source: doc/papers/general/Paper.tex @ 14cbfad

aaron-thesisarm-ehcleanup-dtorsdeferred_resndemanglerjacob/cs343-translationjenkins-sandboxnew-astnew-ast-unique-exprnew-envno_listpersistent-indexerresolv-newwith_gc
Last change on this file since 14cbfad was 14cbfad, checked in by Peter A. Buhr <pabuhr@…>, 4 years ago

complete draft of "with" statement

  • 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;
236Section~\ref{sec:libraries} includes a number of examples of how this overloading simplifies \CFA programming relative to C.
237Code 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.
238As an example:
239
240\begin{cfa}
241int max(int a, int b) { return a < b ? b : a; }  // (1)
242double max(double a, double b) { return a < b ? b : a; }  // (2)
243
244int max = INT_MAX;     // (3)
245double max = DBL_MAX;  // (4)
246
247max(7, -max);   $\C{// uses (1) and (3), by matching int from constant 7}$
248max(max, 3.14); $\C{// uses (2) and (4), by matching double from constant 3.14}$
249
250//max(max, -max);  $\C{// ERROR: ambiguous}$
251int m = max(max, -max); $\C{// uses (1) once and (3) twice, by matching return type}$
252\end{cfa}
253
254\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.
255The 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@.
256Ergonomic 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.
257
258% http://fanf.livejournal.com/144696.html
259% http://www.robertgamble.net/2012/01/c11-generic-selections.html
260% https://abissell.com/2014/01/16/c11s-_generic-keyword-macro-applications-and-performance-impacts/
261
262
263\subsection{\texorpdfstring{\LstKeywordStyle{forall} Functions}{forall Functions}}
264\label{sec:poly-fns}
265
266The 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):
267\begin{lstlisting}
268`forall( otype T )` T identity( T val ) { return val; }
269int forty_two = identity( 42 );                         $\C{// T is bound to int, forty\_two == 42}$
270\end{lstlisting}
271The @identity@ function above can be applied to any complete \newterm{object type} (or @otype@).
272The 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.
273The \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.
274If this extra information is not needed, \eg for a pointer, the type parameter can be declared as a \newterm{data type} (or @dtype@).
275
276In \CFA, the polymorphism runtime-cost is spread over each polymorphic call, due to passing more arguments to polymorphic functions;
277the experiments in Section~\ref{sec:eval} show this overhead is similar to \CC virtual-function calls.
278A design advantage is that, unlike \CC template-functions, \CFA polymorphic-functions are compatible with C \emph{separate compilation}, preventing compilation and code bloat.
279
280Since 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.
281For example, the function @twice@ can be defined using the \CFA syntax for operator overloading:
282\begin{lstlisting}
283forall( otype T `| { T ?+?(T, T); }` ) T twice( T x ) { return x + x; } $\C{// ? denotes operands}$
284int val = twice( twice( 3.7 ) );
285\end{lstlisting}
286which works for any type @T@ with a matching addition operator.
287The 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@.
288There 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.
289The first approach has a late conversion from @double@ to @int@ on the final assignment, while the second has an eager conversion to @int@.
290\CFA minimizes the number of conversions and their potential to lose information, so it selects the first approach, which corresponds with C-programmer intuition.
291
292Crucial to the design of a new programming language are the libraries to access thousands of external software features.
293Like \CC, \CFA inherits a massive compatible library-base, where other programming languages must rewrite or provide fragile inter-language communication with C.
294A simple example is leveraging the existing type-unsafe (@void *@) C @bsearch@ to binary search a sorted float array:
295\begin{lstlisting}
296void * bsearch( const void * key, const void * base, size_t nmemb, size_t size,
297                                int (* compar)( const void *, const void * ));
298
299int comp( const void * t1, const void * t2 ) { return *(double *)t1 < *(double *)t2 ? -1 :
300                                *(double *)t2 < *(double *)t1 ? 1 : 0; }
301
302double key = 5.0, vals[10] = { /* 10 sorted float values */ };
303double * val = (double *)bsearch( &key, vals, 10, sizeof(vals[0]), comp );      $\C{// search sorted array}$
304\end{lstlisting}
305which can be augmented simply with a generalized, type-safe, \CFA-overloaded wrappers:
306\begin{lstlisting}
307forall( otype T | { int ?<?( T, T ); } ) T * bsearch( T key, const T * arr, size_t size ) {
308        int comp( const void * t1, const void * t2 ) { /* as above with double changed to T */ }
309        return (T *)bsearch( &key, arr, size, sizeof(T), comp ); }
310
311forall( otype T | { int ?<?( T, T ); } ) unsigned int bsearch( T key, const T * arr, size_t size ) {
312        T * result = bsearch( key, arr, size ); $\C{// call first version}$
313        return result ? result - arr : size; }  $\C{// pointer subtraction includes sizeof(T)}$
314
315double * val = bsearch( 5.0, vals, 10 );        $\C{// selection based on return type}$
316int posn = bsearch( 5.0, vals, 10 );
317\end{lstlisting}
318The nested function @comp@ provides the hidden interface from typed \CFA to untyped (@void *@) C, plus the cast of the result.
319Providing a hidden @comp@ function in \CC is awkward as lambdas do not use C calling-conventions and template declarations cannot appear at block scope.
320As well, an alternate kind of return is made available: position versus pointer to found element.
321\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@.
322
323\CFA has replacement libraries condensing hundreds of existing C functions into tens of \CFA overloaded functions, all without rewriting the actual computations (see Section~\ref{sec:libraries}).
324For example, it is possible to write a type-safe \CFA wrapper @malloc@ based on the C @malloc@:
325\begin{lstlisting}
326forall( dtype T | sized(T) ) T * malloc( void ) { return (T *)malloc( sizeof(T) ); }
327int * ip = malloc();                                            $\C{// select type and size from left-hand side}$
328double * dp = malloc();
329struct S {...} * sp = malloc();
330\end{lstlisting}
331where the return type supplies the type/size of the allocation, which is impossible in most type systems.
332
333Call-site inferencing and nested functions provide a localized form of inheritance.
334For example, the \CFA @qsort@ only sorts in ascending order using @<@.
335However, it is trivial to locally change this behaviour:
336\begin{lstlisting}
337forall( otype T | { int ?<?( T, T ); } ) void qsort( const T * arr, size_t size ) { /* use C qsort */ }
338{       int ?<?( double x, double y ) { return x `>` y; }       $\C{// locally override behaviour}$
339        qsort( vals, size );                                    $\C{// descending sort}$
340}
341\end{lstlisting}
342Within the block, the nested version of @?<?@ performs @?>?@ and this local version overrides the built-in @?<?@ so it is passed to @qsort@.
343Hence, programmers can easily form local environments, adding and modifying appropriate functions, to maximize reuse of other existing functions and types.
344
345%% Redundant with Section~\ref{sec:libraries} %%
346
347% Finally, \CFA allows variable overloading:
348% \begin{lstlisting}
349% short int MAX = ...;   int MAX = ...;  double MAX = ...;
350% short int s = MAX;    int i = MAX;    double d = MAX;   $\C{// select correct MAX}$
351% \end{lstlisting}
352% Here, the single name @MAX@ replaces all the C type-specific names: @SHRT_MAX@, @INT_MAX@, @DBL_MAX@.
353
354\subsection{Traits}
355
356\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:
357\begin{lstlisting}
358trait `summable`( otype T ) {
359        void ?{}( T *, zero_t );                                $\C{// constructor from 0 literal}$
360        T ?+?( T, T );                                                  $\C{// assortment of additions}$
361        T ?+=?( T *, T );
362        T ++?( T * );
363        T ?++( T * ); };
364
365forall( otype T `| summable( T )` ) T sum( T a[$\,$], size_t size ) {  // use trait
366        `T` total = { `0` };                                    $\C{// instantiate T from 0 by calling its constructor}$
367        for ( unsigned int i = 0; i < size; i += 1 ) total `+=` a[i]; $\C{// select appropriate +}$
368        return total; }
369\end{lstlisting}
370
371In 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:
372\begin{lstlisting}
373trait otype( dtype T | sized(T) ) {  // sized is a pseudo-trait for types with known size and alignment
374        void ?{}( T * );                                                $\C{// default constructor}$
375        void ?{}( T *, T );                                             $\C{// copy constructor}$
376        void ?=?( T *, T );                                             $\C{// assignment operator}$
377        void ^?{}( T * ); };                                    $\C{// destructor}$
378\end{lstlisting}
379Given 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.
380
381In 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.
382Hence, trait names play no part in type equivalence;
383the names are simply macros for a list of polymorphic assertions, which are expanded at usage sites.
384Nevertheless, trait names form a logical subtype-hierarchy with @dtype@ at the top, where traits often contain overlapping assertions, \eg operator @+@.
385Traits are used like interfaces in Java or abstract base-classes in \CC, but without the nominal inheritance-relationships.
386Instead, 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.
387Hence, 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.
388(Nominal inheritance can be approximated with traits using marker variables or functions, as is done in Go.)
389
390% Nominal inheritance can be simulated with traits using marker variables or functions:
391% \begin{lstlisting}
392% trait nominal(otype T) {
393%     T is_nominal;
394% };
395% int is_nominal;                                                               $\C{// int now satisfies the nominal trait}$
396% \end{lstlisting}
397%
398% 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:
399% \begin{lstlisting}
400% trait pointer_like(otype Ptr, otype El) {
401%     lvalue El *?(Ptr);                                                $\C{// Ptr can be dereferenced into a modifiable value of type El}$
402% }
403% struct list {
404%     int value;
405%     list * next;                                                              $\C{// may omit "struct" on type names as in \CC}$
406% };
407% typedef list * list_iterator;
408%
409% lvalue int *?( list_iterator it ) { return it->value; }
410% \end{lstlisting}
411% 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@).
412% 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.
413
414
415\section{Generic Types}
416
417One of the known shortcomings of standard C is that it does not provide reusable type-safe abstractions for generic data structures and algorithms.
418Broadly speaking, there are three approaches to implement abstract data-structures in C.
419One approach is to write bespoke data-structures for each context in which they are needed.
420While 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.
421A 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.
422However, 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.
423A 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.
424Furthermore, writing and using preprocessor macros can be unnatural and inflexible.
425
426\CC, Java, and other languages use \newterm{generic types} to produce type-safe abstract data-types.
427\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.
428However, for known concrete parameters, the generic-type definition can be inlined, like \CC templates.
429
430A 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:
431\begin{lstlisting}
432forall( otype R, otype S ) struct pair {
433        R first;
434        S second;
435};
436forall( otype T ) T value( pair( const char *, T ) p ) { return p.second; }
437forall( dtype F, otype T ) T value_p( pair( F *, T * ) p ) { return * p.second; }
438
439pair( const char *, int ) p = { "magic", 42 };
440int magic = value( p );
441pair( void *, int * ) q = { 0, &p.second };
442magic = value_p( q );
443double d = 1.0;
444pair( double *, double * ) r = { &d, &d };
445d = value_p( r );
446\end{lstlisting}
447
448\CFA classifies generic types as either \newterm{concrete} or \newterm{dynamic}.
449Concrete types have a fixed memory layout regardless of type parameters, while dynamic types vary in memory layout depending on their type parameters.
450A type may have polymorphic parameters but still be concrete, called \newterm{dtype-static}.
451Polymorphic 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.
452
453\CFA generic types also allow checked argument-constraints.
454For example, the following declaration of a sorted set-type ensures the set key supports equality and relational comparison:
455\begin{lstlisting}
456forall( otype Key | { _Bool ?==?(Key, Key); _Bool ?<?(Key, Key); } ) struct sorted_set;
457\end{lstlisting}
458
459
460\subsection{Concrete Generic-Types}
461
462The \CFA translator template-expands concrete generic-types into new structure types, affording maximal inlining.
463To 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.
464A 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.
465For example, the concrete instantiation for @pair( const char *, int )@ is:
466\begin{lstlisting}
467struct _pair_conc1 {
468        const char * first;
469        int second;
470};
471\end{lstlisting}
472
473A concrete generic-type with dtype-static parameters is also expanded to a structure type, but this type is used for all matching instantiations.
474In 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:
475\begin{lstlisting}
476struct _pair_conc0 {
477        void * first;
478        void * second;
479};
480\end{lstlisting}
481
482
483\subsection{Dynamic Generic-Types}
484
485Though \CFA implements concrete generic-types efficiently, it also has a fully general system for dynamic generic types.
486As 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.
487Dynamic generic-types also have an \newterm{offset array} containing structure-member offsets.
488A dynamic generic-union needs no such offset array, as all members are at offset 0, but size and alignment are still necessary.
489Access 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.
490
491The offset arrays are statically generated where possible.
492If 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;
493if 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.
494As 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 )@.
495The offset array @_offsetof_pair@ is generated at the call site as @size_t _offsetof_pair[] = { offsetof(_pair_conc1, first), offsetof(_pair_conc1, second) }@.
496
497In some cases the offset arrays cannot be statically generated.
498For 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.
499\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.
500The \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.
501These 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).
502Results of these layout functions are cached so that they are only computed once per type per function. %, as in the example below for @pair@.
503Layout functions also allow generic types to be used in a function definition without reflecting them in the function signature.
504For 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.
505This function could acquire the layout for @set(T)@ by calling its layout function with the layout of @T@ implicitly passed into the function.
