337 | A function that requires deterministic (ordered) execution acquires mutual
|
---|
338 | exclusion on a monitor object by qualifying an object reference parameter with
|
---|
339 | the @mutex@ qualifier.
|
---|
340 | \begin{cfa}
|
---|
341 | void example(MonitorA & mutex argA, MonitorB & mutex argB);
|
---|
342 | \end{cfa}
|
---|
343 | When the function is called, it implicitly acquires the monitor lock for all of
|
---|
344 | the mutex parameters without deadlock. This semantics means all functions with
|
---|
345 | the same mutex type(s) are part of a critical section for objects of that type
|
---|
346 | and only one runs at a time.
|
---|
347 |
|
---|
348 | \subsection{Thread}
|
---|
349 | Functions, generators and coroutines are sequential, so there is only a single
|
---|
350 | (but potentially sophisticated) execution path in a program. Threads introduce
|
---|
351 | multiple execution paths that continue independently.
|
---|
352 |
|
---|
353 | For threads to work safely with objects requires mutual exclusion using
|
---|
354 | monitors and mutex parameters. For threads to work safely with other threads
|
---|
355 | also requires mutual exclusion in the form of a communication rendezvous, which
|
---|
356 | also supports internal synchronization as for mutex objects. For exceptions,
|
---|
357 | only two basic thread operations are important: fork and join.
|
---|
358 |
|
---|
359 | Threads are created like coroutines with an associated @main@ function:
|
---|
360 | \begin{cfa}
|
---|
361 | thread StringWorker {
|
---|
362 | const char * input;
|
---|
363 | int result;
|
---|
364 | };
|
---|
365 | void main(StringWorker & this) {
|
---|
366 | const char * localCopy = this.input;
|
---|
367 | // ... do some work, perhaps hashing the string ...
|
---|
368 | this.result = result;
|
---|
369 | }
|
---|
370 | {
|
---|
371 | StringWorker stringworker; // fork thread running in "main"
|
---|
372 | } // Implicit call to join(stringworker), waits for completion.
|
---|
373 | \end{cfa}
|
---|
374 | The thread main is where a new thread starts execution after a fork operation
|
---|
375 | and then the thread continues executing until it is finished. If another thread
|
---|
376 | joins with an executing thread, it waits until the executing main completes
|
---|
377 | execution. In other words, everything a thread does is between a fork and join.
|
---|
378 |
|
---|
379 | From the outside, this behaviour is accomplished through creation and
|
---|
380 | destruction of a thread object. Implicitly, fork happens after a thread
|
---|
381 | object's constructor is run and join happens before the destructor runs. Join
|
---|
382 | can also be specified explicitly using the @join@ function to wait for a
|
---|
383 | thread's completion independently from its deallocation (\ie destructor
|
---|
384 | call). If @join@ is called explicitly, the destructor does not implicitly join.
|
---|