source: doc/papers/general/Paper.tex @ 271326e

aaron-thesisarm-ehcleanup-dtorsdeferred_resndemanglerjacob/cs343-translationjenkins-sandboxnew-astnew-ast-unique-exprnew-envno_listpersistent-indexerresolv-newwith_gc
Last change on this file since 271326e was 271326e, checked in by Aaron Moss <a3moss@…>, 4 years ago

Add name overloading subsection, rebuttal of _Generic

  • Property mode set to 100644
File size: 142.7 KB
Line 
1\documentclass{article}
2
3\usepackage{fullpage}
4\usepackage{epic,eepic}
5\usepackage{xspace,calc,comment}
6\usepackage{upquote}                                            % switch curled `'" to straight
7\usepackage{listings}                                           % format program code
8\usepackage{enumitem}
9\usepackage[flushmargin]{footmisc}                      % support label/reference in footnote
10\usepackage{rotating}
11\usepackage[usenames]{color}
12\usepackage{pslatex}                                            % reduce size of san serif font
13\usepackage[plainpages=false,pdfpagelabels,pdfpagemode=UseNone,pagebackref=true,breaklinks=true,colorlinks=true,linkcolor=blue,citecolor=blue,urlcolor=blue]{hyperref}
14
15\setlength{\textheight}{9in}
16%\oddsidemargin 0.0in
17\renewcommand{\topfraction}{0.8}                        % float must be greater than X of the page before it is forced onto its own page
18\renewcommand{\bottomfraction}{0.8}                     % float must be greater than X of the page before it is forced onto its own page
19\renewcommand{\floatpagefraction}{0.8}          % float must be greater than X of the page before it is forced onto its own page
20\renewcommand{\textfraction}{0.0}                       % the entire page maybe devoted to floats with no text on the page at all
21
22\lefthyphenmin=4                                                        % hyphen only after 4 characters
23\righthyphenmin=4
24
25% Names used in the document.
26
27\newcommand{\CFAIcon}{\textsf{C}\raisebox{\depth}{\rotatebox{180}{\textsf{A}}}\xspace} % Cforall symbolic name
28\newcommand{\CFA}{\protect\CFAIcon}             % safe for section/caption
29\newcommand{\CFL}{\textrm{Cforall}\xspace}      % Cforall symbolic name
30\newcommand{\Celeven}{\textrm{C11}\xspace}      % C11 symbolic name
31\newcommand{\CC}{\textrm{C}\kern-.1em\hbox{+\kern-.25em+}\xspace} % C++ symbolic name
32\newcommand{\CCeleven}{\textrm{C}\kern-.1em\hbox{+\kern-.25em+}11\xspace} % C++11 symbolic name
33\newcommand{\CCfourteen}{\textrm{C}\kern-.1em\hbox{+\kern-.25em+}14\xspace} % C++14 symbolic name
34\newcommand{\CCseventeen}{\textrm{C}\kern-.1em\hbox{+\kern-.25em+}17\xspace} % C++17 symbolic name
35\newcommand{\CCtwenty}{\textrm{C}\kern-.1em\hbox{+\kern-.25em+}20\xspace} % C++20 symbolic name
36\newcommand{\CCV}{\rm C\kern-.1em\hbox{+\kern-.25em+}obj\xspace} % C++ virtual symbolic name
37\newcommand{\Csharp}{C\raisebox{-0.7ex}{\Large$^\sharp$}\xspace} % C# symbolic name
38
39%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
40
41\newcommand{\Textbf}[2][red]{{\color{#1}{\textbf{#2}}}}
42\newcommand{\TODO}[1]{\textbf{TODO}: {\itshape #1}} % TODO included
43%\newcommand{\TODO}[1]{} % TODO elided
44
45% Default underscore is too low and wide. Cannot use lstlisting "literate" as replacing underscore
46% removes it as a variable-name character so keywords in variables are highlighted. MUST APPEAR
47% AFTER HYPERREF.
48%\DeclareTextCommandDefault{\textunderscore}{\leavevmode\makebox[1.2ex][c]{\rule{1ex}{0.1ex}}}
49\renewcommand{\textunderscore}{\leavevmode\makebox[1.2ex][c]{\rule{1ex}{0.075ex}}}
50
51\makeatletter
52% parindent is relative, i.e., toggled on/off in environments like itemize, so store the value for
53% use rather than use \parident directly.
54\newlength{\parindentlnth}
55\setlength{\parindentlnth}{\parindent}
56
57\newcommand{\LstKeywordStyle}[1]{{\lst@basicstyle{\lst@keywordstyle{#1}}}}
58\newcommand{\LstCommentStyle}[1]{{\lst@basicstyle{\lst@commentstyle{#1}}}}
59
60\newlength{\gcolumnposn}                                        % temporary hack because lstlisting does not handle tabs correctly
61\newlength{\columnposn}
62\setlength{\gcolumnposn}{2.75in}
63\setlength{\columnposn}{\gcolumnposn}
64\newcommand{\C}[2][\@empty]{\ifx#1\@empty\else\global\setlength{\columnposn}{#1}\global\columnposn=\columnposn\fi\hfill\makebox[\textwidth-\columnposn][l]{\lst@basicstyle{\LstCommentStyle{#2}}}}
65\newcommand{\CRT}{\global\columnposn=\gcolumnposn}
66
67% Denote newterms in particular font and index them without particular font and in lowercase, e.g., \newterm{abc}.
68% The option parameter provides an index term different from the new term, e.g., \newterm[\texttt{abc}]{abc}
69% The star version does not lowercase the index information, e.g., \newterm*{IBM}.
70\newcommand{\newtermFontInline}{\emph}
71\newcommand{\newterm}{\@ifstar\@snewterm\@newterm}
72\newcommand{\@newterm}[2][\@empty]{\lowercase{\def\temp{#2}}{\newtermFontInline{#2}}\ifx#1\@empty\index{\temp}\else\index{#1@{\protect#2}}\fi}
73\newcommand{\@snewterm}[2][\@empty]{{\newtermFontInline{#2}}\ifx#1\@empty\index{#2}\else\index{#1@{\protect#2}}\fi}
74
75% Latin abbreviation
76\newcommand{\abbrevFont}{\textit}                       % set empty for no italics
77\newcommand{\EG}{\abbrevFont{e}.\abbrevFont{g}.}
78\newcommand*{\eg}{%
79        \@ifnextchar{,}{\EG}%
80                {\@ifnextchar{:}{\EG}%
81                        {\EG,\xspace}}%
82}%
83\newcommand{\IE}{\abbrevFont{i}.\abbrevFont{e}.}
84\newcommand*{\ie}{%
85        \@ifnextchar{,}{\IE}%
86                {\@ifnextchar{:}{\IE}%
87                        {\IE,\xspace}}%
88}%
89\newcommand{\ETC}{\abbrevFont{etc}}
90\newcommand*{\etc}{%
91        \@ifnextchar{.}{\ETC}%
92        {\ETC\xspace}%
93}%
94\newcommand{\ETAL}{\abbrevFont{et}\hspace{2pt}\abbrevFont{al}}
95\newcommand*{\etal}{%
96        \@ifnextchar{.}{\protect\ETAL}%
97                {\abbrevFont{\protect\ETAL}.\xspace}%
98}%
99\newcommand{\VIZ}{\abbrevFont{viz}}
100\newcommand*{\viz}{%
101        \@ifnextchar{.}{\VIZ}%
102                {\abbrevFont{\VIZ}.\xspace}%
103}%
104\makeatother
105
106\newenvironment{cquote}{%
107        \list{}{\lstset{resetmargins=true,aboveskip=0pt,belowskip=0pt}\topsep=4pt\parsep=0pt\leftmargin=\parindentlnth\rightmargin\leftmargin}%
108        \item\relax
109}{%
110        \endlist
111}% cquote
112
113% CFA programming language, based on ANSI C (with some gcc additions)
114\lstdefinelanguage{CFA}[ANSI]{C}{
115        morekeywords={
116                _Alignas, _Alignof, __alignof, __alignof__, asm, __asm, __asm__, _At, __attribute,
117                __attribute__, auto, _Bool, catch, catchResume, choose, _Complex, __complex, __complex__,
118                __const, __const__, disable, dtype, enable, __extension__, fallthrough, fallthru,
119                finally, forall, ftype, _Generic, _Imaginary, inline, __label__, lvalue, _Noreturn, one_t,
120                otype, restrict, _Static_assert, throw, throwResume, trait, try, ttype, typeof, __typeof,
121                __typeof__, virtual, with, zero_t},
122        moredirectives={defined,include_next}%
123}%
124
125\lstset{
126language=CFA,
127columns=fullflexible,
128basicstyle=\linespread{0.9}\sf,                                                 % reduce line spacing and use sanserif font
129stringstyle=\tt,                                                                                % use typewriter font
130tabsize=5,                                                                                              % N space tabbing
131xleftmargin=\parindentlnth,                                                             % indent code to paragraph indentation
132%mathescape=true,                                                                               % LaTeX math escape in CFA code $...$
133escapechar=\$,                                                                                  % LaTeX escape in CFA code
134keepspaces=true,                                                                                %
135showstringspaces=false,                                                                 % do not show spaces with cup
136showlines=true,                                                                                 % show blank lines at end of code
137aboveskip=4pt,                                                                                  % spacing above/below code block
138belowskip=3pt,
139% replace/adjust listing characters that look bad in sanserif
140literate={-}{\makebox[1ex][c]{\raisebox{0.4ex}{\rule{0.8ex}{0.1ex}}}}1 {^}{\raisebox{0.6ex}{$\scriptscriptstyle\land\,$}}1
141        {~}{\raisebox{0.3ex}{$\scriptstyle\sim\,$}}1 % {`}{\ttfamily\upshape\hspace*{-0.1ex}`}1
142        {<-}{$\leftarrow$}2 {=>}{$\Rightarrow$}2 {->}{\makebox[1ex][c]{\raisebox{0.4ex}{\rule{0.8ex}{0.075ex}}}\kern-0.2ex\textgreater}2,
143moredelim=**[is][\color{red}]{`}{`},
144}% lstset
145
146% inline code @...@
147\lstMakeShortInline@%
148
149\lstnewenvironment{cfa}[1][]
150{\lstset{#1}}
151{}
152\lstnewenvironment{C++}[1][]                            % use C++ style
153{\lstset{language=C++,moredelim=**[is][\protect\color{red}]{`}{`},#1}\lstset{#1}}
154{}
155
156
157\title{Generic and Tuple Types with Efficient Dynamic Layout in \protect\CFA}
158
159\author{Aaron Moss, Robert Schluntz, Peter Buhr}
160% \email{a3moss@uwaterloo.ca}
161% \email{rschlunt@uwaterloo.ca}
162% \email{pabuhr@uwaterloo.ca}
163% \affiliation{%
164%       \institution{University of Waterloo}
165%       \department{David R. Cheriton School of Computer Science}
166%       \streetaddress{Davis Centre, University of Waterloo}
167%       \city{Waterloo}
168%       \state{ON}
169%       \postcode{N2L 3G1}
170%       \country{Canada}
171% }
172
173%\terms{generic, tuple, variadic, types}
174%\keywords{generic types, tuple types, variadic types, polymorphic functions, C, Cforall}
175
176\begin{document}
177\maketitle
178
179
180\begin{abstract}
181The C programming language is a foundational technology for modern computing with millions of lines of code implementing everything from commercial operating-systems to hobby projects.
182This installation base and the programmers producing it represent a massive software-engineering investment spanning decades and likely to continue for decades more.
183Nonetheless, C, first standardized over thirty years ago, lacks many features that make programming in more modern languages safer and more productive.
184The goal of the \CFA project is to create an extension of C that provides modern safety and productivity features while still ensuring strong backwards compatibility with C and its programmers.
185Prior projects have attempted similar goals but failed to honour C programming-style; for instance, adding object-oriented or functional programming with garbage collection is a non-starter for many C developers.
186Specifically, \CFA is designed to have an orthogonal feature-set based closely on the C programming paradigm, so that \CFA features can be added \emph{incrementally} to existing C code-bases, and C programmers can learn \CFA extensions on an as-needed basis, preserving investment in existing code and engineers.
187This paper describes two \CFA extensions, generic and tuple types, details how their design avoids shortcomings of similar features in C and other C-like languages, and presents experimental results validating the design.
188\end{abstract}
189
190
191\section{Introduction}
192
193The C programming language is a foundational technology for modern computing with millions of lines of code implementing everything from commercial operating-systems to hobby projects.
194This installation base and the programmers producing it represent a massive software-engineering investment spanning decades and likely to continue for decades more.
195The TIOBE~\cite{TIOBE} ranks the top 5 most popular programming languages as: Java 16\%, \Textbf{C 7\%}, \Textbf{\CC 5\%}, \Csharp 4\%, Python 4\% = 36\%, where the next 50 languages are less than 3\% each with a long tail.
196The top 3 rankings over the past 30 years are:
197\begin{center}
198\setlength{\tabcolsep}{10pt}
199\lstDeleteShortInline@%
200\begin{tabular}{@{}rccccccc@{}}
201                & 2017  & 2012  & 2007  & 2002  & 1997  & 1992  & 1987          \\ \hline
202Java    & 1             & 1             & 1             & 1             & 12    & -             & -                     \\
203\Textbf{C}      & \Textbf{2}& \Textbf{2}& \Textbf{2}& \Textbf{2}& \Textbf{1}& \Textbf{1}& \Textbf{1}    \\
204\CC             & 3             & 3             & 3             & 3             & 2             & 2             & 4                     \\
205\end{tabular}
206\lstMakeShortInline@%
207\end{center}
208Love it or hate it, C is extremely popular, highly used, and one of the few systems languages.
209In many cases, \CC is often used solely as a better C.
210Nonetheless, C, first standardized over thirty years ago, lacks many features that make programming in more modern languages safer and more productive.
211
212\CFA (pronounced ``C-for-all'', and written \CFA or Cforall) is an evolutionary extension of the C programming language that aims to add modern language features to C while maintaining both source compatibility with C and a familiar programming model for programmers.
213The four key design goals for \CFA~\cite{Bilson03} are:
214(1) The behaviour of standard C code must remain the same when translated by a \CFA compiler as when translated by a C compiler;
215(2) Standard C code must be as fast and as small when translated by a \CFA compiler as when translated by a C compiler;
216(3) \CFA code must be at least as portable as standard C code;
217(4) Extensions introduced by \CFA must be translated in the most efficient way possible.
218These goals ensure existing C code-bases can be converted to \CFA incrementally with minimal effort, and C programmers can productively generate \CFA code without training beyond the features being used.
219\CC is used similarly, but has the disadvantages of multiple legacy design-choices that cannot be updated and active divergence of the language model from C, requiring significant effort and training to incrementally add \CC to a C-based project.
220
221\CFA is currently implemented as a source-to-source translator from \CFA to the GCC-dialect of C~\cite{GCCExtensions}, allowing it to leverage the portability and code optimizations provided by GCC, meeting goals (1)--(3).
222Ultimately, a compiler is necessary for advanced features and optimal performance.
223
224This paper identifies shortcomings in existing approaches to generic and variadic data types in C-like languages and presents a design for generic and variadic types avoiding those shortcomings.
225Specifically, the solution is both reusable and type-checked, as well as conforming to the design goals of \CFA with ergonomic use of existing C abstractions.
226The new constructs are empirically compared with both standard C and \CC; the results show the new design is comparable in performance.
227
228\section{Polymorphic Functions}
229
230\CFA introduces both ad-hoc and parametric polymorphism to C, with a design originally formalized by Ditchfield~\cite{Ditchfield92}, and first implemented by Bilson~\cite{Bilson03}.
231
232\subsection{Name Overloading}
233
234C already has a limited form of ad-hoc polymorphism in the form of its basic arithmetic operators, which apply to a variety of different types using identical syntax.
235\CFA extends the built-in operator overloading by allowing users to define overloads for any function, not just operators, and even any variable; Section~\ref{sec:libraries} includes a number of examples of how this overloading simplifies \CFA programming relative to C.
236Code generation for these overloaded functions and variables is implemented by the usual approach of mangling the identifier names to include a representation of their type, while \CFA decides which overload to apply based on the same ``usual arithmetic conversions'' used in C to disambiguate operator overloads.
237As an example:
238
239\begin{cfa}
240int max(int a, int b) { return a < b ? b : a; }  // (1)
241double max(double a, double b) { return a < b ? b : a; }  // (2)
242
243int max = INT_MAX;     // (3)
244double max = DBL_MAX;  // (4)
245
246max(7, -max);   $\C{// uses (1) and (3), by matching int from constant 7}$
247max(max, 3.14); $\C{// uses (2) and (4), by matching double from constant 3.14}$
248
249//max(max, -max);  $\C{// ERROR: ambiguous}$
250int m = max(max, -max); $\C{// uses (1) once and (3) twice, by matching return type}$
251\end{cfa}
252
253\Celeven did add @_Generic@ expressions, which can be used in preprocessor macros to provide a form of ad-hoc polymorphism; however, this polymorphism is both functionally and ergonomically inferior to \CFA name overloading.
254The macro wrapping the generic expression imposes some limitations; as an example, it could not implement the example above, because the variables @max@ would collide with the functions @max@.
255Ergonomic limitations of @_Generic@ include the necessity to put a fixed list of supported types in a single place and manually dispatch to appropriate overloads, as well as possible namespace pollution from the functions dispatched to, which must all have distinct names.
256
257\subsection{\texorpdfstring{\LstKeywordStyle{forall} Functions}{forall Functions}}
258\label{sec:poly-fns}
259
260The signature feature of \CFA is parametric-polymorphic functions~\cite{forceone:impl,Cormack90,Duggan96} with functions generalized using a @forall@ clause (giving the language its name):
261\begin{lstlisting}
262`forall( otype T )` T identity( T val ) { return val; }
263int forty_two = identity( 42 );                         $\C{// T is bound to int, forty\_two == 42}$
264\end{lstlisting}
265The @identity@ function above can be applied to any complete \newterm{object type} (or @otype@).
266The type variable @T@ is transformed into a set of additional implicit parameters encoding sufficient information about @T@ to create and return a variable of that type.
267The \CFA implementation passes the size and alignment of the type represented by an @otype@ parameter, as well as an assignment operator, constructor, copy constructor and destructor.
268If this extra information is not needed, \eg for a pointer, the type parameter can be declared as a \newterm{data type} (or @dtype@).
269
270In \CFA, the polymorphism runtime-cost is spread over each polymorphic call, due to passing more arguments to polymorphic functions;
271the experiments in Section~\ref{sec:eval} show this overhead is similar to \CC virtual-function calls.
272A design advantage is that, unlike \CC template-functions, \CFA polymorphic-functions are compatible with C \emph{separate compilation}, preventing compilation and code bloat.
273
274Since bare polymorphic-types provide a restricted set of available operations, \CFA provides a \newterm{type assertion}~\cite[pp.~37-44]{Alphard} mechanism to provide further type information, where type assertions may be variable or function declarations that depend on a polymorphic type-variable.
275For example, the function @twice@ can be defined using the \CFA syntax for operator overloading:
276\begin{lstlisting}
277forall( otype T `| { T ?+?(T, T); }` ) T twice( T x ) { return x + x; } $\C{// ? denotes operands}$
278int val = twice( twice( 3.7 ) );
279\end{lstlisting}
280which works for any type @T@ with a matching addition operator.
281The polymorphism is achieved by creating a wrapper function for calling @+@ with @T@ bound to @double@, then passing this function to the first call of @twice@.
