Ignore:
Timestamp:
Mar 18, 2019, 2:32:26 PM (5 years ago)
Author:
tdelisle <tdelisle@…>
Branches:
ADT, arm-eh, ast-experimental, cleanup-dtors, enum, forall-pointer-decay, jacob/cs343-translation, jenkins-sandbox, master, new-ast, new-ast-unique-expr, pthread-emulation, qualifiedEnum
Children:
78e109f
Parents:
76d3ca6 (diff), a927662 (diff)
Note: this is a merge changeset, the changes displayed below correspond to the merge itself.
Use the (diff) links above to see all the changes relative to each parent.
Message:

Merge branch 'master' of plg.uwaterloo.ca:software/cfa/cfa-cc

File:
1 edited

Legend:

Unmodified
Added
Removed

  • doc/theses/aaron_moss_PhD/phd/background.tex

    r76d3ca6 r466b1c9  
    11\chapter{\CFA{}}
     2\label{cfa-chap}
    23
    34\CFA{} adds a number of features to C, some of them providing significant increases to the expressive power of the language, but all designed to maintain the existing procedural programming paradigm of C and to be as orthogonal as possible to each other.
    45To provide background for the contributions in subsequent chapters, this chapter provides a summary of the features of \CFA{} at the time this work was conducted.
    56
    6 The core design of \CFA{} is laid out in Glen Ditchfield's 1992 PhD thesis, \emph{Contextual Polymorphism}\cite{Ditchfield92}; in that thesis, Ditchfield presents the theoretical underpinnings of the \CFA{} polymorphism model.
    7 Building on Ditchfield's design for contextual polymorphism as well as KW-C\cite{Buhr94a}, an earlier set of (largely syntactic) extensions to C, Richard Bilson\cite{Bilson03} built the first version of the \CFA{} compiler, \CFACC{}, in the early 2000's.
    8 This early \CFACC{} provided basic functionality, but incorporated a number of poor algorithmic choices due to a rushed implementation time frame, and as such lacked the runtime performance required for practical use; this thesis is substantially concerned with rectifying those deficits.
    9 
    10 The \CFA{} project was revived in 2015 with the intention of building a production-ready language and compiler; at the time of this writing, both \CFA{} and \CFACC{} have been under active development continuously since.
    11 As this development has been proceeding concurrently with the work described in this thesis, the state of \CFA{} has been somewhat of a moving target; however, Moss~\etal\cite{Moss18} provides a reasonable summary of the current design.
    12 Notable features added during this period include generic types (Chapter~\ref{generic-chap}), constructors and destructors\cite{Schluntz17}, improved support for tuples\cite{Schluntz17}, reference types\cite{Moss18}, first-class concurrent and parallel programming support\cite{Delisle18}, as well as numerous pieces of syntactic sugar and the start of an idiomatic standard library\cite{Moss18}.
    13 
    14 \section{\CFA{} Features}
     7The core design of \CFA{} is laid out in Glen Ditchfield's 1992 PhD thesis, \emph{Contextual Polymorphism} \cite{Ditchfield92}; in that thesis, Ditchfield presents the theoretical underpinnings of the \CFA{} polymorphism model.
     8Building on Ditchfield's design for contextual polymorphism as well as KW-C \cite{Buhr94a}, an earlier set of (largely syntactic) extensions to C, Richard Bilson \cite{Bilson03} built the first version of the \CFA{} compiler, \CFACC{}, in the early 2000's.
     9This early \CFACC{} provided basic functionality, but incorporated a number of algorithmic choices that have failed to scale as \CFA{} has developed, lacking the runtime performance for practical use; this thesis is substantially concerned with rectifying those deficits.
     10
     11The \CFA{} project was revived in 2015 with the intention of building a production-ready language and compiler; at the time of this writing, both \CFA{} and \CFACC{} remain under active development.
     12As this development has been proceeding concurrently with the work described in this thesis, the state of \CFA{} has been somewhat of a moving target; however, Moss~\etal~\cite{Moss18} provides a reasonable summary of the current design.
     13Notable features added during this period include generic types (Chapter~\ref{generic-chap}), constructors and destructors \cite{Schluntz17}, improved support for tuples \cite{Schluntz17}, reference types \cite{Moss18}, first-class concurrent and parallel programming support \cite{Delisle18}, as well as numerous pieces of syntactic sugar and the start of an idiomatic standard library \cite{Moss18}.
    1514
    1615The selection of features presented in this chapter are chosen to elucidate the design constraints of the work presented in this thesis.
    17 In some cases the interactions of multiple features make this design a significantly more complex problem than any individual feature would; in other cases a feature that does not by itself add any complexity to expression resolution triggers previously rare edge cases more frequently.
    18 
    19 \subsection{Procedural Paradigm}
     16In some cases the interactions of multiple features make this design a significantly more complex problem than any individual feature; in other cases a feature that does not by itself add any complexity to expression resolution triggers previously rare edge cases more frequently.
