source: doc/papers/general/Paper.tex @ bcef6c8

aaron-thesisarm-ehcleanup-dtorsdeferred_resndemanglerenumforall-pointer-decayjacob/cs343-translationjenkins-sandboxnew-astnew-ast-unique-exprnew-envno_listpersistent-indexerresolv-newwith_gc
Last change on this file since bcef6c8 was 6c7c63c, checked in by Peter A. Buhr <pabuhr@…>, 4 years ago

finish changes to Declaration section

  • Property mode set to 100644
File size: 142.5 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}~\abbrevFont{al}}
95\newcommand*{\etal}{%
96        \@ifnextchar{.}{\protect\ETAL}%
97                {\protect\ETAL.\xspace}%
98}%
99\newcommand{\VIZ}{\abbrevFont{viz}}
100\newcommand*{\viz}{%
101        \@ifnextchar{.}{\VIZ}%
102                {\VIZ.\xspace}%
103}%
104\makeatother
105
106\newenvironment{cquote}{%
107        \list{}{\lstset{resetmargins=true,aboveskip=0pt,belowskip=0pt}\topsep=3pt\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        morekeywords=[2]{
123                _Atomic, coroutine, is_coroutine, is_monitor, is_thread, monitor, mutex, nomutex, or,
124                resume, suspend, thread, _Thread_local, waitfor, when, yield},
125        moredirectives={defined,include_next}%
126}
127
128\lstset{
129language=CFA,
130columns=fullflexible,
131basicstyle=\linespread{0.9}\sf,                                                 % reduce line spacing and use sanserif font
132stringstyle=\tt,                                                                                % use typewriter font
133tabsize=5,                                                                                              % N space tabbing
134xleftmargin=\parindentlnth,                                                             % indent code to paragraph indentation
135%mathescape=true,                                                                               % LaTeX math escape in CFA code $...$
136escapechar=\$,                                                                                  % LaTeX escape in CFA code
137keepspaces=true,                                                                                %
138showstringspaces=false,                                                                 % do not show spaces with cup
139showlines=true,                                                                                 % show blank lines at end of code
140aboveskip=4pt,                                                                                  % spacing above/below code block
141belowskip=3pt,
142% replace/adjust listing characters that look bad in sanserif
143literate={-}{\makebox[1ex][c]{\raisebox{0.4ex}{\rule{0.8ex}{0.1ex}}}}1 {^}{\raisebox{0.6ex}{$\scriptscriptstyle\land\,$}}1
144        {~}{\raisebox{0.3ex}{$\scriptstyle\sim\,$}}1 {@}{\small{@}}1 % {`}{\ttfamily\upshape\hspace*{-0.1ex}`}1
145        {<-}{$\leftarrow$}2 {=>}{$\Rightarrow$}2 {->}{\makebox[1ex][c]{\raisebox{0.4ex}{\rule{0.8ex}{0.075ex}}}\kern-0.2ex\textgreater}2,
146moredelim=**[is][\color{red}]{`}{`},
147}% lstset
148
149% inline code @...@
150\lstMakeShortInline@%
151
152\lstnewenvironment{cfa}[1][]
153{\lstset{#1}}
154{}
155\lstnewenvironment{C++}[1][]                            % use C++ style
156{\lstset{language=C++,moredelim=**[is][\protect\color{red}]{`}{`},#1}\lstset{#1}}
157{}
158
159
160\title{\protect\CFA : Adding Modern Programming Language Features to C}
161
162\author{Aaron Moss, Robert Schluntz, Peter Buhr}
163% \email{a3moss@uwaterloo.ca}
164% \email{rschlunt@uwaterloo.ca}
165% \email{pabuhr@uwaterloo.ca}
166% \affiliation{%
167%       \institution{University of Waterloo}
168%       \department{David R. Cheriton School of Computer Science}
169%       \streetaddress{Davis Centre, University of Waterloo}
170%       \city{Waterloo}
171%       \state{ON}
172%       \postcode{N2L 3G1}
173%       \country{Canada}
174% }
175
176%\terms{generic, tuple, variadic, types}
177%\keywords{generic types, tuple types, variadic types, polymorphic functions, C, Cforall}
178
179\begin{document}
180\maketitle
181
182
183\begin{abstract}
184The 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.
185This installation base and the programmers producing it represent a massive software-engineering investment spanning decades and likely to continue for decades more.
186Nevertheless, C, first standardized over thirty years ago, lacks many features that make programming in more modern languages safer and more productive.
187
188The 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.
189Prior 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.
190Specifically, \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 programmers.
191This paper presents a quick tour of \CFA features showing how their design avoids shortcomings of similar features in C and other C-like languages.
192Finally, experimental results are presented to validate several of the new features.
193\end{abstract}
194
195
196\section{Introduction}
197
198The 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.
199This installation base and the programmers producing it represent a massive software-engineering investment spanning decades and likely to continue for decades more.
200The 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.
201The top 3 rankings over the past 30 years are:
202\begin{center}
203\setlength{\tabcolsep}{10pt}
204\lstDeleteShortInline@%
205\begin{tabular}{@{}rccccccc@{}}
206                & 2017  & 2012  & 2007  & 2002  & 1997  & 1992  & 1987          \\ \hline
207Java    & 1             & 1             & 1             & 1             & 12    & -             & -                     \\
208\Textbf{C}      & \Textbf{2}& \Textbf{2}& \Textbf{2}& \Textbf{2}& \Textbf{1}& \Textbf{1}& \Textbf{1}    \\
209\CC             & 3             & 3             & 3             & 3             & 2             & 2             & 4                     \\
210\end{tabular}
211\lstMakeShortInline@%
212\end{center}
213Love it or hate it, C is extremely popular, highly used, and one of the few systems languages.
214In many cases, \CC is often used solely as a better C.
215Nonetheless, C, first standardized over thirty years ago, lacks many features that make programming in more modern languages safer and more productive.
216
217\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.
218The four key design goals for \CFA~\cite{Bilson03} are:
219(1) The behaviour of standard C code must remain the same when translated by a \CFA compiler as when translated by a C compiler;
220(2) Standard C code must be as fast and as small when translated by a \CFA compiler as when translated by a C compiler;
221(3) \CFA code must be at least as portable as standard C code;
222(4) Extensions introduced by \CFA must be translated in the most efficient way possible.
223These 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.
224\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.
225
226\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).
227Ultimately, a compiler is necessary for advanced features and optimal performance.
228
229This 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.
230Specifically, 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.
231The new constructs are empirically compared with both standard C and \CC; the results show the new design is comparable in performance.
232
233\section{Polymorphic Functions}
234
235\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}.
236
237\subsection{Name Overloading}
238
239C 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.
240\CFA extends the built-in operator overloading by allowing users to define overloads for any function, not just operators, and even any variable;
241Section~\ref{sec:libraries} includes a number of examples of how this overloading simplifies \CFA programming relative to C.
242Code 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.
243As an example:
244
245\begin{cfa}
246int max(int a, int b) { return a < b ? b : a; }  // (1)
247double max(double a, double b) { return a < b ? b : a; }  // (2)
248
249int max = INT_MAX;     // (3)
250double max = DBL_MAX;  // (4)
251
252max(7, -max);   $\C{// uses (1) and (3), by matching int from constant 7}$
253max(max, 3.14); $\C{// uses (2) and (4), by matching double from constant 3.14}$
254
255//max(max, -max);  $\C{// ERROR: ambiguous}$
256int m = max(max, -max); $\C{// uses (1) once and (3) twice, by matching return type}$
257\end{cfa}
258
259\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.
260The 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@.
261Ergonomic 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.
262
263% http://fanf.livejournal.com/144696.html
264% http://www.robertgamble.net/2012/01/c11-generic-selections.html
265% https://abissell.com/2014/01/16/c11s-_generic-keyword-macro-applications-and-performance-impacts/
266
267
268\subsection{\texorpdfstring{\LstKeywordStyle{forall} Functions}{forall Functions}}
269\label{sec:poly-fns}
270
271The 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):
272\begin{cfa}
273`forall( otype T )` T identity( T val ) { return val; }
274int forty_two = identity( 42 );                         $\C{// T is bound to int, forty\_two == 42}$
275\end{cfa}
276The @identity@ function above can be applied to any complete \newterm{object type} (or @otype@).
277The 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.
278The \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.
279If this extra information is not needed, \eg for a pointer, the type parameter can be declared as a \newterm{data type} (or @dtype@).
280
281In \CFA, the polymorphism runtime-cost is spread over each polymorphic call, due to passing more arguments to polymorphic functions;
282the experiments in Section~\ref{sec:eval} show this overhead is similar to \CC virtual-function calls.
283A design advantage is that, unlike \CC template-functions, \CFA polymorphic-functions are compatible with C \emph{separate compilation}, preventing compilation and code bloat.
284
285Since 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.
286For example, the function @twice@ can be defined using the \CFA syntax for operator overloading:
287\begin{cfa}
288forall( otype T `| { T ?+?(T, T); }` ) T twice( T x ) { return x + x; } $\C{// ? denotes operands}$
289int val = twice( twice( 3.7 ) );
290\end{cfa}
291which works for any type @T@ with a matching addition operator.
292The 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@.
293There 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.
294The first approach has a late conversion from @double@ to @int@ on the final assignment, while the second has an eager conversion to @int@.
295\CFA minimizes the number of conversions and their potential to lose information, so it selects the first approach, which corresponds with C-programmer intuition.
296
297Crucial to the design of a new programming language are the libraries to access thousands of external software features.
298Like \CC, \CFA inherits a massive compatible library-base, where other programming languages must rewrite or provide fragile inter-language communication with C.
299A simple example is leveraging the existing type-unsafe (@void *@) C @bsearch@ to binary search a sorted float array:
300\begin{cfa}
301void * bsearch( const void * key, const void * base, size_t nmemb, size_t size,
302                                int (* compar)( const void *, const void * ));
303
304int comp( const void * t1, const void * t2 ) { return *(double *)t1 < *(double *)t2 ? -1 :
305                                *(double *)t2 < *(double *)t1 ? 1 : 0; }
306
307double key = 5.0, vals[10] = { /* 10 sorted float values */ };
308double * val = (double *)bsearch( &key, vals, 10, sizeof(vals[0]), comp );      $\C{// search sorted array}$
309\end{cfa}
310which can be augmented simply with a generalized, type-safe, \CFA-overloaded wrappers:
311\begin{cfa}
312forall( otype T | { int ?<?( T, T ); } ) T * bsearch( T key, const T * arr, size_t size ) {
313        int comp( const void * t1, const void * t2 ) { /* as above with double changed to T */ }
314        return (T *)bsearch( &key, arr, size, sizeof(T), comp ); }
315
316forall( otype T | { int ?<?( T, T ); } ) unsigned int bsearch( T key, const T * arr, size_t size ) {
317        T * result = bsearch( key, arr, size ); $\C{// call first version}$
318        return result ? result - arr : size; }  $\C{// pointer subtraction includes sizeof(T)}$
319
320double * val = bsearch( 5.0, vals, 10 );        $\C{// selection based on return type}$
321int posn = bsearch( 5.0, vals, 10 );
322\end{cfa}
323The nested function @comp@ provides the hidden interface from typed \CFA to untyped (@void *@) C, plus the cast of the result.
324Providing a hidden @comp@ function in \CC is awkward as lambdas do not use C calling-conventions and template declarations cannot appear at block scope.
325As well, an alternate kind of return is made available: position versus pointer to found element.
326\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@.
327
328\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}).
329For example, it is possible to write a type-safe \CFA wrapper @malloc@ based on the C @malloc@:
330\begin{cfa}
331forall( dtype T | sized(T) ) T * malloc( void ) { return (T *)malloc( sizeof(T) ); }
332int * ip = malloc();                                            $\C{// select type and size from left-hand side}$
333double * dp = malloc();
334struct S {...} * sp = malloc();
335\end{cfa}
336where the return type supplies the type/size of the allocation, which is impossible in most type systems.
337
338Call-site inferencing and nested functions provide a localized form of inheritance.
339For example, the \CFA @qsort@ only sorts in ascending order using @<@.
340However, it is trivial to locally change this behaviour:
341\begin{cfa}
342forall( otype T | { int ?<?( T, T ); } ) void qsort( const T * arr, size_t size ) { /* use C qsort */ }
343{       int ?<?( double x, double y ) { return x `>` y; }       $\C{// locally override behaviour}$
344        qsort( vals, size );                                    $\C{// descending sort}$
345}
346\end{cfa}
347Within the block, the nested version of @?<?@ performs @?>?@ and this local version overrides the built-in @?<?@ so it is passed to @qsort@.
348Hence, programmers can easily form local environments, adding and modifying appropriate functions, to maximize reuse of other existing functions and types.
349
350
351\subsection{Traits}
352
353\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:
354\begin{cfa}
355trait `summable`( otype T ) {
356        void ?{}( T *, zero_t );                                $\C{// constructor from 0 literal}$
357        T ?+?( T, T );                                                  $\C{// assortment of additions}$
358        T ?+=?( T *, T );
359        T ++?( T * );
360        T ?++( T * ); };
361
362forall( otype T `| summable( T )` ) T sum( T a[$\,$], size_t size ) {  // use trait
363        `T` total = { `0` };                                    $\C{// instantiate T from 0 by calling its constructor}$
364        for ( unsigned int i = 0; i < size; i += 1 ) total `+=` a[i]; $\C{// select appropriate +}$
365        return total; }
366\end{cfa}
367
368In 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:
369\begin{cfa}
370trait otype( dtype T | sized(T) ) {  // sized is a pseudo-trait for types with known size and alignment
371        void ?{}( T * );                                                $\C{// default constructor}$
372        void ?{}( T *, T );                                             $\C{// copy constructor}$
373        void ?=?( T *, T );                                             $\C{// assignment operator}$
374        void ^?{}( T * ); };                                    $\C{// destructor}$
375\end{cfa}
376Given 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.
377
378In 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.
379Hence, trait names play no part in type equivalence;
380the names are simply macros for a list of polymorphic assertions, which are expanded at usage sites.
381Nevertheless, trait names form a logical subtype-hierarchy with @dtype@ at the top, where traits often contain overlapping assertions, \eg operator @+@.
382Traits are used like interfaces in Java or abstract base-classes in \CC, but without the nominal inheritance-relationships.
383Instead, 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.
384Hence, 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.
385(Nominal inheritance can be approximated with traits using marker variables or functions, as is done in Go.)