282There is now the option of using the same @twice@ and converting the result to @int@ on assignment, or creating another @twice@ with type parameter @T@ bound to @int@ because \CFA uses the return type~\cite{Cormack81,Baker82,Ada}, in its type analysis.
283The first approach has a late conversion from @double@ to @int@ on the final assignment, while the second has an eager conversion to @int@.
284\CFA minimizes the number of conversions and their potential to lose information, so it selects the first approach, which corresponds with C-programmer intuition.
285
286Crucial to the design of a new programming language are the libraries to access thousands of external software features.
287Like \CC, \CFA inherits a massive compatible library-base, where other programming languages must rewrite or provide fragile inter-language communication with C.
288A simple example is leveraging the existing type-unsafe (@void *@) C @bsearch@ to binary search a sorted float array:
289\begin{lstlisting}
290void * bsearch( const void * key, const void * base, size_t nmemb, size_t size,
291                                int (* compar)( const void *, const void * ));
292int comp( const void * t1, const void * t2 ) { return *(double *)t1 < *(double *)t2 ? -1 :
293                                *(double *)t2 < *(double *)t1 ? 1 : 0; }
294double key = 5.0, vals[10] = { /* 10 sorted float values */ };
295double * val = (double *)bsearch( &key, vals, 10, sizeof(vals[0]), comp );      $\C{// search sorted array}$
296\end{lstlisting}
297which can be augmented simply with a generalized, type-safe, \CFA-overloaded wrappers:
298\begin{lstlisting}
299forall( otype T | { int ?<?( T, T ); } ) T * bsearch( T key, const T * arr, size_t size ) {
300        int comp( const void * t1, const void * t2 ) { /* as above with double changed to T */ }
301        return (T *)bsearch( &key, arr, size, sizeof(T), comp ); }
302forall( otype T | { int ?<?( T, T ); } ) unsigned int bsearch( T key, const T * arr, size_t size ) {
303        T * result = bsearch( key, arr, size ); $\C{// call first version}$
304        return result ? result - arr : size; }  $\C{// pointer subtraction includes sizeof(T)}$
305double * val = bsearch( 5.0, vals, 10 );        $\C{// selection based on return type}$
306int posn = bsearch( 5.0, vals, 10 );
307\end{lstlisting}
308The nested function @comp@ provides the hidden interface from typed \CFA to untyped (@void *@) C, plus the cast of the result.
309Providing a hidden @comp@ function in \CC is awkward as lambdas do not use C calling-conventions and template declarations cannot appear at block scope.
310As well, an alternate kind of return is made available: position versus pointer to found element.
311\CC's type-system cannot disambiguate between the two versions of @bsearch@ because it does not use the return type in overload resolution, nor can \CC separately compile a templated @bsearch@.
312
313\CFA has replacement libraries condensing hundreds of existing C functions into tens of \CFA overloaded functions, all without rewriting the actual computations.
314For example, it is possible to write a type-safe \CFA wrapper @malloc@ based on the C @malloc@:
315\begin{lstlisting}
316forall( dtype T | sized(T) ) T * malloc( void ) { return (T *)malloc( sizeof(T) ); }
317int * ip = malloc();                                            $\C{// select type and size from left-hand side}$
318double * dp = malloc();
319struct S {...} * sp = malloc();
320\end{lstlisting}
321where the return type supplies the type/size of the allocation, which is impossible in most type systems.
322
323Call-site inferencing and nested functions provide a localized form of inheritance.
324For example, the \CFA @qsort@ only sorts in ascending order using @<@.
325However, it is trivial to locally change this behaviour:
326\begin{lstlisting}
327forall( otype T | { int ?<?( T, T ); } ) void qsort( const T * arr, size_t size ) { /* use C qsort */ }
328{       int ?<?( double x, double y ) { return x `>` y; }       $\C{// locally override behaviour}$
329        qsort( vals, size );                                    $\C{// descending sort}$
330}
331\end{lstlisting}
332Within the block, the nested version of @?<?@ performs @?>?@ and this local version overrides the built-in @?<?@ so it is passed to @qsort@.
333Hence, programmers can easily form local environments, adding and modifying appropriate functions, to maximize reuse of other existing functions and types.
334
335%% Redundant with Section~\ref{sec:libraries} %%
336
337% Finally, \CFA allows variable overloading:
338% \begin{lstlisting}
339% short int MAX = ...;   int MAX = ...;  double MAX = ...;
340% short int s = MAX;    int i = MAX;    double d = MAX;   $\C{// select correct MAX}$
341% \end{lstlisting}
342% Here, the single name @MAX@ replaces all the C type-specific names: @SHRT_MAX@, @INT_MAX@, @DBL_MAX@.
343
344\subsection{Traits}
345
346\CFA provides \newterm{traits} to name a group of type assertions, where the trait name allows specifying the same set of assertions in multiple locations, preventing repetition mistakes at each function declaration:
347\begin{lstlisting}
348trait summable( otype T ) {
349        void ?{}( T *, zero_t );                                $\C{// constructor from 0 literal}$
350        T ?+?( T, T );                                                  $\C{// assortment of additions}$
351        T ?+=?( T *, T );
352        T ++?( T * );
353        T ?++( T * ); };
354forall( otype T `| summable( T )` ) T sum( T a[$\,$], size_t size ) {  // use trait
355        `T` total = { `0` };                                    $\C{// instantiate T from 0 by calling its constructor}$
356        for ( unsigned int i = 0; i < size; i += 1 ) total `+=` a[i]; $\C{// select appropriate +}$
357        return total; }
358\end{lstlisting}
359
360In fact, the set of @summable@ trait operators is incomplete, as it is missing assignment for type @T@, but @otype@ is syntactic sugar for the following implicit trait:
361\begin{lstlisting}
362trait otype( dtype T | sized(T) ) {  // sized is a pseudo-trait for types with known size and alignment
363        void ?{}( T * );                                                $\C{// default constructor}$
364        void ?{}( T *, T );                                             $\C{// copy constructor}$
365        void ?=?( T *, T );                                             $\C{// assignment operator}$
366        void ^?{}( T * ); };                                    $\C{// destructor}$
367\end{lstlisting}
368Given the information provided for an @otype@, variables of polymorphic type can be treated as if they were a complete type: stack-allocatable, default or copy-initialized, assigned, and deleted.
369
370In summation, the \CFA type-system uses \newterm{nominal typing} for concrete types, matching with the C type-system, and \newterm{structural typing} for polymorphic types.
371Hence, trait names play no part in type equivalence;
372the names are simply macros for a list of polymorphic assertions, which are expanded at usage sites.
373Nevertheless, trait names form a logical subtype-hierarchy with @dtype@ at the top, where traits often contain overlapping assertions, \eg operator @+@.
374Traits are used like interfaces in Java or abstract base-classes in \CC, but without the nominal inheritance-relationships.
375Instead, each polymorphic function (or generic type) defines the structural type needed for its execution (polymorphic type-key), and this key is fulfilled at each call site from the lexical environment, which is similar to Go~\cite{Go} interfaces.
376Hence, new lexical scopes and nested functions are used extensively to create local subtypes, as in the @qsort@ example, without having to manage a nominal-inheritance hierarchy.
377(Nominal inheritance can be approximated with traits using marker variables or functions, as is done in Go.)
378
379% Nominal inheritance can be simulated with traits using marker variables or functions:
380% \begin{lstlisting}
381% trait nominal(otype T) {
382%     T is_nominal;
383% };
384% int is_nominal;                                                               $\C{// int now satisfies the nominal trait}$
385% \end{lstlisting}
386%
387% Traits, however, are significantly more powerful than nominal-inheritance interfaces; most notably, traits may be used to declare a relationship \emph{among} multiple types, a property that may be difficult or impossible to represent in nominal-inheritance type systems:
388% \begin{lstlisting}
389% trait pointer_like(otype Ptr, otype El) {
390%     lvalue El *?(Ptr);                                                $\C{// Ptr can be dereferenced into a modifiable value of type El}$
391% }
392% struct list {
393%     int value;
394%     list * next;                                                              $\C{// may omit "struct" on type names as in \CC}$
395% };
396% typedef list * list_iterator;
397%
398% lvalue int *?( list_iterator it ) { return it->value; }
399% \end{lstlisting}
400% In the example above, @(list_iterator, int)@ satisfies @pointer_like@ by the user-defined dereference function, and @(list_iterator, list)@ also satisfies @pointer_like@ by the built-in dereference operator for pointers. Given a declaration @list_iterator it@, @*it@ can be either an @int@ or a @list@, with the meaning disambiguated by context (\eg @int x = *it;@ interprets @*it@ as an @int@, while @(*it).value = 42;@ interprets @*it@ as a @list@).
401% While a nominal-inheritance system with associated types could model one of those two relationships by making @El@ an associated type of @Ptr@ in the @pointer_like@ implementation, few such systems could model both relationships simultaneously.
402
403
404\section{Generic Types}
405
406One of the known shortcomings of standard C is that it does not provide reusable type-safe abstractions for generic data structures and algorithms.
407Broadly speaking, there are three approaches to implement abstract data-structures in C.
408One approach is to write bespoke data-structures for each context in which they are needed.
409While this approach is flexible and supports integration with the C type-checker and tooling, it is also tedious and error-prone, especially for more complex data structures.
410A second approach is to use @void *@--based polymorphism, \eg the C standard-library functions @bsearch@ and @qsort@; an approach which does allow reuse of code for common functionality.
411However, basing all polymorphism on @void *@ eliminates the type-checker's ability to ensure that argument types are properly matched, often requiring a number of extra function parameters, pointer indirection, and dynamic allocation that would not otherwise be needed.
412A third approach to generic code is to use preprocessor macros, which does allow the generated code to be both generic and type-checked, but errors may be difficult to interpret.
413Furthermore, writing and using preprocessor macros can be unnatural and inflexible.
414
415\CC, Java, and other languages use \newterm{generic types} to produce type-safe abstract data-types.
416\CFA also implements generic types that integrate efficiently and naturally with the existing polymorphic functions, while retaining backwards compatibility with C and providing separate compilation.
417However, for known concrete parameters, the generic-type definition can be inlined, like \CC templates.
418
419A generic type can be declared by placing a @forall@ specifier on a @struct@ or @union@ declaration, and instantiated using a parenthesized list of types after the type name:
420\begin{lstlisting}
421forall( otype R, otype S ) struct pair {
422        R first;
423        S second;
424};
425forall( otype T ) T value( pair( const char *, T ) p ) { return p.second; }
426forall( dtype F, otype T ) T value_p( pair( F *, T * ) p ) { return * p.second; }
427pair( const char *, int ) p = { "magic", 42 };
428int magic = value( p );
429pair( void *, int * ) q = { 0, &p.second };
430magic = value_p( q );
431double d = 1.0;
432pair( double *, double * ) r = { &d, &d };
433d = value_p( r );
434\end{lstlisting}
435
436\CFA classifies generic types as either \newterm{concrete} or \newterm{dynamic}.
437Concrete types have a fixed memory layout regardless of type parameters, while dynamic types vary in memory layout depending on their type parameters.
438A type may have polymorphic parameters but still be concrete, called \newterm{dtype-static}.
439Polymorphic pointers are an example of dtype-static types, \eg @forall(dtype T) T *@ is a polymorphic type, but for any @T@, @T *@  is a fixed-sized pointer, and therefore, can be represented by a @void *@ in code generation.
440
441\CFA generic types also allow checked argument-constraints.
442For example, the following declaration of a sorted set-type ensures the set key supports equality and relational comparison:
443\begin{lstlisting}
444forall( otype Key | { _Bool ?==?(Key, Key); _Bool ?<?(Key, Key); } ) struct sorted_set;
445\end{lstlisting}
446
447
448\subsection{Concrete Generic-Types}
449
450The \CFA translator template-expands concrete generic-types into new structure types, affording maximal inlining.
451To enable inter-operation among equivalent instantiations of a generic type, the translator saves the set of instantiations currently in scope and reuses the generated structure declarations where appropriate.
452A function declaration that accepts or returns a concrete generic-type produces a declaration for the instantiated structure in the same scope, which all callers may reuse.
453For example, the concrete instantiation for @pair( const char *, int )@ is:
454\begin{lstlisting}
455struct _pair_conc1 {
456        const char * first;
457        int second;
458};
459\end{lstlisting}
460
461A concrete generic-type with dtype-static parameters is also expanded to a structure type, but this type is used for all matching instantiations.
462In the above example, the @pair( F *, T * )@ parameter to @value_p@ is such a type; its expansion is below and it is used as the type of the variables @q@ and @r@ as well, with casts for member access where appropriate:
463\begin{lstlisting}
464struct _pair_conc0 {
465        void * first;
466        void * second;
467};
468\end{lstlisting}
469
470
471\subsection{Dynamic Generic-Types}
472
473Though \CFA implements concrete generic-types efficiently, it also has a fully general system for dynamic generic types.
474As mentioned in Section~\ref{sec:poly-fns}, @otype@ function parameters (in fact all @sized@ polymorphic parameters) come with implicit size and alignment parameters provided by the caller.
475Dynamic generic-types also have an \newterm{offset array} containing structure-member offsets.
476A dynamic generic-union needs no such offset array, as all members are at offset 0, but size and alignment are still necessary.
477Access to members of a dynamic structure is provided at runtime via base-displacement addressing with the structure pointer and the member offset (similar to the @offsetof@ macro), moving a compile-time offset calculation to runtime.
478
479The offset arrays are statically generated where possible.
480If a dynamic generic-type is declared to be passed or returned by value from a polymorphic function, the translator can safely assume the generic type is complete (\ie has a known layout) at any call-site, and the offset array is passed from the caller;
481if the generic type is concrete at the call site, the elements of this offset array can even be statically generated using the C @offsetof@ macro.
482As an example, @p.second@ in the @value@ function above is implemented as @*(p + _offsetof_pair[1])@, where @p@ is a @void *@, and @_offsetof_pair@ is the offset array passed into @value@ for @pair( const char *, T )@.
483The offset array @_offsetof_pair@ is generated at the call site as @size_t _offsetof_pair[] = { offsetof(_pair_conc1, first), offsetof(_pair_conc1, second) }@.
484
485In some cases the offset arrays cannot be statically generated.
486For instance, modularity is generally provided in C by including an opaque forward-declaration of a structure and associated accessor and mutator functions in a header file, with the actual implementations in a separately-compiled @.c@ file.
487\CFA supports this pattern for generic types, but the caller does not know the actual layout or size of the dynamic generic-type, and only holds it by a pointer.
488The \CFA translator automatically generates \newterm{layout functions} for cases where the size, alignment, and offset array of a generic struct cannot be passed into a function from that function's caller.
489These layout functions take as arguments pointers to size and alignment variables and a caller-allocated array of member offsets, as well as the size and alignment of all @sized@ parameters to the generic structure (un@sized@ parameters are forbidden from being used in a context that affects layout).
490Results of these layout functions are cached so that they are only computed once per type per function. %, as in the example below for @pair@.
491Layout functions also allow generic types to be used in a function definition without reflecting them in the function signature.
492For instance, a function that strips duplicate values from an unsorted @vector(T)@ would likely have a pointer to the vector as its only explicit parameter, but use some sort of @set(T)@ internally to test for duplicate values.
493This function could acquire the layout for @set(T)@ by calling its layout function with the layout of @T@ implicitly passed into the function.
494
495Whether a type is concrete, dtype-static, or dynamic is decided solely on the @forall@'s type parameters.
496This design allows opaque forward declarations of generic types, \eg @forall(otype T)@ @struct Box@ -- like in C, all uses of @Box(T)@ can be separately compiled, and callers from other translation units know the proper calling conventions to use.
497If the definition of a structure type is included in deciding whether a generic type is dynamic or concrete, some further types may be recognized as dtype-static (\eg @forall(otype T)@ @struct unique_ptr { T * p }@ does not depend on @T@ for its layout, but the existence of an @otype@ parameter means that it \emph{could}.), but preserving separate compilation (and the associated C compatibility) in the existing design is judged to be an appropriate trade-off.
498
499
500\subsection{Applications}
501\label{sec:generic-apps}
502
503The reuse of dtype-static structure instantiations enables useful programming patterns at zero runtime cost.
504The most important such pattern is using @forall(dtype T) T *@ as a type-checked replacement for @void *@, \eg creating a lexicographic comparison for pairs of pointers used by @bsearch@ or @qsort@:
505\begin{lstlisting}
506forall(dtype T) int lexcmp( pair( T *, T * ) * a, pair( T *, T * ) * b, int (* cmp)( T *, T * ) ) {
507        return cmp( a->first, b->first ) ? : cmp( a->second, b->second );
508}
509\end{lstlisting}
510Since @pair(T *, T * )@ is a concrete type, there are no implicit parameters passed to @lexcmp@, so the generated code is identical to a function written in standard C using @void *@, yet the \CFA version is type-checked to ensure the fields of both pairs and the arguments to the comparison function match in type.
511
512Another useful pattern enabled by reused dtype-static type instantiations is zero-cost \newterm{tag-structures}.
513Sometimes information is only used for type-checking and can be omitted at runtime, \eg:
514\begin{lstlisting}
515forall(dtype Unit) struct scalar { unsigned long value; };
516struct metres {};
517struct litres {};
518
519forall(dtype U) scalar(U) ?+?( scalar(U) a, scalar(U) b ) {
520        return (scalar(U)){ a.value + b.value };
521}
522scalar(metres) half_marathon = { 21093 };
523scalar(litres) swimming_pool = { 2500000 };
524scalar(metres) marathon = half_marathon + half_marathon;
525scalar(litres) two_pools = swimming_pool + swimming_pool;
526marathon + swimming_pool;                                       $\C{// compilation ERROR}$
527\end{lstlisting}
528@scalar@ is a dtype-static type, so all uses have a single structure definition, containing @unsigned long@, and can share the same implementations of common functions like @?+?@.
529These implementations may even be separately compiled, unlike \CC template functions.
530However, the \CFA type-checker ensures matching types are used by all calls to @?+?@, preventing nonsensical computations like adding a length to a volume.
531
532
533\section{Tuples}
534\label{sec:tuples}
535
536In many languages, functions can return at most one value;
537however, many operations have multiple outcomes, some exceptional.
538Consider C's @div@ and @remquo@ functions, which return the quotient and remainder for a division of integer and float values, respectively.
539\begin{lstlisting}
540typedef struct { int quo, rem; } div_t;         $\C{// from include stdlib.h}$
541div_t div( int num, int den );
542double remquo( double num, double den, int * quo );
543div_t qr = div( 13, 5 );                                        $\C{// return quotient/remainder aggregate}$
544int q;
545double r = remquo( 13.5, 5.2, &q );                     $\C{// return remainder, alias quotient}$
546\end{lstlisting}
547@div@ aggregates the quotient/remainder in a structure, while @remquo@ aliases a parameter to an argument.
548Both approaches are awkward.
549Alternatively, a programming language can directly support returning multiple values, \eg in \CFA:
550\begin{lstlisting}
551[ int, int ] div( int num, int den );           $\C{// return two integers}$
552[ double, double ] div( double num, double den ); $\C{// return two doubles}$
553int q, r;                                                                       $\C{// overloaded variable names}$
554double q, r;
555[ q, r ] = div( 13, 5 );                                        $\C{// select appropriate div and q, r}$
556[ q, r ] = div( 13.5, 5.2 );                            $\C{// assign into tuple}$
557\end{lstlisting}
558Clearly, this approach is straightforward to understand and use;
559therefore, why do few programming languages support this obvious feature or provide it awkwardly?