     17
     18\section{Procedural Paradigm}
    2019
    2120It is important to note that \CFA{} is not an object-oriented language.
    22 This is a deliberate choice intended to maintain the applicability of the mental model and language idioms already possessed by C programmers.
    23 This choice is in marked contrast to \CC{}, which, though it has backward-compatibility with C on the source code level, is a much larger and more complex language, and requires extensive developer re-training to write idiomatic, efficient code in \CC{}'s object-oriented paradigm.
    24 
    25 \CFA{} does have a system of implicit type conversions derived from C's ``usual arithmetic conversions''; while these conversions may be thought of as something like an inheritance hierarchy, the underlying semantics are significantly different and such an analogy is loose at best.
     21This is a deliberate choice intended to maintain the applicability of the programming model and language idioms already possessed by C programmers.
     22This choice is in marked contrast to \CC{}, which is a much larger and more complex language, and requires extensive developer re-training to write idiomatic, efficient code in \CC{}'s object-oriented paradigm.
     23
    2624Particularly, \CFA{} has no concept of \emph{subclass}, and thus no need to integrate an inheritance-based form of polymorphism with its parametric and overloading-based polymorphism.
    27 The graph structure of the \CFA{} type conversions is also markedly different than an inheritance hierarchy; it has neither a top nor a bottom type, and does not satisfy the lattice properties typical of inheritance hierarchies.
     25While \CFA{} does have a system of implicit type conversions derived from C's ``usual arithmetic conversions'' \cite[\S{}6.3.1.8]{C11} and these conversions may be thought of as something like an inheritance hierarchy, the underlying semantics are significantly different and such an analogy is loose at best.
     26The graph structure of the \CFA{} type conversions (discussed in Section~\ref{conv-cost-sec}) is also markedly different than an inheritance hierarchy; it has neither a top nor a bottom type, and does not satisfy the lattice properties typical of inheritance hierarchies.
    2827
    2928Another aspect of \CFA{}'s procedural paradigm is that it retains C's translation-unit-based encapsulation model, rather than class-based encapsulation such as in \CC{}.
    30 This choice implies that that separate compilation must be maintained to allow headers to act as an encapsulation boundary, rather than the header-only libraries used by \CC{} templates.
    31 
    32 \subsection{Name Overloading} \label{overloading-sec}
    33 
    34 In C, no more than one variable or function in the same scope may share the same name\footnote{Technically, C has multiple separated namespaces, one holding \lstinline{struct}, \lstinline{union}, and \lstinline{enum} tags, one holding labels, one holding \lstinline{typedef} names, variable, function, and enumerator identifiers, and one for each \lstinline{struct} and \lstinline{union} type holding the field names\cit{}.}, and variable or function declarations in inner scopes with the same name as a declaration in an outer scope hide the outer declaration.
    35 This restriction makes finding the proper declaration to match to a variable expression or function application a simple matter of symbol-table lookup, which can be easily and efficiently implemented.
     29This design choice implies that separate compilation must be maintained to allow headers to act as an encapsulation boundary, rather than the header-only libraries used by \CC{} templates.
     30
     31\section{Name Overloading} \label{overloading-sec}
     32
     33In C, no more than one variable or function in the same scope may share the same name\footnote{Technically, C has multiple separated namespaces, one holding \lstinline{struct}, \lstinline{union}, and \lstinline{enum} tags, one holding labels, one holding \lstinline{typedef} names, variable, function, and enumerator identifiers, and one for each \lstinline{struct} and \lstinline{union} type holding the field names \cite[\S{}6.2.3]{C11}.}, and variable or function declarations in inner scopes with the same name as a declaration in an outer scope hide the outer declaration.
     34This restriction makes finding the proper declaration to match to a variable expression or function application a simple matter of lexically-scoped name lookup, which can be easily and efficiently implemented.
    3635\CFA{}, on the other hand, allows overloading of variable and function names so long as the overloaded declarations do not have the same type, avoiding the multiplication of variable and function names for different types common in the C standard library, as in the following example:
    3736
    … …  
    5049\end{cfa}
    5150
    52 While the wisdom of giving both the maximum value of a type and the function to take the maximum of two values the same name is debatable, \eg{} some developers may prefer !MAX! for the former, the guiding philosophy of \CFA{} is ``describe, don't prescribe'' --- we prefer to trust programmers with powerful tools, and there is no technical reason to restrict overloading between variables and functions.
    53 However, the expressivity of \CFA{}'s name overloading has the consequence that simple table lookup is insufficient to match identifiers to declarations, and a type-matching algorithm must be part of expression resolution.
    54 
    55 \subsubsection{Operator Overloading}
     51The final expression in the preceding example includes a feature of \CFA{} name overloading not shared by \CC{}, the ability to disambiguate expressions based on their return type. This provides greater flexibility and power than the parameter-based overload selection of \CC{}, though at the cost of greater complexity in the resolution algorithm.