386
387% Nominal inheritance can be simulated with traits using marker variables or functions:
388% \begin{cfa}
389% trait nominal(otype T) {
390%     T is_nominal;
391% };
392% int is_nominal;                                                               $\C{// int now satisfies the nominal trait}$
393% \end{cfa}
394%
395% 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:
396% \begin{cfa}
397% trait pointer_like(otype Ptr, otype El) {
398%     lvalue El *?(Ptr);                                                $\C{// Ptr can be dereferenced into a modifiable value of type El}$
399% }
400% struct list {
401%     int value;
402%     list * next;                                                              $\C{// may omit "struct" on type names as in \CC}$
403% };
404% typedef list * list_iterator;
405%
406% lvalue int *?( list_iterator it ) { return it->value; }
407% \end{cfa}
408% 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@).
409% 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.
410
411
412\section{Generic Types}
413
414One of the known shortcomings of standard C is that it does not provide reusable type-safe abstractions for generic data structures and algorithms.
415Broadly speaking, there are three approaches to implement abstract data-structures in C.
416One approach is to write bespoke data-structures for each context in which they are needed.
417While 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.
418A 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.
419However, 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.
420A 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.
421Furthermore, writing and using preprocessor macros can be unnatural and inflexible.
422
423\CC, Java, and other languages use \newterm{generic types} to produce type-safe abstract data-types.
424\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.
425However, for known concrete parameters, the generic-type definition can be inlined, like \CC templates.
426
427A 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:
428\begin{cfa}
429forall( otype R, otype S ) struct pair {
430        R first;
431        S second;
432};
433forall( otype T ) T value( pair( const char *, T ) p ) { return p.second; }
434forall( dtype F, otype T ) T value_p( pair( F *, T * ) p ) { return * p.second; }
435
436pair( const char *, int ) p = { "magic", 42 };
437int magic = value( p );
438pair( void *, int * ) q = { 0, &p.second };
439magic = value_p( q );
440double d = 1.0;
441pair( double *, double * ) r = { &d, &d };
442d = value_p( r );
443\end{cfa}
444
445\CFA classifies generic types as either \newterm{concrete} or \newterm{dynamic}.
446Concrete types have a fixed memory layout regardless of type parameters, while dynamic types vary in memory layout depending on their type parameters.
447A type may have polymorphic parameters but still be concrete, called \newterm{dtype-static}.
448Polymorphic 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.
449
450\CFA generic types also allow checked argument-constraints.
451For example, the following declaration of a sorted set-type ensures the set key supports equality and relational comparison:
452\begin{cfa}
453forall( otype Key | { _Bool ?==?(Key, Key); _Bool ?<?(Key, Key); } ) struct sorted_set;
454\end{cfa}
455
456
457\subsection{Concrete Generic-Types}
458
459The \CFA translator template-expands concrete generic-types into new structure types, affording maximal inlining.
460To 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.
461A 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.
462For example, the concrete instantiation for @pair( const char *, int )@ is:
463\begin{cfa}
464struct _pair_conc1 {
465        const char * first;
466        int second;
467};
468\end{cfa}
469
470A concrete generic-type with dtype-static parameters is also expanded to a structure type, but this type is used for all matching instantiations.
471In 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:
472\begin{cfa}
473struct _pair_conc0 {
474        void * first;
475        void * second;
476};
477\end{cfa}
478
479
480\subsection{Dynamic Generic-Types}
481
482Though \CFA implements concrete generic-types efficiently, it also has a fully general system for dynamic generic types.
483As 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.
484Dynamic generic-types also have an \newterm{offset array} containing structure-member offsets.
485A dynamic generic-union needs no such offset array, as all members are at offset 0, but size and alignment are still necessary.
486Access 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.
487
488The offset arrays are statically generated where possible.
489If 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;
490if 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.
491As 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 )@.
492The offset array @_offsetof_pair@ is generated at the call site as @size_t _offsetof_pair[] = { offsetof(_pair_conc1, first), offsetof(_pair_conc1, second) }@.
493
494In some cases the offset arrays cannot be statically generated.
495For 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.
496\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.
497The \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.
498These 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).
499Results of these layout functions are cached so that they are only computed once per type per function. %, as in the example below for @pair@.
500Layout functions also allow generic types to be used in a function definition without reflecting them in the function signature.
501For 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.
502This function could acquire the layout for @set(T)@ by calling its layout function with the layout of @T@ implicitly passed into the function.
503
504Whether a type is concrete, dtype-static, or dynamic is decided solely on the @forall@'s type parameters.
505This 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.
506If 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.
507
508
509\subsection{Applications}
510\label{sec:generic-apps}
511
512The reuse of dtype-static structure instantiations enables useful programming patterns at zero runtime cost.
513The 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@:
514\begin{cfa}
515forall(dtype T) int lexcmp( pair( T *, T * ) * a, pair( T *, T * ) * b, int (* cmp)( T *, T * ) ) {
516        return cmp( a->first, b->first ) ? : cmp( a->second, b->second );
517}
518\end{cfa}
519Since @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.
520
521Another useful pattern enabled by reused dtype-static type instantiations is zero-cost \newterm{tag-structures}.
522Sometimes information is only used for type-checking and can be omitted at runtime, \eg:
523\begin{cfa}
524forall(dtype Unit) struct scalar { unsigned long value; };
525struct metres {};
526struct litres {};
527
528forall(dtype U) scalar(U) ?+?( scalar(U) a, scalar(U) b ) {
529        return (scalar(U)){ a.value + b.value };
530}
531scalar(metres) half_marathon = { 21093 };
532scalar(litres) swimming_pool = { 2500000 };
533scalar(metres) marathon = half_marathon + half_marathon;
534scalar(litres) two_pools = swimming_pool + swimming_pool;
535marathon + swimming_pool;                                       $\C{// compilation ERROR}$
536\end{cfa}
537@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 @?+?@.
538These implementations may even be separately compiled, unlike \CC template functions.
539However, the \CFA type-checker ensures matching types are used by all calls to @?+?@, preventing nonsensical computations like adding a length to a volume.
540
541
542\section{Tuples}
543\label{sec:tuples}
544
545In many languages, functions can return at most one value;
546however, many operations have multiple outcomes, some exceptional.
547Consider C's @div@ and @remquo@ functions, which return the quotient and remainder for a division of integer and float values, respectively.
548\begin{cfa}
549typedef struct { int quo, rem; } div_t;         $\C{// from include stdlib.h}$
550div_t div( int num, int den );
551double remquo( double num, double den, int * quo );
552div_t qr = div( 13, 5 );                                        $\C{// return quotient/remainder aggregate}$
553int q;
554double r = remquo( 13.5, 5.2, &q );                     $\C{// return remainder, alias quotient}$
555\end{cfa}
556@div@ aggregates the quotient/remainder in a structure, while @remquo@ aliases a parameter to an argument.
557Both approaches are awkward.
558Alternatively, a programming language can directly support returning multiple values, \eg in \CFA:
559\begin{cfa}
560[ int, int ] div( int num, int den );           $\C{// return two integers}$
561[ double, double ] div( double num, double den ); $\C{// return two doubles}$
562int q, r;                                                                       $\C{// overloaded variable names}$
563double q, r;
564[ q, r ] = div( 13, 5 );                                        $\C{// select appropriate div and q, r}$
565[ q, r ] = div( 13.5, 5.2 );                            $\C{// assign into tuple}$
566\end{cfa}
567Clearly, this approach is straightforward to understand and use;
568therefore, why do few programming languages support this obvious feature or provide it awkwardly?
569The answer is that there are complex consequences that cascade through multiple aspects of the language, especially the type-system.
570This section show these consequences and how \CFA handles them.
571
572
573\subsection{Tuple Expressions}
574
575The addition of multiple-return-value functions (MRVF) are useless without a syntax for accepting multiple values at the call-site.
576The simplest mechanism for capturing the return values is variable assignment, allowing the values to be retrieved directly.
577As 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}.
578
579However, 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:
580\begin{cfa}
581printf( "%d %d\n", div( 13, 5 ) );                      $\C{// return values seperated into arguments}$
582\end{cfa}
583Here, the values returned by @div@ are composed with the call to @printf@ by flattening the tuple into separate arguments.
584However, the \CFA type-system must support significantly more complex composition:
585\begin{cfa}
586[ int, int ] foo$\(_1\)$( int );                        $\C{// overloaded foo functions}$
587[ double ] foo$\(_2\)$( int );
588void bar( int, double, double );
589bar( foo( 3 ), foo( 3 ) );
590\end{cfa}
591The type-resolver only has the tuple return-types to resolve the call to @bar@ as the @foo@ parameters are identical, which involves unifying the possible @foo@ functions with @bar@'s parameter list.
592No combination of @foo@s are an exact match with @bar@'s parameters, so the resolver applies C conversions.
593The minimal cost is @bar( foo@$_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.
594
595
596\subsection{Tuple Variables}
597
598An important observation from function composition is that new variable names are not required to initialize parameters from an MRVF.
599\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:
600\begin{cfa}
601[ int, int ] qr = div( 13, 5 );                         $\C{// tuple-variable declaration and initialization}$
602[ double, double ] qr = div( 13.5, 5.2 );
603\end{cfa}
604where the tuple variable-name serves the same purpose as the parameter name(s).
605Tuple variables can be composed of any types, except for array types, since array sizes are generally unknown in C.
606
607One way to access the tuple-variable components is with assignment or composition:
608\begin{cfa}
609[ q, r ] = qr;                                                          $\C{// access tuple-variable components}$
610printf( "%d %d\n", qr );
611\end{cfa}
612\CFA also supports \newterm{tuple indexing} to access single components of a tuple expression:
613\begin{cfa}
614[int, int] * p = &qr;                                           $\C{// tuple pointer}$
615int rem = qr`.1`;                                                       $\C{// access remainder}$
616int quo = div( 13, 5 )`.0`;                                     $\C{// access quotient}$
617p`->0` = 5;                                                                     $\C{// change quotient}$
618bar( qr`.1`, qr );                                                      $\C{// pass remainder and quotient/remainder}$
619rem = [div( 13, 5 ), 42]`.0.1`;                         $\C{// access 2nd component of 1st component of tuple expression}$
620\end{cfa}
621
622
623\subsection{Flattening and Restructuring}
624
625In function call contexts, tuples support implicit flattening and restructuring conversions.
626Tuple flattening recursively expands a tuple into the list of its basic components.
627Tuple structuring packages a list of expressions into a value of tuple type, \eg:
628%\lstDeleteShortInline@%
629%\par\smallskip
630%\begin{tabular}{@{}l@{\hspace{1.5\parindent}}||@{\hspace{1.5\parindent}}l@{}}
631\begin{cfa}
632int f( int, int );
633int g( [int, int] );
634int h( int, [int, int] );
635[int, int] x;
636int y;
637f( x );                 $\C{// flatten}$
638g( y, 10 );             $\C{// structure}$
639h( x, y );              $\C{// flatten and structure}$
640\end{cfa}
641%\end{cfa}
642%&
643%\begin{cfa}
644%\end{tabular}
645%\smallskip\par\noindent
646%\lstMakeShortInline@%
647In the call to @f@, @x@ is implicitly flattened so the components of @x@ are passed as the two arguments.
648In 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@.
649Finally, 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]@.
650The 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.
651
652
653\subsection{Tuple Assignment}
654
655An assignment where the left side is a tuple type is called \newterm{tuple assignment}.
656There 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.
657%\lstDeleteShortInline@%
658%\par\smallskip
659%\begin{tabular}{@{}l@{\hspace{1.5\parindent}}||@{\hspace{1.5\parindent}}l@{}}
660\begin{cfa}
661int x = 10;
662double y = 3.5;
663[int, double] z;
664z = [x, y];                                                                     $\C{// multiple assignment}$
665[x, y] = z;                                                                     $\C{// multiple assignment}$
666z = 10;                                                                         $\C{// mass assignment}$
667[y, x] = 3.14;                                                          $\C{// mass assignment}$
668\end{cfa}
669%\end{cfa}
670%&
671%\begin{cfa}
672%\end{tabular}
673%\smallskip\par\noindent
674%\lstMakeShortInline@%
675Both kinds of tuple assignment have parallel semantics, so that each value on the left and right side is evaluated before any assignments occur.
676As 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]@.
677This 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.
678For example, @[y, x] = 3.14@ performs the assignments @y = 3.14@ and @x = 3.14@, yielding @y == 3.14@ and @x == 3@;
679whereas, C cascading assignment @y = x = 3.14@ performs the assignments @x = 3.14@ and @y = x@, yielding @3@ in @y@ and @x@.
680Finally, 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.
681This example shows mass, multiple, and cascading assignment used in one expression:
682\begin{cfa}
683void f( [int, int] );
684f( [x, y] = z = 1.5 );                                          $\C{// assignments in parameter list}$
685\end{cfa}
686
687
688\subsection{Member Access}
689
690It is also possible to access multiple fields from a single expression using a \newterm{member-access}.
691The result is a single tuple-valued expression whose type is the tuple of the types of the members, \eg:
692\begin{cfa}
693struct S { int x; double y; char * z; } s;
694s.[x, y, z] = 0;
695\end{cfa}
696Here, the mass assignment sets all members of @s@ to zero.
697Since 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).
698%\lstDeleteShortInline@%
699%\par\smallskip
700%\begin{tabular}{@{}l@{\hspace{1.5\parindent}}||@{\hspace{1.5\parindent}}l@{}}
701\begin{cfa}
702[int, int, long, double] x;
703void f( double, long );
704x.[0, 1] = x.[1, 0];                                            $\C{// rearrange: [x.0, x.1] = [x.1, x.0]}$
705f( x.[0, 3] );                                                          $\C{// drop: f(x.0, x.3)}$
706[int, int, int] y = x.[2, 0, 2];                        $\C{// duplicate: [y.0, y.1, y.2] = [x.2, x.0.x.2]}$
707\end{cfa}
708%\end{cfa}
709%&
710%\begin{cfa}
711%\end{tabular}
712%\smallskip\par\noindent
713%\lstMakeShortInline@%
714It is also possible for a member access to contain other member accesses, \eg:
715\begin{cfa}
716struct A { double i; int j; };
717struct B { int * k; short l; };
718struct C { int x; A y; B z; } v;
719v.[x, y.[i, j], z.k];                                           $\C{// [v.x, [v.y.i, v.y.j], v.z.k]}$
720\end{cfa}
721
722
723\begin{comment}
724\subsection{Casting}
725
726In C, the cast operator is used to explicitly convert between types.
727In \CFA, the cast operator has a secondary use as type ascription.
728That 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:
729\begin{cfa}
730int f();     // (1)
731double f()// (2)
732
733f();       // ambiguous - (1),(2) both equally viable
734(int)f()// choose (2)
735\end{cfa}
736
737Since casting is a fundamental operation in \CFA, casts should be given a meaningful interpretation in the context of tuples.
738Taking a look at standard C provides some guidance with respect to the way casts should work with tuples:
739\begin{cfa}
740int f();
741void g();
742
743(void)f()// (1)
744(int)g()// (2)
745\end{cfa}
746In C, (1) is a valid cast, which calls @f@ and discards its result.