506
507Whether a type is concrete, dtype-static, or dynamic is decided solely on the @forall@'s type parameters.
508This 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.
509If 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.
510
511
512\subsection{Applications}
513\label{sec:generic-apps}
514
515The reuse of dtype-static structure instantiations enables useful programming patterns at zero runtime cost.
516The 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@:
517\begin{lstlisting}
518forall(dtype T) int lexcmp( pair( T *, T * ) * a, pair( T *, T * ) * b, int (* cmp)( T *, T * ) ) {
519        return cmp( a->first, b->first ) ? : cmp( a->second, b->second );
520}
521\end{lstlisting}
522Since @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.
523
524Another useful pattern enabled by reused dtype-static type instantiations is zero-cost \newterm{tag-structures}.
525Sometimes information is only used for type-checking and can be omitted at runtime, \eg:
526\begin{lstlisting}
527forall(dtype Unit) struct scalar { unsigned long value; };
528struct metres {};
529struct litres {};
530
531forall(dtype U) scalar(U) ?+?( scalar(U) a, scalar(U) b ) {
532        return (scalar(U)){ a.value + b.value };
533}
534scalar(metres) half_marathon = { 21093 };
535scalar(litres) swimming_pool = { 2500000 };
536scalar(metres) marathon = half_marathon + half_marathon;
537scalar(litres) two_pools = swimming_pool + swimming_pool;
538marathon + swimming_pool;                                       $\C{// compilation ERROR}$
539\end{lstlisting}
540@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 @?+?@.
541These implementations may even be separately compiled, unlike \CC template functions.
542However, the \CFA type-checker ensures matching types are used by all calls to @?+?@, preventing nonsensical computations like adding a length to a volume.
543
544
545\section{Tuples}
546\label{sec:tuples}
547
548In many languages, functions can return at most one value;
549however, many operations have multiple outcomes, some exceptional.
550Consider C's @div@ and @remquo@ functions, which return the quotient and remainder for a division of integer and float values, respectively.
551\begin{lstlisting}
552typedef struct { int quo, rem; } div_t;         $\C{// from include stdlib.h}$
553div_t div( int num, int den );
554double remquo( double num, double den, int * quo );
555div_t qr = div( 13, 5 );                                        $\C{// return quotient/remainder aggregate}$
556int q;
557double r = remquo( 13.5, 5.2, &q );                     $\C{// return remainder, alias quotient}$
558\end{lstlisting}
559@div@ aggregates the quotient/remainder in a structure, while @remquo@ aliases a parameter to an argument.
560Both approaches are awkward.
561Alternatively, a programming language can directly support returning multiple values, \eg in \CFA:
562\begin{lstlisting}
563[ int, int ] div( int num, int den );           $\C{// return two integers}$
564[ double, double ] div( double num, double den ); $\C{// return two doubles}$
565int q, r;                                                                       $\C{// overloaded variable names}$
566double q, r;
567[ q, r ] = div( 13, 5 );                                        $\C{// select appropriate div and q, r}$
568[ q, r ] = div( 13.5, 5.2 );                            $\C{// assign into tuple}$
569\end{lstlisting}
570Clearly, this approach is straightforward to understand and use;
571therefore, why do few programming languages support this obvious feature or provide it awkwardly?
572The answer is that there are complex consequences that cascade through multiple aspects of the language, especially the type-system.
573This section show these consequences and how \CFA handles them.
574
575
576\subsection{Tuple Expressions}
577
578The addition of multiple-return-value functions (MRVF) are useless without a syntax for accepting multiple values at the call-site.
579The simplest mechanism for capturing the return values is variable assignment, allowing the values to be retrieved directly.
580As 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}.
581
582However, 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:
583\begin{lstlisting}
584printf( "%d %d\n", div( 13, 5 ) );                      $\C{// return values seperated into arguments}$
585\end{lstlisting}
586Here, the values returned by @div@ are composed with the call to @printf@ by flattening the tuple into separate arguments.
587However, the \CFA type-system must support significantly more complex composition:
588\begin{lstlisting}
589[ int, int ] foo$\(_1\)$( int );                        $\C{// overloaded foo functions}$
590[ double ] foo$\(_2\)$( int );
591void bar( int, double, double );
592bar( foo( 3 ), foo( 3 ) );
593\end{lstlisting}
594The type-resolver only has the tuple return-types to resolve the call to @bar@ as the @foo@ parameters are identical, which involves unifying the possible @foo@ functions with @bar@'s parameter list.
595No combination of @foo@s are an exact match with @bar@'s parameters, so the resolver applies C conversions.
596The minimal cost is @bar( foo@$_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.
597
598
599\subsection{Tuple Variables}
600
601An important observation from function composition is that new variable names are not required to initialize parameters from an MRVF.
602\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:
603\begin{lstlisting}
604[ int, int ] qr = div( 13, 5 );                         $\C{// tuple-variable declaration and initialization}$
605[ double, double ] qr = div( 13.5, 5.2 );
606\end{lstlisting}
607where the tuple variable-name serves the same purpose as the parameter name(s).
608Tuple variables can be composed of any types, except for array types, since array sizes are generally unknown in C.
609
610One way to access the tuple-variable components is with assignment or composition:
611\begin{lstlisting}
612[ q, r ] = qr;                                                          $\C{// access tuple-variable components}$
613printf( "%d %d\n", qr );
614\end{lstlisting}
615\CFA also supports \newterm{tuple indexing} to access single components of a tuple expression:
616\begin{lstlisting}
617[int, int] * p = &qr;                                           $\C{// tuple pointer}$
618int rem = qr`.1`;                                                       $\C{// access remainder}$
619int quo = div( 13, 5 )`.0`;                                     $\C{// access quotient}$
620p`->0` = 5;                                                                     $\C{// change quotient}$
621bar( qr`.1`, qr );                                                      $\C{// pass remainder and quotient/remainder}$
622rem = [div( 13, 5 ), 42]`.0.1`;                         $\C{// access 2nd component of 1st component of tuple expression}$
623\end{lstlisting}
624
625
626\subsection{Flattening and Restructuring}
627
628In function call contexts, tuples support implicit flattening and restructuring conversions.
629Tuple flattening recursively expands a tuple into the list of its basic components.
630Tuple structuring packages a list of expressions into a value of tuple type, \eg:
631%\lstDeleteShortInline@%
632%\par\smallskip
633%\begin{tabular}{@{}l@{\hspace{1.5\parindent}}||@{\hspace{1.5\parindent}}l@{}}
634\begin{lstlisting}
635int f( int, int );
636int g( [int, int] );
637int h( int, [int, int] );
638[int, int] x;
639int y;
640f( x );                 $\C{// flatten}$
641g( y, 10 );             $\C{// structure}$
642h( x, y );              $\C{// flatten and structure}$
643\end{lstlisting}
644%\end{lstlisting}
645%&
646%\begin{lstlisting}
647%\end{tabular}
648%\smallskip\par\noindent
649%\lstMakeShortInline@%
650In the call to @f@, @x@ is implicitly flattened so the components of @x@ are passed as the two arguments.
651In 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@.
652Finally, 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]@.
653The 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.
654
655
656\subsection{Tuple Assignment}
657
658An assignment where the left side is a tuple type is called \newterm{tuple assignment}.
659There 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.
660%\lstDeleteShortInline@%
661%\par\smallskip
662%\begin{tabular}{@{}l@{\hspace{1.5\parindent}}||@{\hspace{1.5\parindent}}l@{}}
663\begin{lstlisting}
664int x = 10;
665double y = 3.5;
666[int, double] z;
667z = [x, y];                                                                     $\C{// multiple assignment}$
668[x, y] = z;                                                                     $\C{// multiple assignment}$
669z = 10;                                                                         $\C{// mass assignment}$
670[y, x] = 3.14;                                                          $\C{// mass assignment}$
671\end{lstlisting}
672%\end{lstlisting}
673%&
674%\begin{lstlisting}
675%\end{tabular}
676%\smallskip\par\noindent
677%\lstMakeShortInline@%
678Both kinds of tuple assignment have parallel semantics, so that each value on the left and right side is evaluated before any assignments occur.
679As 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]@.
680This 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.
681For example, @[y, x] = 3.14@ performs the assignments @y = 3.14@ and @x = 3.14@, yielding @y == 3.14@ and @x == 3@;
682whereas, C cascading assignment @y = x = 3.14@ performs the assignments @x = 3.14@ and @y = x@, yielding @3@ in @y@ and @x@.
683Finally, 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.
684This example shows mass, multiple, and cascading assignment used in one expression:
685\begin{lstlisting}
686void f( [int, int] );
687f( [x, y] = z = 1.5 );                                          $\C{// assignments in parameter list}$
688\end{lstlisting}
689
690
691\subsection{Member Access}
692
693It is also possible to access multiple fields from a single expression using a \newterm{member-access}.
694The result is a single tuple-valued expression whose type is the tuple of the types of the members, \eg:
695\begin{lstlisting}
696struct S { int x; double y; char * z; } s;
697s.[x, y, z] = 0;
698\end{lstlisting}
699Here, the mass assignment sets all members of @s@ to zero.
700Since 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).
701%\lstDeleteShortInline@%
702%\par\smallskip
703%\begin{tabular}{@{}l@{\hspace{1.5\parindent}}||@{\hspace{1.5\parindent}}l@{}}
704\begin{lstlisting}
705[int, int, long, double] x;
706void f( double, long );
707x.[0, 1] = x.[1, 0];                                            $\C{// rearrange: [x.0, x.1] = [x.1, x.0]}$
708f( x.[0, 3] );                                                          $\C{// drop: f(x.0, x.3)}$
709[int, int, int] y = x.[2, 0, 2];                        $\C{// duplicate: [y.0, y.1, y.2] = [x.2, x.0.x.2]}$
710\end{lstlisting}
711%\end{lstlisting}
712%&
713%\begin{lstlisting}
714%\end{tabular}
715%\smallskip\par\noindent
716%\lstMakeShortInline@%
717It is also possible for a member access to contain other member accesses, \eg:
718\begin{lstlisting}
719struct A { double i; int j; };
720struct B { int * k; short l; };
721struct C { int x; A y; B z; } v;
722v.[x, y.[i, j], z.k];                                           $\C{// [v.x, [v.y.i, v.y.j], v.z.k]}$
723\end{lstlisting}
724
725
726\begin{comment}
727\subsection{Casting}
728
729In C, the cast operator is used to explicitly convert between types.
730In \CFA, the cast operator has a secondary use as type ascription.
731That 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:
732\begin{lstlisting}
733int f();     // (1)
734double f()// (2)
735
736f();       // ambiguous - (1),(2) both equally viable
737(int)f()// choose (2)
738\end{lstlisting}
739
740Since casting is a fundamental operation in \CFA, casts should be given a meaningful interpretation in the context of tuples.
741Taking a look at standard C provides some guidance with respect to the way casts should work with tuples:
742\begin{lstlisting}
743int f();
744void g();
745
746(void)f()// (1)
747(int)g()// (2)
748\end{lstlisting}
749In C, (1) is a valid cast, which calls @f@ and discards its result.
750On the other hand, (2) is invalid, because @g@ does not produce a result, so requesting an @int@ to materialize from nothing is nonsensical.
751Generalizing 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.
752
753Formally, 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 $T_i$.
754Excess 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.
755This approach follows naturally from the way that a cast to @void@ works in C.
756
757For example, in
758\begin{lstlisting}
759[int, int, int] f();
760[int, [int, int], int] g();
761
762([int, double])f();           $\C{// (1)}$
763([int, int, int])g();         $\C{// (2)}$
764([void, [int, int]])g();      $\C{// (3)}$
765([int, int, int, int])g();    $\C{// (4)}$
766([int, [int, int, int]])g()$\C{// (5)}$
767\end{lstlisting}
768
769(1) discards the last element of the return value and converts the second element to @double@.
770Since @int@ is effectively a 1-element tuple, (2) discards the second component of the second element of the return value of @g@.
771If @g@ is free of side effects, this expression is equivalent to @[(int)(g().0), (int)(g().1.0), (int)(g().2)]@.
772Since @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)]@).
773
774Note 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.}.
775As such, (4) is invalid because the cast target type contains 4 components, while the source type contains only 3.
776Similarly, (5) is invalid because the cast @([int, int, int])(g().1)@ is invalid.
777That is, it is invalid to cast @[int, int]@ to @[int, int, int]@.
778\end{comment}
779
780
781\subsection{Polymorphism}
782
783Tuples also integrate with \CFA polymorphism as a kind of generic type.
784Due 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:
785\begin{lstlisting}
786forall(otype T, dtype U) void f( T x, U * y );
787f( [5, "hello"] );
788\end{lstlisting}
789where @[5, "hello"]@ is flattened, giving argument list @5, "hello"@, and @T@ binds to @int@ and @U@ binds to @const char@.
790Tuples, however, may contain polymorphic components.
791For example, a plus operator can be written to add two triples together.
792\begin{lstlisting}
793forall(otype T | { T ?+?( T, T ); }) [T, T, T] ?+?( [T, T, T] x, [T, T, T] y ) {
794        return [x.0 + y.0, x.1 + y.1, x.2 + y.2];
795}
796[int, int, int] x;
797int i1, i2, i3;
798[i1, i2, i3] = x + ([10, 20, 30]);
799\end{lstlisting}
800
801Flattening and restructuring conversions are also applied to tuple types in polymorphic type assertions.