     52
     53While the wisdom of giving both the maximum value of a type and the function to take the maximum of two values the same name in the example above is debatable, \eg{} some developers may prefer !MAX! for the former, the guiding philosophy of \CFA{} is ``describe, don't prescribe'' --- we prefer to trust programmers with powerful tools, and there is no technical reason to restrict overloading between variables and functions.
     54However, the expressivity of \CFA{}'s name overloading does have the consequence that simple table lookup is insufficient to match identifiers to declarations, and a type-matching algorithm must be part of expression resolution.
     55
     56\subsection{Operator Overloading}
    5657
    5758C does allow name overloading in one context: operator overloading.
    5859From the perspective of the type system, there is nothing special about operators as opposed to other functions, nor is it desirable to restrict the clear and readable syntax of operators to only the built-in types.
    59 For these reasons, \CFA{} also allows overloading of operators by writing specially-named functions where !?! stands in for the operands\footnote{This example uses \CFA{}'s reference types, described in Section~\ref{type-features-sec}}:
     60For these reasons, \CFA{}, like \CC{} and many other programming languages, allows overloading of operators by writing specially-named functions where !?! stands in for the operands.
     61This syntax is more natural than the operator overloading syntax of \CC{}, which requires ``dummy'' parameters to disambiguate overloads of similarly-named pre- and postfix operators\footnote{This example uses \CFA{}'s reference types, described in Section~\ref{type-features-sec}}:
    6062
    6163\begin{cfa}
    … …  
    7173\end{cfa}
    7274
    73 Together, \CFA{}'s backward-compatibility with C and the inclusion of this operator overloading feature imply that \CFA{} must select among function overloads using a method compatible with C's ``usual arithmetic conversions''\cit{}, so as to present user programmers with only a single set of overloading rules.
    74 
    75 \subsubsection{Special Literal Types}
     75Together, \CFA{}'s backward-compatibility with C and the inclusion of this operator overloading feature imply that \CFA{} must select among function overloads using a method compatible with C's ``usual arithmetic conversions'' \cite[\S{}6.3.1.8]{C11}, so as to present user programmers with only a single set of overloading rules.
     76
     77\subsection{Special Literal Types}
    7678
    7779Literal !0! is also used polymorphically in C; it may be either integer zero or the null value of any pointer type.
    78 \CFA{} provides a special type for the !0! literal, !zero_t!, so that users can define a zero value for their own types without being forced to create a conversion from an integer or pointer type (though \CFA{} also includes implicit conversions from !zero_t! to the integer and pointer types for backward compatibility).
    79 
    80 According to the C standard\cit{}, !0! is the only false value; any value that compares equal to zero is false, while any value that does not is true.
    81 By this rule, boolean contexts such as !if ( x )! can always be equivalently rewritten as \lstinline{if ( (x) != 0 )}.
    82 \CFACC{} applies this rewriting in all boolean contexts, so any type !T! can be made ``truthy'' (that is, given a boolean interpretation) in \CFA{} by defining an operator overload \lstinline{int ?!=?(T, zero_t)}; unlike \CC{} prior to the addition of explicit casts in \CCeleven{}, this design does not add comparability or convertablity to arbitrary integer types.
     80\CFA{} provides a special type for the !0! literal, !zero_t!, so that users can define a zero value for their own types without being forced to create a conversion from an integer or pointer type; \CFA{} also includes implicit conversions from !zero_t! to the !int! and pointer type constructors\footnote{See Section~\ref{type-features-sec}} from !zero_t! for backward compatibility.
     81
     82According to the C standard \cite[\S{}6.8.4.1]{C11}, !0! is the only false value; any value that compares equal to zero is false, while any value that does not is true.
     83By this rule, Boolean contexts such as !if ( x )! can always be equivalently rewritten as \lstinline{if ( (x) != 0 )}.
     84\CFACC{} applies this rewriting in all Boolean contexts, so any type !T! can be made ``truthy'' (that is, given a Boolean interpretation) in \CFA{} by defining an operator overload \lstinline{int ?!=?(T, zero_t)}.
     85\CC{} takes a different approach to user-defined truthy types, allowing definition of an implicit conversion operator to !bool!; prior to the introduction of the !explicit! keyword for conversion operators in \CCeleven{} this approach also allowed undesired implicit conversions to all other arithmetic types, a shortcoming not shared by the \CFA{} design.
    8386
    8487\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.
    … …  
    8689
    8790\CFA{} previously allowed !0! and !1! to be the names of polymorphic variables, with separate overloads for !int 0!, !int 1!, and !forall(dtype T) T* 0!.
    88 As revealed in my own work on generic types (Chapter~\ref{generic-chap}), the parameteric polymorphic zero variable was not generalizable to other types; though all null pointers have the same in-memory representation, the same cannot be said of the zero values of arbitrary types.
     91While designing \CFA{} generic types (see Chapter~\ref{generic-chap}), it was discovered that the parametric polymorphic zero variable is not generalizable to other types; though all null pointers have the same in-memory representation, the same cannot be said of the zero values of arbitrary types.