747On the other hand, (2) is invalid, because @g@ does not produce a result, so requesting an @int@ to materialize from nothing is nonsensical.
748Generalizing 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.
749
750Formally, 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 $T_i$.
751Excess 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.
752This approach follows naturally from the way that a cast to @void@ works in C.
753
754For example, in
755\begin{cfa}
756[int, int, int] f();
757[int, [int, int], int] g();
758
759([int, double])f();           $\C{// (1)}$
760([int, int, int])g();         $\C{// (2)}$
761([void, [int, int]])g();      $\C{// (3)}$
762([int, int, int, int])g();    $\C{// (4)}$
763([int, [int, int, int]])g()$\C{// (5)}$
764\end{cfa}
765
766(1) discards the last element of the return value and converts the second element to @double@.
767Since @int@ is effectively a 1-element tuple, (2) discards the second component of the second element of the return value of @g@.
768If @g@ is free of side effects, this expression is equivalent to @[(int)(g().0), (int)(g().1.0), (int)(g().2)]@.
769Since @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)]@).
770
771Note 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.}.
772As such, (4) is invalid because the cast target type contains 4 components, while the source type contains only 3.
773Similarly, (5) is invalid because the cast @([int, int, int])(g().1)@ is invalid.
774That is, it is invalid to cast @[int, int]@ to @[int, int, int]@.
775\end{comment}
776
777
778\subsection{Polymorphism}
779
780Tuples also integrate with \CFA polymorphism as a kind of generic type.
781Due 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:
782\begin{cfa}
783forall(otype T, dtype U) void f( T x, U * y );
784f( [5, "hello"] );
785\end{cfa}
786where @[5, "hello"]@ is flattened, giving argument list @5, "hello"@, and @T@ binds to @int@ and @U@ binds to @const char@.
787Tuples, however, may contain polymorphic components.
788For example, a plus operator can be written to add two triples together.
789\begin{cfa}
790forall(otype T | { T ?+?( T, T ); }) [T, T, T] ?+?( [T, T, T] x, [T, T, T] y ) {
791        return [x.0 + y.0, x.1 + y.1, x.2 + y.2];
792}
793[int, int, int] x;
794int i1, i2, i3;
795[i1, i2, i3] = x + ([10, 20, 30]);
796\end{cfa}
797
798Flattening and restructuring conversions are also applied to tuple types in polymorphic type assertions.
799\begin{cfa}
800int f( [int, double], double );
801forall(otype T, otype U | { T f( T, U, U ); }) void g( T, U );
802g( 5, 10.21 );
803\end{cfa}
804Hence, function parameter and return lists are flattened for the purposes of type unification allowing the example to pass expression resolution.
805This relaxation is possible by extending the thunk scheme described by Bilson~\cite{Bilson03}.
806Whenever 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:
807\begin{cfa}
808int _thunk( int _p0, double _p1, double _p2 ) { return f( [_p0, _p1], _p2 ); }
809\end{cfa}
810so the thunk provides flattening and structuring conversions to inferred functions, improving the compatibility of tuples and polymorphism.
811These thunks take advantage of gcc C nested-functions to produce closures that have the usual function-pointer signature.
812
813
814\subsection{Variadic Tuples}
815\label{sec:variadic-tuples}
816
817To define variadic functions, \CFA adds a new kind of type parameter, @ttype@ (tuple type).
818Matching against a @ttype@ parameter consumes all remaining argument components and packages them into a tuple, binding to the resulting tuple of types.
819In 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.
820As such, @ttype@ variables are also called \newterm{argument packs}.
821
822Like variadic templates, the main way to manipulate @ttype@ polymorphic functions is via recursion.
823Since nothing is known about a parameter pack by default, assertion parameters are key to doing anything meaningful.
824Unlike variadic templates, @ttype@ polymorphic functions can be separately compiled.
825For example, a generalized @sum@ function written using @ttype@:
826\begin{cfa}
827int sum$\(_0\)$() { return 0; }
828forall(ttype Params | { int sum( Params ); } ) int sum$\(_1\)$( int x, Params rest ) {
829        return x + sum( rest );
830}
831sum( 10, 20, 30 );
832\end{cfa}
833Since @sum@\(_0\) does not accept any arguments, it is not a valid candidate function for the call @sum(10, 20, 30)@.
834In 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]@.
835The process continues unitl @Params@ is bound to @[]@, requiring an assertion @int sum()@, which matches @sum@\(_0\) and terminates the recursion.
836Effectively, 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))@.
837
838It is reasonable to take the @sum@ function a step further to enforce a minimum number of arguments:
839\begin{cfa}
840int sum( int x, int y ) { return x + y; }
841forall(ttype Params | { int sum( int, Params ); } ) int sum( int x, int y, Params rest ) {
842        return sum( x + y, rest );
843}
844\end{cfa}
845One more step permits the summation of any summable type with all arguments of the same type:
846\begin{cfa}
847trait summable(otype T) {
848        T ?+?( T, T );
849};
850forall(otype R | summable( R ) ) R sum( R x, R y ) {
851        return x + y;
852}
853forall(otype R, ttype Params | summable(R) | { R sum(R, Params); } ) R sum(R x, R y, Params rest) {
854        return sum( x + y, rest );
855}
856\end{cfa}
857Unlike C variadic functions, it is unnecessary to hard code the number and expected types.
858Furthermore, this code is extendable for any user-defined type with a @?+?@ operator.
859Summing arbitrary heterogeneous lists is possible with similar code by adding the appropriate type variables and addition operators.
860
861It is also possible to write a type-safe variadic print function to replace @printf@:
862\begin{cfa}
863struct S { int x, y; };
864forall(otype T, ttype Params | { void print(T); void print(Params); }) void print(T arg, Params rest) {
865        print(arg);  print(rest);
866}
867void print( char * x ) { printf( "%s", x ); }
868void print( int x ) { printf( "%d", x ); }
869void print( S s ) { print( "{ ", s.x, ",", s.y, " }" ); }
870print( "s = ", (S){ 1, 2 }, "\n" );
871\end{cfa}
872This example showcases a variadic-template-like decomposition of the provided argument list.
873The individual @print@ functions allow printing a single element of a type.
874The polymorphic @print@ allows printing any list of types, where as each individual type has a @print@ function.
875The individual print functions can be used to build up more complicated @print@ functions, such as @S@, which cannot be done with @printf@ in C.
876
877Finally, it is possible to use @ttype@ polymorphism to provide arbitrary argument forwarding functions.
878For example, it is possible to write @new@ as a library function:
879\begin{cfa}
880forall( otype R, otype S ) void ?{}( pair(R, S) *, R, S );
881forall( dtype T, ttype Params | sized(T) | { void ?{}( T *, Params ); } ) T * new( Params p ) {
882        return ((T *)malloc()){ p };                    $\C{// construct into result of malloc}$
883}
884pair( int, char ) * x = new( 42, '!' );
885\end{cfa}
886The @new@ function provides the combination of type-safe @malloc@ with a \CFA constructor call, making it impossible to forget constructing dynamically allocated objects.
887This function provides the type-safety of @new@ in \CC, without the need to specify the allocated type again, thanks to return-type inference.
888
889
890\subsection{Implementation}
891
892Tuples are implemented in the \CFA translator via a transformation into \newterm{generic types}.
893For each $N$, the first time an $N$-tuple is seen in a scope a generic type with $N$ type parameters is generated, \eg:
894\begin{cfa}
895[int, int] f() {
896        [double, double] x;
897        [int, double, int] y;
898}
899\end{cfa}
900is transformed into:
901\begin{cfa}
902forall(dtype T0, dtype T1 | sized(T0) | sized(T1)) struct _tuple2 {
903        T0 field_0;                                                             $\C{// generated before the first 2-tuple}$
904        T1 field_1;
905};
906_tuple2(int, int) f() {
907        _tuple2(double, double) x;
908        forall(dtype T0, dtype T1, dtype T2 | sized(T0) | sized(T1) | sized(T2)) struct _tuple3 {
909                T0 field_0;                                                     $\C{// generated before the first 3-tuple}$
910                T1 field_1;
911                T2 field_2;
912        };
913        _tuple3(int, double, int) y;
914}
915\end{cfa}
916\begin{sloppypar}
917Tuple expressions are then simply converted directly into compound literals, \eg @[5, 'x', 1.24]@ becomes @(_tuple3(int, char, double)){ 5, 'x', 1.24 }@.
918\end{sloppypar}
919
920\begin{comment}
921Since tuples are essentially structures, tuple indexing expressions are just field accesses:
922\begin{cfa}
923void f(int, [double, char]);
924[int, double] x;
925
926x.0+x.1;
927printf("%d %g\n", x);
928f(x, 'z');
929\end{cfa}
930Is transformed into:
931\begin{cfa}
932void f(int, _tuple2(double, char));
933_tuple2(int, double) x;
934
935x.field_0+x.field_1;
936printf("%d %g\n", x.field_0, x.field_1);
937f(x.field_0, (_tuple2){ x.field_1, 'z' });
938\end{cfa}
939Note that due to flattening, @x@ used in the argument position is converted into the list of its fields.
940In the call to @f@, the second and third argument components are structured into a tuple argument.
941Similarly, tuple member expressions are recursively expanded into a list of member access expressions.
942
943Expressions that may contain side effects are made into \newterm{unique expressions} before being expanded by the flattening conversion.
944Each unique expression is assigned an identifier and is guaranteed to be executed exactly once:
945\begin{cfa}
946void g(int, double);
947[int, double] h();
948g(h());
949\end{cfa}
950Internally, this expression is converted to two variables and an expression:
951\begin{cfa}
952void g(int, double);
953[int, double] h();
954
955_Bool _unq0_finished_ = 0;
956[int, double] _unq0;
957g(
958        (_unq0_finished_ ? _unq0 : (_unq0 = f(), _unq0_finished_ = 1, _unq0)).0,
959        (_unq0_finished_ ? _unq0 : (_unq0 = f(), _unq0_finished_ = 1, _unq0)).1,
960);
961\end{cfa}
962Since argument evaluation order is not specified by the C programming language, this scheme is built to work regardless of evaluation order.
963The first time a unique expression is executed, the actual expression is evaluated and the accompanying boolean is set to true.
964Every subsequent evaluation of the unique expression then results in an access to the stored result of the actual expression.
965Tuple member expressions also take advantage of unique expressions in the case of possible impurity.
966
967Currently, the \CFA translator has a very broad, imprecise definition of impurity, where any function call is assumed to be impure.
968This 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.
969
970The various kinds of tuple assignment, constructors, and destructors generate GNU C statement expressions.
971A 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.
972The 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.
973However, 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.
974\end{comment}
975
976
977\section{Control Structures}
978
979\CFA identifies missing and problematic control structures in C, and extends and modifies these control structures to increase functionality and safety.
980
981
982\subsection{\texorpdfstring{Labelled \LstKeywordStyle{continue} / \LstKeywordStyle{break}}{Labelled continue / break}}
983
984While C provides @continue@ and @break@ statements for altering control flow, both are restricted to one level of nesting for a particular control structure.
985Unfortunately, this restriction forces programmers to use @goto@ to achieve the equivalent control-flow for more than one level of nesting.
986To 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.
987For both @continue@ and @break@, the target label must be directly associated with a @for@, @while@ or @do@ statement;
988for @break@, the target label can also be associated with a @switch@, @if@ or compound (@{}@) statement.
989Figure~\ref{f:MultiLevelExit} shows @continue@ and @break@ indicating the specific control structure, and the corresponding C program using only @goto@ and labels.
990The innermost loop has 7 exit points, which cause continuation or termination of one or more of the 7 nested control-structures.
991
992\begin{figure}
993\lstDeleteShortInline@%
994\begin{tabular}{@{\hspace{\parindentlnth}}l@{\hspace{\parindentlnth}}l@{\hspace{\parindentlnth}}l@{}}
995\multicolumn{1}{@{\hspace{\parindentlnth}}c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}   & \multicolumn{1}{@{\hspace{\parindentlnth}}c}{\textbf{C}}      \\
996\begin{cfa}
997`LC:` {
998        ... $declarations$ ...
999        `LS:` switch ( ... ) {
1000          case 3:
1001                `LIF:` if ( ... ) {
1002                        `LF:` for ( ... ) {
1003                                `LW:` while ( ... ) {
1004                                        ... break `LC`; ...
1005                                        ... break `LS`; ...
1006                                        ... break `LIF`; ...
1007                                        ... continue `LF;` ...
1008                                        ... break `LF`; ...
1009                                        ... continue `LW`; ...
1010                                        ... break `LW`; ...
1011                                } // while
1012                        } // for
1013                } else {
1014                        ... break `LIF`; ...
1015                } // if
1016        } // switch
1017} // compound
1018\end{cfa}
1019&
1020\begin{cfa}
1021{
1022        ... $declarations$ ...
1023        switch ( ... ) {
1024          case 3:
1025                if ( ... ) {
1026                        for ( ... ) {
1027                                while ( ... ) {
1028                                        ... goto `LC`; ...
1029                                        ... goto `LS`; ...
1030                                        ... goto `LIF`; ...
1031                                        ... goto `LFC`; ...
1032                                        ... goto `LFB`; ...
1033                                        ... goto `LWC`; ...
1034                                        ... goto `LWB`; ...
1035                                  `LWC`: ; } `LWB:` ;
1036                          `LFC:` ; } `LFB:` ;
1037                } else {
1038                        ... goto `LIF`; ...
1039                } `L3:` ;
1040        } `LS:` ;
1041} `LC:` ;
1042\end{cfa}
1043&
1044\begin{cfa}
1045
1046
1047
1048
1049
1050
1051
1052// terminate compound
1053// terminate switch
1054// terminate if
1055// continue loop
1056// terminate loop
1057// continue loop
1058// terminate loop
1059
1060
1061
1062// terminate if
1063
1064
1065
1066\end{cfa}
1067\end{tabular}
1068\lstMakeShortInline@%
1069\caption{Multi-level Exit}
1070\label{f:MultiLevelExit}
1071\end{figure}
1072
1073Both labelled @continue@ and @break@ are a @goto@ restricted in the following ways:
1074\begin{itemize}[topsep=3pt,itemsep=2pt,parsep=0pt]
1075\item
1076They cannot create a loop, which means only the looping constructs cause looping.
1077This restriction means all situations resulting in repeated execution are clearly delineated.
1078\item
1079They cannot branch into a control structure.
1080This restriction prevents missing declarations and/or initializations at the start of a control structure resulting in undefined behaviour.
1081\end{itemize}
1082The 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.
1083Furthermore, 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.
1084With @goto@, the label is at the end of the control structure, which fails to convey this important clue early enough to the reader.