802\begin{lstlisting}
803int f( [int, double], double );
804forall(otype T, otype U | { T f( T, U, U ); }) void g( T, U );
805g( 5, 10.21 );
806\end{lstlisting}
807Hence, function parameter and return lists are flattened for the purposes of type unification allowing the example to pass expression resolution.
808This relaxation is possible by extending the thunk scheme described by Bilson~\cite{Bilson03}.
809Whenever 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:
810\begin{lstlisting}
811int _thunk( int _p0, double _p1, double _p2 ) { return f( [_p0, _p1], _p2 ); }
812\end{lstlisting}
813so the thunk provides flattening and structuring conversions to inferred functions, improving the compatibility of tuples and polymorphism.
814These thunks take advantage of GCC C nested-functions to produce closures that have the usual function-pointer signature.
815
816
817\subsection{Variadic Tuples}
818\label{sec:variadic-tuples}
819
820To define variadic functions, \CFA adds a new kind of type parameter, @ttype@ (tuple type).
821Matching against a @ttype@ parameter consumes all remaining argument components and packages them into a tuple, binding to the resulting tuple of types.
822In 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.
823As such, @ttype@ variables are also called \newterm{argument packs}.
824
825Like variadic templates, the main way to manipulate @ttype@ polymorphic functions is via recursion.
826Since nothing is known about a parameter pack by default, assertion parameters are key to doing anything meaningful.
827Unlike variadic templates, @ttype@ polymorphic functions can be separately compiled.
828For example, a generalized @sum@ function written using @ttype@:
829\begin{lstlisting}
830int sum$\(_0\)$() { return 0; }
831forall(ttype Params | { int sum( Params ); } ) int sum$\(_1\)$( int x, Params rest ) {
832        return x + sum( rest );
833}
834sum( 10, 20, 30 );
835\end{lstlisting}
836Since @sum@\(_0\) does not accept any arguments, it is not a valid candidate function for the call @sum(10, 20, 30)@.
837In 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]@.
838The process continues unitl @Params@ is bound to @[]@, requiring an assertion @int sum()@, which matches @sum@\(_0\) and terminates the recursion.
839Effectively, 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))@.
840
841It is reasonable to take the @sum@ function a step further to enforce a minimum number of arguments:
842\begin{lstlisting}
843int sum( int x, int y ) { return x + y; }
844forall(ttype Params | { int sum( int, Params ); } ) int sum( int x, int y, Params rest ) {
845        return sum( x + y, rest );
846}
847\end{lstlisting}
848One more step permits the summation of any summable type with all arguments of the same type:
849\begin{lstlisting}
850trait summable(otype T) {
851        T ?+?( T, T );
852};
853forall(otype R | summable( R ) ) R sum( R x, R y ) {
854        return x + y;
855}
856forall(otype R, ttype Params | summable(R) | { R sum(R, Params); } ) R sum(R x, R y, Params rest) {
857        return sum( x + y, rest );
858}
859\end{lstlisting}
860Unlike C variadic functions, it is unnecessary to hard code the number and expected types.
861Furthermore, this code is extendable for any user-defined type with a @?+?@ operator.
862Summing arbitrary heterogeneous lists is possible with similar code by adding the appropriate type variables and addition operators.
863
864It is also possible to write a type-safe variadic print function to replace @printf@:
865\begin{lstlisting}
866struct S { int x, y; };
867forall(otype T, ttype Params | { void print(T); void print(Params); }) void print(T arg, Params rest) {
868        print(arg);  print(rest);
869}
870void print( char * x ) { printf( "%s", x ); }
871void print( int x ) { printf( "%d", x ); }
872void print( S s ) { print( "{ ", s.x, ",", s.y, " }" ); }
873print( "s = ", (S){ 1, 2 }, "\n" );
874\end{lstlisting}
875This example showcases a variadic-template-like decomposition of the provided argument list.
876The individual @print@ functions allow printing a single element of a type.
877The polymorphic @print@ allows printing any list of types, where as each individual type has a @print@ function.
878The individual print functions can be used to build up more complicated @print@ functions, such as @S@, which cannot be done with @printf@ in C.
879
880Finally, it is possible to use @ttype@ polymorphism to provide arbitrary argument forwarding functions.
881For example, it is possible to write @new@ as a library function:
882\begin{lstlisting}
883forall( otype R, otype S ) void ?{}( pair(R, S) *, R, S );
884forall( dtype T, ttype Params | sized(T) | { void ?{}( T *, Params ); } ) T * new( Params p ) {
885        return ((T *)malloc()){ p };                    $\C{// construct into result of malloc}$
886}
887pair( int, char ) * x = new( 42, '!' );
888\end{lstlisting}
889The @new@ function provides the combination of type-safe @malloc@ with a \CFA constructor call, making it impossible to forget constructing dynamically allocated objects.
890This function provides the type-safety of @new@ in \CC, without the need to specify the allocated type again, thanks to return-type inference.
891
892
893\subsection{Implementation}
894
895Tuples are implemented in the \CFA translator via a transformation into \newterm{generic types}.
896For each $N$, the first time an $N$-tuple is seen in a scope a generic type with $N$ type parameters is generated, \eg:
897\begin{lstlisting}
898[int, int] f() {
899        [double, double] x;
900        [int, double, int] y;
901}
902\end{lstlisting}
903is transformed into:
904\begin{lstlisting}
905forall(dtype T0, dtype T1 | sized(T0) | sized(T1)) struct _tuple2 {
906        T0 field_0;                                                             $\C{// generated before the first 2-tuple}$
907        T1 field_1;
908};
909_tuple2(int, int) f() {
910        _tuple2(double, double) x;
911        forall(dtype T0, dtype T1, dtype T2 | sized(T0) | sized(T1) | sized(T2)) struct _tuple3 {
912                T0 field_0;                                                     $\C{// generated before the first 3-tuple}$
913                T1 field_1;
914                T2 field_2;
915        };
916        _tuple3(int, double, int) y;
917}
918\end{lstlisting}
919\begin{sloppypar}
920Tuple expressions are then simply converted directly into compound literals, \eg @[5, 'x', 1.24]@ becomes @(_tuple3(int, char, double)){ 5, 'x', 1.24 }@.
921\end{sloppypar}
922
923\begin{comment}
924Since tuples are essentially structures, tuple indexing expressions are just field accesses:
925\begin{lstlisting}
926void f(int, [double, char]);
927[int, double] x;
928
929x.0+x.1;
930printf("%d %g\n", x);
931f(x, 'z');
932\end{lstlisting}
933Is transformed into:
934\begin{lstlisting}
935void f(int, _tuple2(double, char));
936_tuple2(int, double) x;
937
938x.field_0+x.field_1;
939printf("%d %g\n", x.field_0, x.field_1);
940f(x.field_0, (_tuple2){ x.field_1, 'z' });
941\end{lstlisting}
942Note that due to flattening, @x@ used in the argument position is converted into the list of its fields.
943In the call to @f@, the second and third argument components are structured into a tuple argument.
944Similarly, tuple member expressions are recursively expanded into a list of member access expressions.
945
946Expressions that may contain side effects are made into \newterm{unique expressions} before being expanded by the flattening conversion.
947Each unique expression is assigned an identifier and is guaranteed to be executed exactly once:
948\begin{lstlisting}
949void g(int, double);
950[int, double] h();
951g(h());
952\end{lstlisting}
953Internally, this expression is converted to two variables and an expression:
954\begin{lstlisting}
955void g(int, double);
956[int, double] h();
957
958_Bool _unq0_finished_ = 0;
959[int, double] _unq0;
960g(
961        (_unq0_finished_ ? _unq0 : (_unq0 = f(), _unq0_finished_ = 1, _unq0)).0,
962        (_unq0_finished_ ? _unq0 : (_unq0 = f(), _unq0_finished_ = 1, _unq0)).1,
963);
964\end{lstlisting}
965Since argument evaluation order is not specified by the C programming language, this scheme is built to work regardless of evaluation order.
966The first time a unique expression is executed, the actual expression is evaluated and the accompanying boolean is set to true.
967Every subsequent evaluation of the unique expression then results in an access to the stored result of the actual expression.
968Tuple member expressions also take advantage of unique expressions in the case of possible impurity.
969
970Currently, the \CFA translator has a very broad, imprecise definition of impurity, where any function call is assumed to be impure.
971This 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.
972
973The various kinds of tuple assignment, constructors, and destructors generate GNU C statement expressions.
974A 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.
975The 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.
976However, 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.
977\end{comment}
978
979
980\section{Control Structures}
981
982\CFA identifies missing and problematic control structures in C, and extends and modifies these control structures to increase functionality and safety.
983
984
985\subsection{\texorpdfstring{Labelled \LstKeywordStyle{continue} / \LstKeywordStyle{break}}{Labelled continue / break}}
986
987While C provides @continue@ and @break@ statements for altering control flow, both are restricted to one level of nesting for a particular control structure.
988Unfortunately, this restriction forces programmers to use @goto@ to achieve the equivalent control-flow for more than one level of nesting.
989To 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.
990For both @continue@ and @break@, the target label must be directly associated with a @for@, @while@ or @do@ statement;
991for @break@, the target label can also be associated with a @switch@, @if@ or compound (@{}@) statement.
992Figure~\ref{f:MultiLevelExit} shows @continue@ and @break@ indicating the specific control structure, and the corresponding C program using only @goto@ and labels.
993The innermost loop has 7 exit points, which cause continuation or termination of one or more of the 7 nested control-structures.
994
995\begin{figure}
996\lstDeleteShortInline@%
997\begin{tabular}{@{\hspace{\parindentlnth}}l@{\hspace{\parindentlnth}}l@{\hspace{\parindentlnth}}l@{}}
998\multicolumn{1}{@{\hspace{\parindentlnth}}c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}   & \multicolumn{1}{@{\hspace{\parindentlnth}}c}{\textbf{C}}      \\
999\begin{cfa}
1000`LC:` {
1001        ... $declarations$ ...
1002        `LS:` switch ( ... ) {
1003          case 3:
1004                `LIF:` if ( ... ) {
1005                        `LF:` for ( ... ) {
1006                                `LW:` while ( ... ) {
1007                                        ... break `LC`; ...
1008                                        ... break `LS`; ...
1009                                        ... break `LIF`; ...
1010                                        ... continue `LF;` ...
1011                                        ... break `LF`; ...
1012                                        ... continue `LW`; ...
1013                                        ... break `LW`; ...
1014                                } // while
1015                        } // for
1016                } else {
1017                        ... break `LIF`; ...
1018                } // if
1019        } // switch
1020} // compound
1021\end{cfa}
1022&
1023\begin{cfa}
1024{
1025        ... $declarations$ ...
1026        switch ( ... ) {
1027          case 3:
1028                if ( ... ) {
1029                        for ( ... ) {
1030                                while ( ... ) {
1031                                        ... goto `LC`; ...
1032                                        ... goto `LS`; ...
1033                                        ... goto `LIF`; ...
1034                                        ... goto `LFC`; ...
1035                                        ... goto `LFB`; ...
1036                                        ... goto `LWC`; ...
1037                                        ... goto `LWB`; ...
1038                                  `LWC`: ; } `LWB:` ;
1039                          `LFC:` ; } `LFB:` ;
1040                } else {
1041                        ... goto `LIF`; ...
1042                } `L3:` ;
1043        } `LS:` ;
1044} `LC:` ;
1045\end{cfa}
1046&
1047\begin{cfa}
1048
1049
1050
1051
1052
1053
1054
1055// terminate compound
1056// terminate switch
1057// terminate if
1058// continue loop
1059// terminate loop
1060// continue loop
1061// terminate loop
1062
1063
1064
1065// terminate if
1066
1067
1068
1069\end{cfa}
1070\end{tabular}
1071\lstMakeShortInline@%
1072\caption{Multi-level Exit}
1073\label{f:MultiLevelExit}
1074\end{figure}
1075
1076Both labelled @continue@ and @break@ are a @goto@ restricted in the following ways:
1077\begin{itemize}
1078\item
1079They cannot create a loop, which means only the looping constructs cause looping.
1080This restriction means all situations resulting in repeated execution are clearly delineated.
1081\item
1082They cannot branch into a control structure.
1083This restriction prevents missing declarations and/or initializations at the start of a control structure resulting in undefined behaviour.
1084\end{itemize}
1085The 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.
1086Furthermore, 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.
1087With @goto@, the label is at the end of the control structure, which fails to convey this important clue early enough to the reader.
1088Finally, using an explicit target for the transfer instead of an implicit target allows new constructs to be added or removed without affecting existing constructs.
1089The implicit targets of the current @continue@ and @break@, \ie the closest enclosing loop or @switch@, change as certain constructs are added or removed.
1090
1091
1092\subsection{\texorpdfstring{Enhanced \LstKeywordStyle{switch} Statement}{Enhanced switch Statement}}
1093
1094There 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.
1095
1096C 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.}.