    8992As such, variables that could represent !0! and !1! were phased out in favour of functions that could generate those values for a given type as appropriate.
    9093
    91 \subsection{Polymorphic Functions} \label{poly-func-sec}
    92 
    93 The most significant feature \CFA{} adds is parametric-polymorphic functions.
     94\section{Polymorphic Functions} \label{poly-func-sec}
     95
     96The most significant type-system feature \CFA{} adds is parametric-polymorphic functions.
    9497Such functions are written using a !forall! clause (which gives the language its name):
    9598
    … …  
    102105The type variable !T! is transformed into a set of additional implicit parameters to !identity!, which encode sufficient information about !T! to create and return a variable of that type.
    103106\CFA{} passes the size and alignment of the type represented by an !otype! parameter, as well as a default constructor, copy constructor, assignment operator, and destructor.
    104 Types which do not have one or more of these operators visible cannot be bound to !otype! parameters.
    105 In this design, the runtime cost of polymorphism is spread over each polymorphic call, due to passing more arguments to polymorphic functions; experiments have shown this overhead to be similar to \CC{} virtual function calls. \TODO{rerun experiments, possibly look at vtable variant}
     107Types which do not have one or more of these operators visible cannot be bound to !otype! parameters, but may be bound to un-constrained !dtype! (``data type'') type variables.
     108In this design, the runtime cost of polymorphism is spread over each polymorphic call, due to passing more arguments to polymorphic functions; the experiments in Chapter~\ref{generic-chap} indicate that this overhead is comparable to that of \CC{} virtual function calls.
     109% \TODO{rerun experiments, possibly look at vtable variant}
    106110
    107111One benefit of this design is that it allows polymorphic functions to be separately compiled.
    … …  
    109113The fact that there is only one implementation of each polymorphic function also reduces compile times relative to the template-expansion approach taken by \CC{}, as well as reducing binary sizes and runtime pressure on instruction cache by re-using a single version of each function.
    110114
    111 \subsubsection{Type Assertions}
    112 
    113 Since bare polymorphic types do not provide a great range of available operations, \CFA{} provides a \emph{type assertion} mechanism to provide further information about a type:
     115\subsection{Type Assertions}
     116
     117Since bare polymorphic types do not provide a great range of available operations, \CFA{} provides a \emph{type assertion} mechanism to provide further information about a type\footnote{Subscript not included in source code.\label{sub-foot}}:
    114118
    115119\begin{cfa}
    116120forall(otype T `| { T twice(T); }`)
    117121T four_times(T x) { return twice( twice(x) ); }
    118 double twice(double d) { return d * 2.0; }  $\C[2.75in]{// (1)}$
     122double twice$\(_1\)$(double d) { return d * 2.0; }
    119123
    120124double ans = four_times( 10.5 );  $\C[2.75in]{// T bound to double, ans == 42.0}$
    … …  
    125129
    126130Monomorphic specializations of polymorphic functions can themselves be used to satisfy type assertions.
    127 For instance, !twice! could have been defined like this:
     131For instance, !twice! could have been defined like this\footref{sub-foot}:
    128132
    129133\begin{cfa}
    130134forall(otype S | { S ?+?(S, S); })
    131 S twice(S x) { return x + x; }  $\C[2.75in]{// (2)}
    132 \end{cfa}
    133 
    134 This version of !twice! works for any type !S! that has an addition operator defined for it, and it could be used to satisfy the type assertion on !four_times!.
    135 \CFACC{} accomplishes this by creating a wrapper function calling !twice//(2)! with !S! bound to !double!, then providing this wrapper function to !four_times!\footnote{\lstinline{twice // (2)} could also have had a type parameter named \lstinline{T}; \CFA{} specifies renaming of the type parameters, which would avoid the name conflict with the type variable \lstinline{T} of \lstinline{four_times}}.
     135S twice$\(_2\)$(S x) { return x + x; }
     136\end{cfa}
     137
     138Specializing this polymorphic function with !S = double! produces a monomorphic function which can  be used to satisfy the type assertion on !four_times!.
     139\CFACC{} accomplishes this by creating a wrapper function calling !twice!$_2$ with !S! bound to !double!, then providing this wrapper function to !four_times!\footnote{\lstinline{twice}$_2$ could also have had a type parameter named \lstinline{T}; \CFA{} specifies renaming of the type parameters, which would avoid the name conflict with the type variable \lstinline{T} of \lstinline{four_times}}.
     140However, !twice!$_2$ also works for any type !S! that has an addition operator defined for it.
    136141
    137142Finding appropriate functions to satisfy type assertions is essentially a recursive case of expression resolution, as it takes a name (that of the type assertion) and attempts to match it to a suitable declaration in the current scope.
    138143If a polymorphic function can be used to satisfy one of its own type assertions, this recursion may not terminate, as it is possible that that function is examined as a candidate for its own assertion unboundedly repeatedly.
    139 To avoid such infinite loops, \CFACC{} imposes a fixed limit on the possible depth of recursion, similar to that employed by most \CC{} compilers for template expansion; this restriction means that there are some semantically well-typed expressions that cannot be resolved by \CFACC{}.