1085Finally, using an explicit target for the transfer instead of an implicit target allows new constructs to be added or removed without affecting existing constructs.
1086The implicit targets of the current @continue@ and @break@, \ie the closest enclosing loop or @switch@, change as certain constructs are added or removed.
1087
1088
1089\subsection{\texorpdfstring{Enhanced \LstKeywordStyle{switch} Statement}{Enhanced switch Statement}}
1090
1091There 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.
1092
1093C 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.}.
1094\CFA provides a shorthand for a non-contiguous list:
1095\begin{cquote}
1096\lstDeleteShortInline@%
1097\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1098\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1099\begin{cfa}
1100case 2, 10, 34, 42:
1101\end{cfa}
1102&
1103\begin{cfa}
1104case 2: case 10: case 34: case 42:
1105\end{cfa}
1106\end{tabular}
1107\lstMakeShortInline@%
1108\end{cquote}
1109for a contiguous list:\footnote{gcc provides the same mechanism with awkward syntax, \lstinline@2 ... 42@, where spaces are required around the ellipse.}
1110\begin{cquote}
1111\lstDeleteShortInline@%
1112\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1113\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1114\begin{cfa}
1115case 2~42:
1116\end{cfa}
1117&
1118\begin{cfa}
1119case 2: case 3: ... case 41: case 42:
1120\end{cfa}
1121\end{tabular}
1122\lstMakeShortInline@%
1123\end{cquote}
1124and a combination:
1125\begin{cfa}
1126case -12~-4, -1~5, 14~21, 34~42:
1127\end{cfa}
1128
1129C allows placement of @case@ clauses \emph{within} statements nested in the @switch@ body (see Duff's device~\cite{Duff83});
1130\begin{cfa}
1131switch ( i ) {
1132  case 0:
1133        for ( int i = 0; i < 10; i += 1 ) {
1134                ...
1135  `case 1:`             // no initialization of loop index
1136                ...
1137        }
1138}
1139\end{cfa}
1140\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.
1141
1142C allows placement of declaration within the @switch@ body and unreachable code at the start, resulting in undefined behaviour:
1143\begin{cfa}
1144switch ( x ) {
1145        `int y = 1;`                            $\C{// unreachable initialization}$
1146        `x = 7;`                                        $\C{// unreachable code without label/branch}$
1147  case 0:
1148        ...
1149        `int z = 0;`                            $\C{// unreachable initialization, cannot appear after case}$
1150        z = 2;
1151  case 1:
1152        `x = z;`                                        $\C{// without fall through, z is undefined}$
1153}
1154\end{cfa}
1155\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.
1156\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.
1157
1158C @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};
1159@case@ clauses are made disjoint by the @break@ statement.
1160While 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.
1161For backwards compatibility, \CFA provides a \emph{new} control structure, @choose@, which mimics @switch@, but reverses the meaning of fall through:
1162\begin{cquote}
1163\lstDeleteShortInline@%
1164\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1165\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1166\begin{cfa}
1167`choose` ( day ) {
1168  case Mon~Thu:  // program
1169
1170  case Fri:  // program
1171        wallet += pay;
1172        `fallthrough;`
1173  case Sat:  // party
1174        wallet -= party;
1175
1176  case Sun:  // rest
1177
1178  default:  // error
1179}
1180\end{cfa}
1181&
1182\begin{cfa}
1183switch ( day ) {
1184  case Mon: case Tue: case Wed: case Thu:  // program
1185        `break;`
1186  case Fri:  // program
1187        wallet += pay;
1188
1189  case Sat:  // party
1190        wallet -= party;
1191        `break;`
1192  case Sun:  // rest
1193        `break;`
1194  default:  // error
1195}
1196\end{cfa}
1197\end{tabular}
1198\lstMakeShortInline@%
1199\end{cquote}
1200Collectively, these enhancements reduce programmer burden and increase readability and safety.
1201
1202\begin{comment}
1203Forgotten @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.
1204\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.
1205Since 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.
1206Where 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:
1207
1208\begin{cfa}
1209choose( i ) {
1210        case 2:
1211                printf("even ");
1212                fallthrough;
1213        case 3: case 5: case 7:
1214                printf("small prime\n");
1215        case 4,6,8,9:
1216                printf("small composite\n");
1217        case 13~19:
1218                printf("teen\n");
1219        default:
1220                printf("something else\n");
1221}
1222\end{cfa}
1223\end{comment}
1224
1225
1226\subsection{\texorpdfstring{\LstKeywordStyle{with} Clause / Statement}{with Clause / Statement}}
1227\label{s:WithClauseStatement}
1228
1229Grouping 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:
1230\begin{cfa}
1231struct S {                                                                      $\C{// aggregate}$
1232        char c;                                                                 $\C{// fields}$
1233        int i;
1234        double d;
1235};
1236S s, as[10];
1237\end{cfa}
1238However, routines manipulating aggregates must repeat the aggregate name to access its containing fields:
1239\begin{cfa}
1240void f( S s ) {
1241        `s.`c; `s.`i; `s.`d;                                    $\C{// access containing fields}$
1242}
1243\end{cfa}
1244which extends to multiple levels of qualification for nested aggregates.
1245A similar situation occurs in object-oriented programming, \eg \CC:
1246\begin{C++}
1247class C {
1248        char c;                                                                 $\C{// fields}$
1249        int i;
1250        double d;
1251        int f() {                                                               $\C{// implicit ``this'' aggregate}$
1252                `this->`c; `this->`i; `this->`d;        $\C{// access containing fields}$
1253        }
1254}
1255\end{C++}
1256Object-oriented nesting of member routines in a \lstinline[language=C++]@class@ allows eliding \lstinline[language=C++]@this->@ because of lexical scoping.
1257However, for other aggregate parameters, qualification is necessary:
1258\begin{cfa}
1259struct T { double m, n; };
1260int C::f( T & t ) {                                                     $\C{// multiple aggregate parameters}$
1261        c; i; d;                                                                $\C{\color{red}// this-\textgreater.c, this-\textgreater.i, this-\textgreater.d}$
1262        `t.`m; `t.`n;                                                   $\C{// must qualify}$
1263}
1264\end{cfa}
1265
1266To 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.
1267Hence, the qualified fields become variables with the side-effect that it is easier to optimizing field references in a block.
1268\begin{cfa}
1269void f( S & this ) `with ( this )` {            $\C{// with statement}$
1270        c; i; d;                                                                $\C{\color{red}// this.c, this.i, this.d}$
1271}
1272\end{cfa}
1273with the generality of opening multiple aggregate-parameters:
1274\begin{cfa}
1275int f( S & s, T & t ) `with ( s, t )` {         $\C{// multiple aggregate parameters}$
1276        c; i; d;                                                                $\C{\color{red}// s.c, s.i, s.d}$
1277        m; n;                                                                   $\C{\color{red}// t.m, t.n}$
1278}
1279\end{cfa}
1280
1281In detail, the @with@ statement has the form:
1282\begin{cfa}
1283$\emph{with-statement}$:
1284        'with' '(' $\emph{expression-list}$ ')' $\emph{compound-statement}$
1285\end{cfa}
1286and may appear as the body of a routine or nested within a routine body.
1287Each expression in the expression-list provides a type and object.
1288The type must be an aggregate type.
1289(Enumerations are already opened.)
1290The object is the implicit qualifier for the open structure-fields.
1291
1292All expressions in the expression list are open in parallel within the compound statement.
1293This semantic is different from Pascal, which nests the openings from left to right.
1294The difference between parallel and nesting occurs for fields with the same name and type:
1295\begin{cfa}
1296struct S { int `i`; int j; double m; } s, w;
1297struct T { int `i`; int k; int m; } t, w;
1298with ( s, t ) {
1299        j + k;                                                                  $\C{// unambiguous, s.j + t.k}$
1300        m = 5.0;                                                                $\C{// unambiguous, t.m = 5.0}$
1301        m = 1;                                                                  $\C{// unambiguous, s.m = 1}$
1302        int a = m;                                                              $\C{// unambiguous, a = s.i }$
1303        double b = m;                                                   $\C{// unambiguous, b = t.m}$
1304        int c = s.i + t.i;                                              $\C{// unambiguous, qualification}$
1305        (double)m;                                                              $\C{// unambiguous, cast}$
1306}
1307\end{cfa}
1308For parallel semantics, both @s.i@ and @t.i@ are visible, so @i@ is ambiguous without qualification;
1309for nested semantics, @t.i@ hides @s.i@, so @i@ implies @t.i@.
1310\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.
1311Qualification or a cast is used to disambiguate.
1312
1313There is an interesting problem between parameters and the routine @with@, \eg:
1314\begin{cfa}
1315void ?{}( S & s, int i ) with ( s ) {           $\C{// constructor}$
1316        `s.i = i;`  j = 3;  m = 5.5;                    $\C{// initialize fields}$
1317}
1318\end{cfa}
1319Here, 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@.
1320To solve this problem, parameters are treated like an initialized aggregate:
1321\begin{cfa}
1322struct Params {
1323        S & s;
1324        int i;
1325} params;
1326\end{cfa}
1327and implicitly opened \emph{after} a routine open, to give them higher priority:
1328\begin{cfa}
1329void ?{}( S & s, int i ) with ( s ) `with( $\emph{\color{red}params}$ )` {
1330        s.i = i; j = 3; m = 5.5;
1331}
1332\end{cfa}
1333Finally, a cast may be used to disambiguate among overload variables in a @with@ expression:
1334\begin{cfa}
1335with ( w ) { ... }                                                      $\C{// ambiguous, same name and no context}$
1336with ( (S)w ) { ... }                                           $\C{// unambiguous, cast}$
1337\end{cfa}
1338and @with@ expressions may be complex expressions with type reference (see Section~\ref{s:References}) to aggregate:
1339\begin{cfa}
1340struct S { int i, j; } sv;
1341with ( sv ) {                                                           $\C{implicit reference}$
1342        S & sr = sv;
1343        with ( sr ) {                                                   $\C{explicit reference}$
1344                S * sp = &sv;
1345                with ( *sp ) {                                          $\C{computed reference}$
1346                        i = 3; j = 4;                                   $\C{\color{red}// sp-{\textgreater}i, sp-{\textgreater}j}$
1347                }
1348                i = 2; j = 3;                                           $\C{\color{red}// sr.i, sr.j}$
1349        }
1350        i = 1; j = 2;                                                   $\C{\color{red}// sv.i, sv.j}$
1351}
1352\end{cfa}
1353
1354
1355\subsection{Exception Handling}
1356
1357\CFA provides two forms of exception handling: \newterm{resumption} (fix-up) and \newterm{recovery}.
1358Both mechanisms provide dynamic call to a handler using dynamic name-lookup, where fix-up has dynamic return and recovery has static return from the handler.
1359\begin{cquote}
1360\lstDeleteShortInline@%
1361\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1362\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{Resumption}}       & \multicolumn{1}{c}{\textbf{Recovery}} \\
1363\begin{cfa}
1364_Exception E { int fix; };
1365void f() {
1366        E e;
1367        ... _Resume e;
1368        ... e.fix // control returns here after handler
1369try {
1370        f();
1371} catchResume( E e ) {
1372        ... e.fix = ...; // return correction to raise
1373} // dynamic return to _Resume
1374\end{cfa}
1375&
1376\begin{cfa}
1377_Exception E {};
1378void f() {
1379
1380        ... _Throw E{};
1381        // control does NOT return here after handler
1382try {
1383        f();
1384} catch( E e ) {
1385        ... // recover and continue
1386} // static return to next statement
1387\end{cfa}
1388\end{tabular}
1389\lstMakeShortInline@%
1390\end{cquote}
1391
1392
1393\section{Declarations}
1394
1395Declarations in C have weaknesses and omissions.
1396\CFA attempts to correct and add to C declarations, while ensuring \CFA subjectively ``feels like'' C.
1397An important part of this subjective feel is maintaining C's syntax and procedural paradigm, as opposed to functional and object-oriented approaches in other systems languages such as \CC and Rust.
1398Maintaining the C approach means that C coding-patterns remain not only useable but idiomatic in \CFA, reducing the mental burden of retraining C programmers and switching between C and \CFA development.
1399Nevertheless, some features from other approaches are undeniably convenient;
1400\CFA attempts to adapt these features to the C paradigm.
1401
1402
1403\subsection{Alternative Declaration Syntax}
1404
1405C declaration syntax is notoriously confusing and error prone.
1406For example, many C programmers are confused by a declaration as simple as:
1407\begin{cquote}
1408\lstDeleteShortInline@%
1409\begin{tabular}{@{}ll@{}}
1410\begin{cfa}
1411int * x[5]
1412\end{cfa}
1413&
1414\raisebox{-0.75\totalheight}{\input{Cdecl}}
1415\end{tabular}
1416\lstMakeShortInline@%
1417\end{cquote}
1418Is this an array of 5 pointers to integers or a pointer to an array of 5 integers?
1419If there is any doubt, it implies productivity and safety issues even for basic programs.
1420Another 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.
1421For example, a routine returning a pointer to an array of integers is defined and used in the following way:
1422\begin{cfa}
1423int `(*`f`())[`5`]` {...};                              $\C{// definition}$
1424 ... `(*`f`())[`3`]` += 1;                              $\C{// usage}$
1425\end{cfa}
1426Essentially, the return type is wrapped around the routine name in successive layers (like an onion).
1427While attempting to make the two contexts consistent is a laudable goal, it has not worked out in practice.
1428
1429\CFA provides its own type, variable and routine declarations, using a different syntax.
1430The new declarations place qualifiers to the left of the base type, while C declarations place qualifiers to the right.
1431The qualifiers have the same meaning but are ordered left to right to specify a variable's type.
1432\begin{cquote}
1433\lstDeleteShortInline@%
1434\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{\hspace{\parindentlnth}}l@{}}
1435\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1436\begin{cfa}
1437`[5] *` int x1;
1438`* [5]` int x2;
1439`[* [5] int]` f( int p );
1440\end{cfa}
1441&
1442\begin{cfa}
1443int `*` x1 `[5]`;
1444int `(*`x2`)[5]`;
1445`int (*`f( int p )`)[5]`;
1446\end{cfa}
1447&
1448\begin{cfa}
1449// array of 5 pointers to int
1450// pointer to an array of 5 int
1451// function returning pointer to an array of 5 int and taking an int
1452\end{cfa}
1453\end{tabular}
1454\lstMakeShortInline@%
1455\end{cquote}
1456The only exception is bit field specification, which always appear to the right of the base type.
1457% 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.
1458However, unlike C, \CFA type declaration tokens are distributed across all variables in the declaration list.