560The answer is that there are complex consequences that cascade through multiple aspects of the language, especially the type-system.
561This section show these consequences and how \CFA handles them.
562
563
564\subsection{Tuple Expressions}
565
566The addition of multiple-return-value functions (MRVF) are useless without a syntax for accepting multiple values at the call-site.
567The simplest mechanism for capturing the return values is variable assignment, allowing the values to be retrieved directly.
568As such, \CFA allows assigning multiple values from a function into multiple variables, using a square-bracketed list of lvalue expressions (as above), called a \newterm{tuple}.
569
570However, functions also use \newterm{composition} (nested calls), with the direct consequence that MRVFs must also support composition to be orthogonal with single-returning-value functions (SRVF), \eg:
571\begin{lstlisting}
572printf( "%d %d\n", div( 13, 5 ) );                      $\C{// return values seperated into arguments}$
573\end{lstlisting}
574Here, the values returned by @div@ are composed with the call to @printf@ by flattening the tuple into separate arguments.
575However, the \CFA type-system must support significantly more complex composition:
576\begin{lstlisting}
577[ int, int ] foo$\(_1\)$( int );                        $\C{// overloaded foo functions}$
578[ double ] foo$\(_2\)$( int );
579void bar( int, double, double );
580bar( foo( 3 ), foo( 3 ) );
581\end{lstlisting}
582The type-resolver only has the tuple return-types to resolve the call to @bar@ as the @foo@ parameters are identical, which involves unifying the possible @foo@ functions with @bar@'s parameter list.
583No combination of @foo@s are an exact match with @bar@'s parameters, so the resolver applies C conversions.
584The minimal cost is @bar( foo@$_1$@( 3 ), foo@$_2$@( 3 ) )@, giving (@int@, {\color{ForestGreen}@int@}, @double@) to (@int@, {\color{ForestGreen}@double@}, @double@) with one {\color{ForestGreen}safe} (widening) conversion from @int@ to @double@ versus ({\color{red}@double@}, {\color{ForestGreen}@int@}, {\color{ForestGreen}@int@}) to ({\color{red}@int@}, {\color{ForestGreen}@double@}, {\color{ForestGreen}@double@}) with one {\color{red}unsafe} (narrowing) conversion from @double@ to @int@ and two safe conversions.
585
586
587\subsection{Tuple Variables}
588
589An important observation from function composition is that new variable names are not required to initialize parameters from an MRVF.
590\CFA also allows declaration of tuple variables that can be initialized from an MRVF, since it can be awkward to declare multiple variables of different types, \eg:
591\begin{lstlisting}
592[ int, int ] qr = div( 13, 5 );                         $\C{// tuple-variable declaration and initialization}$
593[ double, double ] qr = div( 13.5, 5.2 );
594\end{lstlisting}
595where the tuple variable-name serves the same purpose as the parameter name(s).
596Tuple variables can be composed of any types, except for array types, since array sizes are generally unknown in C.
597
598One way to access the tuple-variable components is with assignment or composition:
599\begin{lstlisting}
600[ q, r ] = qr;                                                          $\C{// access tuple-variable components}$
601printf( "%d %d\n", qr );
602\end{lstlisting}
603\CFA also supports \newterm{tuple indexing} to access single components of a tuple expression:
604\begin{lstlisting}
605[int, int] * p = &qr;                                           $\C{// tuple pointer}$
606int rem = qr`.1`;                                                       $\C{// access remainder}$
607int quo = div( 13, 5 )`.0`;                                     $\C{// access quotient}$
608p`->0` = 5;                                                                     $\C{// change quotient}$
609bar( qr`.1`, qr );                                                      $\C{// pass remainder and quotient/remainder}$
610rem = [div( 13, 5 ), 42]`.0.1`;                         $\C{// access 2nd component of 1st component of tuple expression}$
611\end{lstlisting}
612
613
614\subsection{Flattening and Restructuring}
615
616In function call contexts, tuples support implicit flattening and restructuring conversions.
617Tuple flattening recursively expands a tuple into the list of its basic components.
618Tuple structuring packages a list of expressions into a value of tuple type, \eg:
619%\lstDeleteShortInline@%
620%\par\smallskip
621%\begin{tabular}{@{}l@{\hspace{1.5\parindent}}||@{\hspace{1.5\parindent}}l@{}}
622\begin{lstlisting}
623int f( int, int );
624int g( [int, int] );
625int h( int, [int, int] );
626[int, int] x;
627int y;
628f( x );                 $\C{// flatten}$
629g( y, 10 );             $\C{// structure}$
630h( x, y );              $\C{// flatten and structure}$
631\end{lstlisting}
632%\end{lstlisting}
633%&
634%\begin{lstlisting}
635%\end{tabular}
636%\smallskip\par\noindent
637%\lstMakeShortInline@%
638In the call to @f@, @x@ is implicitly flattened so the components of @x@ are passed as the two arguments.
639In the call to @g@, the values @y@ and @10@ are structured into a single argument of type @[int, int]@ to match the parameter type of @g@.
640Finally, in the call to @h@, @x@ is flattened to yield an argument list of length 3, of which the first component of @x@ is passed as the first parameter of @h@, and the second component of @x@ and @y@ are structured into the second argument of type @[int, int]@.
641The flexible structure of tuples permits a simple and expressive function call syntax to work seamlessly with both SRVF and MRVF, and with any number of arguments of arbitrarily complex structure.
642
643
644\subsection{Tuple Assignment}
645
646An assignment where the left side is a tuple type is called \newterm{tuple assignment}.
647There are two kinds of tuple assignment depending on whether the right side of the assignment operator has a tuple type or a non-tuple type, called \newterm{multiple} and \newterm{mass assignment}, respectively.
648%\lstDeleteShortInline@%
649%\par\smallskip
650%\begin{tabular}{@{}l@{\hspace{1.5\parindent}}||@{\hspace{1.5\parindent}}l@{}}
651\begin{lstlisting}
652int x = 10;
653double y = 3.5;
654[int, double] z;
655z = [x, y];                                                                     $\C{// multiple assignment}$
656[x, y] = z;                                                                     $\C{// multiple assignment}$
657z = 10;                                                                         $\C{// mass assignment}$
658[y, x] = 3.14;                                                          $\C{// mass assignment}$
659\end{lstlisting}
660%\end{lstlisting}
661%&
662%\begin{lstlisting}
663%\end{tabular}
664%\smallskip\par\noindent
665%\lstMakeShortInline@%
666Both kinds of tuple assignment have parallel semantics, so that each value on the left and right side is evaluated before any assignments occur.
667As a result, it is possible to swap the values in two variables without explicitly creating any temporary variables or calling a function, \eg, @[x, y] = [y, x]@.
668This semantics means mass assignment differs from C cascading assignment (\eg @a = b = c@) in that conversions are applied in each individual assignment, which prevents data loss from the chain of conversions that can happen during a cascading assignment.
669For example, @[y, x] = 3.14@ performs the assignments @y = 3.14@ and @x = 3.14@, yielding @y == 3.14@ and @x == 3@;
670whereas, C cascading assignment @y = x = 3.14@ performs the assignments @x = 3.14@ and @y = x@, yielding @3@ in @y@ and @x@.
671Finally, tuple assignment is an expression where the result type is the type of the left-hand side of the assignment, just like all other assignment expressions in C.
672This example shows mass, multiple, and cascading assignment used in one expression:
673\begin{lstlisting}
674void f( [int, int] );
675f( [x, y] = z = 1.5 );                                          $\C{// assignments in parameter list}$
676\end{lstlisting}
677
678
679\subsection{Member Access}
680
681It is also possible to access multiple fields from a single expression using a \newterm{member-access}.
682The result is a single tuple-valued expression whose type is the tuple of the types of the members, \eg:
683\begin{lstlisting}
684struct S { int x; double y; char * z; } s;
685s.[x, y, z] = 0;
686\end{lstlisting}
687Here, the mass assignment sets all members of @s@ to zero.
688Since tuple-index expressions are a form of member-access expression, it is possible to use tuple-index expressions in conjunction with member tuple expressions to manually restructure a tuple (\eg rearrange, drop, and duplicate components).
689%\lstDeleteShortInline@%
690%\par\smallskip
691%\begin{tabular}{@{}l@{\hspace{1.5\parindent}}||@{\hspace{1.5\parindent}}l@{}}
692\begin{lstlisting}
693[int, int, long, double] x;
694void f( double, long );
695x.[0, 1] = x.[1, 0];                                            $\C{// rearrange: [x.0, x.1] = [x.1, x.0]}$
696f( x.[0, 3] );                                                          $\C{// drop: f(x.0, x.3)}$
697[int, int, int] y = x.[2, 0, 2];                        $\C{// duplicate: [y.0, y.1, y.2] = [x.2, x.0.x.2]}$
698\end{lstlisting}
699%\end{lstlisting}
700%&
701%\begin{lstlisting}
702%\end{tabular}
703%\smallskip\par\noindent
704%\lstMakeShortInline@%
705It is also possible for a member access to contain other member accesses, \eg:
706\begin{lstlisting}
707struct A { double i; int j; };
708struct B { int * k; short l; };
709struct C { int x; A y; B z; } v;
710v.[x, y.[i, j], z.k];                                           $\C{// [v.x, [v.y.i, v.y.j], v.z.k]}$
711\end{lstlisting}
712
713
714\begin{comment}
715\subsection{Casting}
716
717In C, the cast operator is used to explicitly convert between types.
718In \CFA, the cast operator has a secondary use as type ascription.
719That is, a cast can be used to select the type of an expression when it is ambiguous, as in the call to an overloaded function:
720\begin{lstlisting}
721int f();     // (1)
722double f()// (2)
723
724f();       // ambiguous - (1),(2) both equally viable
725(int)f()// choose (2)
726\end{lstlisting}
727
728Since casting is a fundamental operation in \CFA, casts should be given a meaningful interpretation in the context of tuples.
729Taking a look at standard C provides some guidance with respect to the way casts should work with tuples:
730\begin{lstlisting}
731int f();
732void g();
733
734(void)f()// (1)
735(int)g()// (2)
736\end{lstlisting}
737In C, (1) is a valid cast, which calls @f@ and discards its result.
738On the other hand, (2) is invalid, because @g@ does not produce a result, so requesting an @int@ to materialize from nothing is nonsensical.
739Generalizing these principles, any cast wherein the number of components increases as a result of the cast is invalid, while casts that have the same or fewer number of components may be valid.
740
741Formally, a cast to tuple type is valid when $T_n \leq S_m$, where $T_n$ is the number of components in the target type and $S_m$ is the number of components in the source type, and for each $i$ in $[0, n)$, $S_i$ can be cast to $T_i$.
742Excess elements ($S_j$ for all $j$ in $[n, m)$) are evaluated, but their values are discarded so that they are not included in the result expression.
743This approach follows naturally from the way that a cast to @void@ works in C.
744
745For example, in
746\begin{lstlisting}
747[int, int, int] f();
748[int, [int, int], int] g();
749
750([int, double])f();           $\C{// (1)}$
751([int, int, int])g();         $\C{// (2)}$
752([void, [int, int]])g();      $\C{// (3)}$
753([int, int, int, int])g();    $\C{// (4)}$
754([int, [int, int, int]])g()$\C{// (5)}$
755\end{lstlisting}
756
757(1) discards the last element of the return value and converts the second element to @double@.
758Since @int@ is effectively a 1-element tuple, (2) discards the second component of the second element of the return value of @g@.
759If @g@ is free of side effects, this expression is equivalent to @[(int)(g().0), (int)(g().1.0), (int)(g().2)]@.
760Since @void@ is effectively a 0-element tuple, (3) discards the first and third return values, which is effectively equivalent to @[(int)(g().1.0), (int)(g().1.1)]@).
761
762Note that a cast is not a function call in \CFA, so flattening and structuring conversions do not occur for cast expressions\footnote{User-defined conversions have been considered, but for compatibility with C and the existing use of casts as type ascription, any future design for such conversions would require more precise matching of types than allowed for function arguments and parameters.}.
763As such, (4) is invalid because the cast target type contains 4 components, while the source type contains only 3.
764Similarly, (5) is invalid because the cast @([int, int, int])(g().1)@ is invalid.
765That is, it is invalid to cast @[int, int]@ to @[int, int, int]@.
766\end{comment}
767
768
769\subsection{Polymorphism}
770
771Tuples also integrate with \CFA polymorphism as a kind of generic type.
772Due to the implicit flattening and structuring conversions involved in argument passing, @otype@ and @dtype@ parameters are restricted to matching only with non-tuple types, \eg:
773\begin{lstlisting}
774forall(otype T, dtype U) void f( T x, U * y );
775f( [5, "hello"] );
776\end{lstlisting}
777where @[5, "hello"]@ is flattened, giving argument list @5, "hello"@, and @T@ binds to @int@ and @U@ binds to @const char@.
778Tuples, however, may contain polymorphic components.
779For example, a plus operator can be written to add two triples together.
780\begin{lstlisting}
781forall(otype T | { T ?+?( T, T ); }) [T, T, T] ?+?( [T, T, T] x, [T, T, T] y ) {
782        return [x.0 + y.0, x.1 + y.1, x.2 + y.2];
783}
784[int, int, int] x;
785int i1, i2, i3;
786[i1, i2, i3] = x + ([10, 20, 30]);
787\end{lstlisting}
788
789Flattening and restructuring conversions are also applied to tuple types in polymorphic type assertions.
790\begin{lstlisting}
791int f( [int, double], double );
792forall(otype T, otype U | { T f( T, U, U ); }) void g( T, U );
793g( 5, 10.21 );
794\end{lstlisting}
795Hence, function parameter and return lists are flattened for the purposes of type unification allowing the example to pass expression resolution.
796This relaxation is possible by extending the thunk scheme described by Bilson~\cite{Bilson03}.
797Whenever a candidate's parameter structure does not exactly match the formal parameter's structure, a thunk is generated to specialize calls to the actual function:
798\begin{lstlisting}
799int _thunk( int _p0, double _p1, double _p2 ) { return f( [_p0, _p1], _p2 ); }
800\end{lstlisting}
801so the thunk provides flattening and structuring conversions to inferred functions, improving the compatibility of tuples and polymorphism.
802These thunks take advantage of GCC C nested-functions to produce closures that have the usual function-pointer signature.
803
804
805\subsection{Variadic Tuples}
806\label{sec:variadic-tuples}
807
808To define variadic functions, \CFA adds a new kind of type parameter, @ttype@ (tuple type).
809Matching against a @ttype@ parameter consumes all remaining argument components and packages them into a tuple, binding to the resulting tuple of types.
810In a given parameter list, there must be at most one @ttype@ parameter that occurs last, which matches normal variadic semantics, with a strong feeling of similarity to \CCeleven variadic templates.
811As such, @ttype@ variables are also called \newterm{argument packs}.
812
813Like variadic templates, the main way to manipulate @ttype@ polymorphic functions is via recursion.
814Since nothing is known about a parameter pack by default, assertion parameters are key to doing anything meaningful.
815Unlike variadic templates, @ttype@ polymorphic functions can be separately compiled.
816For example, a generalized @sum@ function written using @ttype@:
817\begin{lstlisting}
818int sum$\(_0\)$() { return 0; }
819forall(ttype Params | { int sum( Params ); } ) int sum$\(_1\)$( int x, Params rest ) {
820        return x + sum( rest );
821}
822sum( 10, 20, 30 );
823\end{lstlisting}
824Since @sum@\(_0\) does not accept any arguments, it is not a valid candidate function for the call @sum(10, 20, 30)@.
825In order to call @sum@\(_1\), @10@ is matched with @x@, and the argument resolution moves on to the argument pack @rest@, which consumes the remainder of the argument list and @Params@ is bound to @[20, 30]@.
826The process continues unitl @Params@ is bound to @[]@, requiring an assertion @int sum()@, which matches @sum@\(_0\) and terminates the recursion.
827Effectively, this algorithm traces as @sum(10, 20, 30)@ $\rightarrow$ @10 + sum(20, 30)@ $\rightarrow$ @10 + (20 + sum(30))@ $\rightarrow$ @10 + (20 + (30 + sum()))@ $\rightarrow$ @10 + (20 + (30 + 0))@.
828
829It is reasonable to take the @sum@ function a step further to enforce a minimum number of arguments:
830\begin{lstlisting}
831int sum( int x, int y ) { return x + y; }
832forall(ttype Params | { int sum( int, Params ); } ) int sum( int x, int y, Params rest ) {
833        return sum( x + y, rest );
834}
835\end{lstlisting}
836One more step permits the summation of any summable type with all arguments of the same type:
837\begin{lstlisting}
838trait summable(otype T) {
839        T ?+?( T, T );
840};
841forall(otype R | summable( R ) ) R sum( R x, R y ) {
842        return x + y;
843}
844forall(otype R, ttype Params | summable(R) | { R sum(R, Params); } ) R sum(R x, R y, Params rest) {
845        return sum( x + y, rest );
846}
847\end{lstlisting}
848Unlike C variadic functions, it is unnecessary to hard code the number and expected types.
849Furthermore, this code is extendable for any user-defined type with a @?+?@ operator.
850Summing arbitrary heterogeneous lists is possible with similar code by adding the appropriate type variables and addition operators.
851
852It is also possible to write a type-safe variadic print function to replace @printf@:
853\begin{lstlisting}
854struct S { int x, y; };
855forall(otype T, ttype Params | { void print(T); void print(Params); }) void print(T arg, Params rest) {
856        print(arg);  print(rest);
857}
858void print( char * x ) { printf( "%s", x ); }
859void print( int x ) { printf( "%d", x ); }
860void print( S s ) { print( "{ ", s.x, ",", s.y, " }" ); }
861print( "s = ", (S){ 1, 2 }, "\n" );
862\end{lstlisting}
863This example showcases a variadic-template-like decomposition of the provided argument list.
864The individual @print@ functions allow printing a single element of a type.
865The polymorphic @print@ allows printing any list of types, where as each individual type has a @print@ function.
866The individual print functions can be used to build up more complicated @print@ functions, such as @S@, which cannot be done with @printf@ in C.
867
868Finally, it is possible to use @ttype@ polymorphism to provide arbitrary argument forwarding functions.
869For example, it is possible to write @new@ as a library function:
870\begin{lstlisting}
871forall( otype R, otype S ) void ?{}( pair(R, S) *, R, S );
872forall( dtype T, ttype Params | sized(T) | { void ?{}( T *, Params ); } ) T * new( Params p ) {
873        return ((T *)malloc()){ p };                    $\C{// construct into result of malloc}$
874}
875pair( int, char ) * x = new( 42, '!' );
876\end{lstlisting}
877The @new@ function provides the combination of type-safe @malloc@ with a \CFA constructor call, making it impossible to forget constructing dynamically allocated objects.
878This function provides the type-safety of @new@ in \CC, without the need to specify the allocated type again, thanks to return-type inference.
879
880
881\subsection{Implementation}
882
883Tuples are implemented in the \CFA translator via a transformation into \newterm{generic types}.