    140 \TODO{Update this with final state} One contribution made in the course of this thesis was modifying \CFACC{} to use the more flexible expression resolution algorithm for assertion matching, rather than the simpler but limited previous approach of unification on the types of the functions.
    141 
    142 \subsubsection{Deleted Declarations}
    143 
    144 Particular type combinations can be exempted from matching a given polymorphic function through use of a \emph{deleted function declaration}:
    145 
    146 \begin{cfa}
    147 int somefn(char) = void;
    148 \end{cfa}
    149 
    150 This feature is based on a \CCeleven{} feature typically used to make a type non-copyable by deleting its copy constructor and assignment operator\footnote{In previous versions of \CC{}, a type could be made non-copyable by declaring a private copy constructor and assignment operator, but not defining either. This idiom is well-known, but depends on some rather subtle and \CC{}-specific rules about private members and implicitly-generated functions; the deleted function form is both clearer and less verbose.} or forbidding some interpretations of a polymorphic function by specifically deleting the forbidden overloads\footnote{Specific polymorphic function overloads can also be forbidden in previous \CC{} versions through use of template metaprogramming techniques, though this advanced usage is beyond the skills of many programmers. A similar effect can be produced on an ad-hoc basis at the appropriate call sites through use of casts to determine the function type. In both cases, the deleted-function form is clearer and more concise.}.
    151 Deleted function declarations are implemented in \CFACC{} by adding them to the symbol table as usual, but with a flag set that indicates that the function is deleted.
    152 If this deleted declaration is selected as the unique minimal-cost interpretation of an expression than an error is produced. \TODO{Check this is implemented at print.}
    153 
    154 \subsubsection{Traits}
    155 
    156 \CFA{} provides \emph{traits} as a means to name a group of type assertions, as in the example below\footnote{This example uses \CFA{}'s reference types, constructors, and zero type, described in Section~\ref{type-features-sec}.}:
     144To avoid such infinite loops, \CFACC{} imposes a fixed limit on the possible depth of recursion, similar to that employed by most \CC{} compilers for template expansion; this restriction means that there are some otherwise semantically well-typed expressions that cannot be resolved by \CFACC{}.
     145
     146% \subsection{Deleted Declarations}
     147
     148% Particular type combinations can be exempted from matching a given polymorphic function through use of a \emph{deleted function declaration}:
     149
     150% \begin{cfa}
     151% int somefn(char) = void;
     152% \end{cfa}
     153
     154% This feature is based on a \CCeleven{} feature typically used to make a type non-copyable by deleting its copy constructor and assignment operator\footnote{In previous versions of \CC{}, a type could be made non-copyable by declaring a private copy constructor and assignment operator, but not defining either. This idiom is well-known, but depends on some rather subtle and \CC{}-specific rules about private members and implicitly-generated functions; the deleted function form is both clearer and less verbose.} or forbidding some interpretations of a polymorphic function by specifically deleting the forbidden overloads\footnote{Specific polymorphic function overloads can also be forbidden in previous \CC{} versions through use of template metaprogramming techniques, though this advanced usage is beyond the skills of many programmers. A similar effect can be produced on an ad-hoc basis at the appropriate call sites through use of casts to determine the function type. In both cases, the deleted-function form is clearer and more concise.}.
     155% Deleted function declarations are implemented in \CFACC{} by adding them to the symbol table as usual, but with a flag set that indicates that the function is deleted.
     156% If this deleted declaration is selected as the unique minimal-cost interpretation of an expression than an error is produced.
     157
     158\subsection{Traits}
     159
     160\CFA{} provides \emph{traits} as a means to name a group of type assertions, as in the example below\footnote{This example uses \CFA{}'s reference types and constructors, described in Section~\ref{type-features-sec}.}:
    157161
    158162\begin{cfa}
    … …  
    171175
    172176Semantically, traits are simply a named list of type assertions, but they may be used for many of the same purposes that interfaces in Java or abstract base classes in \CC{} are used for.
    173 Unlike Java interfaces or \CC{} base classes, \CFA{} types do not explicitly state any inheritance relationship to traits they satisfy; this can be considered a form a structural inheritance, similar to interface implementation in Go, as opposed to the nominal inheritance model of Java and \CC{}.
     177Unlike Java interfaces or \CC{} base classes, \CFA{} types do not explicitly state any inheritance relationship to traits they satisfy; this can be considered a form of structural inheritance, similar to interface implementation in Go, as opposed to the nominal inheritance model of Java and \CC{}.
    174178Nominal inheritance can be simulated in \CFA{} using marker variables in traits:
    175179
    … …  
    186190
    187191Traits, however, are significantly more powerful than nominal-inheritance interfaces; firstly, due to the scoping rules of the declarations that satisfy a trait's type assertions, a type may not satisfy a trait everywhere that that type is declared, as with !char! and the !nominal! trait above.