1459For instance, variables @x@ and @y@ of type pointer to integer are defined in \CFA as follows:
1460\begin{cquote}
1461\lstDeleteShortInline@%
1462\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1463\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1464\begin{cfa}
1465`*` int x, y;
1466int y;
1467\end{cfa}
1468&
1469\begin{cfa}
1470int `*`x, `*`y;
1471
1472\end{cfa}
1473\end{tabular}
1474\lstMakeShortInline@%
1475\end{cquote}
1476The downside of the \CFA semantics is the need to separate regular and pointer declarations.
1477
1478\begin{comment}
1479Other examples are:
1480\begin{cquote}
1481\lstDeleteShortInline@%
1482\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{\hspace{\parindentlnth}}l@{}}
1483\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{C}}      \\
1484\begin{cfa}
1485[ 5 ] int z;
1486[ 5 ] * char w;
1487* [ 5 ] double v;
1488struct s {
1489        int f0:3;
1490        * int f1;
1491        [ 5 ] * int f2;
1492};
1493\end{cfa}
1494&
1495\begin{cfa}
1496int z[ 5 ];
1497char * w[ 5 ];
1498double (* v)[ 5 ];
1499struct s {
1500        int f0:3;
1501        int * f1;
1502        int * f2[ 5 ]
1503};
1504\end{cfa}
1505&
1506\begin{cfa}
1507// array of 5 integers
1508// array of 5 pointers to char
1509// pointer to array of 5 doubles
1510
1511// common bit field syntax
1512
1513
1514
1515\end{cfa}
1516\end{tabular}
1517\lstMakeShortInline@%
1518\end{cquote}
1519\end{comment}
1520
1521All specifiers (@extern@, @static@, \etc) and qualifiers (@const@, @volatile@, \etc) are used in the normal way with the new declarations and also appear left to right, \eg:
1522\begin{cquote}
1523\lstDeleteShortInline@%
1524\begin{tabular}{@{}l@{\hspace{1em}}l@{\hspace{1em}}l@{}}
1525\multicolumn{1}{c@{\hspace{1em}}}{\textbf{\CFA}}        & \multicolumn{1}{c@{\hspace{1em}}}{\textbf{C}} \\
1526\begin{cfa}
1527extern const * const int x;
1528static const * [ 5 ] const int y;
1529\end{cfa}
1530&
1531\begin{cfa}
1532int extern const * const x;
1533static const int (* const y)[ 5 ]
1534\end{cfa}
1535&
1536\begin{cfa}
1537// external const pointer to const int
1538// internal const pointer to array of 5 const int
1539\end{cfa}
1540\end{tabular}
1541\lstMakeShortInline@%
1542\end{cquote}
1543Specifiers must appear at the start of a \CFA routine declaration\footnote{\label{StorageClassSpecifier}.
1544The 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}}.
1545
1546The new declaration syntax can be used in other contexts where types are required, \eg casts and the pseudo-routine @sizeof@:
1547\begin{cquote}
1548\lstDeleteShortInline@%
1549\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1550\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1551\begin{cfa}
1552y = (* int)x;
1553i = sizeof([ 5 ] * int);
1554\end{cfa}
1555&
1556\begin{cfa}
1557y = (int *)x;
1558i = sizeof(int * [ 5 ]);
1559\end{cfa}
1560\end{tabular}
1561\lstMakeShortInline@%
1562\end{cquote}
1563
1564The syntax of the new routine prototype declaration follows directly from the new routine definition syntax;
1565as well, parameter names are optional, \eg:
1566\begin{cfa}
1567[ int x ] f ( /* void */ );                             $\C{// returning int with no parameters}$
1568[ int x ] f (...);                                              $\C{// returning int with unknown parameters}$
1569[ * int ] g ( int y );                                  $\C{// returning pointer to int with int parameter}$
1570[ void ] h ( int, char );                               $\C{// returning no result with int and char parameters}$
1571[ * int, int ] j ( int );                               $\C{// returning pointer to int and int, with int parameter}$
1572\end{cfa}
1573This syntax allows a prototype declaration to be created by cutting and pasting source text from the routine definition header (or vice versa).
1574Like 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:
1575\begin{cquote}
1576\lstDeleteShortInline@%
1577\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1578\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1579\begin{cfa}
1580[double] foo(), foo( int ), foo( double ) {...}
1581\end{cfa}
1582&
1583\begin{cfa}
1584double foo1( void ), foo2( int ), foo3( double );
1585\end{cfa}
1586\end{tabular}
1587\lstMakeShortInline@%
1588\end{cquote}
1589where \CFA allows the last routine in the list to define its body.
1590
1591The syntax for pointers to \CFA routines specifies the pointer name on the right, \eg:
1592\begin{cfa}
1593* [ int x ] () fp;                                              $\C{// pointer to routine returning int with no parameters}$
1594* [ * int ] ( int y ) gp;                               $\C{// pointer to routine returning pointer to int with int parameter}$
1595* [ ] ( int, char ) hp;                                 $\C{// pointer to routine returning no result with int and char parameters}$
1596* [ * int, int ] ( int ) jp;                    $\C{// pointer to routine returning pointer to int and int, with int parameter}$
1597\end{cfa}
1598Note, \emph{a routine name cannot be specified}:
1599\begin{cfa}
1600* [ int x ] f () fp;                                    $\C{// routine name "f" is disallowed}$
1601\end{cfa}
1602
1603Finally, new \CFA declarations may appear together with C declarations in the same program block, but cannot be mixed within a specific declaration.
1604Therefore, a programmer has the option of either continuing to use traditional C declarations or take advantage of the new style.
1605Clearly, both styles need to be supported for some time due to existing C-style header-files, particularly for UNIX-like systems.
1606
1607
1608\subsection{References}
1609\label{s:References}
1610
1611All variables in C have an \newterm{address}, a \newterm{value}, and a \newterm{type};
1612at 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.
1613The C type-system does not always track the relationship between a value and its address;
1614a 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'').
1615For 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.
1616Despite 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.
1617
1618Within a lexical scope, lvalue expressions have an \newterm{address interpretation} for writing a value or a \newterm{value interpretation} to read a value.
1619For example, in @x = y@, @x@ has an address interpretation, while @y@ has a value interpretation.
1620While 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.
1621In C, for any type @T@ there is a pointer type @T *@, the value of which is the address of a value of type @T@.
1622A 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 @&?@.
1623
1624\begin{cfa}
1625int x = 1, y = 2, * p1, * p2, ** p3;
1626p1 = &x;                                                                $\C{// p1 points to x}$
1627p2 = &y;                                                                $\C{// p2 points to y}$
1628p3 = &p1;                                                               $\C{// p3 points to p1}$
1629*p2 = ((*p1 + *p2) * (**p3 - *p1)) / (**p3 - 15);
1630\end{cfa}
1631
1632Unfortunately, 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 because of pointer arithmetic.
1633For both brevity and clarity, it is desirable for the compiler to figure out how to elide the dereference operators in a complex expression such as the assignment to @*p2@ above.
1634However, 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 programmer may write one when they mean the other, and precluding any simple algorithm for elision of dereference operators.
1635To solve these problems, \CFA introduces reference types @T &@;
1636a @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:
1637
1638\begin{cfa}
1639int x = 1, y = 2, & r1, & r2, && r3;
1640&r1 = &x;  $\C{// r1 points to x}$
1641&r2 = &y;  $\C{// r2 points to y}$
1642&&r3 = &&r1;  $\C{// r3 points to r2}$
1643r2 = ((r1 + r2) * (r3 - r1)) / (r3 - 15);       $\C{// implicit dereferencing}$
1644\end{cfa}
1645
1646Except for auto-dereferencing by the compiler, this reference example is exactly the same as the previous pointer example.
1647Hence, 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.
1648One 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:
1649
1650\begin{cfa}
1651`*`r2 = ((`*`r1 + `*`r2) * (`**`r3 - `*`r1)) / (`**`r3 - 15);
1652\end{cfa}
1653
1654References in \CFA are similar to those in \CC, with important improvements, which can be seen in the example above.
1655Firstly, \CFA does not forbid references to references.
1656This provides a much more orthogonal design for library implementors, obviating the need for workarounds such as @std::reference_wrapper@.
1657Secondly, \CFA references are rebindable, whereas \CC references have a fixed address.
1658\newsavebox{\LstBox}
1659\begin{lrbox}{\LstBox}
1660\lstset{basicstyle=\footnotesize\linespread{0.9}\sf}
1661\begin{cfa}
1662int & r = *new( int );
1663...
1664delete &r;
1665r += 1;                 // undefined reference
1666\end{cfa}
1667\end{lrbox}
1668Rebinding allows \CFA references to be default-initialized (\eg to a null pointer\footnote{
1669While effort has been put into non-null reference checking in \CC and Java, the exercise seems moot for any non-managed languages (C/\CC), given that it only handles one of many different error situations:
1670\begin{cquote}
1671\usebox{\LstBox}
1672\end{cquote}
1673}%
1674) and point to different addresses throughout their lifetime, like pointers.
1675Rebinding is accomplished by extending the existing syntax and semantics of the address-of operator in C.
1676
1677In 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.
1678In \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.
1679The 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.
1680This rebinding occurs to an arbitrary depth of reference nesting;
1681loosely speaking, nested address-of operators produce a nested lvalue pointer up to the depth of the reference.
1682These 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@.
1683More precisely:
1684\begin{itemize}[topsep=3pt,itemsep=2pt,parsep=0pt]
1685\item
1686if @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).
1687       
1688\item
1689if @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).
1690\end{itemize}
1691Since 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 is made solely on convenience, \eg many pointer or value accesses.
1692
1693By analogy to pointers, \CFA references also allow cv-qualifiers such as @const@:
1694\begin{cfa}
1695const int cx = 5;                                                       $\C{// cannot change cx}$
1696const int & cr = cx;                                            $\C{// cannot change cr's referred value}$
1697&cr = &cx;                                                                      $\C{// rebinding cr allowed}$
1698cr = 7;                                                                         $\C{// ERROR, cannot change cr}$
1699int & const rc = x;                                                     $\C{// must be initialized, like in \CC}$
1700&rc = &x;                                                                       $\C{// ERROR, cannot rebind rc}$
1701rc = 7;                                                                         $\C{// x now equal to 7}$
1702\end{cfa}
1703Given that a reference is meant to represent a lvalue, \CFA provides some syntactic shortcuts when initializing references.
1704There are three initialization contexts in \CFA: declaration initialization, argument/parameter binding, and return/temporary binding.
1705In each of these contexts, the address-of operator on the target lvalue is elided.
1706The syntactic motivation 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@.
1707
1708More 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;
1709this conversion is used in any context in \CFA where an implicit conversion is allowed.
1710Similarly, 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.
1711The final reference conversion included in \CFA is ``rvalue-to-reference'' conversion, implemented by means of an implicit temporary.
1712When 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.
1713\begin{cfa}
1714struct S { double x, y; };
1715int i, j;
1716void f( int & i, int & j, S & s, int v[] );
1717f( 3, i + j, (S){ 1.0, 7.0 }, (int [3]){ 1, 2, 3 } );   $\C{// pass rvalue to lvalue \(\Rightarrow\) implicit temporary}$
1718\end{cfa}
1719This 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.
1720\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.
1721
1722
1723\subsection{Type Nesting}
1724
1725Nested types provide a mechanism to organize associated types and refactor a subset of fields into a named aggregate (\eg sub-aggregates @name@, @address@, @department@, within aggregate @employe@).
1726Java nested types are dynamic (apply to objects), \CC are static (apply to the \lstinline[language=C++]@class@), and C hoists (refactors) nested types into the enclosing scope, meaning there is no need for type qualification.
1727Since \CFA in not object-oriented, adopting dynamic scoping does not make sense;
1728instead \CFA adopts \CC static nesting, using the field-selection operator ``@.@'' for type qualification, as does Java, rather than the \CC type-selection operator ``@::@'' (see Figure~\ref{f:TypeNestingQualification}).
1729\begin{figure}
1730\centering
1731\lstDeleteShortInline@%
1732\begin{tabular}{@{}l@{\hspace{3em}}l|l@{}}
1733\multicolumn{1}{c@{\hspace{3em}}}{\textbf{C Type Nesting}}      & \multicolumn{1}{c}{\textbf{C Implicit Hoisting}}      & \multicolumn{1}{|c}{\textbf{\CFA}}    \\
1734\hline
1735\begin{cfa}
1736struct S {
1737        enum C { R, G, B };
1738        struct T {
1739                union U { int i, j; };
1740                enum C c;
1741                short int i, j;
1742        };
1743        struct T t;
1744} s;
1745
1746int rtn() {
1747        s.t.c = R;
1748        struct T t = { R, 1, 2 };
1749        enum C c;
1750        union U u;
1751}
1752\end{cfa}
1753&
1754\begin{cfa}
1755enum C { R, G, B };
1756union U { int i, j; };
1757struct T {
1758        enum C c;
1759        short int i, j;
1760};
1761struct S {
1762        struct T t;
1763} s;
1764       
1765
1766
1767
1768
1769
1770
1771\end{cfa}
1772&
1773\begin{cfa}
1774struct S {
1775        enum C { R, G, B };
1776        struct T {
1777                union U { int i, j; };
1778                enum C c;
1779                short int i, j;
1780        };
1781        struct T t;
1782} s;
1783
1784int rtn() {
1785        s.t.c = `S.`R;  // type qualification
1786        struct `S.`T t = { `S.`R, 1, 2 };
1787        enum `S.`C c;
1788        union `S.T.`U u;
1789}
1790\end{cfa}
1791\end{tabular}
1792\lstMakeShortInline@%
1793\caption{Type Nesting / Qualification}
1794\label{f:TypeNestingQualification}
1795\end{figure}
1796In the left example in C, types @C@, @U@ and @T@ are implicitly hoisted outside of type @S@ into the containing block scope.
1797In the right example in \CFA, the types are not hoisted and accessed .
1798
1799
1800\subsection{Constructors and Destructors}
1801
1802One of the strengths (and weaknesses) of C is memory-management control, allowing resource release to be precisely specified versus unknown release with garbage-collected memory-management.
1803However, this manual approach is often verbose, and it is useful to manage resources other than memory (\eg file handles) using the same mechanism as memory.
1804\CC addresses these issues using Resource Aquisition Is Initialization (RAII), implemented by means of \newterm{constructor} and \newterm{destructor} functions;
1805\CFA adopts constructors and destructors (and @finally@) to facilitate RAII.
1806While constructors and destructors are a common feature of object-oriented programming-languages, they are an independent capability allowing \CFA to adopt them while retaining a procedural paradigm.
1807Specifically, \CFA constructors and destructors are denoted by name and first parameter-type versus name and nesting in an aggregate type.
1808Constructor calls seamlessly integrate with existing C initialization syntax, providing a simple and familiar syntax to C programmers and allowing constructor calls to be inserted into legacy C code with minimal code changes.