884For each $N$, the first time an $N$-tuple is seen in a scope a generic type with $N$ type parameters is generated, \eg:
885\begin{lstlisting}
886[int, int] f() {
887        [double, double] x;
888        [int, double, int] y;
889}
890\end{lstlisting}
891is transformed into:
892\begin{lstlisting}
893forall(dtype T0, dtype T1 | sized(T0) | sized(T1)) struct _tuple2 {
894        T0 field_0;                                                             $\C{// generated before the first 2-tuple}$
895        T1 field_1;
896};
897_tuple2(int, int) f() {
898        _tuple2(double, double) x;
899        forall(dtype T0, dtype T1, dtype T2 | sized(T0) | sized(T1) | sized(T2)) struct _tuple3 {
900                T0 field_0;                                                     $\C{// generated before the first 3-tuple}$
901                T1 field_1;
902                T2 field_2;
903        };
904        _tuple3(int, double, int) y;
905}
906\end{lstlisting}
907\begin{sloppypar}
908Tuple expressions are then simply converted directly into compound literals, \eg @[5, 'x', 1.24]@ becomes @(_tuple3(int, char, double)){ 5, 'x', 1.24 }@.
909\end{sloppypar}
910
911\begin{comment}
912Since tuples are essentially structures, tuple indexing expressions are just field accesses:
913\begin{lstlisting}
914void f(int, [double, char]);
915[int, double] x;
916
917x.0+x.1;
918printf("%d %g\n", x);
919f(x, 'z');
920\end{lstlisting}
921Is transformed into:
922\begin{lstlisting}
923void f(int, _tuple2(double, char));
924_tuple2(int, double) x;
925
926x.field_0+x.field_1;
927printf("%d %g\n", x.field_0, x.field_1);
928f(x.field_0, (_tuple2){ x.field_1, 'z' });
929\end{lstlisting}
930Note that due to flattening, @x@ used in the argument position is converted into the list of its fields.
931In the call to @f@, the second and third argument components are structured into a tuple argument.
932Similarly, tuple member expressions are recursively expanded into a list of member access expressions.
933
934Expressions that may contain side effects are made into \newterm{unique expressions} before being expanded by the flattening conversion.
935Each unique expression is assigned an identifier and is guaranteed to be executed exactly once:
936\begin{lstlisting}
937void g(int, double);
938[int, double] h();
939g(h());
940\end{lstlisting}
941Internally, this expression is converted to two variables and an expression:
942\begin{lstlisting}
943void g(int, double);
944[int, double] h();
945
946_Bool _unq0_finished_ = 0;
947[int, double] _unq0;
948g(
949        (_unq0_finished_ ? _unq0 : (_unq0 = f(), _unq0_finished_ = 1, _unq0)).0,
950        (_unq0_finished_ ? _unq0 : (_unq0 = f(), _unq0_finished_ = 1, _unq0)).1,
951);
952\end{lstlisting}
953Since argument evaluation order is not specified by the C programming language, this scheme is built to work regardless of evaluation order.
954The first time a unique expression is executed, the actual expression is evaluated and the accompanying boolean is set to true.
955Every subsequent evaluation of the unique expression then results in an access to the stored result of the actual expression.
956Tuple member expressions also take advantage of unique expressions in the case of possible impurity.
957
958Currently, the \CFA translator has a very broad, imprecise definition of impurity, where any function call is assumed to be impure.
959This notion could be made more precise for certain intrinsic, auto-generated, and builtin functions, and could analyze function bodies when they are available to recursively detect impurity, to eliminate some unique expressions.
960
961The various kinds of tuple assignment, constructors, and destructors generate GNU C statement expressions.
962A variable is generated to store the value produced by a statement expression, since its fields may need to be constructed with a non-trivial constructor and it may need to be referred to multiple time, \eg in a unique expression.
963The use of statement expressions allows the translator to arbitrarily generate additional temporary variables as needed, but binds the implementation to a non-standard extension of the C language.
964However, there are other places where the \CFA translator makes use of GNU C extensions, such as its use of nested functions, so this restriction is not new.
965\end{comment}
966
967
968\section{Control Structures}
969
970\CFA identifies missing and problematic control structures in C, and extends and modifies these control structures to increase functionality and safety.
971
972
973\subsection{\texorpdfstring{Labelled \LstKeywordStyle{continue} / \LstKeywordStyle{break}}{Labelled continue / break}}
974
975While C provides @continue@ and @break@ statements for altering control flow, both are restricted to one level of nesting for a particular control structure.
976Unfortunately, this restriction forces programmers to use @goto@ to achieve the equivalent control-flow for more than one level of nesting.
977To prevent having to switch to the @goto@, \CFA extends the @continue@ and @break@ with a target label to support static multi-level exit~\cite{Buhr85}, as in Java.
978For both @continue@ and @break@, the target label must be directly associated with a @for@, @while@ or @do@ statement;
979for @break@, the target label can also be associated with a @switch@, @if@ or compound (@{}@) statement.
980Figure~\ref{f:MultiLevelExit} shows @continue@ and @break@ indicating the specific control structure, and the corresponding C program using only @goto@ and labels.
981The innermost loop has 7 exit points, which cause continuation or termination of one or more of the 7 nested control-structures.
982
983\begin{figure}
984\lstDeleteShortInline@%
985\begin{tabular}{@{\hspace{\parindentlnth}}l@{\hspace{\parindentlnth}}l@{\hspace{\parindentlnth}}l@{}}
986\multicolumn{1}{@{\hspace{\parindentlnth}}c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}   & \multicolumn{1}{@{\hspace{\parindentlnth}}c}{\textbf{C}}      \\
987\begin{cfa}
988`LC:` {
989        ... $declarations$ ...
990        `LS:` switch ( ... ) {
991          case 3:
992                `LIF:` if ( ... ) {
993                        `LF:` for ( ... ) {
994                                `LW:` while ( ... ) {
995                                        ... break `LC`; ...
996                                        ... break `LS`; ...
997                                        ... break `LIF`; ...
998                                        ... continue `LF;` ...
999                                        ... break `LF`; ...
1000                                        ... continue `LW`; ...
1001                                        ... break `LW`; ...
1002                                } // while
1003                        } // for
1004                } else {
1005                        ... break `LIF`; ...
1006                } // if
1007        } // switch
1008} // compound
1009\end{cfa}
1010&
1011\begin{cfa}
1012{
1013        ... $declarations$ ...
1014        switch ( ... ) {
1015          case 3:
1016                if ( ... ) {
1017                        for ( ... ) {
1018                                while ( ... ) {
1019                                        ... goto `LC`; ...
1020                                        ... goto `LS`; ...
1021                                        ... goto `LIF`; ...
1022                                        ... goto `LFC`; ...
1023                                        ... goto `LFB`; ...
1024                                        ... goto `LWC`; ...
1025                                        ... goto `LWB`; ...
1026                                  `LWC`: ; } `LWB:` ;
1027                          `LFC:` ; } `LFB:` ;
1028                } else {
1029                        ... goto `LIF`; ...
1030                } `L3:` ;
1031        } `LS:` ;
1032} `LC:` ;
1033\end{cfa}
1034&
1035\begin{cfa}
1036
1037
1038
1039
1040
1041
1042
1043// terminate compound
1044// terminate switch
1045// terminate if
1046// continue loop
1047// terminate loop
1048// continue loop
1049// terminate loop
1050
1051
1052
1053// terminate if
1054
1055
1056
1057\end{cfa}
1058\end{tabular}
1059\lstMakeShortInline@%
1060\caption{Multi-level Exit}
1061\label{f:MultiLevelExit}
1062\end{figure}
1063
1064Both labelled @continue@ and @break@ are a @goto@ restricted in the following ways:
1065\begin{itemize}
1066\item
1067They cannot create a loop, which means only the looping constructs cause looping.
1068This restriction means all situations resulting in repeated execution are clearly delineated.
1069\item
1070They cannot branch into a control structure.
1071This restriction prevents missing declarations and/or initializations at the start of a control structure resulting in undefined behaviour.
1072\end{itemize}
1073The advantage of the labelled @continue@/@break@ is allowing static multi-level exits without having to use the @goto@ statement, and tying control flow to the target control structure rather than an arbitrary point in a program.
1074Furthermore, the location of the label at the \emph{beginning} of the target control structure informs the reader (eye candy) that complex control-flow is occurring in the body of the control structure.
1075With @goto@, the label is at the end of the control structure, which fails to convey this important clue early enough to the reader.
1076Finally, using an explicit target for the transfer instead of an implicit target allows new constructs to be added or removed without affecting existing constructs.
1077The implicit targets of the current @continue@ and @break@, \ie the closest enclosing loop or @switch@, change as certain constructs are added or removed.
1078
1079
1080\subsection{\texorpdfstring{Enhanced \LstKeywordStyle{switch} Statement}{Enhanced switch Statement}}
1081
1082There are a number of deficiencies with the C @switch@ statements: enumerating @case@ lists, placement of @case@ clauses, scope of the switch body, and fall through between case clauses.
1083
1084C has no shorthand for specifying a list of case values, whether the list is non-contiguous or contiguous\footnote{C provides this mechanism via fall through.}.
1085\CFA provides a shorthand for a non-contiguous list:
1086\begin{cquote}
1087\lstDeleteShortInline@%
1088\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1089\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1090\begin{cfa}
1091case 2, 10, 34, 42:
1092\end{cfa}
1093&
1094\begin{cfa}
1095case 2: case 10: case 34: case 42:
1096\end{cfa}
1097\end{tabular}
1098\lstMakeShortInline@%
1099\end{cquote}
1100for a contiguous list:\footnote{gcc provides the same mechanism with awkward syntax, \lstinline@2 ... 42@, where spaces are required around the ellipse.}
1101\begin{cquote}
1102\lstDeleteShortInline@%
1103\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1104\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1105\begin{cfa}
1106case 2~42:
1107\end{cfa}
1108&
1109\begin{cfa}
1110case 2: case 3: ... case 41: case 42:
1111\end{cfa}
1112\end{tabular}
1113\lstMakeShortInline@%
1114\end{cquote}
1115and a combination:
1116\begin{cfa}
1117case -12~-4, -1~5, 14~21, 34~42:
1118\end{cfa}
1119
1120C allows placement of @case@ clauses \emph{within} statements nested in the @switch@ body (see Duff's device~\cite{Duff83});
1121\begin{cfa}
1122switch ( i ) {
1123  case 0:
1124        for ( int i = 0; i < 10; i += 1 ) {
1125                ...
1126  `case 1:`             // no initialization of loop index
1127                ...
1128        }
1129}
1130\end{cfa}
1131\CFA precludes this form of transfer into a control structure because it causes undefined behaviour, especially with respect to missed initialization, and provides very limited functionality.
1132
1133C allows placement of declaration within the @switch@ body and unreachable code at the start, resulting in undefined behaviour:
1134\begin{cfa}
1135switch ( x ) {
1136        `int y = 1;`                            $\C{// unreachable initialization}$
1137        `x = 7;`                                        $\C{// unreachable code without label/branch}$
1138  case 0:
1139        ...
1140        `int z = 0;`                            $\C{// unreachable initialization, cannot appear after case}$
1141        z = 2;
1142  case 1:
1143        `x = z;`                                        $\C{// without fall through, z is undefined}$
1144}
1145\end{cfa}
1146\CFA allows the declaration of local variables, \eg @y@, at the start of the @switch@ with scope across the entire @switch@ body, \ie all @case@ clauses, but no statements.
1147\CFA disallows the declaration of local variable, \eg @z@, directly within the @switch@ body, because a declaration cannot occur immediately after a @case@ since a label can only be attached to a statement, and the use of @z@ is undefined in @case 1@ as neither storage allocation nor initialization may have occurred.
1148
1149C @switch@ provides multiple entry points into the statement body, but once an entry point is selected, control continues across \emph{all} @case@ clauses until the end of the @switch@ body, called \newterm{fall through};
1150@case@ clauses are made disjoint by the @break@ statement.
1151While the ability to fall through \emph{is} a useful form of control flow, it does not match well with programmer intuition, resulting in many errors from missing @break@ statements.
1152\CFA provides a new control structure, @choose@, which mimics @switch@, but reverses the meaning of fall through:
1153\begin{cquote}
1154\lstDeleteShortInline@%
1155\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1156\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1157\begin{cfa}
1158`choose` ( day ) {
1159  case Mon~Thu:
1160        // program
1161
1162  case Fri:
1163        // program
1164        wallet += pay;
1165        `fallthrough;`
1166  case Sat:
1167        // party
1168        wallet -= party;
1169
1170  case Sun:
1171        // rest
1172
1173  default:
1174        // error
1175}
1176\end{cfa}
1177&
1178\begin{cfa}
1179switch ( day ) {
1180  case Mon: case Tue: case Wed: case Thu:
1181        // program
1182        `break;`
1183  case Fri:
1184        // program
1185        wallet += pay;
1186
1187  case Sat:
1188        // party
1189        wallet -= party;
1190        `break;`
1191  case Sun:
1192        // rest
1193        `break;`
1194  default:
1195        // error
1196}
1197\end{cfa}
1198\end{tabular}
1199\lstMakeShortInline@%
1200\end{cquote}
1201Collectively, these enhancements reduce programmer burden and increase readability and safety.
1202
1203\begin{comment}
1204Forgotten @break@ statements at the end of @switch@ cases are a persistent sort of programmer error in C, and the @break@ statements themselves introduce visual clutter and an un-C-like keyword-based block delimiter.
1205\CFA addresses this error by introducing a @choose@ statement, which works identically to a @switch@ except that its default end-of-case behaviour is to break rather than to fall through for all non-empty cases.
1206Since empty cases like @case 7:@ in @case 7: case 11:@ still have fall-through semantics and explicit @break@ is still allowed at the end of a @choose@ case, many idiomatic uses of @switch@ in standard C can be converted to @choose@ statements by simply changing the keyword.
1207Where fall-through is desired for a non-empty case, it can be specified with the new @fallthrough@ statement, making @choose@ equivalently powerful to @switch@, but more concise in the common case where most non-empty cases end with a @break@ statement, as in the example below:
1208
1209\begin{cfa}
1210choose( i ) {
1211        case 2:
1212                printf("even ");
1213                fallthrough;
1214        case 3: case 5: case 7:
1215                printf("small prime\n");
1216        case 4,6,8,9:
1217                printf("small composite\n");
1218        case 13~19:
1219                printf("teen\n");
1220        default:
1221                printf("something else\n");
1222}
1223\end{cfa}
1224\end{comment}
1225
1226
1227\subsection{\texorpdfstring{\LstKeywordStyle{with} Clause / Statement}{with Clause / Statement}}
1228\label{s:WithClauseStatement}
1229
1230Grouping heterogenous data into \newterm{aggregate}s (structure/union) is a common programming practice, and an aggregate can be further organized into more complex structures, such as arrays and containers:
1231\begin{cfa}
1232struct S {                                                                      $\C{// aggregate}$
1233        char c;                                                                 $\C{// fields}$
1234        int i;
1235        double d;
1236};
1237S s, as[10];
1238\end{cfa}
1239However, routines manipulating aggregates must repeat the aggregate name to access its containing fields:
1240\begin{cfa}
1241void f( S s ) {
1242        `s.`c; `s.`i; `s.`d;                                    $\C{// access containing fields}$
1243}
1244\end{cfa}
1245A similar situation occurs in object-oriented programming, \eg \CC:
1246\begin{C++}
1247class C {
1248        char c;                                                                 $\C{// fields}$
1249        int i;
1250        double d;
1251        int mem() {                                                             $\C{// implicit "this" parameter}$
1252                `this->`c; `this->`i; `this->`d;        $\C{// access containing fields}$
1253        }
1254}
1255\end{C++}
1256Nesting of member routines in a \lstinline[language=C++]@class@ allows eliding \lstinline[language=C++]@this->@ because of lexical scoping.
1257However, for other aggregate parameters, qualification is necessary:
1258\begin{cfa}
1259struct T { double m, n; };
1260int C::mem( T & t ) {                                           $\C{// multiple aggregate parameters}$
1261        c; i; d;                                                                $\C{\color{red}// this-\textgreater.c, this-\textgreater.i, this-\textgreater.d}$
1262        `t.`m; `t.`n;                                                   $\C{// must qualify}$
1263}
1264\end{cfa}
1265
1266% In object-oriented programming, there is an implicit first parameter, often names @self@ or @this@, which is elided.
1267% In any programming language, some functions have a naturally close relationship with a particular data type.
1268% Object-oriented programming allows this close relationship to be codified in the language by making such functions \newterm{class methods} of their related data type.
1269% Class methods have certain privileges with respect to their associated data type, notably un-prefixed access to the fields of that data type.
1270% When writing C functions in an object-oriented style, this un-prefixed access is swiftly missed, as access to fields of a @Foo* f@ requires an extra three characters @f->@ every time, which disrupts coding flow and clutters the produced code.
1271%
1272% \TODO{Fill out section. Be sure to mention arbitrary expressions in with-blocks, recent change driven by Thierry to prioritize field name over parameters.}
1273
1274To simplify the programmer experience, \CFA provides a @with@ clause/statement (see Pascal~\cite[\S~4.F]{Pascal}) to elide aggregate qualification to fields by opening a scope containing the field identifiers.
1275Hence, the qualified fields become variables with the side-effect that it is easier to optimizing field references in a block.
1276\begin{cfa}
1277void f( S s ) `with( s )` {                                     $\C{// with clause}$
1278        c; i; d;                                                                $\C{\color{red}// s.c, s.i, s.d}$
1279}
1280\end{cfa}
1281and the equivalence for object-style programming is:
1282\begin{cfa}
1283int mem( S & this ) `with( this )` {            $\C{// with clause}$
1284        c; i; d;                                                                $\C{\color{red}// this.c, this.i, this.d}$
1285}
1286\end{cfa}
1287with the generality of opening multiple aggregate-parameters:
1288\begin{cfa}
1289int mem( S & s, T & t ) `with( s, t )` {        $\C{// multiple aggregate parameters}$
1290        c; i; d;                                                                $\C{\color{red}// s.c, s.i, s.d}$
1291        m; n;                                                                   $\C{\color{red}// t.m, t.n}$
1292}
1293\end{cfa}
1294
1295In detail, the @with@ clause/statement has the form:
1296\begin{cfa}
1297$\emph{with-statement}$:
1298        'with' '(' $\emph{expression-list}$ ')' $\emph{compound-statement}$
1299\end{cfa}
1300and may appear as the body of a routine or nested within a routine body.
1301Each expression in the expression-list provides a type and object.
1302The type must be an aggregate type.
1303(Enumerations are already opened.)
1304The object is the implicit qualifier for the open structure-fields.
1305
1306All expressions in the expression list are open in ``parallel'' within the compound statement.
1307This semantic is different from Pascal, which nests the openings.
1308The difference between parallel and nesting occurs for fields with the same name but different type:
1309\begin{cfa}
1310struct S { int i; int j; double m; } s, w;
1311struct T { int i; int k; int m } t, w;
1312with( s, t ) {
1313        j + k;                                                                  $\C{// unambiguous, s.j + t.m}$
1314        m = 5.0;                                                                $\C{// unambiguous, t.m = 5.0}$
1315        m = 1;                                                                  $\C{// unambiguous, s.m = 1}$
1316        int a = s.i + m;                                                $\C{// unambiguous, a = s.i + t.i}$
1317        int b = s.i + t.i;                                              $\C{// unambiguous, qualification}$
1318        sout | (double)m | endl;                                $\C{// unambiguous, cast}$
1319        i;                                                                              $\C{// ambiguous}$
1320}
1321\end{cfa}
1322\CFA's ability to overload variables means usages of field with the same names can be automatically disambiguated, eliminating most qualification.
1323Qualification or a cast is used to disambiguate.