    188 Secondly, because \CFA{} is not object-oriented and types do not have a closed set of methods, existing C library types can be extended to implement a trait simply by writing the requisite functions.
     192Secondly, because \CFA{} is not object-oriented and types do not have a closed set of methods, existing C library types can be extended to implement a trait simply by writing the requisite functions\footnote{\CC{} only allows partial extension of C types, because constructors, destructors, conversions, and the assignment, indexing, and function-call operators may only be defined in a class; \CFA{} does not have any of these restrictions.}.
    189193Finally, traits may be used to declare a relationship among multiple types, a property that may be difficult or impossible to represent in nominal-inheritance type systems\footnote{This example uses \CFA{}'s reference types, described in Section~\ref{type-features-sec}.}:
    190194
    … …  
    206210\end{cfa}
    207211
    208 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.
     212In this 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.
    209213Given 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 !int!, while !(*it).value = 42;! interprets !*it! as !list!).
    210214While 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.
    211215
    212216The flexibility of \CFA{}'s implicit trait-satisfaction mechanism provides programmers with a great deal of power, but also blocks some optimization approaches for expression resolution.
    213 The ability of types to begin or cease to satisfy traits when declarations go into or out of scope makes caching of trait satisfaction judgements difficult, and the ability of traits to take multiple type parameters can lead to a combinatorial explosion of work in any attempt to pre-compute trait satisfaction relationships.
    214 
    215 \subsection{Implicit Conversions} \label{implicit-conv-sec}
    216 
    217 In addition to the multiple interpretations of an expression produced by name overloading and polymorphic functions, for backward compatibility \CFA{} must support all of the implicit conversions present in C, producing further candidate interpretations for expressions.
    218 As mentioned above, C does not have an inheritance hierarchy of types, but the C standard's rules for the ``usual arithmetic conversions'\cit{} define which of the built-in types are implicitly convertible to which other types, and the relative cost of any pair of such conversions from a single source type.
     217The ability of types to begin or cease to satisfy traits when declarations go into or out of scope makes caching of trait satisfaction judgments difficult, and the ability of traits to take multiple type parameters can lead to a combinatorial explosion of work in any attempt to pre-compute trait satisfaction relationships.
     218
     219\section{Implicit Conversions} \label{implicit-conv-sec}
     220
     221In addition to the multiple interpretations of an expression produced by name overloading and polymorphic functions, \CFA{} must support all of the implicit conversions present in C for backward compatibility, producing further candidate interpretations for expressions.
     222As mentioned above, C does not have an inheritance hierarchy of types, but the C standard's rules for the ``usual arithmetic conversions'' \cite[\S{}6.3.1.8]{C11} define which of the built-in types are implicitly convertible to which other types, as well as which implicit conversions to apply for mixed arguments to binary operators.
    219223\CFA{} adds rules to the usual arithmetic conversions defining the cost of binding a polymorphic type variable in a function call; such bindings are cheaper than any \emph{unsafe} (narrowing) conversion, \eg{} !int! to !char!, but more expensive than any \emph{safe} (widening) conversion, \eg{} !int! to !double!.
    220 One contribution of this thesis, discussed in Section \TODO{add to resolution chapter}, is a number of refinements to this cost model to more efficiently resolve polymorphic function calls.
    221 
    222 The expression resolution problem which is the focus of Chapter~\ref{resolution-chap} is to find the unique minimal-cost interpretation of each expression in the program, where all identifiers must be matched to a declaration, and implicit conversions or polymorphic bindings of the result of an expression may increase the cost of the expression.
    223 While semantically valid \CFA{} code always has such a unique minimal-cost interpretation, \CFACC{} must also be able to detect and report as errors expressions which have either no interpretation or multiple ambiguous minimal-cost interpretations.
     224One contribution of this thesis, discussed in Section~\ref{conv-cost-sec}, is a number of refinements to this cost model to more efficiently resolve polymorphic function calls.
     225
     226The expression resolution problem, which is the focus of Chapter~\ref{resolution-chap}, is to find the unique minimal-cost interpretation of each expression in the program, where all identifiers must be matched to a declaration, and implicit conversions or polymorphic bindings of the result of an expression may increase the cost of the expression.
     227While semantically valid \CFA{} code always has such a unique minimal-cost interpretation, \CFACC{} must also be able to detect and report as errors expressions that have either no interpretation or multiple ambiguous minimal-cost interpretations.
    224228Note that which subexpression interpretation is minimal-cost may require contextual information to disambiguate.
    225229For instance, in the example in Section~\ref{overloading-sec}, !max(max, -max)! cannot be unambiguously resolved, but !int m = max(max, -max)! has a single minimal-cost resolution.
    226 While the interpretation !int m = (int)max((double)max, -(double)max)! is also a valid interpretation, it is not minimal-cost due to the unsafe cast from the !double! result of !max! to !int!\footnote{The two \lstinline{double} casts function as type ascriptions selecting \lstinline{double max} rather than casts from \lstinline{int max} to \lstinline{double}, and as such are zero-cost.}.