1809
1810In \CFA, a constructor is named @?{}@ and a destructor is named @^?{}@.
1811The name @{}@ comes from the syntax for the initializer: @struct S { int i, j; } s = `{` 2, 3 `}`@.
1812The symbol \lstinline+^+ is used because it was the last remaining binary operator that could be used in a unary context.
1813Like other \CFA operators, these names represent the syntax used to call the constructor or destructor, \eg @?{}(x, ...)@ or @^{}(x, ...)@.
1814The constructor and destructor have return type @void@ and a first parameter of reference to the object type to be constructed or destructs.
1815While the first parameter is informally called the @this@ parameter, as in object-oriented languages, any variable name may be used.
1816Both constructors and destructors allow additional parametes after the @this@ parameter for specifying values for initialization/de-initialization\footnote{
1817Destruction parameters are useful for specifying storage-management actions, such as de-initialize but not deallocate.}.
1818\begin{cfa}
1819struct VLA {
1820        int len, * data;
1821};
1822void ?{}( VLA & vla ) with ( vla ) {            $\C{// default constructor}$
1823        len = 10;  data = alloc( len );
1824}
1825void ^?{}( VLA & vla ) with ( vla ) {           $\C{// destructor}$
1826        free( data );
1827}
1828{
1829        VLA x;                                                                  $\C{// implicit:  ?\{\}( x );}$
1830}                                                                                       $\C{// implicit:  ?\^{}\{\}( x );}$
1831\end{cfa}
1832@VLA@ is a \newterm{managed type}\footnote{
1833A managed type affects the runtime environment versus a self-contained type.}: a type requiring a non-trivial constructor or destructor, or with a field of a managed type.
1834A managed type is implicitly constructed upon allocation and destructed upon deallocation to ensure proper interaction with runtime resources, in this case the @data@ array in the heap.
1835For details of the placement of implicit constructor and destructor calls among complex executable statements see~\cite[\S~2.2]{Schluntz17}.
1836
1837\CFA also provides syntax for \newterm{initialization} and \newterm{copy}:
1838\begin{cfa}
1839void ?{}( VLA & vla, int size, char fill ) with ( vla ) {       $\C{// initialization}$
1840        len = size;  data = alloc( len, fill );
1841}
1842void ?{}( VLA & vla, VLA other ) {                      $\C{// copy}$
1843        vla.len = other.len;  vla.data = other.data;
1844}
1845\end{cfa}
1846An initialization constructor-call has the same syntax as a C initializer, except the initialization values are passed as arguments to a matching constructor (number and type of paremeters).
1847\begin{cfa}
1848VLA va = `{` 20, 0 `}`,  * arr = alloc()`{` 5, 0 `}`;
1849\end{cfa}
1850Note, the use of a \newterm{constructor expression} to initialize the storage from the dynamic storage-allocation.
1851Like \CC, the copy constructor has two parameters, the second of which is a value parameter with the same type as the first parameter;
1852appropriate care is taken to not recursively call the copy constructor when initializing the second parameter.
1853
1854\CFA constructors may be explicitly call, like Java, and destructors may be explicitly called, like \CC.
1855Explicit calls to constructors double as a \CC-style \emph{placement syntax}, useful for construction of member fields in user-defined constructors and reuse of existing storage allocations.
1856While existing call syntax works for explicit calls to constructors and destructors, \CFA also provides a more concise \newterm{operator syntax} for both:
1857\begin{cfa}
1858{
1859        VLA x,  y = { 20, 0x01 },  z = y;
1860        // implicit:  ?{}( x );  ?{}( y, 20, 0x01 );  ?{}( z, y ); z points to y
1861        ^x{};                                                                   $\C{// deallocate x}$
1862        x{};                                                                    $\C{// reallocate x}$
1863        z{ 5, 0xff };                                                   $\C{// reallocate z, not pointing to y}$
1864        ^y{};                                                                   $\C{// deallocate y}$
1865        y{ x };                                                                 $\C{// reallocate y, points to x}$
1866        x{};                                                                    $\C{// reallocate x, not pointing to y}$
1867        // implicit:  ^?{}(z);  ^?{}(y);  ^?{}(x);
1868}
1869\end{cfa}
1870
1871To 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.
1872These default functions can be overridden by user-generated versions.
1873For 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;
1874if default zero-initialization is desired, the default constructors can be overridden.
1875For user-generated types, the four functions are also automatically generated.
1876@enum@ types are handled the same as their underlying integral type, and unions are also bitwise copied and no-op initialized and destructed.
1877For compatibility with C, a copy constructor from the first union member type is also defined.
1878For @struct@ types, each of the four functions are implicitly defined to call their corresponding functions on each member of the struct.
1879To better simulate the behaviour of C initializers, a set of \newterm{field constructors} is also generated for structures.
1880A constructor is generated for each non-empty prefix of a structure's member-list to copy-construct the members passed as parameters and default-construct the remaining members.
1881To 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;
1882similarly, the generated default constructor is hidden upon declaration of any constructor.
1883These semantics closely mirror the rule for implicit declaration of constructors in \CC\cite[p.~186]{ANSI98:C++}.
1884
1885In some circumstance programmers may not wish to have constructor and destructor calls.
1886In these cases, \CFA provides the initialization syntax \lstinline|S x @= {}|, and the object becomes unmanaged, so constructors and destructors calls are not generated.
1887Any C initializer can be the right-hand side of an \lstinline|@=| initializer, \eg \lstinline|VLA a @= { 0, 0x0 }|, with the usual C initialization semantics.
1888The point of \lstinline|@=| is to provide a migration path from legacy C code to \CFA, by providing a mechanism to incrementally convert to implicit initialization.
1889
1890
1891% \subsection{Default Parameters}
1892
1893
1894\section{Literals}
1895
1896C 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.
1897In 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.
1898
1899
1900\subsection{0/1}
1901
1902In 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.
1903As 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.
1904The 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.
1905To 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.
1906With 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)@.
1907\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.
1908
1909\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.
1910The addition of @one_t@ allows generic algorithms to handle the unit value uniformly for types where that is meaningful.
1911\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.
1912
1913
1914\subsection{Units}
1915
1916Alternative call syntax (literal argument before routine name) to convert basic literals into user literals.
1917
1918{\lstset{language=CFA,deletedelim=**[is][]{`}{`},moredelim=**[is][\color{red}]{@}{@}}
1919\begin{cfa}
1920struct Weight { double stones; };
1921
1922void ?{}( Weight & w ) { w.stones = 0; }        $\C{// operations}$
1923void ?{}( Weight & w, double w ) { w.stones = w; }
1924Weight ?+?( Weight l, Weight r ) { return (Weight){ l.stones + r.stones }; }
1925
1926Weight @?`st@( double w ) { return (Weight){ w }; } $\C{// backquote for units}$
1927Weight @?`lb@( double w ) { return (Weight){ w / 14.0 }; }
1928Weight @?`kg@( double w ) { return (Weight) { w * 0.1575}; }
1929
1930int main() {
1931        Weight w, hw = { 14 };                                  $\C{// 14 stone}$
1932        w = 11@`st@ + 1@`lb@;
1933        w = 70.3@`kg@;
1934        w = 155@`lb@;
1935        w = 0x_9b_u@`lb@;                                               $\C{// hexadecimal unsigned weight (155)}$
1936        w = 0_233@`lb@;                                                 $\C{// octal weight (155)}$
1937        w = 5@`st@ + 8@`kg@ + 25@`lb@ + hw;
1938}
1939\end{cfa}
1940}%
1941
1942
1943\section{Libraries}
1944\label{sec:libraries}
1945
1946As 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.
1947\CFA has replacement libraries condensing hundreds of existing C names into tens of \CFA overloaded names, all without rewriting the actual computations.
1948In many cases, the interface is an inline wrapper providing overloading during compilation but zero cost at runtime.
1949The following sections give a glimpse of the interface reduction to many C libraries.
1950In 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.
1951
1952
1953\subsection{Limits}
1954
1955C library @limits.h@ provides lower and upper bound constants for the basic types.
1956\CFA name overloading is used to condense these typed constants, \eg:
1957\begin{cquote}
1958\lstDeleteShortInline@%
1959\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1960\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{Definition}}       & \multicolumn{1}{c}{\textbf{Usage}}    \\
1961\begin{cfa}
1962const short int `MIN` = -32768;
1963const int `MIN` = -2147483648;
1964const long int `MIN` = -9223372036854775808L;
1965\end{cfa}
1966&
1967\begin{cfa}
1968short int si = `MIN`;
1969int i = `MIN`;
1970long int li = `MIN`;
1971\end{cfa}
1972\end{tabular}
1973\lstMakeShortInline@%
1974\end{cquote}
1975The result is a significant reduction in names to access typed constants, \eg:
1976\begin{cquote}
1977\lstDeleteShortInline@%
1978\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
1979\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
1980\begin{cfa}
1981MIN
1982MAX
1983M_PI
1984M_E
1985\end{cfa}
1986&
1987\begin{cfa}
1988SCHAR_MIN, CHAR_MIN, SHRT_MIN, INT_MIN, LONG_MIN, LLONG_MIN,
1989SCHAR_MAX, UCHAR_MAX, SHRT_MAX, INT_MAX, LONG_MAX, LLONG_MAX,
1990M_PI, M_PIl, M_CPI, M_CPIl,
1991M_E, M_El, M_CE, M_CEl
1992\end{cfa}
1993\end{tabular}
1994\lstMakeShortInline@%
1995\end{cquote}
1996
1997
1998\subsection{Math}
1999
2000C library @math.h@ provides many mathematical routines.
2001\CFA routine overloading is used to condense these mathematical routines, \eg:
2002\begin{cquote}
2003\lstDeleteShortInline@%
2004\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
2005\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{Definition}}       & \multicolumn{1}{c}{\textbf{Usage}}    \\
2006\begin{cfa}
2007float `log`( float x );
2008double `log`( double );
2009double _Complex `log`( double _Complex x );
2010\end{cfa}
2011&
2012\begin{cfa}
2013float f = `log`( 3.5 );
2014double d = `log`( 3.5 );
2015double _Complex dc = `log`( 3.5+0.5I );
2016\end{cfa}
2017\end{tabular}
2018\lstMakeShortInline@%
2019\end{cquote}
2020The result is a significant reduction in names to access math routines, \eg:
2021\begin{cquote}
2022\lstDeleteShortInline@%
2023\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
2024\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
2025\begin{cfa}
2026log
2027sqrt
2028sin
2029\end{cfa}
2030&
2031\begin{cfa}
2032logf, log, logl, clogf, clog, clogl
2033sqrtf, sqrt, sqrtl, csqrtf, csqrt, csqrtl
2034sinf, sin, sinl, csinf, csin, csinl
2035\end{cfa}
2036\end{tabular}
2037\lstMakeShortInline@%
2038\end{cquote}
2039While \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).
2040For example, it is not possible to overload @atan@ for both one and two arguments;
2041instead the names @atan@ and @atan2@ are required.
2042The 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.
2043
2044
2045\subsection{Standard}
2046
2047C library @stdlib.h@ provides many general routines.
2048\CFA routine overloading is used to condense these utility routines, \eg:
2049\begin{cquote}
2050\lstDeleteShortInline@%
2051\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
2052\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{Definition}}       & \multicolumn{1}{c}{\textbf{Usage}}    \\
2053\begin{cfa}
2054unsigned int `abs`( int );
2055double `abs`( double );
2056double abs( double _Complex );
2057\end{cfa}
2058&
2059\begin{cfa}
2060unsigned int i = `abs`( -1 );
2061double d = `abs`( -1.5 );
2062double d = `abs`( -1.5+0.5I );
2063\end{cfa}
2064\end{tabular}
2065\lstMakeShortInline@%
2066\end{cquote}
2067The result is a significant reduction in names to access utility routines, \eg:
2068\begin{cquote}
2069\lstDeleteShortInline@%
2070\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
2071\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
2072\begin{cfa}
2073abs
2074strto
2075random
2076\end{cfa}
2077&
2078\begin{cfa}
2079abs, labs, llabs, fabsf, fabs, fabsl, cabsf, cabs, cabsl
2080strtol, strtoul, strtoll, strtoull, strtof, strtod, strtold
2081srand48, mrand48, lrand48, drand48
2082\end{cfa}
2083\end{tabular}
2084\lstMakeShortInline@%
2085\end{cquote}
2086In additon, there are polymorphic routines, like @min@ and @max@, which work on any type with operators @?<?@ or @?>?@.
2087
2088The following shows one example where \CFA \emph{extends} an existing standard C interface to reduce complexity and provide safety.
2089C/\Celeven provide a number of complex and overlapping storage-management operation to support the following capabilities:
2090\begin{description}[topsep=3pt,itemsep=2pt,parsep=0pt]
2091\item[fill]
2092after allocation the storage is filled with a specified character.
2093\item[resize]
2094an existing allocation is decreased or increased in size.
2095In 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.
2096For an increase in storage size, new storage after the copied data may be filled.
2097\item[alignment]
2098an allocation starts on a specified memory boundary, \eg, an address multiple of 64 or 128 for cache-line purposes.
2099\item[array]
2100the allocation size is scaled to the specified number of array elements.
2101An array may be filled, resized, or aligned.
2102\end{description}
2103Table~\ref{t:StorageManagementOperations} shows the capabilities provided by C/\Celeven allocation-routines and how all the capabilities can be combined into two \CFA routines.
2104
2105\CFA storage-management routines extend the C equivalents by overloading, providing shallow type-safety, and removing the need to specify the base allocation-size.
2106The following example contrasts \CFA and C storage-allocation operation performing the same operations with the same type safety:
2107\begin{cquote}
2108\begin{cfa}[aboveskip=0pt]
2109size_t  dim = 10;                                                       $\C{// array dimension}$
2110char fill = '\xff';                                                     $\C{// initialization fill value}$
2111int * ip;
2112\end{cfa}
2113\lstDeleteShortInline@%
2114\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
2115\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{C}}        \\
2116\begin{cfa}
2117ip = alloc();
2118ip = alloc( fill );
2119ip = alloc( dim );
2120ip = alloc( dim, fill );
2121ip = alloc( ip, 2 * dim );
2122ip = alloc( ip, 4 * dim, fill );
2123\end{cfa}
2124&
2125\begin{cfa}
2126ip = (int *)malloc( sizeof( int ) );
2127ip = (int *)malloc( sizeof( int ) ); memset( ip, fill, sizeof( int ) );
2128ip = (int *)malloc( dim * sizeof( int ) );
2129ip = (int *)malloc( sizeof( int ) ); memset( ip, fill, dim * sizeof( int ) );
2130ip = (int *)realloc( ip, 2 * dim * sizeof( int ) );
2131ip = (int *)realloc( ip, 4 * dim * sizeof( int ) ); memset( ip, fill, 4 * dim * sizeof( int ) );
2132\end{cfa}
2133\end{tabular}
2134\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
2135\begin{cfa}
2136ip = align_alloc( 16 );
2137ip = align_alloc( 16, fill );
2138ip = align_alloc( 16, dim );
2139ip = align_alloc( 16, dim, fill );
2140\end{cfa}
2141&
2142\begin{cfa}
2143ip = memalign( 16, sizeof( int ) );
2144ip = memalign( 16, sizeof( int ) ); memset( ip, fill, sizeof( int ) );
2145ip = memalign( 16, dim * sizeof( int ) );
2146ip = memalign( 16, dim * sizeof( int ) ); memset( ip, fill, dim * sizeof( int ) );
2147\end{cfa}
2148\end{tabular}
2149\lstMakeShortInline@%
2150\end{cquote}
2151Variadic @new@ (see Section~\ref{sec:variadic-tuples}) cannot support the same overloading because extra parameters are for initialization.