1324A cast may be necessary to disambiguate between the overload variables in a @with@ expression:
1325\begin{cfa}
1326with( w ) { ... }                                                       $\C{// ambiguous, same name and no context}$
1327with( (S)w ) { ... }                                            $\C{// unambiguous}$
1328\end{cfa}
1329
1330\begin{cfa}
1331struct S { int i, j; } sv;
1332with( sv ) {
1333        S & sr = sv;
1334        with( sr ) {
1335                S * sp = &sv;
1336                with( *sp ) {
1337                        i = 3; j = 4;                                   $\C{\color{red}// sp-{\textgreater}i, sp-{\textgreater}j}$
1338                }
1339                i = 3; j = 4;                                           $\C{\color{red}// sr.i, sr.j}$
1340        }
1341        i = 3; j = 4;                                                   $\C{\color{red}// sv.i, sv.j}$
1342}
1343\end{cfa}
1344
1345The statement form is used within a block:
1346\begin{cfa}
1347int foo() {
1348        struct S1 { ... } s1;
1349        struct S2 { ... } s2;
1350        `with( s1 )` {                                                  $\C{// with statement}$
1351                // access fields of s1 without qualification
1352                `with( s2 )` {                                          $\C{// nesting}$
1353                        // access fields of s1 and s2 without qualification
1354                }
1355        }
1356        `with( s1, s2 )` {
1357                // access unambiguous fields of s1 and s2 without qualification
1358        }
1359}
1360\end{cfa}
1361
1362% \subsection{Exception Handling ???}
1363
1364\section{Declarations}
1365
1366It is important to the design team that \CFA subjectively ``feel like'' C to user programmers.
1367An important part of this subjective feel is maintaining C's procedural programming paradigm, as opposed to the object-oriented paradigm of other systems languages such as \CC and Rust.
1368Maintaining this procedural paradigm means that coding patterns that work in C will remain not only functional but idiomatic in \CFA, reducing the mental burden of retraining C programmers and switching between C and \CFA development.
1369Nonetheless, some features of object-oriented languages are undeniably convienient, and the \CFA design team has attempted to adapt them to a procedural paradigm so as to incorporate their benefits into \CFA; two of these features are resource management and name scoping.
1370
1371
1372\subsection{Alternative Declaration Syntax}
1373
1374\newcommand{\R}[1]{\Textbf{#1}}
1375\newcommand{\B}[1]{{\Textbf[blue]{#1}}}
1376\newcommand{\G}[1]{{\Textbf[OliveGreen]{#1}}}
1377
1378C declaration syntax is notoriously confusing and error prone.
1379For example, many C programmers are confused by a declaration as simple as:
1380\begin{cquote}
1381\lstDeleteShortInline@%
1382\begin{tabular}{@{}ll@{}}
1383\begin{cfa}
1384int * x[5]
1385\end{cfa}
1386&
1387\raisebox{-0.75\totalheight}{\input{Cdecl}}
1388\end{tabular}
1389\lstMakeShortInline@%
1390\end{cquote}
1391Is this an array of 5 pointers to integers or a pointer to an array of 5 integers?
1392If there is any doubt, it implies productivity and safety issues even for basic programs.
1393Another example of confusion results from the fact that a routine name and its parameters are embedded within the return type, mimicking the way the return value is used at the routine's call site.
1394For example, a routine returning a pointer to an array of integers is defined and used in the following way:
1395\begin{cfa}
1396int `(*`f`())[`5`]` {...};                              $\C{// definition}$
1397 ... `(*`f`())[`3`]` += 1;                              $\C{// usage}$
1398\end{cfa}
1399Essentially, the return type is wrapped around the routine name in successive layers (like an onion).
1400While attempting to make the two contexts consistent is a laudable goal, it has not worked out in practice.
1401
1402\CFA provides its own type, variable and routine declarations, using a different syntax.
1403The new declarations place qualifiers to the left of the base type, while C declarations place qualifiers to the right of the base type.
1404In the following example, \R{red} is the base type and \B{blue} is qualifiers.
1405The \CFA declarations move the qualifiers to the left of the base type, \ie move the blue to the left of the red, while the qualifiers have the same meaning but are ordered left to right to specify a variable's type.
1406\begin{cquote}
1407\lstDeleteShortInline@%
1408\lstset{moredelim=**[is][\color{blue}]{+}{+}}
1409\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1410\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1411\begin{cfa}
1412+[5] *+ `int` x1;
1413+* [5]+ `int` x2;
1414`[* [5] int]` f+( int p )+;
1415\end{cfa}
1416&
1417\begin{cfa}
1418`int` +*+ x1 +[5]+;
1419`int` +(*+x2+)[5]+;
1420`int (*`f+( int p )+`)[5]`;
1421\end{cfa}
1422\end{tabular}
1423\lstMakeShortInline@%
1424\end{cquote}
1425The only exception is bit field specification, which always appear to the right of the base type.
1426% Specifically, the character @*@ is used to indicate a pointer, square brackets @[@\,@]@ are used to represent an array or function return value, and parentheses @()@ are used to indicate a routine parameter.
1427However, unlike C, \CFA type declaration tokens are distributed across all variables in the declaration list.
1428For instance, variables @x@ and @y@ of type pointer to integer are defined in \CFA as follows:
1429\begin{cquote}
1430\lstDeleteShortInline@%
1431\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1432\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1433\begin{cfa}
1434`*` int x, y;
1435\end{cfa}
1436&
1437\begin{cfa}
1438int `*`x, `*`y;
1439\end{cfa}
1440\end{tabular}
1441\lstMakeShortInline@%
1442\end{cquote}
1443The downside of this semantics is the need to separate regular and pointer declarations:
1444\begin{cquote}
1445\lstDeleteShortInline@%
1446\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1447\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1448\begin{cfa}
1449`*` int x;
1450int y;
1451\end{cfa}
1452&
1453\begin{cfa}
1454int `*`x, y;
1455
1456\end{cfa}
1457\end{tabular}
1458\lstMakeShortInline@%
1459\end{cquote}
1460which is prescribing a safety benefit.
1461Other examples are:
1462\begin{cquote}
1463\lstDeleteShortInline@%
1464\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{\hspace{\parindentlnth}}l@{}}
1465\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{C}}      \\
1466\begin{cfa}
1467[ 5 ] int z;
1468[ 5 ] * char w;
1469* [ 5 ] double v;
1470struct s {
1471        int f0:3;
1472        * int f1;
1473        [ 5 ] * int f2;
1474};
1475\end{cfa}
1476&
1477\begin{cfa}
1478int z[ 5 ];
1479char * w[ 5 ];
1480double (* v)[ 5 ];
1481struct s {
1482        int f0:3;
1483        int * f1;
1484        int * f2[ 5 ]
1485};
1486\end{cfa}
1487&
1488\begin{cfa}
1489// array of 5 integers
1490// array of 5 pointers to char
1491// pointer to array of 5 doubles
1492
1493// common bit field syntax
1494
1495
1496
1497\end{cfa}
1498\end{tabular}
1499\lstMakeShortInline@%
1500\end{cquote}
1501
1502All type qualifiers, \eg @const@, @volatile@, etc., are used in the normal way with the new declarations and also appear left to right, \eg:
1503\begin{cquote}
1504\lstDeleteShortInline@%
1505\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{\hspace{\parindentlnth}}l@{}}
1506\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{C}}      \\
1507\begin{cfa}
1508const * const int x;
1509const * [ 5 ] const int y;
1510\end{cfa}
1511&
1512\begin{cfa}
1513int const * const x;
1514const int (* const y)[ 5 ]
1515\end{cfa}
1516&
1517\begin{cfa}
1518// const pointer to const integer
1519// const pointer to array of 5 const integers
1520\end{cfa}
1521\end{tabular}
1522\lstMakeShortInline@%
1523\end{cquote}
1524All declaration qualifiers, \eg @extern@, @static@, etc., are used in the normal way with the new declarations but can only appear at the start of a \CFA routine declaration,\footnote{\label{StorageClassSpecifier}
1525The placement of a storage-class specifier other than at the beginning of the declaration specifiers in a declaration is an obsolescent feature.~\cite[\S~6.11.5(1)]{C11}} \eg:
1526\begin{cquote}
1527\lstDeleteShortInline@%
1528\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{\hspace{\parindentlnth}}l@{}}
1529\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{C}}      \\
1530\begin{cfa}
1531extern [ 5 ] int x;
1532static * const int y;
1533\end{cfa}
1534&
1535\begin{cfa}
1536int extern x[ 5 ];
1537const int static * y;
1538\end{cfa}
1539&
1540\begin{cfa}
1541// externally visible array of 5 integers
1542// internally visible pointer to constant int
1543\end{cfa}
1544\end{tabular}
1545\lstMakeShortInline@%
1546\end{cquote}
1547
1548The new declaration syntax can be used in other contexts where types are required, \eg casts and the pseudo-routine @sizeof@:
1549\begin{cquote}
1550\lstDeleteShortInline@%
1551\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1552\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1553\begin{cfa}
1554y = (* int)x;
1555i = sizeof([ 5 ] * int);
1556\end{cfa}
1557&
1558\begin{cfa}
1559y = (int *)x;
1560i = sizeof(int * [ 5 ]);
1561\end{cfa}
1562\end{tabular}
1563\lstMakeShortInline@%
1564\end{cquote}
1565
1566Finally, new \CFA declarations may appear together with C declarations in the same program block, but cannot be mixed within a specific declaration.
1567Therefore, a programmer has the option of either continuing to use traditional C declarations or take advantage of the new style.
1568Clearly, both styles need to be supported for some time due to existing C-style header-files, particularly for UNIX-like systems.
1569
1570The syntax of the new routine prototype declaration follows directly from the new routine definition syntax;
1571as well, parameter names are optional, \eg:
1572\begin{cfa}
1573[ int x ] f ();                                                 $\C{// returning int with no parameters}$
1574[ * int ] g (int y);                                    $\C{// returning pointer to int with int parameter}$
1575[ ] h ( int, char );                                    $\C{// returning no result with int and char parameters}$
1576[ * int, int ] j ( int );                               $\C{// returning pointer to int and int, with int parameter}$
1577\end{cfa}
1578This syntax allows a prototype declaration to be created by cutting and pasting source text from the routine definition header (or vice versa).
1579Like C, it is possible to declare multiple routine-prototypes in a single declaration, where the return type is distributed across \emph{all} routine names in the declaration list, \eg:
1580\begin{cquote}
1581\lstDeleteShortInline@%
1582\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1583\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1584\begin{cfa}
1585[double] foo(), foo( int ), foo( double ) {...}
1586\end{cfa}
1587&
1588\begin{cfa}
1589double foo1(), foo2( int ), foo3( double );
1590\end{cfa}
1591\end{tabular}
1592\lstMakeShortInline@%
1593\end{cquote}
1594\CFA allows the last routine in the list to define its body.
1595
1596Declaration qualifiers can only appear at the start of a \CFA routine declaration,\footref{StorageClassSpecifier} \eg:
1597\begin{cfa}
1598extern [ int ] f ( int );
1599static [ int ] g ( int );
1600\end{cfa}
1601
1602The syntax for pointers to \CFA routines specifies the pointer name on the right, \eg:
1603\begin{cfa}
1604* [ int x ] () fp;                                              $\C{// pointer to routine returning int with no parameters}$
1605* [ * int ] (int y) gp;                                 $\C{// pointer to routine returning pointer to int with int parameter}$
1606* [ ] (int,char) hp;                                    $\C{// pointer to routine returning no result with int and char parameters}$
1607* [ * int,int ] ( int ) jp;                             $\C{// pointer to routine returning pointer to int and int, with int parameter}$
1608\end{cfa}
1609While parameter names are optional, \emph{a routine name cannot be specified};
1610for example, the following is incorrect:
1611\begin{cfa}
1612* [ int x ] f () fp;                                    $\C{// routine name "f" is not allowed}$
1613\end{cfa}
1614
1615
1616\subsection{References}
1617
1618All variables in C have an \newterm{address}, a \newterm{value}, and a \newterm{type};
1619at the position in the program's memory denoted by the address, there exists a sequence of bits (the value), with the length and semantic meaning of this bit sequence defined by the type.
1620The C type-system does not always track the relationship between a value and its address;
1621a value that does not have a corresponding address is called a \newterm{rvalue} (for ``right-hand value''), while a value that does have an address is called a \newterm{lvalue} (for ``left-hand value'').
1622For example, in @int x; x = 42;@ the variable expression @x@ on the left-hand-side of the assignment is a lvalue, while the constant expression @42@ on the right-hand-side of the assignment is a rvalue.
1623Despite the nomenclature of ``left-hand'' and ``right-hand'', an expression's classification as lvalue or rvalue is entirely dependent on whether it has an address or not; in imperative programming, the address of a value is used for both reading and writing (mutating) a value, and as such lvalues can be converted to rvalues and read from, but rvalues cannot be mutated because they lack a location to store the updated value.
1624
1625Within a lexical scope, lvalue expressions have an \newterm{address interpretation} for writing a value or a \newterm{value interpretation} to read a value.
1626For example, in @x = y@, @x@ has an address interpretation, while @y@ has a value interpretation.
1627Though this duality of interpretation is useful, C lacks a direct mechanism to pass lvalues between contexts, instead relying on \newterm{pointer types} to serve a similar purpose.
1628In C, for any type @T@ there is a pointer type @T *@, the value of which is the address of a value of type @T@.
1629A pointer rvalue can be explicitly \newterm{dereferenced} to the pointed-to lvalue with the dereference operator @*?@, while the rvalue representing the address of a lvalue can be obtained with the address-of operator @&?@.
1630
1631\begin{cfa}
1632int x = 1, y = 2, * p1, * p2, ** p3;
1633p1 = &x;                                                                $\C{// p1 points to x}$
1634p2 = &y;                                                                $\C{// p2 points to y}$
1635p3 = &p1;                                                               $\C{// p3 points to p1}$
1636*p2 = ((*p1 + *p2) * (**p3 - *p1)) / (**p3 - 15);
1637\end{cfa}
1638
1639Unfortunately, the dereference and address-of operators introduce a great deal of syntactic noise when dealing with pointed-to values rather than pointers, as well as the potential for subtle bugs.
1640For both brevity and clarity, it would be desirable to have the compiler figure out how to elide the dereference operators in a complex expression such as the assignment to @*p2@ above.
1641However, since C defines a number of forms of \newterm{pointer arithmetic}, two similar expressions involving pointers to arithmetic types (\eg @*p1 + x@ and @p1 + x@) may each have well-defined but distinct semantics, introducing the possibility that a user programmer may write one when they mean the other, and precluding any simple algorithm for elision of dereference operators.
1642To solve these problems, \CFA introduces reference types @T&@; a @T&@ has exactly the same value as a @T*@, but where the @T*@ takes the address interpretation by default, a @T&@ takes the value interpretation by default, as below:
1643
1644\begin{cfa}
1645int x = 1, y = 2, & r1, & r2, && r3;
1646&r1 = &x;  $\C{// r1 points to x}$
1647&r2 = &y;  $\C{// r2 points to y}$
1648&&r3 = &&r1;  $\C{// r3 points to r2}$
1649r2 = ((r1 + r2) * (r3 - r1)) / (r3 - 15);  $\C{// implicit dereferencing}$
1650\end{cfa}
1651
1652Except for auto-dereferencing by the compiler, this reference example is exactly the same as the previous pointer example.
1653Hence, a reference behaves like a variable name -- an lvalue expression which is interpreted as a value, but also has the type system track the address of that value.
1654One way to conceptualize a reference is via a rewrite rule, where the compiler inserts a dereference operator before the reference variable for each reference qualifier in the reference variable declaration, so the previous example implicitly acts like:
1655
1656\begin{cfa}
1657`*`r2 = ((`*`r1 + `*`r2) * (`**`r3 - `*`r1)) / (`**`r3 - 15);
1658\end{cfa}
1659
1660References in \CFA are similar to those in \CC, but with a couple important improvements, both of which can be seen in the example above.
1661Firstly, \CFA does not forbid references to references, unlike \CC.
1662This provides a much more orthogonal design for library implementors, obviating the need for workarounds such as @std::reference_wrapper@.
1663
1664Secondly, unlike the references in \CC which always point to a fixed address, \CFA references are rebindable.
1665This allows \CFA references to be default-initialized (\eg to a null pointer), and also to point to different addresses throughout their lifetime.
1666This rebinding is accomplished without adding any new syntax to \CFA, but simply by extending the existing semantics of the address-of operator in C.
1667
1668In C, the address of a lvalue is always a rvalue, as in general that address is not stored anywhere in memory, and does not itself have an address.
1669In \CFA, the address of a @T&@ is a lvalue @T*@, as the address of the underlying @T@ is stored in the reference, and can thus be mutated there.
1670The result of this rule is that any reference can be rebound using the existing pointer assignment semantics by assigning a compatible pointer into the address of the reference, \eg @&r1 = &x;@ above.
1671This rebinding can occur to an arbitrary depth of reference nesting; loosely speaking, nested address-of operators will produce an lvalue nested pointer up to as deep as the reference they're applied to.
1672These explicit address-of operators can be thought of as ``cancelling out'' the implicit dereference operators, \eg @(&`*`)r1 = &x@ or @(&(&`*`)`*`)r3 = &(&`*`)r1@ or even @(&`*`)r2 = (&`*`)`*`r3@ for @&r2 = &r3@.
1673More precisely:
1674\begin{itemize}
1675        \item
1676        if @R@ is an rvalue of type {@T &@$_1 \cdots$@ &@$_r$} where $r \ge 1$ references (@&@ symbols) than @&R@ has type {@T `*`&@$_{\color{red}2} \cdots$@ &@$_{\color{red}r}$}, \\ \ie @T@ pointer with $r-1$ references (@&@ symbols).
1677       
1678        \item
1679        if @L@ is an lvalue of type {@T &@$_1 \cdots$@ &@$_l$} where $l \ge 0$ references (@&@ symbols) then @&L@ has type {@T `*`&@$_{\color{red}1} \cdots$@ &@$_{\color{red}l}$}, \\ \ie @T@ pointer with $l$ references (@&@ symbols).
1680\end{itemize}
1681Since pointers and references share the same internal representation, code using either is equally performant; in fact the \CFA compiler converts references to pointers internally, and the choice between them in user code can be made based solely on convenience.
1682
1683By analogy to pointers, \CFA references also allow cv-qualifiers such as @const@:
1684
1685\begin{cfa}
1686const int cx = 5;               $\C{// cannot change cx}$
1687const int & cr = cx;    $\C{// cannot change cr's referred value}$
1688&cr = &cx;                              $\C{// rebinding cr allowed}$
1689cr = 7;                                 $\C{// ERROR, cannot change cr}$
1690int & const rc = x;             $\C{// must be initialized, like in \CC}$
1691&rc = &x;                               $\C{// ERROR, cannot rebind rc}$
1692rc = 7;                                 $\C{// x now equal to 7}$
1693\end{cfa}
1694
1695Given that a reference is meant to represent a lvalue, \CFA provides some syntactic shortcuts when initializing references.
1696There are three initialization contexts in \CFA: declaration initialization, argument/parameter binding, and return/temporary binding.
1697In each of these contexts, the address-of operator on the target lvalue may (in fact, must) be elided.
1698The syntactic motivation for this is clearest when considering overloaded operator-assignment, \eg @int ?+=?(int &, int)@; given @int x, y@, the expected call syntax is @x += y@, not @&x += y@.