     230While the interpretation !int m = (int)max((double)max, -(double)max)! is also a valid interpretation, it is not minimal-cost due to the unsafe cast from the !double! result of !max! to !int!\footnote{The two \lstinline{double} casts function as type ascriptions selecting \lstinline{double max} rather than casts from \lstinline{int max} to \lstinline{double}, and as such are zero-cost. The \lstinline{int} to \lstinline{double} conversion could be forced if desired with two casts: \lstinline{(double)(int)max}}.
    227231These contextual effects make the expression resolution problem for \CFA{} both theoretically and practically difficult, but the observation driving the work in Chapter~\ref{resolution-chap} is that of the many top-level expressions in a given program, most are straightforward and idiomatic so that programmers writing and maintaining the code can easily understand them; it follows that effective heuristics for common cases can bring down compiler runtime enough that a small proportion of harder-to-resolve expressions does not inordinately increase overall compiler runtime or memory usage.
    228232
    229 \subsection{Type Features} \label{type-features-sec}
    230 
    231 The name overloading and polymorphism features of \CFA{} have the greatest effect on language design and compiler runtime, but there are a number of other features in the type system which have a smaller effect but are useful for code examples.
     233\section{Type Features} \label{type-features-sec}
     234
     235The name overloading and polymorphism features of \CFA{} have the greatest effect on language design and compiler runtime, but there are a number of other features in the type system that have a smaller effect but are useful for code examples.
    232236These features are described here.
    233237
    234 \subsubsection{Reference Types}
    235 
    236 One of the key ergonomic improvements in \CFA{} is reference types, designed and implemented by Robert Schluntz\cite{Schluntz17}.
     238\subsection{Reference Types}
     239
     240One of the key ergonomic improvements in \CFA{} is reference types, designed and implemented by Robert Schluntz \cite{Schluntz17}.
    237241Given some type !T!, a !T&! (``reference to !T!'') is essentially an automatically dereferenced pointer.
    238 These types allow seamless pass-by-reference for function parameters, without the extraneous dereferencing syntax present in C; they also allow easy easy aliasing of nested values with a similarly convenient syntax.
    239 A particular improvement is removing syntactic special cases for operators which take or return mutable values; for example, the use !a += b! of a compound assignment operator now matches its signature, !int& ?+=?(int&, int)!, as opposed to the previous syntactic special cases to automatically take the address of the first argument to !+=! and to mark its return value as mutable.
    240 
    241 The C standard makes heavy use of the concept of \emph{lvalue}, an expression with a memory address; its complement, \emph{rvalue} (a non-addressable expression) is not explicitly named.
    242 In \CFA{}, the distinction between lvalue and rvalue can be reframed in terms of reference and non-reference types, with the benefit of being able to express the difference in user code.
     242These types allow seamless pass-by-reference for function parameters, without the extraneous dereferencing syntax present in C; they also allow easy aliasing of nested values with a similarly convenient syntax.
     243A particular improvement is removing syntactic special cases for operators that take or return mutable values; for example, the use !a += b! of a compound assignment operator now matches its signature, !int& ?+=?(int&, int)!, as opposed to the syntactic special cases in earlier versions of \CFA{} to automatically take the address of the first argument to !+=! and to mark its return value as mutable.
     244
     245The C standard makes heavy use of the concept of \emph{lvalue}, an expression with a memory address; its complement, \emph{rvalue} (a non-addressable expression) is not explicitly named in the standard.
     246In \CFA{}, the distinction between lvalue and rvalue can be re-framed in terms of reference and non-reference types, with the benefit of being able to express the difference in user code.
    243247\CFA{} references preserve the existing qualifier-dropping implicit lvalue-to-rvalue conversion from C (\eg{} a !const volatile int&! can be implicitly copied to a bare !int!)
    244 To make reference types more easily usable in legacy pass-by-value code, \CFA{} also adds an implicit rvalue-to-lvalue conversion, implemented by storing the value in a fresh compiler-generated temporary variable and passing a reference to that temporary.
    245 To mitigate the ``!const! hell'' problem present in \CC{}, there is also a qualifier-dropping lvalue-to-lvalue conversion, also implemented by copying into a temporary:
     248To make reference types more easily usable in legacy pass-by-value code, \CFA{} also adds an implicit rvalue-to-lvalue conversion, implemented by storing the value in a compiler-generated temporary variable and passing a reference to that temporary.
     249To mitigate the ``!const! hell'' problem present in \CC{}, there is also a qualifier-dropping lvalue-to-lvalue conversion implemented by copying into a temporary:
    246250
    247251\begin{cfa}
    … …  
    259263The primary issue with \CC{} references is that it is impossible to extract the address of the reference variable rather than the address of the referred-to variable.
    260264This breaks a number of the usual compositional properties of the \CC{} type system, \eg{} a reference cannot be re-bound to another variable, nor is it possible to take a pointer to, array of, or reference to a reference.
    261 \CFA{} supports all of these use cases \TODO{test array} without further added syntax.