2152Hence, there are @new@ and @anew@ routines for single and array variables, and the fill value is the arguments to the constructor, \eg:
2153\begin{cfa}
2154struct S { int i, j; };
2155void ?{}( S & s, int i, int j ) { s.i = i; s.j = j; }
2156S * s = new( 2, 3 );                                            $\C{// allocate storage and run constructor}$
2157S * as = anew( dim, 2, 3 );                                     $\C{// each array element initialized to 2, 3}$
2158\end{cfa}
2159Note, \CC can only initialization array elements via the default constructor.
2160
2161Finally, the \CFA memory-allocator has \newterm{sticky properties} for dynamic storage: fill and alignment are remembered with an object's storage in the heap.
2162When a @realloc@ is performed, the sticky properties are respected, so that new storage is correctly aligned and initialized with the fill character.
2163
2164\begin{table}
2165\centering
2166\lstDeleteShortInline@%
2167\lstMakeShortInline~%
2168\begin{tabular}{@{}r|r|l|l|l|l@{}}
2169\multicolumn{1}{c}{}&           & \multicolumn{1}{c|}{fill}     & resize        & alignment     & array \\
2170\hline
2171C               & ~malloc~                      & no                    & no            & no            & no    \\
2172                & ~calloc~                      & yes (0 only)  & no            & no            & yes   \\
2173                & ~realloc~                     & no/copy               & yes           & no            & no    \\
2174                & ~memalign~            & no                    & no            & yes           & no    \\
2175                & ~posix_memalign~      & no                    & no            & yes           & no    \\
2176\hline
2177C11             & ~aligned_alloc~       & no                    & no            & yes           & no    \\
2178\hline
2179\CFA    & ~alloc~                       & yes/copy              & no/yes        & no            & yes   \\
2180                & ~align_alloc~         & yes                   & no            & yes           & yes   \\
2181\end{tabular}
2182\lstDeleteShortInline~%
2183\lstMakeShortInline@%
2184\caption{Storage-Management Operations}
2185\label{t:StorageManagementOperations}
2186\end{table}
2187
2188
2189\subsection{I/O}
2190\label{s:IOLibrary}
2191
2192The goal of \CFA I/O is to simplify the common cases, while fully supporting polymorphism and user defined types in a consistent way.
2193The approach combines ideas from \CC and Python.
2194The \CFA header file for the I/O library is @fstream@.
2195
2196The common case is printing out a sequence of variables separated by whitespace.
2197\begin{cquote}
2198\lstDeleteShortInline@%
2199\begin{tabular}{@{}l@{\hspace{\parindentlnth}}l@{}}
2200\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{c}{\textbf{\CC}}      \\
2201\begin{cfa}
2202int x = 1, y = 2, z = 3;
2203sout | x `|` y `|` z | endl;
2204\end{cfa}
2205&
2206\begin{cfa}
2207
2208cout << x `<< " "` << y `<< " "` << z << endl;
2209\end{cfa}
2210\\
2211\begin{cfa}[showspaces=true,aboveskip=0pt,belowskip=0pt]
22121` `2` `3
2213\end{cfa}
2214&
2215\begin{cfa}[showspaces=true,aboveskip=0pt,belowskip=0pt]
22161 2 3
2217\end{cfa}
2218\end{tabular}
2219\lstMakeShortInline@%
2220\end{cquote}
2221The \CFA form has half the characters of the \CC form, and is similar to Python I/O with respect to implicit separators.
2222Similar simplification occurs for tuple I/O, which prints all tuple values separated by ``\lstinline[showspaces=true]@, @''.
2223\begin{cfa}
2224[int, [ int, int ] ] t1 = [ 1, [ 2, 3 ] ], t2 = [ 4, [ 5, 6 ] ];
2225sout | t1 | t2 | endl;                                  $\C{// print tuples}$
2226\end{cfa}
2227\begin{cfa}[showspaces=true,aboveskip=0pt]
22281`, `2`, `3 4`, `5`, `6
2229\end{cfa}
2230Finally, \CFA uses the logical-or operator for I/O as it is the lowest-priority overloadable operator, other than assignment.
2231Therefore, fewer output expressions require parenthesis.
2232\begin{cquote}
2233\lstDeleteShortInline@%
2234\begin{tabular}{@{}ll@{}}
2235\textbf{\CFA:}
2236&
2237\begin{cfa}
2238sout | x * 3 | y + 1 | z << 2 | x == y | (x | y) | (x || y) | (x > z ? 1 : 2) | endl;
2239\end{cfa}
2240\\
2241\textbf{\CC:}
2242&
2243\begin{cfa}
2244cout << x * 3 << y + 1 << `(`z << 2`)` << `(`x == y`)` << (x | y) << (x || y) << (x > z ? 1 : 2) << endl;
2245\end{cfa}
2246\\
2247\textbf{output:}
2248&
2249\begin{cfa}[showspaces=true,aboveskip=0pt]
22503 3 12 0 3 1 2
2251\end{cfa}
2252\end{tabular}
2253\lstMakeShortInline@%
2254\end{cquote}
2255There 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.
2256
2257The implicit separator character (space/blank) is a separator not a terminator.
2258The rules for implicitly adding the separator are:
2259\begin{itemize}[topsep=3pt,itemsep=2pt,parsep=0pt]
2260\item
2261A separator does not appear at the start or end of a line.
2262\item
2263A separator does not appear before or after a character literal or variable.
2264\item
2265A separator does not appear before or after a null (empty) C string, which is a local mechanism to disable insertion of the separator character.
2266\item
2267A separator does not appear before a C string starting with the characters: \lstinline[mathescape=off,basicstyle=\tt]@([{=$@
2268\item
2269A separator does not appear after a C string ending with the characters: \lstinline[basicstyle=\tt]@,.;!?)]}%@
2270\item
2271{\lstset{language=CFA,deletedelim=**[is][]{`}{`}}
2272A separator does not appear before or after a C string beginning/ending with the quote or whitespace characters: \lstinline[basicstyle=\tt,showspaces=true]@`'": \t\v\f\r\n@
2273}%
2274\item
2275There are routines to set and get the separator string, and manipulators to toggle separation on and off in the middle of output.
2276\end{itemize}
2277
2278
2279\subsection{Multi-precision Integers}
2280\label{s:MultiPrecisionIntegers}
2281
2282\CFA has an interface to the GMP multi-precision signed-integers~\cite{GMP}, similar to the \CC interface provided by GMP.
2283The \CFA interface wraps GMP routines into operator routines to make programming with multi-precision integers identical to using fixed-sized integers.
2284The \CFA type name for multi-precision signed-integers is @Int@ and the header file is @gmp@.
2285The following multi-precision factorial programs contrast using GMP with the \CFA and C interfaces.
2286\begin{cquote}
2287\lstDeleteShortInline@%
2288\begin{tabular}{@{}l@{\hspace{\parindentlnth}}@{\hspace{\parindentlnth}}l@{}}
2289\multicolumn{1}{c@{\hspace{\parindentlnth}}}{\textbf{\CFA}}     & \multicolumn{1}{@{\hspace{\parindentlnth}}c}{\textbf{C}}      \\
2290\begin{cfa}
2291#include <gmp>
2292int main( void ) {
2293        sout | "Factorial Numbers" | endl;
2294        Int fact = 1;
2295
2296        sout | 0 | fact | endl;
2297        for ( unsigned int i = 1; i <= 40; i += 1 ) {
2298                fact *= i;
2299                sout | i | fact | endl;
2300        }
2301}
2302\end{cfa}
2303&
2304\begin{cfa}
2305#include <gmp.h>
2306int main( void ) {
2307        `gmp_printf`( "Factorial Numbers\n" );
2308        `mpz_t` fact;
2309        `mpz_init_set_ui`( fact, 1 );
2310        `gmp_printf`( "%d %Zd\n", 0, fact );
2311        for ( unsigned int i = 1; i <= 40; i += 1 ) {
2312                `mpz_mul_ui`( fact, fact, i );
2313                `gmp_printf`( "%d %Zd\n", i, fact );
2314        }
2315}
2316\end{cfa}
2317\end{tabular}
2318\lstMakeShortInline@%
2319\end{cquote}
2320
2321
2322\section{Evaluation}
2323\label{sec:eval}
2324
2325Though \CFA provides significant added functionality over C, these features have a low runtime penalty.
2326In fact, \CFA's features for generic programming can enable faster runtime execution than idiomatic @void *@-based C code.
2327This 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}).
2328Since 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.
2329A more illustrative benchmark measures the costs of idiomatic usage of each language's features.
2330Figure~\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}.
2331The benchmark test is similar for C and \CC.
2332The 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.
2333
2334\begin{figure}
2335\begin{cfa}[xleftmargin=3\parindentlnth,aboveskip=0pt,belowskip=0pt]
2336int main( int argc, char * argv[] ) {
2337        FILE * out = fopen( "cfa-out.txt", "w" );
2338        int maxi = 0, vali = 42;
2339        stack(int) si, ti;
2340
2341        REPEAT_TIMED( "push_int", N, push( &si, vali ); )
2342        TIMED( "copy_int", ti = si; )
2343        TIMED( "clear_int", clear( &si ); )
2344        REPEAT_TIMED( "pop_int", N,
2345                int xi = pop( &ti ); if ( xi > maxi ) { maxi = xi; } )
2346        REPEAT_TIMED( "print_int", N/2, print( out, vali, ":", vali, "\n" ); )
2347
2348        pair(_Bool, char) maxp = { (_Bool)0, '\0' }, valp = { (_Bool)1, 'a' };
2349        stack(pair(_Bool, char)) sp, tp;
2350
2351        REPEAT_TIMED( "push_pair", N, push( &sp, valp ); )
2352        TIMED( "copy_pair", tp = sp; )
2353        TIMED( "clear_pair", clear( &sp ); )
2354        REPEAT_TIMED( "pop_pair", N,
2355                pair(_Bool, char) xp = pop( &tp ); if ( xp > maxp ) { maxp = xp; } )
2356        REPEAT_TIMED( "print_pair", N/2, print( out, valp, ":", valp, "\n" ); )
2357        fclose(out);
2358}
2359\end{cfa}
2360\caption{\protect\CFA Benchmark Test}
2361\label{fig:BenchmarkTest}
2362\end{figure}
2363
2364The 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.
2365The \CCV variant illustrates an alternative object-oriented idiom where all objects inherit from a base @object@ class, mimicking a Java-like interface;
2366hence runtime checks are necessary to safely down-cast objects.
2367The 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.
2368For 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.
2369Note, the C benchmark uses unchecked casts as there is no runtime mechanism to perform such checks, while \CFA and \CC provide type-safety statically.
2370
2371Figure~\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.
2372The graph plots the median of 5 consecutive runs of each program, with an initial warm-up run omitted.
2373All code is compiled at \texttt{-O2} by gcc or g++ 6.2.0, with all \CC code compiled as \CCfourteen.
2374The 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.
2375
2376\begin{figure}
2377\centering
2378\input{timing}
2379\caption{Benchmark Timing Results (smaller is better)}
2380\label{fig:eval}
2381\end{figure}
2382
2383\begin{table}
2384\centering
2385\caption{Properties of benchmark code}
2386\label{tab:eval}
2387\newcommand{\CT}[1]{\multicolumn{1}{c}{#1}}
2388\begin{tabular}{rrrrr}
2389                                                                        & \CT{C}        & \CT{\CFA}     & \CT{\CC}      & \CT{\CCV}             \\ \hline
2390maximum memory usage (MB)                       & 10001         & 2502          & 2503          & 11253                 \\
2391source code size (lines)                        & 247           & 222           & 165           & 339                   \\
2392redundant type annotations (lines)      & 39            & 2                     & 2                     & 15                    \\
2393binary size (KB)                                        & 14            & 229           & 18            & 38                    \\
2394\end{tabular}
2395\end{table}
2396
2397The 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;
2398this inefficiency is exacerbated by the second level of generic types in the pair-based benchmarks.
2399By 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.
2400\CCV is slower than C largely due to the cost of runtime type-checking of down-casts (implemented with @dynamic_cast@);
2401There are two outliers in the graph for \CFA: all prints and pop of @pair@.
2402Both 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.
2403A compiler designed for \CFA could easily perform these optimizations.
2404Finally, the binary size for \CFA is larger because of static linking with the \CFA libraries.
2405
2406\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.
2407On 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.
2408\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;
2409with their omission, the \CCV line count is similar to C.
2410We 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.
2411
2412Raw line-count, however, is a fairly rough measure of code complexity;
2413another important factor is how much type information the programmer must manually specify, especially where that information is not checked by the compiler.
2414Such 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@).
2415To 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.
2416The \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.
2417The 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).
2418These uses are similar to the @new@ expressions in \CC, though the \CFA compiler's type resolver should shortly render even these type casts superfluous.
2419
2420
2421\section{Related Work}
2422
2423
2424\subsection{Polymorphism}
2425
2426\CC is the most similar language to \CFA;
2427both are extensions to C with source and runtime backwards compatibility.
2428The fundamental difference is in their engineering approach to C compatibility and programmer expectation.
2429While \CC provides good backwards compatibility with C, it has a steep learning curve for many of its extensions.
2430For example, polymorphism is provided via three disjoint mechanisms: overloading, inheritance, and templates.
2431The 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.
2432In 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.
2433The key mechanism to support separate compilation is \CFA's \emph{explicit} use of assumed properties for a type.
2434Until \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;
2435furthermore, \CC concepts are restricted to template polymorphism.
2436
2437Cyclone~\cite{Grossman06} also provides capabilities for polymorphic functions and existential types, similar to \CFA's @forall@ functions and generic types.
2438Cyclone 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.
2439Furthermore, 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@.
2440In \CFA terms, all Cyclone polymorphism must be dtype-static.