1699
1700More generally, this initialization of references from lvalues rather than pointers is an instance of a ``lvalue-to-reference'' conversion rather than an elision of the address-of operator; this conversion can actually be used in any context in \CFA an implicit conversion would be allowed.
1701Similarly, use of a the value pointed to by a reference in an rvalue context can be thought of as a ``reference-to-rvalue'' conversion, and \CFA also includes a qualifier-adding ``reference-to-reference'' conversion, analogous to the @T *@ to @const T *@ conversion in standard C.
1702The final reference conversion included in \CFA is ``rvalue-to-reference'' conversion, implemented by means of an implicit temporary.
1703When an rvalue is used to initialize a reference, it is instead used to initialize a hidden temporary value with the same lexical scope as the reference, and the reference is initialized to the address of this temporary.
1704This allows complex values to be succinctly and efficiently passed to functions, without the syntactic overhead of explicit definition of a temporary variable or the runtime cost of pass-by-value.
1705\CC allows a similar binding, but only for @const@ references; the more general semantics of \CFA are an attempt to avoid the \newterm{const hell} problem, in which addition of a @const@ qualifier to one reference requires a cascading chain of added qualifiers.
1706
1707
1708\subsection{Constructors and Destructors}
1709
1710One of the strengths of C is the control over memory management it gives programmers, allowing resource release to be more consistent and precisely timed than is possible with garbage-collected memory management.
1711However, this manual approach to memory management is often verbose, and it is useful to manage resources other than memory (\eg file handles) using the same mechanism as memory.
1712\CC is well-known for an approach to manual memory management that addresses both these issues, Resource Aquisition Is Initialization (RAII), implemented by means of special \newterm{constructor} and \newterm{destructor} functions; we have implemented a similar feature in \CFA.
1713While RAII is a common feature of object-oriented programming languages, its inclusion in \CFA does not violate the design principle that \CFA retain the same procedural paradigm as C.
1714In particular, \CFA does not implement class-based encapsulation: neither the constructor nor any other function has privileged access to the implementation details of a type, except through the translation-unit-scope method of opaque structs provided by C.
1715
1716In \CFA, a constructor is a function named @?{}@, while a destructor is a function named @^?{}@; like other \CFA operators, these names represent the syntax used to call the constructor or destructor, \eg @x{ ... };@ or @^x{};@.
1717Every constructor and destructor must have a return type of @void@, and its first parameter must have a reference type whose base type is the type of the object the function constructs or destructs.
1718This first parameter is informally called the @this@ parameter, as in many object-oriented languages, though a programmer may give it an arbitrary name.
1719Destructors must have exactly one parameter, while constructors allow passing of zero or more additional arguments along with the @this@ parameter.
1720
1721\begin{cfa}
1722struct Array {
1723        int * data;
1724        int len;
1725};
1726
1727void ?{}( Array& arr ) {
1728        arr.len = 10;
1729        arr.data = calloc( arr.len, sizeof(int) );
1730}
1731
1732void ^?{}( Array& arr ) {
1733        free( arr.data );
1734}
1735
1736{
1737        Array x;
1738        `?{}(x);`       $\C{// implicitly compiler-generated}$
1739        // ... use x
1740        `^?{}(x);`      $\C{// implicitly compiler-generated}$
1741}
1742\end{cfa}
1743
1744In the example above, a \newterm{default constructor} (\ie one with no parameters besides the @this@ parameter) and destructor are defined for the @Array@ struct, a dynamic array of @int@.
1745@Array@ is an example of a \newterm{managed type} in \CFA, a type with a non-trivial constructor or destructor, or with a field of a managed type.
1746As in the example, all instances of managed types are implicitly constructed upon allocation, and destructed upon deallocation; this ensures proper initialization and cleanup of resources contained in managed types, in this case the @data@ array on the heap.
1747The exact details of the placement of these implicit constructor and destructor calls are omitted here for brevity, the interested reader should consult \cite{Schluntz17}.
1748
1749Constructor calls are intended to seamlessly integrate with existing C initialization syntax, providing a simple and familiar syntax to veteran C programmers and allowing constructor calls to be inserted into legacy C code with minimal code changes.
1750As such, \CFA also provides syntax for \newterm{copy initialization} and \newterm{initialization parameters}:
1751
1752\begin{cfa}
1753void ?{}( Array& arr, Array other );
1754
1755void ?{}( Array& arr, int size, int fill );
1756
1757Array y = { 20, 0xDEADBEEF }, z = y;
1758\end{cfa}
1759
1760Copy constructors have exactly two parameters, the second of which has the same type as the base type of the @this@ parameter; appropriate care is taken in the implementation to avoid recursive calls to the copy constructor when initializing this second parameter.
1761Other constructor calls look just like C initializers, except rather than using field-by-field initialization (as in C), an initialization which matches a defined constructor will call the constructor instead.
1762
1763In addition to initialization syntax, \CFA provides two ways to explicitly call constructors and destructors.
1764Explicit calls to constructors double as a placement syntax, useful for construction of member fields in user-defined constructors and reuse of large storage allocations.
1765While the existing function-call syntax works for explicit calls to constructors and destructors, \CFA also provides a more concise \newterm{operator syntax} for both:
1766
1767\begin{cfa}
1768Array a, b;
1769a{};                            $\C{// default construct}$
1770b{ a };                         $\C{// copy construct}$
1771^a{};                           $\C{// destruct}$
1772a{ 5, 0xFFFFFFFF };     $\C{// explicit constructor call}$
1773\end{cfa}
1774
1775To provide a uniform type interface for @otype@ polymorphism, the \CFA compiler automatically generates a default constructor, copy constructor, assignment operator, and destructor for all types.
1776These default functions can be overridden by user-generated versions of them.
1777For compatibility with the standard behaviour of C, the default constructor and destructor for all basic, pointer, and reference types do nothing, while the copy constructor and assignment operator are bitwise copies; if default zero-initialization is desired, the default constructors can be overridden.
1778For user-generated types, the four functions are also automatically generated.
1779@enum@ types are handled the same as their underlying integral type, and unions are also bitwise copied and no-op initialized and destructed.
1780For compatibility with C, a copy constructor from the first union member type is also defined.
1781For @struct@ types, each of the four functions are implicitly defined to call their corresponding functions on each member of the struct.
1782To better simulate the behaviour of C initializers, a set of \newterm{field constructors} is also generated for structures.
1783A constructor is generated for each non-empty prefix of a structure's member-list which copy-constructs the members passed as parameters and default-constructs the remaining members.
1784To allow users to limit the set of constructors available for a type, when a user declares any constructor or destructor, the corresponding generated function and all field constructors for that type are hidden from expression resolution; similarly, the generated default constructor is hidden upon declaration of any constructor.
1785These semantics closely mirror the rule for implicit declaration of constructors in \CC\cite[p.~186]{ANSI98:C++}.
1786
1787In rare situations user programmers may not wish to have constructors and destructors called; in these cases, \CFA provides an ``escape hatch'' to not call them.
1788If a variable is initialized using the syntax \lstinline|S x @= {}| it will be an \newterm{unmanaged object}, and will not have constructors or destructors called.
1789Any C initializer can be the right-hand side of an \lstinline|@=| initializer, \eg  \lstinline|Array a @= { 0, 0x0 }|, with the usual C initialization semantics.
1790In addition to the expressive power, \lstinline|@=| provides a simple path for migrating legacy C code to \CFA, by providing a mechanism to incrementally convert initializers; the \CFA design team decided to introduce a new syntax for this escape hatch because we believe that our RAII implementation will handle the vast majority of code in a desirable way, and we wished to maintain familiar syntax for this common case.
1791
1792
1793\subsection{Type Nesting}
1794
1795\CFA allows \newterm{type nesting}, and type qualification of the nested types (see Figure~\ref{f:TypeNestingQualification}), where as C hoists (refactors) nested types into the enclosing scope and has no type qualification.
1796\begin{figure}
1797\centering
1798\lstDeleteShortInline@%
1799\begin{tabular}{@{}l@{\hspace{3em}}l|l@{}}
1800\multicolumn{1}{c@{\hspace{3em}}}{\textbf{C Type Nesting}}      & \multicolumn{1}{c}{\textbf{C Implicit Hoisting}}      & \multicolumn{1}{|c}{\textbf{\CFA}}    \\
1801\hline
1802\begin{cfa}
1803struct S {
1804        enum C { R, G, B };
1805        struct T {
1806                union U { int i, j; };
1807                enum C c;
1808                short int i, j;
1809        };
1810        struct T t;
1811} s;
1812
1813int rtn() {
1814        s.t.c = R;
1815        struct T t = { R, 1, 2 };
1816        enum C c;
1817        union U u;
1818}
1819\end{cfa}
1820&
1821\begin{cfa}
1822enum C { R, G, B };
1823union U { int i, j; };
1824struct T {
1825        enum C c;
1826        short int i, j;
1827};
1828struct S {
1829        struct T t;
1830} s;
1831       
1832
1833
1834
1835
1836
1837
1838\end{cfa}
1839&
1840\begin{cfa}
1841struct S {
1842        enum C { R, G, B };
1843        struct T {
1844                union U { int i, j; };
1845                enum C c;
1846                short int i, j;
1847        };
1848        struct T t;
1849} s;
1850
1851int rtn() {
1852        s.t.c = `S.`R;  // type qualification
1853        struct `S.`T t = { `S.`R, 1, 2 };
1854        enum `S.`C c;
1855        union `S.T.`U u;
1856}
1857\end{cfa}
1858\end{tabular}
1859\lstMakeShortInline@%
1860\caption{Type Nesting / Qualification}
1861\label{f:TypeNestingQualification}
1862\end{figure}
1863In the left example in C, types @C@, @U@ and @T@ are implicitly hoisted outside of type @S@ into the containing block scope.
1864In the right example in \CFA, the types are not hoisted and accessed using the field-selection operator ``@.@'' for type qualification, as does Java, rather than the \CC type-selection operator ``@::@''.
1865
1866
1867\subsection{Default Parameters}
1868
1869
1870\section{Literals}
1871
1872C already includes limited polymorphism for literals -- @0@ can be either an integer or a pointer literal, depending on context, while the syntactic forms of literals of the various integer and float types are very similar, differing from each other only in suffix.
1873In keeping with the general \CFA approach of adding features while respecting ``the C way'' of doing things, we have extended both C's polymorphic zero and typed literal syntax to interoperate with user-defined types, while maintaining a backwards-compatible semantics.
1874
1875\subsection{0/1}
1876
1877In C, @0@ has the special property that it is the only ``false'' value; by the standard, any value which compares equal to @0@ is false, while any value that compares unequal to @0@ is true.
1878As such, an expression @x@ in any boolean context (such as the condition of an @if@ or @while@ statement, or the arguments to an @&&@, @||@, or ternary operator) can be rewritten as @x != 0@ without changing its semantics.
1879The operator overloading feature of \CFA provides a natural means to implement this truth value comparison for arbitrary types, but the C type system is not precise enough to distinguish an equality comparison with @0@ from an equality comparison with an arbitrary integer or pointer.
1880To provide this precision, \CFA introduces a new type @zero_t@ as type type of literal @0@ (somewhat analagous to @nullptr_t@ and @nullptr@ in \CCeleven); @zero_t@ can only take the value @0@, but has implicit conversions to the integer and pointer types so that standard C code involving @0@ continues to work properly.
1881With this addition, the \CFA compiler rewrites @if (x)@ and similar expressions to @if ((x) != 0)@ or the appropriate analogue, and any type @T@ can be made ``truthy'' by defining an operator overload @int ?!=?(T, zero_t)@.
1882\CC makes types truthy by adding a conversion to @bool@; prior to the addition of explicit cast operators in \CCeleven this approach had the pitfall of making truthy types transitively convertable to any numeric type; our design for \CFA avoids this issue.
1883
1884\CFA also includes a special type for @1@, @one_t@; like @zero_t@, @one_t@ has built-in implicit conversions to the various integral types so that @1@ maintains its expected semantics in legacy code.
1885The addition of @one_t@ allows generic algorithms to handle the unit value uniformly for types where that is meaningful.
1886\TODO{Make this sentence true} In particular, polymorphic functions in the \CFA prelude define @++x@ and @x++@ in terms of @x += 1@, allowing users to idiomatically define all forms of increment for a type @T@ by defining the single function @T& ?+=(T&, one_t)@; analogous overloads for the decrement operators are present as well.
1887
1888\subsection{Units}
1889
1890Alternative call syntax (literal argument before routine name) to convert basic literals into user literals.
1891
1892{\lstset{language=CFA,deletedelim=**[is][]{`}{`},moredelim=**[is][\color{red}]{@}{@}}
1893\begin{cfa}
1894struct Weight { double stones; };
1895
1896void ?{}( Weight & w ) { w.stones = 0; }        $\C{// operations}$
1897void ?{}( Weight & w, double w ) { w.stones = w; }
1898Weight ?+?( Weight l, Weight r ) { return (Weight){ l.stones + r.stones }; }
1899
1900Weight @?`st@( double w ) { return (Weight){ w }; } $\C{// backquote for units}$
1901Weight @?`lb@( double w ) { return (Weight){ w / 14.0 }; }
1902Weight @?`kg@( double w ) { return (Weight) { w * 0.1575}; }
1903
1904int main() {
1905        Weight w, hw = { 14 };                                  $\C{// 14 stone}$
1906        w = 11@`st@ + 1@`lb@;
1907        w = 70.3@`kg@;
1908        w = 155@`lb@;
1909        w = 0x_9b_u@`lb@;                                               $\C{// hexadecimal unsigned weight (155)}$
1910        w = 0_233@`lb@;                                                 $\C{// octal weight (155)}$
1911        w = 5@`st@ + 8@`kg@ + 25@`lb@ + hw;
1912}
1913\end{cfa}
1914}%
1915
1916
1917\section{Libraries}
1918\label{sec:libraries}
1919
1920As stated in Section~\ref{sec:poly-fns}, \CFA inherits a large corpus of library code, where other programming languages must rewrite or provide fragile inter-language communication with C.
1921\CFA has replacement libraries condensing hundreds of existing C names into tens of \CFA overloaded names, all without rewriting the actual computations.
1922In many cases, the interface is an inline wrapper providing overloading during compilation but zero cost at runtime.
1923The following sections give a glimpse of the interface reduction to many C libraries.
1924In many cases, @signed@/@unsigned@ @char@ and @short@ routines are available (but not shown) to ensure expression computations remain in a single type, as conversions can distort results.
1925
1926
1927\subsection{Limits}
1928
1929C library @limits.h@ provides lower and upper bound constants for the basic types.
1930\CFA name overloading is used to condense these typed constants, \eg:
1931\begin{cquote}
1932\lstDeleteShortInline@%
1933\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1934\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{Definition}}       & \multicolumn{1}{c}{\textbf{Usage}}    \\
1935\begin{cfa}
1936const short int `MIN` = -32768;
1937const int `MIN` = -2147483648;
1938const long int `MIN` = -9223372036854775808L;
1939\end{cfa}
1940&
1941\begin{cfa}
1942short int si = `MIN`;
1943int i = `MIN`;
1944long int li = `MIN`;
1945\end{cfa}
1946\end{tabular}
1947\lstMakeShortInline@%
1948\end{cquote}
1949The result is a significant reduction in names to access typed constants, \eg:
1950\begin{cquote}
1951\lstDeleteShortInline@%
1952\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1953\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1954\begin{cfa}
1955MIN
1956MAX
1957M_PI
1958M_E
1959\end{cfa}
1960&
1961\begin{cfa}
1962SCHAR_MIN, CHAR_MIN, SHRT_MIN, INT_MIN, LONG_MIN, LLONG_MIN,
1963SCHAR_MAX, UCHAR_MAX, SHRT_MAX, INT_MAX, LONG_MAX, LLONG_MAX,
1964M_PI, M_PIl, M_CPI, M_CPIl,
1965M_E, M_El, M_CE, M_CEl
1966\end{cfa}
1967\end{tabular}
1968\lstMakeShortInline@%
1969\end{cquote}
1970
1971
1972\subsection{Math}
1973
1974C library @math.h@ provides many mathematical routines.
1975\CFA routine overloading is used to condense these mathematical routines, \eg:
1976\begin{cquote}
1977\lstDeleteShortInline@%
1978\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1979\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{Definition}}       & \multicolumn{1}{c}{\textbf{Usage}}    \\
1980\begin{cfa}
1981float `log`( float x );
1982double `log`( double );
1983double _Complex `log`( double _Complex x );
1984\end{cfa}
1985&
1986\begin{cfa}
1987float f = `log`( 3.5 );
1988double d = `log`( 3.5 );
1989double _Complex dc = `log`( 3.5+0.5I );
1990\end{cfa}
1991\end{tabular}
1992\lstMakeShortInline@%
1993\end{cquote}
1994The result is a significant reduction in names to access math routines, \eg:
1995\begin{cquote}
1996\lstDeleteShortInline@%
1997\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1998\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1999\begin{cfa}
2000log
2001sqrt
2002sin
2003\end{cfa}
2004&
2005\begin{cfa}
2006logf, log, logl, clogf, clog, clogl
2007sqrtf, sqrt, sqrtl, csqrtf, csqrt, csqrtl
2008sinf, sin, sinl, csinf, csin, csinl
2009\end{cfa}
2010\end{tabular}
2011\lstMakeShortInline@%
2012\end{cquote}
2013While \Celeven has type-generic math~\cite[\S~7.25]{C11} in @tgmath.h@ to provide a similar mechanism, these macros are limited, matching a routine name with a single set of floating type(s).
2014For example, it is not possible to overload @atan@ for both one and two arguments;
2015instead the names @atan@ and @atan2@ are required.
2016The key observation is that only a restricted set of type-generic macros are provided for a limited set of routine names, which do not generalize across the type system, as in \CFA.
2017
2018
2019\subsection{Standard}
2020
2021C library @stdlib.h@ provides many general routines.
2022\CFA routine overloading is used to condense these utility routines, \eg:
2023\begin{cquote}
2024\lstDeleteShortInline@%
2025\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
2026\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{Definition}}       & \multicolumn{1}{c}{\textbf{Usage}}    \\
2027\begin{cfa}
2028unsigned int `abs`( int );
2029double `abs`( double );
2030double abs( double _Complex );
2031\end{cfa}
2032&
2033\begin{cfa}
2034unsigned int i = `abs`( -1 );
2035double d = `abs`( -1.5 );
2036double d = `abs`( -1.5+0.5I );
2037\end{cfa}
2038\end{tabular}
2039\lstMakeShortInline@%
2040\end{cquote}
2041The result is a significant reduction in names to access utility routines, \eg:
2042\begin{cquote}
2043\lstDeleteShortInline@%
2044\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
2045\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
2046\begin{cfa}
2047abs
2048strto
2049random
2050\end{cfa}
2051&
2052\begin{cfa}
2053abs, labs, llabs, fabsf, fabs, fabsl, cabsf, cabs, cabsl
2054strtol, strtoul, strtoll, strtoull, strtof, strtod, strtold
2055srand48, mrand48, lrand48, drand48
2056\end{cfa}
2057\end{tabular}
2058\lstMakeShortInline@%
2059\end{cquote}
2060In additon, there are polymorphic routines, like @min@ and @max@, which work on any type with operators @?<?@ or @?>?@.
2061
2062The following shows one example where \CFA \emph{extends} an existing standard C interface to reduce complexity and provide safety.
2063C/\Celeven provide a number of complex and overlapping storage-management operation to support the following capabilities:
2064\begin{description}[itemsep=2pt,parsep=0pt]
2065\item[fill]
2066after allocation the storage is filled with a specified character.