1097\CFA provides a shorthand for a non-contiguous list:
1098\begin{cquote}
1099\lstDeleteShortInline@%
1100\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1101\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1102\begin{cfa}
1103case 2, 10, 34, 42:
1104\end{cfa}
1105&
1106\begin{cfa}
1107case 2: case 10: case 34: case 42:
1108\end{cfa}
1109\end{tabular}
1110\lstMakeShortInline@%
1111\end{cquote}
1112for a contiguous list:\footnote{gcc provides the same mechanism with awkward syntax, \lstinline@2 ... 42@, where spaces are required around the ellipse.}
1113\begin{cquote}
1114\lstDeleteShortInline@%
1115\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1116\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1117\begin{cfa}
1118case 2~42:
1119\end{cfa}
1120&
1121\begin{cfa}
1122case 2: case 3: ... case 41: case 42:
1123\end{cfa}
1124\end{tabular}
1125\lstMakeShortInline@%
1126\end{cquote}
1127and a combination:
1128\begin{cfa}
1129case -12~-4, -1~5, 14~21, 34~42:
1130\end{cfa}
1131
1132C allows placement of @case@ clauses \emph{within} statements nested in the @switch@ body (see Duff's device~\cite{Duff83});
1133\begin{cfa}
1134switch ( i ) {
1135  case 0:
1136        for ( int i = 0; i < 10; i += 1 ) {
1137                ...
1138  `case 1:`             // no initialization of loop index
1139                ...
1140        }
1141}
1142\end{cfa}
1143\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.
1144
1145C allows placement of declaration within the @switch@ body and unreachable code at the start, resulting in undefined behaviour:
1146\begin{cfa}
1147switch ( x ) {
1148        `int y = 1;`                            $\C{// unreachable initialization}$
1149        `x = 7;`                                        $\C{// unreachable code without label/branch}$
1150  case 0:
1151        ...
1152        `int z = 0;`                            $\C{// unreachable initialization, cannot appear after case}$
1153        z = 2;
1154  case 1:
1155        `x = z;`                                        $\C{// without fall through, z is undefined}$
1156}
1157\end{cfa}
1158\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.
1159\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.
1160
1161C @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};
1162@case@ clauses are made disjoint by the @break@ statement.
1163While 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.
1164For backwards compatibility, \CFA provides a \emph{new} control structure, @choose@, which mimics @switch@, but reverses the meaning of fall through:
1165\begin{cquote}
1166\lstDeleteShortInline@%
1167\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1168\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1169\begin{cfa}
1170`choose` ( day ) {
1171  case Mon~Thu:  // program
1172
1173  case Fri:  // program
1174        wallet += pay;
1175        `fallthrough;`
1176  case Sat:  // party
1177        wallet -= party;
1178
1179  case Sun:  // rest
1180
1181  default:  // error
1182}
1183\end{cfa}
1184&
1185\begin{cfa}
1186switch ( day ) {
1187  case Mon: case Tue: case Wed: case Thu:  // program
1188        `break;`
1189  case Fri:  // program
1190        wallet += pay;
1191
1192  case Sat:  // party
1193        wallet -= party;
1194        `break;`
1195  case Sun:  // rest
1196        `break;`
1197  default:  // error
1198}
1199\end{cfa}
1200\end{tabular}
1201\lstMakeShortInline@%
1202\end{cquote}
1203Collectively, these enhancements reduce programmer burden and increase readability and safety.
1204
1205\begin{comment}
1206Forgotten @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.
1207\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.
1208Since 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.
1209Where 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:
1210
1211\begin{cfa}
1212choose( i ) {
1213        case 2:
1214                printf("even ");
1215                fallthrough;
1216        case 3: case 5: case 7:
1217                printf("small prime\n");
1218        case 4,6,8,9:
1219                printf("small composite\n");
1220        case 13~19:
1221                printf("teen\n");
1222        default:
1223                printf("something else\n");
1224}
1225\end{cfa}
1226\end{comment}
1227
1228
1229\subsection{\texorpdfstring{\LstKeywordStyle{with} Clause / Statement}{with Clause / Statement}}
1230\label{s:WithClauseStatement}
1231
1232Grouping heterogeneous 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:
1233\begin{cfa}
1234struct S {                                                                      $\C{// aggregate}$
1235        char c;                                                                 $\C{// fields}$
1236        int i;
1237        double d;
1238};
1239S s, as[10];
1240\end{cfa}
1241However, routines manipulating aggregates must repeat the aggregate name to access its containing fields:
1242\begin{cfa}
1243void f( S s ) {
1244        `s.`c; `s.`i; `s.`d;                                    $\C{// access containing fields}$
1245}
1246\end{cfa}
1247which extends to multiple levels of qualification for nested aggregates.
1248A similar situation occurs in object-oriented programming, \eg \CC:
1249\begin{C++}
1250class C {
1251        char c;                                                                 $\C{// fields}$
1252        int i;
1253        double d;
1254        int f() {                                                               $\C{// implicit "this" parameter}$
1255                `this->`c; `this->`i; `this->`d;        $\C{// access containing fields}$
1256        }
1257}
1258\end{C++}
1259Object-oriented nesting of member routines in a \lstinline[language=C++]@class@ allows eliding \lstinline[language=C++]@this->@ because of lexical scoping.
1260However, for other aggregate parameters, qualification is necessary:
1261\begin{cfa}
1262struct T { double m, n; };
1263int C::f( T & t ) {                                                     $\C{// multiple aggregate parameters}$
1264        c; i; d;                                                                $\C{\color{red}// this-\textgreater.c, this-\textgreater.i, this-\textgreater.d}$
1265        `t.`m; `t.`n;                                                   $\C{// must qualify}$
1266}
1267\end{cfa}
1268
1269To simplify the programmer experience, \CFA provides a @with@ statement (see Pascal~\cite[\S~4.F]{Pascal}) to elide aggregate qualification to fields by opening a scope containing the field identifiers.
1270Hence, the qualified fields become variables with the side-effect that it is easier to optimizing field references in a block.
1271\begin{cfa}
1272void f( S & this ) `with ( this )` {            $\C{// with statement}$
1273        c; i; d;                                                                $\C{\color{red}// this.c, this.i, this.d}$
1274}
1275\end{cfa}
1276with the generality of opening multiple aggregate-parameters:
1277\begin{cfa}
1278int f( S & s, T & t ) `with ( s, t )` {         $\C{// multiple aggregate parameters}$
1279        c; i; d;                                                                $\C{\color{red}// s.c, s.i, s.d}$
1280        m; n;                                                                   $\C{\color{red}// t.m, t.n}$
1281}
1282\end{cfa}
1283
1284In detail, the @with@ statement has the form:
1285\begin{cfa}
1286$\emph{with-statement}$:
1287        'with' '(' $\emph{expression-list}$ ')' $\emph{compound-statement}$
1288\end{cfa}
1289and may appear as the body of a routine or nested within a routine body.
1290Each expression in the expression-list provides a type and object.
1291The type must be an aggregate type.
1292(Enumerations are already opened.)
1293The object is the implicit qualifier for the open structure-fields.
1294
1295All expressions in the expression list are open in ``parallel'' within the compound statement.
1296This semantic is different from Pascal, which nests the openings from left to right.
1297The difference between parallel and nesting occurs for fields with the same name and type:
1298\begin{cfa}
1299struct S { int `i`; int j; double m; } s, w;
1300struct T { int `i`; int k; int m; } t, w;
1301with ( s, t ) {
1302        j + k;                                                                  $\C{// unambiguous, s.j + t.k}$
1303        m = 5.0;                                                                $\C{// unambiguous, t.m = 5.0}$
1304        m = 1;                                                                  $\C{// unambiguous, s.m = 1}$
1305        int a = m;                                                              $\C{// unambiguous, a = s.i }$
1306        double b = m;                                                   $\C{// unambiguous, b = t.m}$
1307        int c = s.i + t.i;                                              $\C{// unambiguous, qualification}$
1308        (double)m;                                                              $\C{// unambiguous, cast}$
1309}
1310\end{cfa}
1311For parallel semantics, both @s.i@ and @t.i@ are visible, so @i@ is ambiguous without qualification;
1312for nested semantics, @t.i@ hides @s.i@, so @i@ implies @t.i@.
1313\CFA's ability to overload variables means fields with the same name but different types are automatically disambiguated, eliminating most qualification when opening multiple aggregates.
1314Qualification or a cast is used to disambiguate.
1315
1316There is an interesting problem between parameters and the routine @with@, \eg:
1317\begin{cfa}
1318void ?{}( S & s, int i ) with ( s ) {           $\C{// constructor}$
1319        `s.i = i;` j = 3; m = 5.5;
1320}
1321\end{cfa}
1322Here, the assignment @s.i = i@ means @s.i = s.i@, which is meaningless, and there is no mechanism to qualify the parameter @i@, making the assignment impossible using the routine @with@.
1323To solve this problem, parameters are treated like an initialized aggregate:
1324\begin{cfa}
1325struct Params {
1326        S & s;
1327        int i;
1328} params;
1329\end{cfa}
1330and implicitly opened \emph{after} a routine open, to give them higher priority:
1331\begin{cfa}
1332void ?{}( S & s, int i ) with ( s ) `with( $\emph{\color{red}params}$ )` {
1333        s.i = i; j = 3; m = 5.5;
1334}
1335\end{cfa}
1336Finally, a cast may be used to disambiguate among overload variables in a @with@ expression:
1337\begin{cfa}
1338with ( w ) { ... }                                                      $\C{// ambiguous, same name and no context}$
1339with ( (S)w ) { ... }                                           $\C{// unambiguous, cast}$
1340\end{cfa}
1341and @with@ expressions may be pointers and references (see Section~\ref{s:References}) to aggregates:
1342\begin{cfa}
1343struct S { int i, j; } sv;
1344with ( sv ) {                                                           $\C{variable}$
1345        S & sr = sv;
1346        with ( sr ) {                                                   $\C{reference}$
1347                S * sp = &sv;
1348                with ( *sp ) {                                          $\C{pointer}$
1349                        i = 3; j = 4;                                   $\C{\color{red}// sp-{\textgreater}i, sp-{\textgreater}j}$
1350                }
1351                i = 3; j = 4;                                           $\C{\color{red}// sr.i, sr.j}$
1352        }
1353        i = 3; j = 4;                                                   $\C{\color{red}// sv.i, sv.j}$
1354}
1355\end{cfa}
1356
1357
1358% \subsection{Exception Handling ???}
1359
1360
1361\section{Declarations}
1362
1363It is important to the design team that \CFA subjectively ``feel like'' C to user programmers.
1364An 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.
1365Maintaining 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.
1366Nonetheless, 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.
1367
1368
1369\subsection{Alternative Declaration Syntax}
1370
1371\newcommand{\R}[1]{\Textbf{#1}}
1372\newcommand{\B}[1]{{\Textbf[blue]{#1}}}
1373\newcommand{\G}[1]{{\Textbf[OliveGreen]{#1}}}
1374
1375C declaration syntax is notoriously confusing and error prone.
1376For example, many C programmers are confused by a declaration as simple as:
1377\begin{cquote}
1378\lstDeleteShortInline@%
1379\begin{tabular}{@{}ll@{}}
1380\begin{cfa}
1381int * x[5]
1382\end{cfa}
1383&
1384\raisebox{-0.75\totalheight}{\input{Cdecl}}
1385\end{tabular}
1386\lstMakeShortInline@%
1387\end{cquote}
1388Is this an array of 5 pointers to integers or a pointer to an array of 5 integers?
1389If there is any doubt, it implies productivity and safety issues even for basic programs.
1390Another 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.
1391For example, a routine returning a pointer to an array of integers is defined and used in the following way:
1392\begin{cfa}
1393int `(*`f`())[`5`]` {...};                              $\C{// definition}$
1394 ... `(*`f`())[`3`]` += 1;                              $\C{// usage}$
1395\end{cfa}
1396Essentially, the return type is wrapped around the routine name in successive layers (like an onion).
1397While attempting to make the two contexts consistent is a laudable goal, it has not worked out in practice.
1398
1399\CFA provides its own type, variable and routine declarations, using a different syntax.
1400The new declarations place qualifiers to the left of the base type, while C declarations place qualifiers to the right of the base type.
1401In the following example, \R{red} is the base type and \B{blue} is qualifiers.
1402The \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.
1403\begin{cquote}
1404\lstDeleteShortInline@%
1405\lstset{moredelim=**[is][\color{blue}]{+}{+}}
1406\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1407\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1408\begin{cfa}
1409+[5] *+ `int` x1;
1410+* [5]+ `int` x2;
1411`[* [5] int]` f+( int p )+;
1412\end{cfa}
1413&
1414\begin{cfa}
1415`int` +*+ x1 +[5]+;
1416`int` +(*+x2+)[5]+;
1417`int (*`f+( int p )+`)[5]`;
1418\end{cfa}
1419\end{tabular}
1420\lstMakeShortInline@%
1421\end{cquote}
1422The only exception is bit field specification, which always appear to the right of the base type.
1423% 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.
1424However, unlike C, \CFA type declaration tokens are distributed across all variables in the declaration list.
1425For instance, variables @x@ and @y@ of type pointer to integer are defined in \CFA as follows:
1426\begin{cquote}
1427\lstDeleteShortInline@%
1428\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1429\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1430\begin{cfa}
1431`*` int x, y;
1432\end{cfa}
1433&
1434\begin{cfa}
1435int `*`x, `*`y;
1436\end{cfa}
1437\end{tabular}
1438\lstMakeShortInline@%
1439\end{cquote}
1440The downside of this semantics is the need to separate regular and pointer declarations:
1441\begin{cquote}
1442\lstDeleteShortInline@%
1443\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1444\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1445\begin{cfa}
1446`*` int x;
1447int y;
1448\end{cfa}
1449&
1450\begin{cfa}
1451int `*`x, y;
1452
1453\end{cfa}
1454\end{tabular}
1455\lstMakeShortInline@%
1456\end{cquote}
1457which is prescribing a safety benefit.