     265\CFA{} supports all of these use cases without further added syntax.
    262266The key to this syntax-free feature support is an observation made by the author that the address of a reference is a lvalue.
    263267In C, the address-of operator !&x! can only be applied to lvalue expressions, and always produces an immutable rvalue; \CFA{} supports reference re-binding by assignment to the address of a reference, and pointers to references by repeating the address-of operator:
    … …  
    272276For better compatibility with C, the \CFA{} team has chosen not to differentiate function overloads based on top-level reference types, and as such their contribution to the difficulty of \CFA{} expression resolution is largely restricted to the implementation details of normalization conversions and adapters.
    273277
    274 \subsubsection{Resource Management}
    275 
    276 \CFA{} also supports the RAII (``Resource Acquisition is Initialization'') idiom originated by \CC{}, thanks to the object lifetime work of Robert Schluntz\cite{Schluntz17}.
     278\subsection{Resource Management}
     279
     280\CFA{} also supports the RAII (``Resource Acquisition is Initialization'') idiom originated by \CC{}, thanks to the object lifetime work of Robert Schluntz \cite{Schluntz17}.
    277281This idiom allows a safer and more principled approach to resource management by tying acquisition of a resource to object initialization, with the corresponding resource release executed automatically at object finalization.
    278282A wide variety of conceptual resources may be conveniently managed by this scheme, including heap memory, file handles, and software locks.
    … …  
    325329\end{cfa}
    326330
    327 \subsubsection{Tuple Types}
     331\subsection{Tuple Types}
    328332
    329333\CFA{} adds \emph{tuple types} to C, a syntactic facility for referring to lists of values anonymously or with a single identifier.
    330334An identifier may name a tuple, a function may return one, and a tuple may be implicitly \emph{destructured} into its component values.
    331335The implementation of tuples in \CFACC{}'s code generation is based on the generic types introduced in Chapter~\ref{generic-chap}, with one compiler-generated generic type for each tuple arity.
    332 This allows tuples to take advantage of the same runtime optimizations available to generic types, while reducing code bloat.
    333 An extended presentation of the tuple features of \CFA{} can be found in \cite{Moss18}, but the following example shows the basics:
    334 
    335 \begin{cfa}
    336 [char, char] x = ['!', '?']; $\C{// (1); tuple type and expression syntax}$
    337 int x = 2; $\C{// (2)}$
     336This reuse allows tuples to take advantage of the same runtime optimizations available to generic types, while reducing code bloat.
     337An extended presentation of the tuple features of \CFA{} can be found in \cite{Moss18}, but the following example demonstrates the basic features:
     338
     339\begin{cfa}
     340[char, char] x$\(_1\)$ = ['!', '?']; $\C{// tuple type and expression syntax}$
     341int x$\(_2\)$ = 2;
    338342
    339343forall(otype T)
    340 [T, T] swap( T a, T b ) { $\C{// (3)}$
     344[T, T] swap$\(_1\)$( T a, T b ) {
    341345        return [b, a]; $\C{// one-line swap syntax}$
    342346}
    343347
    344 x = swap( x ); $\C{// destructure [char, char] x into two elements}$
    345 $\C{// cannot use int x, not enough arguments}$
    346 
    347 void swap( int, char, char ); $\C{// (4)}$
    348 
    349 swap( x, x ); $\C{// (4) on (2), (1)}$
    350 $\C{// not (3) on (2), (2) due to polymorphism cost}$
     348x = swap( x ); $\C{// destructure x\(_1\) into two elements}$
     349$\C{// cannot use x\(_2\), not enough arguments}$
     350
     351void swap$\(_2\)$( int, char, char );
     352
     353swap( x, x ); $\C{// swap\(_2\)( x\(_2\), x\(_1\) )}$
     354$\C{// not swap\(_1\)( x\(_2\), x\(_2\) ) due to polymorphism cost}$
    351355\end{cfa}
    352356
    353357Tuple destructuring breaks the one-to-one relationship between identifiers and values.
    354 This precludes some argument-parameter matching strategies for expression resolution, as well as cheap interpretation filters based on comparing number of parameters and arguments.
    355 As an example, in the call to !swap( x, x )! above, the second !x! can be resolved starting at the second or third parameter of !swap!, depending which interpretation of !x! was chosen for the first argument.
     358Hence, some argument-parameter matching strategies for expression resolution are precluded, as well as cheap interpretation filters based on comparing number of parameters and arguments.
     359As an example, in the call to !swap( x, x )! above, the second !x! can be resolved starting at the second or third parameter of !swap!$_2$, depending which interpretation of !x! is chosen for the first argument.
     360
     361\section{Conclusion}
     362
     363\CFA{} adds a significant number of features to standard C, increasing the expressivity and re-usability of \CFA{} code while maintaining backwards compatibility for both code and larger language paradigms.
     364This flexibility does incur significant added compilation costs, however, the mitigation of which are the primary concern of this thesis.
Note: See TracChangeset for help on using the changeset viewer.