2067\item[resize]
2068an existing allocation is decreased or increased in size.
2069In either case, new storage may or may not be allocated and, if there is a new allocation, as much data from the existing allocation is copied.
2070For an increase in storage size, new storage after the copied data may be filled.
2071\item[alignment]
2072an allocation starts on a specified memory boundary, \eg, an address multiple of 64 or 128 for cache-line purposes.
2073\item[array]
2074the allocation size is scaled to the specified number of array elements.
2075An array may be filled, resized, or aligned.
2076\end{description}
2077Table~\ref{t:StorageManagementOperations} shows the capabilities provided by C/\Celeven allocation-routines and how all the capabilities can be combined into two \CFA routines.
2078
2079\CFA storage-management routines extend the C equivalents by overloading, providing shallow type-safety, and removing the need to specify the base allocation-size.
2080The following example contrasts \CFA and C storage-allocation operation performing the same operations with the same type safety:
2081\begin{cquote}
2082\begin{cfa}[aboveskip=0pt]
2083size_t  dim = 10;                                                       $\C{// array dimension}$
2084char fill = '\xff';                                                     $\C{// initialization fill value}$
2085int * ip;
2086\end{cfa}
2087\lstDeleteShortInline@%
2088\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
2089\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
2090\begin{cfa}
2091ip = alloc();
2092ip = alloc( fill );
2093ip = alloc( dim );
2094ip = alloc( dim, fill );
2095ip = alloc( ip, 2 * dim );
2096ip = alloc( ip, 4 * dim, fill );
2097
2098ip = align_alloc( 16 );
2099ip = align_alloc( 16, fill );
2100ip = align_alloc( 16, dim );
2101ip = align_alloc( 16, dim, fill );
2102\end{cfa}
2103&
2104\begin{cfa}
2105ip = (int *)malloc( sizeof( int ) );
2106ip = (int *)malloc( sizeof( int ) ); memset( ip, fill, sizeof( int ) );
2107ip = (int *)malloc( dim * sizeof( int ) );
2108ip = (int *)malloc( sizeof( int ) ); memset( ip, fill, dim * sizeof( int ) );
2109ip = (int *)realloc( ip, 2 * dim * sizeof( int ) );
2110ip = (int *)realloc( ip, 4 * dim * sizeof( int ) ); memset( ip, fill, 4 * dim * sizeof( int ) );
2111
2112ip = memalign( 16, sizeof( int ) );
2113ip = memalign( 16, sizeof( int ) ); memset( ip, fill, sizeof( int ) );
2114ip = memalign( 16, dim * sizeof( int ) );
2115ip = memalign( 16, dim * sizeof( int ) ); memset( ip, fill, dim * sizeof( int ) );
2116\end{cfa}
2117\end{tabular}
2118\lstMakeShortInline@%
2119\end{cquote}
2120Variadic @new@ (see Section~\ref{sec:variadic-tuples}) cannot support the same overloading because extra parameters are for initialization.
2121Hence, there are @new@ and @anew@ routines for single and array variables, and the fill value is the arguments to the constructor, \eg:
2122\begin{cfa}
2123struct S { int i, j; };
2124void ?{}( S & s, int i, int j ) { s.i = i; s.j = j; }
2125S * s = new( 2, 3 );                                            $\C{// allocate storage and run constructor}$
2126S * as = anew( dim, 2, 3 );                                     $\C{// each array element initialized to 2, 3}$
2127\end{cfa}
2128Note, \CC can only initialization array elements via the default constructor.
2129
2130Finally, the \CFA memory-allocator has \newterm{sticky properties} for dynamic storage: fill and alignment are remembered with an object's storage in the heap.
2131When a @realloc@ is performed, the sticky properties are respected, so that new storage is correctly aligned and initialized with the fill character.
2132
2133\begin{table}
2134\centering
2135\lstDeleteShortInline@%
2136\lstMakeShortInline~%
2137\begin{tabular}{@{}r|r|l|l|l|l@{}}
2138\multicolumn{1}{c}{}&           & \multicolumn{1}{c|}{fill}     & resize        & alignment     & array \\
2139\hline
2140C               & ~malloc~                      & no                    & no            & no            & no    \\
2141                & ~calloc~                      & yes (0 only)  & no            & no            & yes   \\
2142                & ~realloc~                     & no/copy               & yes           & no            & no    \\
2143                & ~memalign~            & no                    & no            & yes           & no    \\
2144                & ~posix_memalign~      & no                    & no            & yes           & no    \\
2145\hline
2146C11             & ~aligned_alloc~       & no                    & no            & yes           & no    \\
2147\hline
2148\CFA    & ~alloc~                       & yes/copy              & no/yes        & no            & yes   \\
2149                & ~align_alloc~         & yes                   & no            & yes           & yes   \\
2150\end{tabular}
2151\lstDeleteShortInline~%
2152\lstMakeShortInline@%
2153\caption{Storage-Management Operations}
2154\label{t:StorageManagementOperations}
2155\end{table}
2156
2157
2158\subsection{I/O}
2159\label{s:IOLibrary}
2160
2161The goal of \CFA I/O is to simplify the common cases, while fully supporting polymorphism and user defined types in a consistent way.
2162The approach combines ideas from \CC and Python.
2163The \CFA header file for the I/O library is @fstream@.
2164
2165The common case is printing out a sequence of variables separated by whitespace.
2166\begin{cquote}
2167\lstDeleteShortInline@%
2168\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
2169\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{\CC}}      \\
2170\begin{cfa}
2171int x = 1, y = 2, z = 3;
2172sout | x `|` y `|` z | endl;
2173\end{cfa}
2174&
2175\begin{cfa}
2176
2177cout << x `<< " "` << y `<< " "` << z << endl;
2178\end{cfa}
2179\\
2180\begin{cfa}[showspaces=true,aboveskip=0pt,belowskip=0pt]
21811` `2` `3
2182\end{cfa}
2183&
2184\begin{cfa}[showspaces=true,aboveskip=0pt,belowskip=0pt]
21851 2 3
2186\end{cfa}
2187\end{tabular}
2188\lstMakeShortInline@%
2189\end{cquote}
2190The \CFA form has half the characters of the \CC form, and is similar to Python I/O with respect to implicit separators.
2191Similar simplification occurs for tuple I/O, which prints all tuple values separated by ``\lstinline[showspaces=true]@, @''.
2192\begin{cfa}
2193[int, [ int, int ] ] t1 = [ 1, [ 2, 3 ] ], t2 = [ 4, [ 5, 6 ] ];
2194sout | t1 | t2 | endl;                                  $\C{// print tuples}$
2195\end{cfa}
2196\begin{cfa}[showspaces=true,aboveskip=0pt]
21971`, `2`, `3 4`, `5`, `6
2198\end{cfa}
2199Finally, \CFA uses the logical-or operator for I/O as it is the lowest-priority overloadable operator, other than assignment.
2200Therefore, fewer output expressions require parenthesis.
2201\begin{cquote}
2202\lstDeleteShortInline@%
2203\begin{tabular}{@{}ll@{}}
2204\textbf{\CFA:}
2205&
2206\begin{cfa}
2207sout | x * 3 | y + 1 | z << 2 | x == y | (x | y) | (x || y) | (x > z ? 1 : 2) | endl;
2208\end{cfa}
2209\\
2210\textbf{\CC:}
2211&
2212\begin{cfa}
2213cout << x * 3 << y + 1 << `(`z << 2`)` << `(`x == y`)` << (x | y) << (x || y) << (x > z ? 1 : 2) << endl;
2214\end{cfa}
2215\\
2216\textbf{output:}
2217&
2218\begin{cfa}[showspaces=true,aboveskip=0pt]
22193 3 12 0 3 1 2
2220\end{cfa}
2221\end{tabular}
2222\lstMakeShortInline@%
2223\end{cquote}
2224There is a weak similarity between the \CFA logical-or operator and the Shell pipe-operator for moving data, where data flows in the correct direction for input but the opposite direction for output.
2225
2226The implicit separator character (space/blank) is a separator not a terminator.
2227The rules for implicitly adding the separator are:
2228\begin{itemize}[itemsep=2pt,parsep=0pt]
2229\item
2230A separator does not appear at the start or end of a line.
2231\item
2232A separator does not appear before or after a character literal or variable.
2233\item
2234A separator does not appear before or after a null (empty) C string, which is a local mechanism to disable insertion of the separator character.
2235\item
2236A separator does not appear before a C string starting with the characters: \lstinline[mathescape=off,basicstyle=\tt]@([{=$@
2237\item
2238A seperator does not appear after a C string ending with the characters: \lstinline[basicstyle=\tt]@,.;!?)]}%@
2239\item
2240{\lstset{language=CFA,deletedelim=**[is][]{`}{`}}
2241A seperator does not appear before or after a C string begining/ending with the quote or whitespace characters: \lstinline[basicstyle=\tt,showspaces=true]@`'": \t\v\f\r\n@
2242}%
2243\item
2244There are routines to set and get the separator string, and manipulators to toggle separation on and off in the middle of output.
2245\end{itemize}
2246
2247
2248\subsection{Multi-precision Integers}
2249\label{s:MultiPrecisionIntegers}
2250
2251\CFA has an interface to the GMP multi-precision signed-integers~\cite{GMP}, similar to the \CC interface provided by GMP.
2252The \CFA interface wraps GMP routines into operator routines to make programming with multi-precision integers identical to using fixed-sized integers.
2253The \CFA type name for multi-precision signed-integers is @Int@ and the header file is @gmp@.
2254The following multi-precision factorial programs contrast using GMP with the \CFA and C interfaces.
2255\begin{cquote}
2256\lstDeleteShortInline@%
2257\begin{tabular}{@{}l@{\hspace{\parindentlnth}}@{\hspace{\parindentlnth}}l@{}}
2258\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{@{\hspace{\parindentlnth}}c}{\textbf{C}}      \\
2259\begin{cfa}
2260#include <gmp>
2261int main( void ) {
2262        sout | "Factorial Numbers" | endl;
2263        Int fact = 1;
2264
2265        sout | 0 | fact | endl;
2266        for ( unsigned int i = 1; i <= 40; i += 1 ) {
2267                fact *= i;
2268                sout | i | fact | endl;
2269        }
2270}
2271\end{cfa}
2272&
2273\begin{cfa}
2274#include <gmp.h>
2275int main( void ) {
2276        `gmp_printf`( "Factorial Numbers\n" );
2277        `mpz_t` fact;
2278        `mpz_init_set_ui`( fact, 1 );
2279        `gmp_printf`( "%d %Zd\n", 0, fact );
2280        for ( unsigned int i = 1; i <= 40; i += 1 ) {
2281                `mpz_mul_ui`( fact, fact, i );
2282                `gmp_printf`( "%d %Zd\n", i, fact );
2283        }
2284}
2285\end{cfa}
2286\end{tabular}
2287\lstMakeShortInline@%
2288\end{cquote}
2289
2290
2291\section{Evaluation}
2292\label{sec:eval}
2293
2294Though \CFA provides significant added functionality over C, these features have a low runtime penalty.
2295In fact, \CFA's features for generic programming can enable faster runtime execution than idiomatic @void *@-based C code.
2296This claim is demonstrated through a set of generic-code-based micro-benchmarks in C, \CFA, and \CC (see stack implementations in Appendix~\ref{sec:BenchmarkStackImplementation}).
2297Since all these languages share a subset essentially comprising standard C, maximal-performance benchmarks would show little runtime variance, other than in length and clarity of source code.
2298A more illustrative benchmark measures the costs of idiomatic usage of each language's features.
2299Figure~\ref{fig:BenchmarkTest} shows the \CFA benchmark tests for a generic stack based on a singly linked-list, a generic pair-data-structure, and a variadic @print@ routine similar to that in Section~\ref{sec:variadic-tuples}.
2300The benchmark test is similar for C and \CC.
2301The experiment uses element types @int@ and @pair(_Bool, char)@, and pushes $N=40M$ elements on a generic stack, copies the stack, clears one of the stacks, finds the maximum value in the other stack, and prints $N/2$ (to reduce graph height) constants.
2302
2303\begin{figure}
2304\begin{lstlisting}[xleftmargin=3\parindentlnth,aboveskip=0pt,belowskip=0pt]
2305int main( int argc, char * argv[] ) {
2306        FILE * out = fopen( "cfa-out.txt", "w" );
2307        int maxi = 0, vali = 42;
2308        stack(int) si, ti;
2309
2310        REPEAT_TIMED( "push_int", N, push( &si, vali ); )
2311        TIMED( "copy_int", ti = si; )
2312        TIMED( "clear_int", clear( &si ); )
2313        REPEAT_TIMED( "pop_int", N,
2314                int xi = pop( &ti ); if ( xi > maxi ) { maxi = xi; } )
2315        REPEAT_TIMED( "print_int", N/2, print( out, vali, ":", vali, "\n" ); )
2316
2317        pair(_Bool, char) maxp = { (_Bool)0, '\0' }, valp = { (_Bool)1, 'a' };
2318        stack(pair(_Bool, char)) sp, tp;
2319
2320        REPEAT_TIMED( "push_pair", N, push( &sp, valp ); )
2321        TIMED( "copy_pair", tp = sp; )
2322        TIMED( "clear_pair", clear( &sp ); )
2323        REPEAT_TIMED( "pop_pair", N,
2324                pair(_Bool, char) xp = pop( &tp ); if ( xp > maxp ) { maxp = xp; } )
2325        REPEAT_TIMED( "print_pair", N/2, print( out, valp, ":", valp, "\n" ); )
2326        fclose(out);
2327}
2328\end{lstlisting}
2329\caption{\protect\CFA Benchmark Test}
2330\label{fig:BenchmarkTest}
2331\end{figure}
2332
2333The structure of each benchmark implemented is: C with @void *@-based polymorphism, \CFA with the presented features, \CC with templates, and \CC using only class inheritance for polymorphism, called \CCV.
2334The \CCV variant illustrates an alternative object-oriented idiom where all objects inherit from a base @object@ class, mimicking a Java-like interface;
2335hence runtime checks are necessary to safely down-cast objects.
2336The most notable difference among the implementations is in memory layout of generic types: \CFA and \CC inline the stack and pair elements into corresponding list and pair nodes, while C and \CCV lack such a capability and instead must store generic objects via pointers to separately-allocated objects.
2337For the print benchmark, idiomatic printing is used: the C and \CFA variants used @stdio.h@, while the \CC and \CCV variants used @iostream@; preliminary tests show this distinction has negligible runtime impact.
2338Note, the C benchmark uses unchecked casts as there is no runtime mechanism to perform such checks, while \CFA and \CC provide type-safety statically.
2339
2340Figure~\ref{fig:eval} and Table~\ref{tab:eval} show the results of running the benchmark in Figure~\ref{fig:BenchmarkTest} and its C, \CC, and \CCV equivalents.
2341The graph plots the median of 5 consecutive runs of each program, with an initial warm-up run omitted.
2342All code is compiled at \texttt{-O2} by GCC or G++ 6.2.0, with all \CC code compiled as \CCfourteen.
2343The benchmarks are run on an Ubuntu 16.04 workstation with 16 GB of RAM and a 6-core AMD FX-6300 CPU with 3.5 GHz maximum clock frequency.
2344
2345\begin{figure}
2346\centering
2347\input{timing}
2348\caption{Benchmark Timing Results (smaller is better)}
2349\label{fig:eval}
2350\end{figure}
2351
2352\begin{table}
2353\centering
2354\caption{Properties of benchmark code}
2355\label{tab:eval}
2356\newcommand{\CT}[1]{\multicolumn{1}{c}{#1}}
2357\begin{tabular}{rrrrr}
2358                                                                        & \CT{C}        & \CT{\CFA}     & \CT{\CC}      & \CT{\CCV}             \\ \hline
2359maximum memory usage (MB)                       & 10001         & 2502          & 2503          & 11253                 \\
2360source code size (lines)                        & 247           & 222           & 165           & 339                   \\
2361redundant type annotations (lines)      & 39            & 2                     & 2                     & 15                    \\
2362binary size (KB)                                        & 14            & 229           & 18            & 38                    \\
2363\end{tabular}
2364\end{table}
2365
2366The C and \CCV variants are generally the slowest with the largest memory footprint, because of their less-efficient memory layout and the pointer-indirection necessary to implement generic types;
2367this inefficiency is exacerbated by the second level of generic types in the pair-based benchmarks.
2368By contrast, the \CFA and \CC variants run in roughly equivalent time for both the integer and pair of @_Bool@ and @char@ because the storage layout is equivalent, with the inlined libraries (\ie no separate compilation) and greater maturity of the \CC compiler contributing to its lead.
2369\CCV is slower than C largely due to the cost of runtime type-checking of down-casts (implemented with @dynamic_cast@);
2370There are two outliers in the graph for \CFA: all prints and pop of @pair@.
2371Both of these cases result from the complexity of the C-generated polymorphic code, so that the GCC compiler is unable to optimize some dead code and condense nested calls.
2372A compiler designed for \CFA could easily perform these optimizations.
2373Finally, the binary size for \CFA is larger because of static linking with the \CFA libraries.
2374
2375\CFA is also competitive in terms of source code size, measured as a proxy for programmer effort. The line counts in Table~\ref{tab:eval} include implementations of @pair@ and @stack@ types for all four languages for purposes of direct comparison, though it should be noted that \CFA and \CC have pre-written data structures in their standard libraries that programmers would generally use instead. Use of these standard library types has minimal impact on the performance benchmarks, but shrinks the \CFA and \CC benchmarks to 73 and 54 lines, respectively.
2376On the other hand, C does not have a generic collections-library in its standard distribution, resulting in frequent reimplementation of such collection types by C programmers.
2377\CCV does not use the \CC standard template library by construction, and in fact includes the definition of @object@ and wrapper classes for @bool@, @char@, @int@, and @const char *@ in its line count, which inflates this count somewhat, as an actual object-oriented language would include these in the standard library;
2378with their omission, the \CCV line count is similar to C.
2379We justify the given line count by noting that many object-oriented languages do not allow implementing new interfaces on library types without subclassing or wrapper types, which may be similarly verbose.
2380
2381Raw line-count, however, is a fairly rough measure of code complexity;
2382another important factor is how much type information the programmer must manually specify, especially where that information is not checked by the compiler.
2383Such unchecked type information produces a heavier documentation burden and increased potential for runtime bugs, and is much less common in \CFA than C, with its manually specified function pointers arguments and format codes, or \CCV, with its extensive use of un-type-checked downcasts (\eg @object@ to @integer@ when popping a stack, or @object@ to @printable@ when printing the elements of a @pair@).
2384To quantify this, the ``redundant type annotations'' line in Table~\ref{tab:eval} counts the number of lines on which the type of a known variable is re-specified, either as a format specifier, explicit downcast, type-specific function, or by name in a @sizeof@, struct literal, or @new@ expression.
2385The \CC benchmark uses two redundant type annotations to create a new stack nodes, while the C and \CCV benchmarks have several such annotations spread throughout their code.
2386The two instances in which the \CFA benchmark still uses redundant type specifiers are to cast the result of a polymorphic @malloc@ call (the @sizeof@ argument is inferred by the compiler).
2387These uses are similar to the @new@ expressions in \CC, though the \CFA compiler's type resolver should shortly render even these type casts superfluous.
2388
2389
2390\section{Related Work}
2391
2392
2393\subsection{Polymorphism}
2394
2395\CC is the most similar language to \CFA;
2396both are extensions to C with source and runtime backwards compatibility.