1458Other examples are:
1459\begin{cquote}
1460\lstDeleteShortInline@%
1461\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{\hspace{\parindentlnth}}l@{}}
1462\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{C}}      \\
1463\begin{cfa}
1464[ 5 ] int z;
1465[ 5 ] * char w;
1466* [ 5 ] double v;
1467struct s {
1468        int f0:3;
1469        * int f1;
1470        [ 5 ] * int f2;
1471};
1472\end{cfa}
1473&
1474\begin{cfa}
1475int z[ 5 ];
1476char * w[ 5 ];
1477double (* v)[ 5 ];
1478struct s {
1479        int f0:3;
1480        int * f1;
1481        int * f2[ 5 ]
1482};
1483\end{cfa}
1484&
1485\begin{cfa}
1486// array of 5 integers
1487// array of 5 pointers to char
1488// pointer to array of 5 doubles
1489
1490// common bit field syntax
1491
1492
1493
1494\end{cfa}
1495\end{tabular}
1496\lstMakeShortInline@%
1497\end{cquote}
1498
1499All type qualifiers, \eg @const@, @volatile@, etc., are used in the normal way with the new declarations and also appear left to right, \eg:
1500\begin{cquote}
1501\lstDeleteShortInline@%
1502\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{\hspace{\parindentlnth}}l@{}}
1503\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{C}}      \\
1504\begin{cfa}
1505const * const int x;
1506const * [ 5 ] const int y;
1507\end{cfa}
1508&
1509\begin{cfa}
1510int const * const x;
1511const int (* const y)[ 5 ]
1512\end{cfa}
1513&
1514\begin{cfa}
1515// const pointer to const integer
1516// const pointer to array of 5 const integers
1517\end{cfa}
1518\end{tabular}
1519\lstMakeShortInline@%
1520\end{cquote}
1521All 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}
1522The 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:
1523\begin{cquote}
1524\lstDeleteShortInline@%
1525\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{\hspace{\parindentlnth}}l@{}}
1526\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{C}}      \\
1527\begin{cfa}
1528extern [ 5 ] int x;
1529static * const int y;
1530\end{cfa}
1531&
1532\begin{cfa}
1533int extern x[ 5 ];
1534const int static * y;
1535\end{cfa}
1536&
1537\begin{cfa}
1538// externally visible array of 5 integers
1539// internally visible pointer to constant int
1540\end{cfa}
1541\end{tabular}
1542\lstMakeShortInline@%
1543\end{cquote}
1544
1545The new declaration syntax can be used in other contexts where types are required, \eg casts and the pseudo-routine @sizeof@:
1546\begin{cquote}
1547\lstDeleteShortInline@%
1548\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1549\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1550\begin{cfa}
1551y = (* int)x;
1552i = sizeof([ 5 ] * int);
1553\end{cfa}
1554&
1555\begin{cfa}
1556y = (int *)x;
1557i = sizeof(int * [ 5 ]);
1558\end{cfa}
1559\end{tabular}
1560\lstMakeShortInline@%
1561\end{cquote}
1562
1563Finally, new \CFA declarations may appear together with C declarations in the same program block, but cannot be mixed within a specific declaration.
1564Therefore, a programmer has the option of either continuing to use traditional C declarations or take advantage of the new style.
1565Clearly, both styles need to be supported for some time due to existing C-style header-files, particularly for UNIX-like systems.
1566
1567The syntax of the new routine prototype declaration follows directly from the new routine definition syntax;
1568as well, parameter names are optional, \eg:
1569\begin{cfa}
1570[ int x ] f ();                                                 $\C{// returning int with no parameters}$
1571[ * int ] g (int y);                                    $\C{// returning pointer to int with int parameter}$
1572[ ] h ( int, char );                                    $\C{// returning no result with int and char parameters}$
1573[ * int, int ] j ( int );                               $\C{// returning pointer to int and int, with int parameter}$
1574\end{cfa}
1575This syntax allows a prototype declaration to be created by cutting and pasting source text from the routine definition header (or vice versa).
1576Like 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:
1577\begin{cquote}
1578\lstDeleteShortInline@%
1579\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1580\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1581\begin{cfa}
1582[double] foo(), foo( int ), foo( double ) {...}
1583\end{cfa}
1584&
1585\begin{cfa}
1586double foo1(), foo2( int ), foo3( double );
1587\end{cfa}
1588\end{tabular}
1589\lstMakeShortInline@%
1590\end{cquote}
1591\CFA allows the last routine in the list to define its body.
1592
1593Declaration qualifiers can only appear at the start of a \CFA routine declaration,\footref{StorageClassSpecifier} \eg:
1594\begin{cfa}
1595extern [ int ] f ( int );
1596static [ int ] g ( int );
1597\end{cfa}
1598
1599The syntax for pointers to \CFA routines specifies the pointer name on the right, \eg:
1600\begin{cfa}
1601* [ int x ] () fp;                                              $\C{// pointer to routine returning int with no parameters}$
1602* [ * int ] (int y) gp;                                 $\C{// pointer to routine returning pointer to int with int parameter}$
1603* [ ] (int,char) hp;                                    $\C{// pointer to routine returning no result with int and char parameters}$
1604* [ * int,int ] ( int ) jp;                             $\C{// pointer to routine returning pointer to int and int, with int parameter}$
1605\end{cfa}
1606While parameter names are optional, \emph{a routine name cannot be specified};
1607for example, the following is incorrect:
1608\begin{cfa}
1609* [ int x ] f () fp;                                    $\C{// routine name "f" is not allowed}$
1610\end{cfa}
1611
1612
1613\subsection{References}
1614\label{s:References}
1615
1616All variables in C have an \newterm{address}, a \newterm{value}, and a \newterm{type};
1617at 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.
1618The C type-system does not always track the relationship between a value and its address;
1619a 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'').
1620For 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.
1621Despite 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.
1622
1623Within a lexical scope, lvalue expressions have an \newterm{address interpretation} for writing a value or a \newterm{value interpretation} to read a value.
1624For example, in @x = y@, @x@ has an address interpretation, while @y@ has a value interpretation.
1625Though 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.
1626In C, for any type @T@ there is a pointer type @T *@, the value of which is the address of a value of type @T@.
1627A 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 @&?@.
1628
1629\begin{cfa}
1630int x = 1, y = 2, * p1, * p2, ** p3;
1631p1 = &x;                                                                $\C{// p1 points to x}$
1632p2 = &y;                                                                $\C{// p2 points to y}$
1633p3 = &p1;                                                               $\C{// p3 points to p1}$
1634*p2 = ((*p1 + *p2) * (**p3 - *p1)) / (**p3 - 15);
1635\end{cfa}
1636
1637Unfortunately, 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.
1638For 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.
1639However, 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.
1640To 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:
1641
1642\begin{cfa}
1643int x = 1, y = 2, & r1, & r2, && r3;
1644&r1 = &x;  $\C{// r1 points to x}$
1645&r2 = &y;  $\C{// r2 points to y}$
1646&&r3 = &&r1;  $\C{// r3 points to r2}$
1647r2 = ((r1 + r2) * (r3 - r1)) / (r3 - 15);  $\C{// implicit dereferencing}$
1648\end{cfa}
1649
1650Except for auto-dereferencing by the compiler, this reference example is exactly the same as the previous pointer example.
1651Hence, 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.
1652One 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:
1653
1654\begin{cfa}
1655`*`r2 = ((`*`r1 + `*`r2) * (`**`r3 - `*`r1)) / (`**`r3 - 15);
1656\end{cfa}
1657
1658References in \CFA are similar to those in \CC, but with a couple important improvements, both of which can be seen in the example above.
1659Firstly, \CFA does not forbid references to references, unlike \CC.
1660This provides a much more orthogonal design for library implementors, obviating the need for workarounds such as @std::reference_wrapper@.
1661
1662Secondly, unlike the references in \CC which always point to a fixed address, \CFA references are rebindable.
1663This allows \CFA references to be default-initialized (\eg to a null pointer), and also to point to different addresses throughout their lifetime.
1664This rebinding is accomplished without adding any new syntax to \CFA, but simply by extending the existing semantics of the address-of operator in C.
1665
1666In 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.
1667In \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.
1668The 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.
1669This 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.
1670These 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@.
1671More precisely:
1672\begin{itemize}
1673        \item
1674        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).
1675       
1676        \item
1677        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).
1678\end{itemize}
1679Since 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.
1680
1681By analogy to pointers, \CFA references also allow cv-qualifiers such as @const@:
1682
1683\begin{cfa}
1684const int cx = 5;               $\C{// cannot change cx}$
1685const int & cr = cx;    $\C{// cannot change cr's referred value}$
1686&cr = &cx;                              $\C{// rebinding cr allowed}$
1687cr = 7;                                 $\C{// ERROR, cannot change cr}$
1688int & const rc = x;             $\C{// must be initialized, like in \CC}$
1689&rc = &x;                               $\C{// ERROR, cannot rebind rc}$
1690rc = 7;                                 $\C{// x now equal to 7}$
1691\end{cfa}
1692
1693Given that a reference is meant to represent a lvalue, \CFA provides some syntactic shortcuts when initializing references.
1694There are three initialization contexts in \CFA: declaration initialization, argument/parameter binding, and return/temporary binding.
1695In each of these contexts, the address-of operator on the target lvalue may (in fact, must) be elided.
1696The 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@.
1697
1698More 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.
1699Similarly, 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.
1700The final reference conversion included in \CFA is ``rvalue-to-reference'' conversion, implemented by means of an implicit temporary.
1701When 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.
1702This 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.
1703\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.
1704
1705
1706\subsection{Constructors and Destructors}
1707
1708One 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.
1709However, 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.
1710\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.
1711While 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.
1712In 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.
1713
1714In \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{};@.
1715Every 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.
1716This first parameter is informally called the @this@ parameter, as in many object-oriented languages, though a programmer may give it an arbitrary name.
1717Destructors must have exactly one parameter, while constructors allow passing of zero or more additional arguments along with the @this@ parameter.
1718
1719\begin{cfa}
1720struct Array {
1721        int * data;
1722        int len;
1723};
1724
1725void ?{}( Array& arr ) {
1726        arr.len = 10;
1727        arr.data = calloc( arr.len, sizeof(int) );
1728}
1729
1730void ^?{}( Array& arr ) {
1731        free( arr.data );
1732}
1733
1734{
1735        Array x;
1736        `?{}(x);`       $\C{// implicitly compiler-generated}$
1737        // ... use x
1738        `^?{}(x);`      $\C{// implicitly compiler-generated}$
1739}
1740\end{cfa}
1741
1742In 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@.
1743@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.
1744As 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.
1745The exact details of the placement of these implicit constructor and destructor calls are omitted here for brevity, the interested reader should consult \cite{Schluntz17}.
1746
1747Constructor 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.
1748As such, \CFA also provides syntax for \newterm{copy initialization} and \newterm{initialization parameters}:
1749
1750\begin{cfa}
1751void ?{}( Array& arr, Array other );
1752
1753void ?{}( Array& arr, int size, int fill );
1754
1755Array y = { 20, 0xDEADBEEF }, z = y;
1756\end{cfa}
1757
1758Copy 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.
1759Other 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.
1760
1761In addition to initialization syntax, \CFA provides two ways to explicitly call constructors and destructors.
1762Explicit calls to constructors double as a placement syntax, useful for construction of member fields in user-defined constructors and reuse of large storage allocations.
1763While the existing function-call syntax works for explicit calls to constructors and destructors, \CFA also provides a more concise \newterm{operator syntax} for both:
1764
1765\begin{cfa}
1766Array a, b;
1767a{};                            $\C{// default construct}$
1768b{ a };                         $\C{// copy construct}$
1769^a{};                           $\C{// destruct}$
1770a{ 5, 0xFFFFFFFF };     $\C{// explicit constructor call}$
1771\end{cfa}
1772
1773To 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.
1774These default functions can be overridden by user-generated versions of them.
1775For 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.
1776For user-generated types, the four functions are also automatically generated.
1777@enum@ types are handled the same as their underlying integral type, and unions are also bitwise copied and no-op initialized and destructed.
1778For compatibility with C, a copy constructor from the first union member type is also defined.
1779For @struct@ types, each of the four functions are implicitly defined to call their corresponding functions on each member of the struct.
1780To better simulate the behaviour of C initializers, a set of \newterm{field constructors} is also generated for structures.
1781A 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.
1782To 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.
1783These semantics closely mirror the rule for implicit declaration of constructors in \CC\cite[p.~186]{ANSI98:C++}.
1784
1785In 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.
1786If 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.
1787Any C initializer can be the right-hand side of an \lstinline|@=| initializer, \eg  \lstinline|Array a @= { 0, 0x0 }|, with the usual C initialization semantics.
1788In 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.
1789
1790
1791\subsection{Type Nesting}
1792
1793\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.