2441While 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.
2442Smith 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.
2443
2444Objective-C~\cite{obj-c-book} is an industrially successful extension to C.
2445However, Objective-C is a radical departure from C, using an object-oriented model with message-passing.
2446Objective-C did not support type-checked generics until recently \cite{xcode7}, historically using less-efficient runtime checking of object types.
2447The GObject~\cite{GObject} framework also adds object-oriented programming with runtime type-checking and reference-counting garbage-collection to C;
2448these features are more intrusive additions than those provided by \CFA, in addition to the runtime overhead of reference-counting.
2449Vala~\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.
2450Java~\cite{Java8} included generic types in Java~5, which are type-checked at compilation and type-erased at runtime, similar to \CFA's.
2451However, in Java, each object carries its own table of method pointers, while \CFA passes the method pointers separately to maintain a C-compatible layout.
2452Java is also a garbage-collected, object-oriented language, with the associated resource usage and C-interoperability burdens.
2453
2454D~\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.
2455However, 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.
2456D and Go are garbage-collected languages, imposing the associated runtime overhead.
2457The necessity of accounting for data transfer between managed runtimes and the unmanaged C runtime complicates foreign-function interfaces to C.
2458Furthermore, 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.
2459D 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.
2460Rust also possesses much more powerful abstraction capabilities for writing generic code than Go.
2461On 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.
2462\CFA, with its more modest safety features, allows direct ports of C code while maintaining the idiomatic style of the original source.
2463
2464
2465\subsection{Tuples/Variadics}
2466
2467Many programming languages have some form of tuple construct and/or variadic functions, \eg SETL, C, KW-C, \CC, D, Go, Java, ML, and Scala.
2468SETL~\cite{SETL} is a high-level mathematical programming language, with tuples being one of the primary data types.
2469Tuples in SETL allow subscripting, dynamic expansion, and multiple assignment.
2470C 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.
2471KW-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.
2472The main contributions of that work were adding MRVF, tuple mass and multiple assignment, and record-field access.
2473\CCeleven introduced @std::tuple@ as a library variadic template structure.
2474Tuples are a generalization of @std::pair@, in that they allow for arbitrary length, fixed-size aggregation of heterogeneous values.
2475Operations include @std::get<N>@ to extract values, @std::tie@ to create a tuple of references used for assignment, and lexicographic comparisons.
2476\CCseventeen proposes \emph{structured bindings}~\cite{Sutter15} to eliminate pre-declaring variables and use of @std::tie@ for binding the results.
2477This 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.
2478Furthermore, structured bindings are not a full replacement for @std::tie@, as it always declares new variables.
2479Like \CC, D provides tuples through a library variadic-template structure.
2480Go does not have tuples but supports MRVF.
2481Java'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.
2482Tuples are a fundamental abstraction in most functional programming languages, such as Standard ML~\cite{sml} and~\cite{Scala}, which decompose tuples using pattern matching.
2483
2484
2485\section{Conclusion and Future Work}
2486
2487The 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.
2488While 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.
2489The 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.
2490The contributions are a powerful type-system using parametric polymorphism and overloading, generic types, and tuples, which all have complex interactions.
2491The work is a challenging design, engineering, and implementation exercise.
2492On the surface, the project may appear as a rehash of similar mechanisms in \CC.
2493However, 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.
2494All of these new features are being used by the \CFA development-team to build the \CFA runtime-system.
2495Finally, we demonstrate that \CFA performance for some idiomatic cases is better than C and close to \CC, showing the design is practically applicable.
2496
2497There is ongoing work on a wide range of \CFA feature extensions, including arrays with size, exceptions, concurrent primitives, modules, and user-defined conversions.
2498(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.)
2499In addition, there are interesting future directions for the polymorphism design.
2500Notably, \CC template functions trade compile time and code bloat for optimal runtime of individual instantiations of polymorphic functions.
2501\CFA polymorphic functions use dynamic virtual-dispatch;
2502the 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.
2503Two 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).
2504These approaches are not mutually exclusive and allow performance optimizations to be applied only when necessary, without suffering global code-bloat.
2505In general, we believe separate compilation, producing smaller code, works well with loaded hardware-caches, which may offset the benefit of larger inlined-code.
2506
2507
2508\section{Acknowledgments}
2509
2510The 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.
2511%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.
2512% 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.
2513
2514
2515\bibliographystyle{plain}
2516\bibliography{pl}
2517
2518
2519\appendix
2520
2521\section{Benchmark Stack Implementation}
2522\label{sec:BenchmarkStackImplementation}
2523
2524\lstset{basicstyle=\linespread{0.9}\sf\small}
2525
2526Throughout, @/***/@ designates a counted redundant type annotation.
2527
2528\smallskip\noindent
2529\CFA
2530\begin{cfa}[xleftmargin=2\parindentlnth,aboveskip=0pt,belowskip=0pt]
2531forall(otype T) struct stack_node {
2532        T value;
2533        stack_node(T) * next;
2534};
2535forall(otype T) void ?{}(stack(T) * s) { (&s->head){ 0 }; }
2536forall(otype T) void ?{}(stack(T) * s, stack(T) t) {
2537        stack_node(T) ** crnt = &s->head;
2538        for ( stack_node(T) * next = t.head; next; next = next->next ) {
2539                *crnt = ((stack_node(T) *)malloc()){ next->value }; /***/
2540                stack_node(T) * acrnt = *crnt;
2541                crnt = &acrnt->next;
2542        }
2543        *crnt = 0;
2544}
2545forall(otype T) stack(T) ?=?(stack(T) * s, stack(T) t) {
2546        if ( s->head == t.head ) return *s;
2547        clear(s);
2548        s{ t };
2549        return *s;
2550}
2551forall(otype T) void ^?{}(stack(T) * s) { clear(s); }
2552forall(otype T) _Bool empty(const stack(T) * s) { return s->head == 0; }
2553forall(otype T) void push(stack(T) * s, T value) {
2554        s->head = ((stack_node(T) *)malloc()){ value, s->head }; /***/
2555}
2556forall(otype T) T pop(stack(T) * s) {
2557        stack_node(T) * n = s->head;
2558        s->head = n->next;
2559        T x = n->value;
2560        ^n{};
2561        free(n);
2562        return x;
2563}
2564forall(otype T) void clear(stack(T) * s) {
2565        for ( stack_node(T) * next = s->head; next; ) {
2566                stack_node(T) * crnt = next;
2567                next = crnt->next;
2568                delete(crnt);
2569        }
2570        s->head = 0;
2571}
2572\end{cfa}
2573
2574\medskip\noindent
2575\CC
2576\begin{cfa}[xleftmargin=2\parindentlnth,aboveskip=0pt,belowskip=0pt]
2577template<typename T> class stack {
2578        struct node {
2579                T value;
2580                node * next;
2581                node( const T & v, node * n = nullptr ) : value(v), next(n) {}
2582        };
2583        node * head;
2584        void copy(const stack<T>& o) {
2585                node ** crnt = &head;
2586                for ( node * next = o.head;; next; next = next->next ) {
2587                        *crnt = new node{ next->value }; /***/
2588                        crnt = &(*crnt)->next;
2589                }
2590                *crnt = nullptr;
2591        }
2592  public:
2593        stack() : head(nullptr) {}
2594        stack(const stack<T>& o) { copy(o); }
2595        stack(stack<T> && o) : head(o.head) { o.head = nullptr; }
2596        ~stack() { clear(); }
2597        stack & operator= (const stack<T>& o) {
2598                if ( this == &o ) return *this;
2599                clear();
2600                copy(o);
2601                return *this;
2602        }
2603        stack & operator= (stack<T> && o) {
2604                if ( this == &o ) return *this;
2605                head = o.head;
2606                o.head = nullptr;
2607                return *this;
2608        }
2609        bool empty() const { return head == nullptr; }
2610        void push(const T & value) { head = new node{ value, head };  /***/ }
2611        T pop() {
2612                node * n = head;
2613                head = n->next;
2614                T x = std::move(n->value);
2615                delete n;
2616                return x;
2617        }
2618        void clear() {
2619                for ( node * next = head; next; ) {
2620                        node * crnt = next;
2621                        next = crnt->next;
2622                        delete crnt;
2623                }
2624                head = nullptr;
2625        }
2626};
2627\end{cfa}
2628
2629\medskip\noindent
2630C
2631\begin{cfa}[xleftmargin=2\parindentlnth,aboveskip=0pt,belowskip=0pt]
2632struct stack_node {
2633        void * value;
2634        struct stack_node * next;
2635};
2636struct stack new_stack() { return (struct stack){ NULL }; /***/ }
2637void copy_stack(struct stack * s, const struct stack * t, void * (*copy)(const void *)) {
2638        struct stack_node ** crnt = &s->head;
2639        for ( struct stack_node * next = t->head; next; next = next->next ) {
2640                *crnt = malloc(sizeof(struct stack_node)); /***/
2641                **crnt = (struct stack_node){ copy(next->value) }; /***/
2642                crnt = &(*crnt)->next;
2643        }
2644        *crnt = 0;
2645}
2646_Bool stack_empty(const struct stack * s) { return s->head == NULL; }
2647void push_stack(struct stack * s, void * value) {
2648        struct stack_node * n = malloc(sizeof(struct stack_node)); /***/
2649        *n = (struct stack_node){ value, s->head }; /***/
2650        s->head = n;
2651}
2652void * pop_stack(struct stack * s) {
2653        struct stack_node * n = s->head;
2654        s->head = n->next;
2655        void * x = n->value;
2656        free(n);
2657        return x;
2658}
2659void clear_stack(struct stack * s, void (*free_el)(void *)) {
2660        for ( struct stack_node * next = s->head; next; ) {
2661                struct stack_node * crnt = next;
2662                next = crnt->next;
2663                free_el(crnt->value);
2664                free(crnt);
2665        }
2666        s->head = NULL;
2667}
2668\end{cfa}
2669
2670\medskip\noindent
2671\CCV
2672\begin{cfa}[xleftmargin=2\parindentlnth,aboveskip=0pt,belowskip=0pt]
2673stack::node::node( const object & v, node * n ) : value( v.new_copy() ), next( n ) {}
2674void stack::copy(const stack & o) {
2675        node ** crnt = &head;
2676        for ( node * next = o.head; next; next = next->next ) {
2677                *crnt = new node{ *next->value };
2678                crnt = &(*crnt)->next;
2679        }
2680        *crnt = nullptr;
2681}
2682stack::stack() : head(nullptr) {}
2683stack::stack(const stack & o) { copy(o); }
2684stack::stack(stack && o) : head(o.head) { o.head = nullptr; }
2685stack::~stack() { clear(); }
2686stack & stack::operator= (const stack & o) {
2687        if ( this == &o ) return *this;
2688        clear();
2689        copy(o);
2690        return *this;
2691}
2692stack & stack::operator= (stack && o) {
2693        if ( this == &o ) return *this;
2694        head = o.head;
2695        o.head = nullptr;
2696        return *this;
2697}
2698bool stack::empty() const { return head == nullptr; }
2699void stack::push(const object & value) { head = new node{ value, head }; /***/ }
2700ptr<object> stack::pop() {
2701        node * n = head;
2702        head = n->next;
2703        ptr<object> x = std::move(n->value);
2704        delete n;
2705        return x;
2706}
2707void stack::clear() {
2708        for ( node * next = head; next; ) {
2709                node * crnt = next;
2710                next = crnt->next;
2711                delete crnt;
2712        }
2713        head = nullptr;
2714}
2715\end{cfa}
2716
2717
2718\begin{comment}
2719
2720\subsubsection{bench.h}
2721(\texttt{bench.hpp} is similar.)
2722
2723\lstinputlisting{evaluation/bench.h}
2724
2725\subsection{C}
2726
2727\subsubsection{c-stack.h} ~
2728
2729\lstinputlisting{evaluation/c-stack.h}
2730
2731\subsubsection{c-stack.c} ~
2732
2733\lstinputlisting{evaluation/c-stack.c}
2734
2735\subsubsection{c-pair.h} ~
2736
2737\lstinputlisting{evaluation/c-pair.h}
2738
2739\subsubsection{c-pair.c} ~
2740
2741\lstinputlisting{evaluation/c-pair.c}
2742
2743\subsubsection{c-print.h} ~
2744
2745\lstinputlisting{evaluation/c-print.h}
2746
2747\subsubsection{c-print.c} ~
2748
2749\lstinputlisting{evaluation/c-print.c}
2750
2751\subsubsection{c-bench.c} ~
2752
2753\lstinputlisting{evaluation/c-bench.c}
2754
2755\subsection{\CFA}
2756
2757\subsubsection{cfa-stack.h} ~
2758
2759\lstinputlisting{evaluation/cfa-stack.h}
2760
2761\subsubsection{cfa-stack.c} ~
2762
2763\lstinputlisting{evaluation/cfa-stack.c}
2764
2765\subsubsection{cfa-print.h} ~
2766
2767\lstinputlisting{evaluation/cfa-print.h}
2768
2769\subsubsection{cfa-print.c} ~
2770
2771\lstinputlisting{evaluation/cfa-print.c}
2772
2773\subsubsection{cfa-bench.c} ~
2774
2775\lstinputlisting{evaluation/cfa-bench.c}
2776
2777\subsection{\CC}
2778
2779\subsubsection{cpp-stack.hpp} ~
2780
2781\lstinputlisting[language=c++]{evaluation/cpp-stack.hpp}
2782
2783\subsubsection{cpp-print.hpp} ~
2784
2785\lstinputlisting[language=c++]{evaluation/cpp-print.hpp}
2786
2787\subsubsection{cpp-bench.cpp} ~
2788
2789\lstinputlisting[language=c++]{evaluation/cpp-bench.cpp}
2790
2791\subsection{\CCV}
2792
2793\subsubsection{object.hpp} ~
2794
2795\lstinputlisting[language=c++]{evaluation/object.hpp}
2796
2797\subsubsection{cpp-vstack.hpp} ~
2798
2799\lstinputlisting[language=c++]{evaluation/cpp-vstack.hpp}
2800
2801\subsubsection{cpp-vstack.cpp} ~
2802
2803\lstinputlisting[language=c++]{evaluation/cpp-vstack.cpp}
2804
2805\subsubsection{cpp-vprint.hpp} ~
2806
2807\lstinputlisting[language=c++]{evaluation/cpp-vprint.hpp}
2808
2809\subsubsection{cpp-vbench.cpp} ~
2810
2811\lstinputlisting[language=c++]{evaluation/cpp-vbench.cpp}
2812\end{comment}
2813
2814\end{document}
2815
2816% Local Variables: %
2817% tab-width: 4 %
2818% compile-command: "make" %
2819% End: %
Note: See TracBrowser for help on using the repository browser.