2397The fundamental difference is in their engineering approach to C compatibility and programmer expectation.
2398While \CC provides good backwards compatibility with C, it has a steep learning curve for many of its extensions.
2399For example, polymorphism is provided via three disjoint mechanisms: overloading, inheritance, and templates.
2400The overloading is restricted because resolution does not use the return type, inheritance requires learning object-oriented programming and coping with a restricted nominal-inheritance hierarchy, templates cannot be separately compiled resulting in compilation/code bloat and poor error messages, and determining how these mechanisms interact and which to use is confusing.
2401In contrast, \CFA has a single facility for polymorphic code supporting type-safe separate-compilation of polymorphic functions and generic (opaque) types, which uniformly leverage the C procedural paradigm.
2402The key mechanism to support separate compilation is \CFA's \emph{explicit} use of assumed properties for a type.
2403Until \CC concepts~\cite{C++Concepts} are standardized (anticipated for \CCtwenty), \CC provides no way to specify the requirements of a generic function in code beyond compilation errors during template expansion;
2404furthermore, \CC concepts are restricted to template polymorphism.
2405
2406Cyclone~\cite{Grossman06} also provides capabilities for polymorphic functions and existential types, similar to \CFA's @forall@ functions and generic types.
2407Cyclone existential types can include function pointers in a construct similar to a virtual function-table, but these pointers must be explicitly initialized at some point in the code, a tedious and potentially error-prone process.
2408Furthermore, Cyclone's polymorphic functions and types are restricted to abstraction over types with the same layout and calling convention as @void *@, \ie only pointer types and @int@.
2409In \CFA terms, all Cyclone polymorphism must be dtype-static.
2410While the Cyclone design provides the efficiency benefits discussed in Section~\ref{sec:generic-apps} for dtype-static polymorphism, it is more restrictive than \CFA's general model.
2411Smith and Volpano~\cite{Smith98} present Polymorphic C, an ML dialect with polymorphic functions, C-like syntax, and pointer types; it lacks many of C's features, however, most notably structure types, and so is not a practical C replacement.
2412
2413Objective-C~\cite{obj-c-book} is an industrially successful extension to C.
2414However, Objective-C is a radical departure from C, using an object-oriented model with message-passing.
2415Objective-C did not support type-checked generics until recently \cite{xcode7}, historically using less-efficient runtime checking of object types.
2416The GObject~\cite{GObject} framework also adds object-oriented programming with runtime type-checking and reference-counting garbage-collection to C;
2417these features are more intrusive additions than those provided by \CFA, in addition to the runtime overhead of reference-counting.
2418Vala~\cite{Vala} compiles to GObject-based C, adding the burden of learning a separate language syntax to the aforementioned demerits of GObject as a modernization path for existing C code-bases.
2419Java~\cite{Java8} included generic types in Java~5, which are type-checked at compilation and type-erased at runtime, similar to \CFA's.
2420However, in Java, each object carries its own table of method pointers, while \CFA passes the method pointers separately to maintain a C-compatible layout.
2421Java is also a garbage-collected, object-oriented language, with the associated resource usage and C-interoperability burdens.
2422
2423D~\cite{D}, Go, and Rust~\cite{Rust} are modern, compiled languages with abstraction features similar to \CFA traits, \emph{interfaces} in D and Go and \emph{traits} in Rust.
2424However, each language represents a significant departure from C in terms of language model, and none has the same level of compatibility with C as \CFA.
2425D and Go are garbage-collected languages, imposing the associated runtime overhead.
2426The necessity of accounting for data transfer between managed runtimes and the unmanaged C runtime complicates foreign-function interfaces to C.
2427Furthermore, while generic types and functions are available in Go, they are limited to a small fixed set provided by the compiler, with no language facility to define more.
2428D restricts garbage collection to its own heap by default, while Rust is not garbage-collected, and thus has a lighter-weight runtime more interoperable with C.
2429Rust also possesses much more powerful abstraction capabilities for writing generic code than Go.
2430On the other hand, Rust's borrow-checker provides strong safety guarantees but is complex and difficult to learn and imposes a distinctly idiomatic programming style.
2431\CFA, with its more modest safety features, allows direct ports of C code while maintaining the idiomatic style of the original source.
2432
2433
2434\subsection{Tuples/Variadics}
2435
2436Many programming languages have some form of tuple construct and/or variadic functions, \eg SETL, C, KW-C, \CC, D, Go, Java, ML, and Scala.
2437SETL~\cite{SETL} is a high-level mathematical programming language, with tuples being one of the primary data types.
2438Tuples in SETL allow subscripting, dynamic expansion, and multiple assignment.
2439C provides variadic functions through @va_list@ objects, but the programmer is responsible for managing the number of arguments and their types, so the mechanism is type unsafe.
2440KW-C~\cite{Buhr94a}, a predecessor of \CFA, introduced tuples to C as an extension of the C syntax, taking much of its inspiration from SETL.
2441The main contributions of that work were adding MRVF, tuple mass and multiple assignment, and record-field access.
2442\CCeleven introduced @std::tuple@ as a library variadic template structure.
2443Tuples are a generalization of @std::pair@, in that they allow for arbitrary length, fixed-size aggregation of heterogeneous values.
2444Operations include @std::get<N>@ to extract values, @std::tie@ to create a tuple of references used for assignment, and lexicographic comparisons.
2445\CCseventeen proposes \emph{structured bindings}~\cite{Sutter15} to eliminate pre-declaring variables and use of @std::tie@ for binding the results.
2446This extension requires the use of @auto@ to infer the types of the new variables, so complicated expressions with a non-obvious type must be documented with some other mechanism.
2447Furthermore, structured bindings are not a full replacement for @std::tie@, as it always declares new variables.
2448Like \CC, D provides tuples through a library variadic-template structure.
2449Go does not have tuples but supports MRVF.
2450Java's variadic functions appear similar to C's but are type-safe using homogeneous arrays, which are less useful than \CFA's heterogeneously-typed variadic functions.
2451Tuples are a fundamental abstraction in most functional programming languages, such as Standard ML~\cite{sml} and~\cite{Scala}, which decompose tuples using pattern matching.
2452
2453
2454\section{Conclusion and Future Work}
2455
2456The goal of \CFA is to provide an evolutionary pathway for large C development-environments to be more productive and safer, while respecting the talent and skill of C programmers.
2457While other programming languages purport to be a better C, they are in fact new and interesting languages in their own right, but not C extensions.
2458The purpose of this paper is to introduce \CFA, and showcase language features that illustrate the \CFA type-system and approaches taken to achieve the goal of evolutionary C extension.
2459The contributions are a powerful type-system using parametric polymorphism and overloading, generic types, and tuples, which all have complex interactions.
2460The work is a challenging design, engineering, and implementation exercise.
2461On the surface, the project may appear as a rehash of similar mechanisms in \CC.
2462However, every \CFA feature is different than its \CC counterpart, often with extended functionality, better integration with C and its programmers, and always supporting separate compilation.
2463All of these new features are being used by the \CFA development-team to build the \CFA runtime-system.
2464Finally, we demonstrate that \CFA performance for some idiomatic cases is better than C and close to \CC, showing the design is practically applicable.
2465
2466There is ongoing work on a wide range of \CFA feature extensions, including arrays with size, exceptions, concurrent primitives, modules, and user-defined conversions.
2467(While all examples in the paper compile and run, a public beta-release of \CFA will take another 8--12 months to finalize these additional extensions.)
2468In addition, there are interesting future directions for the polymorphism design.
2469Notably, \CC template functions trade compile time and code bloat for optimal runtime of individual instantiations of polymorphic functions.
2470\CFA polymorphic functions use dynamic virtual-dispatch;
2471the runtime overhead of this approach is low, but not as low as inlining, and it may be beneficial to provide a mechanism for performance-sensitive code.
2472Two promising approaches are an @inline@ annotation at polymorphic function call sites to create a template-specialization of the function (provided the code is visible) or placing an @inline@ annotation on polymorphic function-definitions to instantiate a specialized version for some set of types (\CC template specialization).
2473These approaches are not mutually exclusive and allow performance optimizations to be applied only when necessary, without suffering global code-bloat.
2474In general, we believe separate compilation, producing smaller code, works well with loaded hardware-caches, which may offset the benefit of larger inlined-code.
2475
2476
2477\section{Acknowledgments}
2478
2479The authors would like to recognize the design assistance of Glen Ditchfield, Richard Bilson, and Thierry Delisle on the features described in this paper, and thank Magnus Madsen and the three anonymous reviewers for valuable feedback.
2480%This work is supported in part by a corporate partnership with \grantsponsor{Huawei}{Huawei Ltd.}{http://www.huawei.com}, and Aaron Moss and Peter Buhr are funded by the \grantsponsor{Natural Sciences and Engineering Research Council} of Canada.
2481% the first author's \grantsponsor{NSERC-PGS}{NSERC PGS D}{http://www.nserc-crsng.gc.ca/Students-Etudiants/PG-CS/BellandPostgrad-BelletSuperieures_eng.asp} scholarship.
2482
2483
2484\bibliographystyle{plain}
2485\bibliography{pl}
2486
2487
2488\appendix
2489
2490\section{Benchmark Stack Implementation}
2491\label{sec:BenchmarkStackImplementation}
2492
2493\lstset{basicstyle=\linespread{0.9}\sf\small}
2494
2495Throughout, @/***/@ designates a counted redundant type annotation.
2496
2497\smallskip\noindent
2498\CFA
2499\begin{lstlisting}[xleftmargin=2\parindentlnth,aboveskip=0pt,belowskip=0pt]
2500forall(otype T) struct stack_node {
2501        T value;
2502        stack_node(T) * next;
2503};
2504forall(otype T) void ?{}(stack(T) * s) { (&s->head){ 0 }; }
2505forall(otype T) void ?{}(stack(T) * s, stack(T) t) {
2506        stack_node(T) ** crnt = &s->head;
2507        for ( stack_node(T) * next = t.head; next; next = next->next ) {
2508                *crnt = ((stack_node(T) *)malloc()){ next->value }; /***/
2509                stack_node(T) * acrnt = *crnt;
2510                crnt = &acrnt->next;
2511        }
2512        *crnt = 0;
2513}
2514forall(otype T) stack(T) ?=?(stack(T) * s, stack(T) t) {
2515        if ( s->head == t.head ) return *s;
2516        clear(s);
2517        s{ t };
2518        return *s;
2519}
2520forall(otype T) void ^?{}(stack(T) * s) { clear(s); }
2521forall(otype T) _Bool empty(const stack(T) * s) { return s->head == 0; }
2522forall(otype T) void push(stack(T) * s, T value) {
2523        s->head = ((stack_node(T) *)malloc()){ value, s->head }; /***/
2524}
2525forall(otype T) T pop(stack(T) * s) {
2526        stack_node(T) * n = s->head;
2527        s->head = n->next;
2528        T x = n->value;
2529        ^n{};
2530        free(n);
2531        return x;
2532}
2533forall(otype T) void clear(stack(T) * s) {
2534        for ( stack_node(T) * next = s->head; next; ) {
2535                stack_node(T) * crnt = next;
2536                next = crnt->next;
2537                delete(crnt);
2538        }
2539        s->head = 0;
2540}
2541\end{lstlisting}
2542
2543\medskip\noindent
2544\CC
2545\begin{lstlisting}[xleftmargin=2\parindentlnth,aboveskip=0pt,belowskip=0pt]
2546template<typename T> class stack {
2547        struct node {
2548                T value;
2549                node * next;
2550                node( const T & v, node * n = nullptr ) : value(v), next(n) {}
2551        };
2552        node * head;
2553        void copy(const stack<T>& o) {
2554                node ** crnt = &head;
2555                for ( node * next = o.head;; next; next = next->next ) {
2556                        *crnt = new node{ next->value }; /***/
2557                        crnt = &(*crnt)->next;
2558                }
2559                *crnt = nullptr;
2560        }
2561  public:
2562        stack() : head(nullptr) {}
2563        stack(const stack<T>& o) { copy(o); }
2564        stack(stack<T> && o) : head(o.head) { o.head = nullptr; }
2565        ~stack() { clear(); }
2566        stack & operator= (const stack<T>& o) {
2567                if ( this == &o ) return *this;
2568                clear();
2569                copy(o);
2570                return *this;
2571        }
2572        stack & operator= (stack<T> && o) {
2573                if ( this == &o ) return *this;
2574                head = o.head;
2575                o.head = nullptr;
2576                return *this;
2577        }
2578        bool empty() const { return head == nullptr; }
2579        void push(const T & value) { head = new node{ value, head };  /***/ }
2580        T pop() {
2581                node * n = head;
2582                head = n->next;
2583                T x = std::move(n->value);
2584                delete n;
2585                return x;
2586        }
2587        void clear() {
2588                for ( node * next = head; next; ) {
2589                        node * crnt = next;
2590                        next = crnt->next;
2591                        delete crnt;
2592                }
2593                head = nullptr;
2594        }
2595};
2596\end{lstlisting}
2597
2598\medskip\noindent
2599C
2600\begin{lstlisting}[xleftmargin=2\parindentlnth,aboveskip=0pt,belowskip=0pt]
2601struct stack_node {
2602        void * value;
2603        struct stack_node * next;
2604};
2605struct stack new_stack() { return (struct stack){ NULL }; /***/ }
2606void copy_stack(struct stack * s, const struct stack * t, void * (*copy)(const void *)) {
2607        struct stack_node ** crnt = &s->head;
2608        for ( struct stack_node * next = t->head; next; next = next->next ) {
2609                *crnt = malloc(sizeof(struct stack_node)); /***/
2610                **crnt = (struct stack_node){ copy(next->value) }; /***/
2611                crnt = &(*crnt)->next;
2612        }
2613        *crnt = 0;
2614}
2615_Bool stack_empty(const struct stack * s) { return s->head == NULL; }
2616void push_stack(struct stack * s, void * value) {
2617        struct stack_node * n = malloc(sizeof(struct stack_node)); /***/
2618        *n = (struct stack_node){ value, s->head }; /***/
2619        s->head = n;
2620}
2621void * pop_stack(struct stack * s) {
2622        struct stack_node * n = s->head;
2623        s->head = n->next;
2624        void * x = n->value;
2625        free(n);
2626        return x;
2627}
2628void clear_stack(struct stack * s, void (*free_el)(void *)) {
2629        for ( struct stack_node * next = s->head; next; ) {
2630                struct stack_node * crnt = next;
2631                next = crnt->next;
2632                free_el(crnt->value);
2633                free(crnt);
2634        }
2635        s->head = NULL;
2636}
2637\end{lstlisting}
2638
2639\medskip\noindent
2640\CCV
2641\begin{lstlisting}[xleftmargin=2\parindentlnth,aboveskip=0pt,belowskip=0pt]
2642stack::node::node( const object & v, node * n ) : value( v.new_copy() ), next( n ) {}
2643void stack::copy(const stack & o) {
2644        node ** crnt = &head;
2645        for ( node * next = o.head; next; next = next->next ) {
2646                *crnt = new node{ *next->value };
2647                crnt = &(*crnt)->next;
2648        }
2649        *crnt = nullptr;
2650}
2651stack::stack() : head(nullptr) {}
2652stack::stack(const stack & o) { copy(o); }
2653stack::stack(stack && o) : head(o.head) { o.head = nullptr; }
2654stack::~stack() { clear(); }
2655stack & stack::operator= (const stack & o) {
2656        if ( this == &o ) return *this;
2657        clear();
2658        copy(o);
2659        return *this;
2660}
2661stack & stack::operator= (stack && o) {
2662        if ( this == &o ) return *this;
2663        head = o.head;
2664        o.head = nullptr;
2665        return *this;
2666}
2667bool stack::empty() const { return head == nullptr; }
2668void stack::push(const object & value) { head = new node{ value, head }; /***/ }
2669ptr<object> stack::pop() {
2670        node * n = head;
2671        head = n->next;
2672        ptr<object> x = std::move(n->value);
2673        delete n;
2674        return x;
2675}
2676void stack::clear() {
2677        for ( node * next = head; next; ) {
2678                node * crnt = next;
2679                next = crnt->next;
2680                delete crnt;
2681        }
2682        head = nullptr;
2683}
2684\end{lstlisting}
2685
2686
2687\begin{comment}
2688
2689\subsubsection{bench.h}
2690(\texttt{bench.hpp} is similar.)
2691
2692\lstinputlisting{evaluation/bench.h}
2693
2694\subsection{C}
2695
2696\subsubsection{c-stack.h} ~
2697
2698\lstinputlisting{evaluation/c-stack.h}
2699
2700\subsubsection{c-stack.c} ~
2701
2702\lstinputlisting{evaluation/c-stack.c}
2703
2704\subsubsection{c-pair.h} ~
2705
2706\lstinputlisting{evaluation/c-pair.h}
2707
2708\subsubsection{c-pair.c} ~
2709
2710\lstinputlisting{evaluation/c-pair.c}
2711
2712\subsubsection{c-print.h} ~
2713
2714\lstinputlisting{evaluation/c-print.h}
2715
2716\subsubsection{c-print.c} ~
2717
2718\lstinputlisting{evaluation/c-print.c}
2719
2720\subsubsection{c-bench.c} ~
2721
2722\lstinputlisting{evaluation/c-bench.c}
2723
2724\subsection{\CFA}
2725
2726\subsubsection{cfa-stack.h} ~
2727
2728\lstinputlisting{evaluation/cfa-stack.h}
2729
2730\subsubsection{cfa-stack.c} ~
2731
2732\lstinputlisting{evaluation/cfa-stack.c}
2733
2734\subsubsection{cfa-print.h} ~
2735
2736\lstinputlisting{evaluation/cfa-print.h}
2737
2738\subsubsection{cfa-print.c} ~
2739
2740\lstinputlisting{evaluation/cfa-print.c}
2741
2742\subsubsection{cfa-bench.c} ~
2743
2744\lstinputlisting{evaluation/cfa-bench.c}
2745
2746\subsection{\CC}
2747
2748\subsubsection{cpp-stack.hpp} ~
2749
2750\lstinputlisting[language=c++]{evaluation/cpp-stack.hpp}
2751
2752\subsubsection{cpp-print.hpp} ~
2753
2754\lstinputlisting[language=c++]{evaluation/cpp-print.hpp}
2755
2756\subsubsection{cpp-bench.cpp} ~
2757
2758\lstinputlisting[language=c++]{evaluation/cpp-bench.cpp}
2759
2760\subsection{\CCV}
2761
2762\subsubsection{object.hpp} ~
2763
2764\lstinputlisting[language=c++]{evaluation/object.hpp}
2765
2766\subsubsection{cpp-vstack.hpp} ~
2767
2768\lstinputlisting[language=c++]{evaluation/cpp-vstack.hpp}
2769
2770\subsubsection{cpp-vstack.cpp} ~
2771
2772\lstinputlisting[language=c++]{evaluation/cpp-vstack.cpp}
2773
2774\subsubsection{cpp-vprint.hpp} ~
2775
2776\lstinputlisting[language=c++]{evaluation/cpp-vprint.hpp}
2777
2778\subsubsection{cpp-vbench.cpp} ~
2779
2780\lstinputlisting[language=c++]{evaluation/cpp-vbench.cpp}
2781\end{comment}
2782
2783\end{document}
2784
2785% Local Variables: %
2786% tab-width: 4 %
2787% compile-command: "make" %
2788% End: %
Note: See TracBrowser for help on using the repository browser.