1794\begin{figure}
1795\centering
1796\lstDeleteShortInline@%
1797\begin{tabular}{@{}l@{\hspace{3em}}l|l@{}}
1798\multicolumn{1}{c@{\hspace{3em}}}{\textbf{C Type Nesting}}      & \multicolumn{1}{c}{\textbf{C Implicit Hoisting}}      & \multicolumn{1}{|c}{\textbf{\CFA}}    \\
1799\hline
1800\begin{cfa}
1801struct S {
1802        enum C { R, G, B };
1803        struct T {
1804                union U { int i, j; };
1805                enum C c;
1806                short int i, j;
1807        };
1808        struct T t;
1809} s;
1810
1811int rtn() {
1812        s.t.c = R;
1813        struct T t = { R, 1, 2 };
1814        enum C c;
1815        union U u;
1816}
1817\end{cfa}
1818&
1819\begin{cfa}
1820enum C { R, G, B };
1821union U { int i, j; };
1822struct T {
1823        enum C c;
1824        short int i, j;
1825};
1826struct S {
1827        struct T t;
1828} s;
1829       
1830
1831
1832
1833
1834
1835
1836\end{cfa}
1837&
1838\begin{cfa}
1839struct S {
1840        enum C { R, G, B };
1841        struct T {
1842                union U { int i, j; };
1843                enum C c;
1844                short int i, j;
1845        };
1846        struct T t;
1847} s;
1848
1849int rtn() {
1850        s.t.c = `S.`R;  // type qualification
1851        struct `S.`T t = { `S.`R, 1, 2 };
1852        enum `S.`C c;
1853        union `S.T.`U u;
1854}
1855\end{cfa}
1856\end{tabular}
1857\lstMakeShortInline@%
1858\caption{Type Nesting / Qualification}
1859\label{f:TypeNestingQualification}
1860\end{figure}
1861In the left example in C, types @C@, @U@ and @T@ are implicitly hoisted outside of type @S@ into the containing block scope.
1862In 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 ``@::@''.
1863
1864
1865\subsection{Default Parameters}
1866
1867
1868\section{Literals}
1869
1870C 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.
1871In 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.
1872
1873\subsection{0/1}
1874
1875In 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.
1876As 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.
1877The 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.
1878To 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.
1879With 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)@.
1880\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.
1881
1882\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.
1883The addition of @one_t@ allows generic algorithms to handle the unit value uniformly for types where that is meaningful.
1884\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.
1885
1886\subsection{Units}
1887
1888Alternative call syntax (literal argument before routine name) to convert basic literals into user literals.
1889
1890{\lstset{language=CFA,deletedelim=**[is][]{`}{`},moredelim=**[is][\color{red}]{@}{@}}
1891\begin{cfa}
1892struct Weight { double stones; };
1893
1894void ?{}( Weight & w ) { w.stones = 0; }        $\C{// operations}$
1895void ?{}( Weight & w, double w ) { w.stones = w; }
1896Weight ?+?( Weight l, Weight r ) { return (Weight){ l.stones + r.stones }; }
1897
1898Weight @?`st@( double w ) { return (Weight){ w }; } $\C{// backquote for units}$
1899Weight @?`lb@( double w ) { return (Weight){ w / 14.0 }; }
1900Weight @?`kg@( double w ) { return (Weight) { w * 0.1575}; }
1901
1902int main() {
1903        Weight w, hw = { 14 };                                  $\C{// 14 stone}$
1904        w = 11@`st@ + 1@`lb@;
1905        w = 70.3@`kg@;
1906        w = 155@`lb@;
1907        w = 0x_9b_u@`lb@;                                               $\C{// hexadecimal unsigned weight (155)}$
1908        w = 0_233@`lb@;                                                 $\C{// octal weight (155)}$
1909        w = 5@`st@ + 8@`kg@ + 25@`lb@ + hw;
1910}
1911\end{cfa}
1912}%
1913
1914
1915\section{Libraries}
1916\label{sec:libraries}
1917
1918As 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.
1919\CFA has replacement libraries condensing hundreds of existing C names into tens of \CFA overloaded names, all without rewriting the actual computations.
1920In many cases, the interface is an inline wrapper providing overloading during compilation but zero cost at runtime.
1921The following sections give a glimpse of the interface reduction to many C libraries.
1922In 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.
1923
1924
1925\subsection{Limits}
1926
1927C library @limits.h@ provides lower and upper bound constants for the basic types.
1928\CFA name overloading is used to condense these typed constants, \eg:
1929\begin{cquote}
1930\lstDeleteShortInline@%
1931\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1932\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{Definition}}       & \multicolumn{1}{c}{\textbf{Usage}}    \\
1933\begin{cfa}
1934const short int `MIN` = -32768;
1935const int `MIN` = -2147483648;
1936const long int `MIN` = -9223372036854775808L;
1937\end{cfa}
1938&
1939\begin{cfa}
1940short int si = `MIN`;
1941int i = `MIN`;
1942long int li = `MIN`;
1943\end{cfa}
1944\end{tabular}
1945\lstMakeShortInline@%
1946\end{cquote}
1947The result is a significant reduction in names to access typed constants, \eg:
1948\begin{cquote}
1949\lstDeleteShortInline@%
1950\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1951\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1952\begin{cfa}
1953MIN
1954MAX
1955M_PI
1956M_E
1957\end{cfa}
1958&
1959\begin{cfa}
1960SCHAR_MIN, CHAR_MIN, SHRT_MIN, INT_MIN, LONG_MIN, LLONG_MIN,
1961SCHAR_MAX, UCHAR_MAX, SHRT_MAX, INT_MAX, LONG_MAX, LLONG_MAX,
1962M_PI, M_PIl, M_CPI, M_CPIl,
1963M_E, M_El, M_CE, M_CEl
1964\end{cfa}
1965\end{tabular}
1966\lstMakeShortInline@%
1967\end{cquote}
1968
1969
1970\subsection{Math}
1971
1972C library @math.h@ provides many mathematical routines.
1973\CFA routine overloading is used to condense these mathematical routines, \eg:
1974\begin{cquote}
1975\lstDeleteShortInline@%
1976\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1977\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{Definition}}       & \multicolumn{1}{c}{\textbf{Usage}}    \\
1978\begin{cfa}
1979float `log`( float x );
1980double `log`( double );
1981double _Complex `log`( double _Complex x );
1982\end{cfa}
1983&
1984\begin{cfa}
1985float f = `log`( 3.5 );
1986double d = `log`( 3.5 );
1987double _Complex dc = `log`( 3.5+0.5I );
1988\end{cfa}
1989\end{tabular}
1990\lstMakeShortInline@%
1991\end{cquote}
1992The result is a significant reduction in names to access math routines, \eg:
1993\begin{cquote}
1994\lstDeleteShortInline@%
1995\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1996\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1997\begin{cfa}
1998log
1999sqrt
2000sin
2001\end{cfa}
2002&
2003\begin{cfa}
2004logf, log, logl, clogf, clog, clogl
2005sqrtf, sqrt, sqrtl, csqrtf, csqrt, csqrtl
2006sinf, sin, sinl, csinf, csin, csinl
2007\end{cfa}
2008\end{tabular}
2009\lstMakeShortInline@%
2010\end{cquote}
2011While \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).
2012For example, it is not possible to overload @atan@ for both one and two arguments;
2013instead the names @atan@ and @atan2@ are required.
2014The 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.
2015
2016
2017\subsection{Standard}
2018
2019C library @stdlib.h@ provides many general routines.
2020\CFA routine overloading is used to condense these utility routines, \eg:
2021\begin{cquote}
2022\lstDeleteShortInline@%
2023\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
2024\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{Definition}}       & \multicolumn{1}{c}{\textbf{Usage}}    \\
2025\begin{cfa}
2026unsigned int `abs`( int );
2027double `abs`( double );
2028double abs( double _Complex );
2029\end{cfa}
2030&
2031\begin{cfa}
2032unsigned int i = `abs`( -1 );
2033double d = `abs`( -1.5 );
2034double d = `abs`( -1.5+0.5I );
2035\end{cfa}
2036\end{tabular}
2037\lstMakeShortInline@%
2038\end{cquote}
2039The result is a significant reduction in names to access utility routines, \eg:
2040\begin{cquote}
2041\lstDeleteShortInline@%
2042\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
2043\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
2044\begin{cfa}
2045abs
2046strto
2047random
2048\end{cfa}
2049&
2050\begin{cfa}
2051abs, labs, llabs, fabsf, fabs, fabsl, cabsf, cabs, cabsl
2052strtol, strtoul, strtoll, strtoull, strtof, strtod, strtold
2053srand48, mrand48, lrand48, drand48
2054\end{cfa}
2055\end{tabular}
2056\lstMakeShortInline@%
2057\end{cquote}
2058In additon, there are polymorphic routines, like @min@ and @max@, which work on any type with operators @?<?@ or @?>?@.
2059
2060The following shows one example where \CFA \emph{extends} an existing standard C interface to reduce complexity and provide safety.
2061C/\Celeven provide a number of complex and overlapping storage-management operation to support the following capabilities:
2062\begin{description}[itemsep=2pt,parsep=0pt]
2063\item[fill]
2064after allocation the storage is filled with a specified character.
2065\item[resize]
2066an existing allocation is decreased or increased in size.
2067In 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.
2068For an increase in storage size, new storage after the copied data may be filled.
2069\item[alignment]
2070an allocation starts on a specified memory boundary, \eg, an address multiple of 64 or 128 for cache-line purposes.
2071\item[array]
2072the allocation size is scaled to the specified number of array elements.
2073An array may be filled, resized, or aligned.
2074\end{description}
2075Table~\ref{t:StorageManagementOperations} shows the capabilities provided by C/\Celeven allocation-routines and how all the capabilities can be combined into two \CFA routines.
2076
2077\CFA storage-management routines extend the C equivalents by overloading, providing shallow type-safety, and removing the need to specify the base allocation-size.
2078The following example contrasts \CFA and C storage-allocation operation performing the same operations with the same type safety:
2079\begin{cquote}
2080\begin{cfa}[aboveskip=0pt]
2081size_t  dim = 10;                                                       $\C{// array dimension}$
2082char fill = '\xff';                                                     $\C{// initialization fill value}$
2083int * ip;
2084\end{cfa}
2085\lstDeleteShortInline@%
2086\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
2087\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
2088\begin{cfa}
2089ip = alloc();
2090ip = alloc( fill );
2091ip = alloc( dim );
2092ip = alloc( dim, fill );
2093ip = alloc( ip, 2 * dim );
2094ip = alloc( ip, 4 * dim, fill );
2095
2096ip = align_alloc( 16 );
2097ip = align_alloc( 16, fill );
2098ip = align_alloc( 16, dim );
2099ip = align_alloc( 16, dim, fill );
2100\end{cfa}
2101&
2102\begin{cfa}
2103ip = (int *)malloc( sizeof( int ) );
2104ip = (int *)malloc( sizeof( int ) ); memset( ip, fill, sizeof( int ) );
2105ip = (int *)malloc( dim * sizeof( int ) );
2106ip = (int *)malloc( sizeof( int ) ); memset( ip, fill, dim * sizeof( int ) );
2107ip = (int *)realloc( ip, 2 * dim * sizeof( int ) );
2108ip = (int *)realloc( ip, 4 * dim * sizeof( int ) ); memset( ip, fill, 4 * dim * sizeof( int ) );
2109
2110ip = memalign( 16, sizeof( int ) );
2111ip = memalign( 16, sizeof( int ) ); memset( ip, fill, sizeof( int ) );
2112ip = memalign( 16, dim * sizeof( int ) );
2113ip = memalign( 16, dim * sizeof( int ) ); memset( ip, fill, dim * sizeof( int ) );
2114\end{cfa}
2115\end{tabular}
2116\lstMakeShortInline@%
2117\end{cquote}
2118Variadic @new@ (see Section~\ref{sec:variadic-tuples}) cannot support the same overloading because extra parameters are for initialization.
2119Hence, there are @new@ and @anew@ routines for single and array variables, and the fill value is the arguments to the constructor, \eg:
2120\begin{cfa}
2121struct S { int i, j; };
2122void ?{}( S & s, int i, int j ) { s.i = i; s.j = j; }
2123S * s = new( 2, 3 );                                            $\C{// allocate storage and run constructor}$
2124S * as = anew( dim, 2, 3 );                                     $\C{// each array element initialized to 2, 3}$
2125\end{cfa}
2126Note, \CC can only initialization array elements via the default constructor.
2127
2128Finally, the \CFA memory-allocator has \newterm{sticky properties} for dynamic storage: fill and alignment are remembered with an object's storage in the heap.
2129When a @realloc@ is performed, the sticky properties are respected, so that new storage is correctly aligned and initialized with the fill character.
2130
2131\begin{table}
2132\centering
2133\lstDeleteShortInline@%
2134\lstMakeShortInline~%
2135\begin{tabular}{@{}r|r|l|l|l|l@{}}
2136\multicolumn{1}{c}{}&           & \multicolumn{1}{c|}{fill}     & resize        & alignment     & array \\
2137\hline
2138C               & ~malloc~                      & no                    & no            & no            & no    \\
2139                & ~calloc~                      & yes (0 only)  & no            & no            & yes   \\
2140                & ~realloc~                     & no/copy               & yes           & no            & no    \\
2141                & ~memalign~            & no                    & no            & yes           & no    \\
2142                & ~posix_memalign~      & no                    & no            & yes           & no    \\
2143\hline
2144C11             & ~aligned_alloc~       & no                    & no            & yes           & no    \\
2145\hline
2146\CFA    & ~alloc~                       & yes/copy              & no/yes        & no            & yes   \\
2147                & ~align_alloc~         & yes                   & no            & yes           & yes   \\
2148\end{tabular}
2149\lstDeleteShortInline~%
2150\lstMakeShortInline@%
2151\caption{Storage-Management Operations}
2152\label{t:StorageManagementOperations}
2153\end{table}
2154
2155
2156\subsection{I/O}
2157\label{s:IOLibrary}
2158
2159The goal of \CFA I/O is to simplify the common cases, while fully supporting polymorphism and user defined types in a consistent way.
2160The approach combines ideas from \CC and Python.
2161The \CFA header file for the I/O library is @fstream@.
2162
2163The common case is printing out a sequence of variables separated by whitespace.
2164\begin{cquote}
2165\lstDeleteShortInline@%
2166\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
2167\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{\CC}}      \\
2168\begin{cfa}
2169int x = 1, y = 2, z = 3;
2170sout | x `|` y `|` z | endl;
2171\end{cfa}
2172&
2173\begin{cfa}
2174
2175cout << x `<< " "` << y `<< " "` << z << endl;
2176\end{cfa}
2177\\
2178\begin{cfa}[showspaces=true,aboveskip=0pt,belowskip=0pt]
21791` `2` `3
2180\end{cfa}
2181&
2182\begin{cfa}[showspaces=true,aboveskip=0pt,belowskip=0pt]
21831 2 3
2184\end{cfa}
2185\end{tabular}
2186\lstMakeShortInline@%
2187\end{cquote}
2188The \CFA form has half the characters of the \CC form, and is similar to Python I/O with respect to implicit separators.
2189Similar simplification occurs for tuple I/O, which prints all tuple values separated by ``\lstinline[showspaces=true]@, @''.
2190\begin{cfa}
2191[int, [ int, int ] ] t1 = [ 1, [ 2, 3 ] ], t2 = [ 4, [ 5, 6 ] ];
2192sout | t1 | t2 | endl;                                  $\C{// print tuples}$
2193\end{cfa}
2194\begin{cfa}[showspaces=true,aboveskip=0pt]
21951`, `2`, `3 4`, `5`, `6
2196\end{cfa}
2197Finally, \CFA uses the logical-or operator for I/O as it is the lowest-priority overloadable operator, other than assignment.
2198Therefore, fewer output expressions require parenthesis.
2199\begin{cquote}
2200\lstDeleteShortInline@%
2201\begin{tabular}{@{}ll@{}}
2202\textbf{\CFA:}
2203&
2204\begin{cfa}
2205sout | x * 3 | y + 1 | z << 2 | x == y | (x | y) | (x || y) | (x > z ? 1 : 2) | endl;
2206\end{cfa}
2207\\
2208\textbf{\CC:}
2209&
2210\begin{cfa}
2211cout << x * 3 << y + 1 << `(`z << 2`)` << `(`x == y`)` << (x | y) << (x || y) << (x > z ? 1 : 2) << endl;
2212\end{cfa}
2213\\
2214\textbf{output:}
2215&
2216\begin{cfa}[showspaces=true,aboveskip=0pt]
22173 3 12 0 3 1 2
2218\end{cfa}
2219\end{tabular}
2220\lstMakeShortInline@%
2221\end{cquote}
2222There 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.
2223
2224The implicit separator character (space/blank) is a separator not a terminator.
2225The rules for implicitly adding the separator are:
2226\begin{itemize}[itemsep=2pt,parsep=0pt]
2227\item
2228A separator does not appear at the start or end of a line.
2229\item
2230A separator does not appear before or after a character literal or variable.
2231\item
2232A separator does not appear before or after a null (empty) C string, which is a local mechanism to disable insertion of the separator character.
2233\item
2234A separator does not appear before a C string starting with the characters: \lstinline[mathescape=off,basicstyle=\tt]@([{=$@
2235\item
2236A seperator does not appear after a C string ending with the characters: \lstinline[basicstyle=\tt]@,.;!?)]}%@
2237\item
2238{\lstset{language=CFA,deletedelim=**[is][]{`}{`}}
2239A 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@
2240}%
2241\item
2242There are routines to set and get the separator string, and manipulators to toggle separation on and off in the middle of output.
2243\end{itemize}
2244
2245
2246\subsection{Multi-precision Integers}
2247\label{s:MultiPrecisionIntegers}
2248
2249\CFA has an interface to the GMP multi-precision signed-integers~\cite{GMP}, similar to the \CC interface provided by GMP.
2250The \CFA interface wraps GMP routines into operator routines to make programming with multi-precision integers identical to using fixed-sized integers.
2251The \CFA type name for multi-precision signed-integers is @Int@ and the header file is @gmp@.
2252The following multi-precision factorial programs contrast using GMP with the \CFA and C interfaces.
2253\begin{cquote}
2254\lstDeleteShortInline@%
2255\begin{tabular}{@{}l@{\hspace{\parindentlnth}}@{\hspace{\parindentlnth}}l@{}}
2256\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{@{\hspace{\parindentlnth}}c}{\textbf{C}}      \\
2257\begin{cfa}
2258#include <gmp>
2259int main( void ) {
2260        sout | "Factorial Numbers" | endl;
2261        Int fact = 1;
2262
2263        sout | 0 | fact | endl;
2264        for ( unsigned int i = 1; i <= 40; i += 1 ) {
2265                fact *= i;
2266                sout | i | fact | endl;
2267        }
2268}
2269\end{cfa}
2270&
2271\begin{cfa}
2272#include <gmp.h>
2273int main( void ) {
2274        `gmp_printf`( "Factorial Numbers\n" );
2275        `mpz_t` fact;
2276        `mpz_init_set_ui`( fact, 1 );
2277        `gmp_printf`( "%d %Zd\n", 0, fact );
2278        for ( unsigned int i = 1; i <= 40; i += 1 ) {
2279                `mpz_mul_ui`( fact, fact, i );
2280                `gmp_printf`( "%d %Zd\n", i, fact );
2281        }
2282}
2283\end{cfa}
2284\end{tabular}
2285\lstMakeShortInline@%
2286\end{cquote}
2287
2288
2289\section{Evaluation}
2290\label{sec:eval}
2291
2292Though \CFA provides significant added functionality over C, these features have a low runtime penalty.
2293In fact, \CFA's features for generic programming can enable faster runtime execution than idiomatic @void *@-based C code.
2294This 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}).
2295Since 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.
2296A more illustrative benchmark measures the costs of idiomatic usage of each language's features.
2297Figure~\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}.
2298The benchmark test is similar for C and \CC.
2299The 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.
2300
2301\begin{figure}
2302\begin{lstlisting}[xleftmargin=3\parindentlnth,aboveskip=0pt,belowskip=0pt]
2303int main( int argc, char * argv[] ) {
2304        FILE * out = fopen( "cfa-out.txt", "w" );
2305        int maxi = 0, vali = 42;
2306        stack(int) si, ti;
2307
2308        REPEAT_TIMED( "push_int", N, push( &si, vali ); )
2309        TIMED( "copy_int", ti = si; )
2310        TIMED( "clear_int", clear( &si ); )
2311        REPEAT_TIMED( "pop_int", N,
2312                int xi = pop( &ti ); if ( xi > maxi ) { maxi = xi; } )
2313        REPEAT_TIMED( "print_int", N/2, print( out, vali, ":", vali, "\n" ); )
2314
2315        pair(_Bool, char) maxp = { (_Bool)0, '\0' }, valp = { (_Bool)1, 'a' };
2316        stack(pair(_Bool, char)) sp, tp;
2317
2318        REPEAT_TIMED( "push_pair", N, push( &sp, valp ); )
2319        TIMED( "copy_pair", tp = sp; )
2320        TIMED( "clear_pair", clear( &sp ); )
2321        REPEAT_TIMED( "pop_pair", N,
2322                pair(_Bool, char) xp = pop( &tp ); if ( xp > maxp ) { maxp = xp; } )
2323        REPEAT_TIMED( "print_pair", N/2, print( out, valp, ":", valp, "\n" ); )
2324        fclose(out);
2325}
2326\end{lstlisting}
2327\caption{\protect\CFA Benchmark Test}
2328\label{fig:BenchmarkTest}
2329\end{figure}
2330
2331The 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.
2332The \CCV variant illustrates an alternative object-oriented idiom where all objects inherit from a base @object@ class, mimicking a Java-like interface;
2333hence runtime checks are necessary to safely down-cast objects.
2334The 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.
2335For 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.
2336Note, the C benchmark uses unchecked casts as there is no runtime mechanism to perform such checks, while \CFA and \CC provide type-safety statically.
2337
2338Figure~\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.
2339The graph plots the median of 5 consecutive runs of each program, with an initial warm-up run omitted.
2340All code is compiled at \texttt{-O2} by GCC or G++ 6.2.0, with all \CC code compiled as \CCfourteen.
2341The 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.
2342
2343\begin{figure}
2344\centering
2345\input{timing}
2346\caption{Benchmark Timing Results (smaller is better)}
2347\label{fig:eval}
2348\end{figure}
2349
2350\begin{table}
2351\centering
2352\caption{Properties of benchmark code}
2353\label{tab:eval}
2354\newcommand{\CT}[1]{\multicolumn{1}{c}{#1}}
2355\begin{tabular}{rrrrr}
2356                                                                        & \CT{C}        & \CT{\CFA}     & \CT{\CC}      & \CT{\CCV}             \\ \hline
2357maximum memory usage (MB)                       & 10001         & 2502          & 2503          & 11253                 \\
2358source code size (lines)                        & 247           & 222           & 165           & 339                   \\
2359redundant type annotations (lines)      & 39            & 2                     & 2                     & 15                    \\
2360binary size (KB)                                        & 14            & 229           & 18            & 38                    \\
2361\end{tabular}
2362\end{table}
2363
2364The 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;
2365this inefficiency is exacerbated by the second level of generic types in the pair-based benchmarks.
2366By 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.
2367\CCV is slower than C largely due to the cost of runtime type-checking of down-casts (implemented with @dynamic_cast@);
2368There are two outliers in the graph for \CFA: all prints and pop of @pair@.
2369Both 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.
2370A compiler designed for \CFA could easily perform these optimizations.
2371Finally, the binary size for \CFA is larger because of static linking with the \CFA libraries.
2372
2373\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.
2374On 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.
2375\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;
2376with their omission, the \CCV line count is similar to C.
2377We 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.
2378
2379Raw line-count, however, is a fairly rough measure of code complexity;
2380another important factor is how much type information the programmer must manually specify, especially where that information is not checked by the compiler.
2381Such 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@).
2382To 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.
2383The \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.
2384The 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).
2385These uses are similar to the @new@ expressions in \CC, though the \CFA compiler's type resolver should shortly render even these type casts superfluous.
2386
2387
2388\section{Related Work}
2389
2390
2391\subsection{Polymorphism}
2392
2393\CC is the most similar language to \CFA;
2394both are extensions to C with source and runtime backwards compatibility.
2395The fundamental difference is in their engineering approach to C compatibility and programmer expectation.
2396While \CC provides good backwards compatibility with C, it has a steep learning curve for many of its extensions.
2397For example, polymorphism is provided via three disjoint mechanisms: overloading, inheritance, and templates.
2398The 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.
2399In 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.
2400The key mechanism to support separate compilation is \CFA's \emph{explicit} use of assumed properties for a type.
2401Until \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;
2402furthermore, \CC concepts are restricted to template polymorphism.
2403
2404Cyclone~\cite{Grossman06} also provides capabilities for polymorphic functions and existential types, similar to \CFA's @forall@ functions and generic types.
2405Cyclone 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.
2406Furthermore, 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@.
2407In \CFA terms, all Cyclone polymorphism must be dtype-static.
2408While 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.
2409Smith 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.
2410
2411Objective-C~\cite{obj-c-book} is an industrially successful extension to C.
2412However, Objective-C is a radical departure from C, using an object-oriented model with message-passing.
2413Objective-C did not support type-checked generics until recently \cite{xcode7}, historically using less-efficient runtime checking of object types.
2414The GObject~\cite{GObject} framework also adds object-oriented programming with runtime type-checking and reference-counting garbage-collection to C;
2415these features are more intrusive additions than those provided by \CFA, in addition to the runtime overhead of reference-counting.
2416Vala~\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.
2417Java~\cite{Java8} included generic types in Java~5, which are type-checked at compilation and type-erased at runtime, similar to \CFA's.
2418However, in Java, each object carries its own table of method pointers, while \CFA passes the method pointers separately to maintain a C-compatible layout.
2419Java is also a garbage-collected, object-oriented language, with the associated resource usage and C-interoperability burdens.
2420
2421D~\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.
2422However, 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.
2423D and Go are garbage-collected languages, imposing the associated runtime overhead.
2424The necessity of accounting for data transfer between managed runtimes and the unmanaged C runtime complicates foreign-function interfaces to C.
2425Furthermore, 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.
2426D 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.
2427Rust also possesses much more powerful abstraction capabilities for writing generic code than Go.
2428On 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.
2429\CFA, with its more modest safety features, allows direct ports of C code while maintaining the idiomatic style of the original source.
2430
2431
2432\subsection{Tuples/Variadics}
2433
2434Many programming languages have some form of tuple construct and/or variadic functions, \eg SETL, C, KW-C, \CC, D, Go, Java, ML, and Scala.
2435SETL~\cite{SETL} is a high-level mathematical programming language, with tuples being one of the primary data types.
2436Tuples in SETL allow subscripting, dynamic expansion, and multiple assignment.
2437C 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.
2438KW-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.
2439The main contributions of that work were adding MRVF, tuple mass and multiple assignment, and record-field access.
2440\CCeleven introduced @std::tuple@ as a library variadic template structure.
2441Tuples are a generalization of @std::pair@, in that they allow for arbitrary length, fixed-size aggregation of heterogeneous values.
2442Operations include @std::get<N>@ to extract values, @std::tie@ to create a tuple of references used for assignment, and lexicographic comparisons.
2443\CCseventeen proposes \emph{structured bindings}~\cite{Sutter15} to eliminate pre-declaring variables and use of @std::tie@ for binding the results.
2444This 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.
2445Furthermore, structured bindings are not a full replacement for @std::tie@, as it always declares new variables.
2446Like \CC, D provides tuples through a library variadic-template structure.
2447Go does not have tuples but supports MRVF.
2448Java'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.
2449Tuples are a fundamental abstraction in most functional programming languages, such as Standard ML~\cite{sml} and~\cite{Scala}, which decompose tuples using pattern matching.
2450
2451
2452\section{Conclusion and Future Work}
2453
2454The 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.
2455While 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.
2456The 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.
2457The contributions are a powerful type-system using parametric polymorphism and overloading, generic types, and tuples, which all have complex interactions.
2458The work is a challenging design, engineering, and implementation exercise.
2459On the surface, the project may appear as a rehash of similar mechanisms in \CC.
2460However, 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.
2461All of these new features are being used by the \CFA development-team to build the \CFA runtime-system.
2462Finally, we demonstrate that \CFA performance for some idiomatic cases is better than C and close to \CC, showing the design is practically applicable.
2463
2464There is ongoing work on a wide range of \CFA feature extensions, including arrays with size, exceptions, concurrent primitives, modules, and user-defined conversions.
2465(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.)
2466In addition, there are interesting future directions for the polymorphism design.
2467Notably, \CC template functions trade compile time and code bloat for optimal runtime of individual instantiations of polymorphic functions.
2468\CFA polymorphic functions use dynamic virtual-dispatch;
2469the 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.
2470Two 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).
2471These approaches are not mutually exclusive and allow performance optimizations to be applied only when necessary, without suffering global code-bloat.
2472In general, we believe separate compilation, producing smaller code, works well with loaded hardware-caches, which may offset the benefit of larger inlined-code.
2473
2474
2475\section{Acknowledgments}
2476
2477The 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.
2478%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.
2479% 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.
2480
2481
2482\bibliographystyle{plain}
2483\bibliography{pl}
2484
2485
2486\appendix
2487
2488\section{Benchmark Stack Implementation}
2489\label{sec:BenchmarkStackImplementation}
2490
2491\lstset{basicstyle=\linespread{0.9}\sf\small}
2492
2493Throughout, @/***/@ designates a counted redundant type annotation.
2494
2495\smallskip\noindent
2496\CFA
2497\begin{lstlisting}[xleftmargin=2\parindentlnth,aboveskip=0pt,belowskip=0pt]
2498forall(otype T) struct stack_node {
2499        T value;
2500        stack_node(T) * next;
2501};
2502forall(otype T) void ?{}(stack(T) * s) { (&s->head){ 0 }; }
2503forall(otype T) void ?{}(stack(T) * s, stack(T) t) {
2504        stack_node(T) ** crnt = &s->head;
2505        for ( stack_node(T) * next = t.head; next; next = next->next ) {
2506                *crnt = ((stack_node(T) *)malloc()){ next->value }; /***/
2507                stack_node(T) * acrnt = *crnt;
2508                crnt = &acrnt->next;
2509        }
2510        *crnt = 0;
2511}
2512forall(otype T) stack(T) ?=?(stack(T) * s, stack(T) t) {
2513        if ( s->head == t.head ) return *s;
2514        clear(s);
2515        s{ t };
2516        return *s;
2517}
2518forall(otype T) void ^?{}(stack(T) * s) { clear(s); }
2519forall(otype T) _Bool empty(const stack(T) * s) { return s->head == 0; }
2520forall(otype T) void push(stack(T) * s, T value) {
2521        s->head = ((stack_node(T) *)malloc()){ value, s->head }; /***/
2522}
2523forall(otype T) T pop(stack(T) * s) {
2524        stack_node(T) * n = s->head;
2525        s->head = n->next;
2526        T x = n->value;
2527        ^n{};
2528        free(n);
2529        return x;
2530}
2531forall(otype T) void clear(stack(T) * s) {
2532        for ( stack_node(T) * next = s->head; next; ) {
2533                stack_node(T) * crnt = next;
2534                next = crnt->next;
2535                delete(crnt);
2536        }
2537        s->head = 0;
2538}
2539\end{lstlisting}
2540
2541\medskip\noindent
2542\CC
2543\begin{lstlisting}[xleftmargin=2\parindentlnth,aboveskip=0pt,belowskip=0pt]
2544template<typename T> class stack {
2545        struct node {
2546                T value;
2547                node * next;
2548                node( const T & v, node * n = nullptr ) : value(v), next(n) {}
2549        };
2550        node * head;
2551        void copy(const stack<T>& o) {
2552                node ** crnt = &head;
2553                for ( node * next = o.head;; next; next = next->next ) {
2554                        *crnt = new node{ next->value }; /***/
2555                        crnt = &(*crnt)->next;
2556                }
2557                *crnt = nullptr;
2558        }
2559  public:
2560        stack() : head(nullptr) {}
2561        stack(const stack<T>& o) { copy(o); }
2562        stack(stack<T> && o) : head(o.head) { o.head = nullptr; }
2563        ~stack() { clear(); }
2564        stack & operator= (const stack<T>& o) {
2565                if ( this == &o ) return *this;
2566                clear();
2567                copy(o);
2568                return *this;
2569        }
2570        stack & operator= (stack<T> && o) {
2571                if ( this == &o ) return *this;
2572                head = o.head;
2573                o.head = nullptr;
2574                return *this;
2575        }
2576        bool empty() const { return head == nullptr; }
2577        void push(const T & value) { head = new node{ value, head };  /***/ }
2578        T pop() {
2579                node * n = head;
2580                head = n->next;
2581                T x = std::move(n->value);
2582                delete n;
2583                return x;
2584        }
2585        void clear() {
2586                for ( node * next = head; next; ) {
2587                        node * crnt = next;
2588                        next = crnt->next;
2589                        delete crnt;
2590                }
2591                head = nullptr;
2592        }
2593};
2594\end{lstlisting}
2595
2596\medskip\noindent
2597C
2598\begin{lstlisting}[xleftmargin=2\parindentlnth,aboveskip=0pt,belowskip=0pt]
2599struct stack_node {
2600        void * value;
2601        struct stack_node * next;
2602};
2603struct stack new_stack() { return (struct stack){ NULL }; /***/ }
2604void copy_stack(struct stack * s, const struct stack * t, void * (*copy)(const void *)) {
2605        struct stack_node ** crnt = &s->head;
2606        for ( struct stack_node * next = t->head; next; next = next->next ) {
2607                *crnt = malloc(sizeof(struct stack_node)); /***/
2608                **crnt = (struct stack_node){ copy(next->value) }; /***/
2609                crnt = &(*crnt)->next;
2610        }
2611        *crnt = 0;
2612}
2613_Bool stack_empty(const struct stack * s) { return s->head == NULL; }
2614void push_stack(struct stack * s, void * value) {
2615        struct stack_node * n = malloc(sizeof(struct stack_node)); /***/
2616        *n = (struct stack_node){ value, s->head }; /***/
2617        s->head = n;
2618}
2619void * pop_stack(struct stack * s) {
2620        struct stack_node * n = s->head;
2621        s->head = n->next;
2622        void * x = n->value;
2623        free(n);
2624        return x;
2625}
2626void clear_stack(struct stack * s, void (*free_el)(void *)) {
2627        for ( struct stack_node * next = s->head; next; ) {
2628                struct stack_node * crnt = next;
2629                next = crnt->next;
2630                free_el(crnt->value);
2631                free(crnt);
2632        }
2633        s->head = NULL;
2634}
2635\end{lstlisting}
2636
2637\medskip\noindent
2638\CCV
2639\begin{lstlisting}[xleftmargin=2\parindentlnth,aboveskip=0pt,belowskip=0pt]
2640stack::node::node( const object & v, node * n ) : value( v.new_copy() ), next( n ) {}
2641void stack::copy(const stack & o) {
2642        node ** crnt = &head;
2643        for ( node * next = o.head; next; next = next->next ) {
2644                *crnt = new node{ *next->value };
2645                crnt = &(*crnt)->next;
2646        }
2647        *crnt = nullptr;
2648}
2649stack::stack() : head(nullptr) {}
2650stack::stack(const stack & o) { copy(o); }
2651stack::stack(stack && o) : head(o.head) { o.head = nullptr; }
2652stack::~stack() { clear(); }
2653stack & stack::operator= (const stack & o) {
2654        if ( this == &o ) return *this;
2655        clear();
2656        copy(o);
2657        return *this;
2658}
2659stack & stack::operator= (stack && o) {
2660        if ( this == &o ) return *this;
2661        head = o.head;
2662        o.head = nullptr;
2663        return *this;
2664}
2665bool stack::empty() const { return head == nullptr; }
2666void stack::push(const object & value) { head = new node{ value, head }; /***/ }
2667ptr<object> stack::pop() {
2668        node * n = head;
2669        head = n->next;
2670        ptr<object> x = std::move(n->value);
2671        delete n;
2672        return x;
2673}
2674void stack::clear() {
2675        for ( node * next = head; next; ) {
2676                node * crnt = next;
2677                next = crnt->next;
2678                delete crnt;
2679        }
2680        head = nullptr;
2681}
2682\end{lstlisting}
2683
2684
2685\begin{comment}
2686
2687\subsubsection{bench.h}
2688(\texttt{bench.hpp} is similar.)
2689
2690\lstinputlisting{evaluation/bench.h}
2691
2692\subsection{C}
2693
2694\subsubsection{c-stack.h} ~
2695
2696\lstinputlisting{evaluation/c-stack.h}
2697
2698\subsubsection{c-stack.c} ~
2699
2700\lstinputlisting{evaluation/c-stack.c}
2701
2702\subsubsection{c-pair.h} ~
2703
2704\lstinputlisting{evaluation/c-pair.h}
2705
2706\subsubsection{c-pair.c} ~
2707
2708\lstinputlisting{evaluation/c-pair.c}
2709
2710\subsubsection{c-print.h} ~
2711
2712\lstinputlisting{evaluation/c-print.h}
2713
2714\subsubsection{c-print.c} ~
2715
2716\lstinputlisting{evaluation/c-print.c}
2717
2718\subsubsection{c-bench.c} ~
2719
2720\lstinputlisting{evaluation/c-bench.c}
2721
2722\subsection{\CFA}
2723
2724\subsubsection{cfa-stack.h} ~
2725
2726\lstinputlisting{evaluation/cfa-stack.h}
2727
2728\subsubsection{cfa-stack.c} ~
2729
2730\lstinputlisting{evaluation/cfa-stack.c}
2731
2732\subsubsection{cfa-print.h} ~
2733
2734\lstinputlisting{evaluation/cfa-print.h}
2735
2736\subsubsection{cfa-print.c} ~
2737
2738\lstinputlisting{evaluation/cfa-print.c}
2739
2740\subsubsection{cfa-bench.c} ~
2741
2742\lstinputlisting{evaluation/cfa-bench.c}
2743
2744\subsection{\CC}
2745
2746\subsubsection{cpp-stack.hpp} ~
2747
2748\lstinputlisting[language=c++]{evaluation/cpp-stack.hpp}
2749
2750\subsubsection{cpp-print.hpp} ~
2751
2752\lstinputlisting[language=c++]{evaluation/cpp-print.hpp}
2753
2754\subsubsection{cpp-bench.cpp} ~
2755
2756\lstinputlisting[language=c++]{evaluation/cpp-bench.cpp}
2757
2758\subsection{\CCV}
2759
2760\subsubsection{object.hpp} ~
2761
2762\lstinputlisting[language=c++]{evaluation/object.hpp}
2763
2764\subsubsection{cpp-vstack.hpp} ~
2765
2766\lstinputlisting[language=c++]{evaluation/cpp-vstack.hpp}
2767
2768\subsubsection{cpp-vstack.cpp} ~
2769
2770\lstinputlisting[language=c++]{evaluation/cpp-vstack.cpp}
2771
2772\subsubsection{cpp-vprint.hpp} ~
2773
2774\lstinputlisting[language=c++]{evaluation/cpp-vprint.hpp}
2775
2776\subsubsection{cpp-vbench.cpp} ~
2777
2778\lstinputlisting[language=c++]{evaluation/cpp-vbench.cpp}
2779\end{comment}
2780
2781\end{document}
2782
2783% Local Variables: %
2784% tab-width: 4 %
2785% compile-command: "make" %
2786% End: %
Note: See TracBrowser for help on using the repository browser.