Diff [eaeca5f23e76ee6a6e339574fa4dcb93d881eef0:1d402bee5423cb538e4289c1671feede9350f2bd] for / – Cforall

doc/theses/andrew_beach_MMath/conclusion.tex

-              reaeca5f
+              r1d402be
 \chapter{Conclusion}
-\label{c:conclusion}
 % Just a little knot to tie the paper together.
+In the previous chapters this thesis presents the design and implementation
+of \CFA's exception handling mechanism (EHM).
+Both the design and implementation are based off of tools and
+In the previous chapters this thesis presents the design and implementation of
+\CFA's EHM.  Both the design and implementation are based off of tools and
 techniques developed for other programming languages but they were adapted to
 better fit \CFA's feature set and add a few features that do not exist in
+other EHMs;
 including conditional matching, default handlers for unhandled exceptions
 and cancellation though coroutines and threads back to the program main stack.
+better fit \CFA's feature set and add a few features that do not exist in other
+EHMs, like conditional catch, default handlers, implicitly changing resumption
+into termination in the resumption default handler, and cancellation through
+coroutines and threads back to program main.
 The resulting features cover all of the major use cases of the most popular
 …
 such as virtuals independent of traditional objects.
+The \CFA project's test suite has been expanded to test the EHM.
+The implementation's performance has also been
+compared to other implementations with a small set of targeted
+micro-benchmarks.
+The implementation has been tested through a set of small but interesting micro-benchmarks
+and compared to other implementations.
 The results, while not cutting edge, are good enough for prototyping, which
 is \CFA's current stage of development.
+This initial EHM will bring valuable new features to \CFA in its own right
+but also serves as a tool and motivation for other developments in the
+language.
+This initial EHM is a valuable new feature for \CFA in its own right but also serves
+as a tool and motivation for other developments in the language.

doc/theses/andrew_beach_MMath/existing.tex

-              reaeca5f
+              r1d402be
 Only those \CFA features pertaining to this thesis are discussed.
+% Also, only new features of \CFA will be discussed,
 A familiarity with
 C or C-like languages is assumed.
 …
 \CFA has extensive overloading, allowing multiple definitions of the same name
 to be defined~\cite{Moss18}.
 \begin{cfa}
 char i; int i; double i;
 int f(); double f();
 void g( int ); void g( double );
 \end{cfa}
+\begin{lstlisting}[language=CFA,{moredelim=**[is][\color{red}]{@}{@}}]
+char @i@; int @i@; double @i@;
+int @f@(); double @f@();
+void @g@( int ); void @g@( double );
+\end{lstlisting}
 This feature requires name mangling so the assembly symbols are unique for
 different overloads. For compatibility with names in C, there is also a syntax
 …
 int && rri = ri;
 rri = 3;
 &ri = &j;
+&ri = &j; // rebindable
 ri = 5;
 \end{cfa}
 …
 \end{minipage}
+References are intended to be used when the indirection of a pointer is
+required, but the address is not as important as the value and dereferencing
+is the common usage.
+References are intended for pointer situations where dereferencing is the common usage,
+\ie the value is more important than the pointer.
 Mutable references may be assigned to by converting them to a pointer
+with a @&@ and then assigning a pointer to them, as in @&ri = &j;@ above.
+% ???
+with a @&@ and then assigning a pointer to them, as in @&ri = &j;@ above
 \section{Operators}
 \CFA implements operator overloading by providing special names, where
 operator expressions are translated into function calls using these names.
+operator usages are translated into function calls using these names.
 An operator name is created by taking the operator symbols and joining them with
 @?@s to show where the arguments go.
 …
 This syntax make it easy to tell the difference between prefix operations
 (such as @++?@) and post-fix operations (@?++@).
+As an example, here are the addition and equality operators for a point type.
+For example, plus and equality operators are defined for a point type.
 \begin{cfa}
 point ?+?(point a, point b) { return point{a.x + b.x, a.y + b.y}; }
 …
+}
 \end{cfa}
+Note that this syntax works effectively but a textual transformation,
+the compiler converts all operators into functions and then resolves them
+normally. This means any combination of types may be used,
+although nonsensical ones (like @double ?==?(point, int);@) are discouraged.
+This feature is also used for all builtin operators as well,
+although those are implicitly provided by the language.
+Note these special names are not limited to builtin
+operators, and hence, may be used with arbitrary types.
+\begin{cfa}
+double ?+?( int x, point y ); // arbitrary types
+\end{cfa}
+% Some ``near misses", that are that do not match an operator form but looks like
+% it may have been supposed to, will generate warning but otherwise they are
+% left alone.
+Because operators are never part of the type definition they may be added
+at any time, including on built-in types.
 %\subsection{Constructors and Destructors}
+In \CFA, constructors and destructors are operators, which means they are
+functions with special operator names rather than type names in \Cpp.
+Both constructors and destructors can be implicity called by the compiler,
+however the operator names allow explicit calls.
+% Placement new means that this is actually equivant to C++.
+\CFA also provides constructors and destructors as operators, which means they
+are functions with special operator names rather than type names in \Cpp.
+While constructors and destructions are normally called implicitly by the compiler,
+the special operator names, allow explicit calls.
+% Placement new means that this is actually equivalent to C++.
 The special name for a constructor is @?{}@, which comes from the
 …
 struct Example { ... };
 void ?{}(Example & this) { ... }
+{
-        Example a;
-        Example b = {};
+}
 void ?{}(Example & this, char first, int num) { ... }
+{
         Example c = {'a', 2};
+}
 \end{cfa}
 Both @a@ and @b@ will be initalized with the first constructor,
 @b@ because of the explicit call and @a@ implicitly.
 @c@ will be initalized with the second constructor.
+Currently, there is no general way to skip initialation.
+% I don't use @= anywhere in the thesis.
+Example a;              // implicit constructor calls
+Example b = {};
+Example c = {'a', 2};
+\end{cfa}
+Both @a@ and @b@ are initialized with the first constructor,
+while @c@ is initialized with the second.
+Constructor calls can be replaced with C initialization using special operator \lstinline{@=}.
+\begin{cfa}
+Example d @= {42};
+\end{cfa}
 % I don't like the \^{} symbol but $^\wedge$ isn't better.
 Similarly, destructors use the special name @^?{}@ (the @^@ has no special
 meaning).
+% These are a normally called implicitly called on a variable when it goes out
+% of scope. They can be called explicitly as well.
 \begin{cfa}
 void ^?{}(Example & this) { ... }
+{
+        Example d;
+        ^?{}(d);
+        Example e;
+} // Implicit call of ^?{}(e);
+        Example e;      // implicit constructor call
+        ^?{}(e);                // explicit destructor call
+        ?{}(e);         // explicit constructor call
+} // implicit destructor call
 \end{cfa}
 …
 The global definition of @do_once@ is ignored, however if quadruple took a
 @double@ argument, then the global definition would be used instead as it
 would then be a better match.
+\todo{cite Aaron's thesis (maybe)}
 To avoid typing long lists of assertions, constraints can be collected into
 convenient a package called a @trait@, which can then be used in an assertion
+is a better match.
+% Aaron's thesis might be a good reference here.
+To avoid typing long lists of assertions, constraints can be collect into
+convenient package called a @trait@, which can then be used in an assertion
 instead of the individual constraints.
 \begin{cfa}
 …
         node(T) * next;
         T * data;
 };
+}
 node(int) inode;
 \end{cfa}
 …
 };
 CountUp countup;
+for (10) sout | resume(countup).next; // print 10 values
 \end{cfa}
 Each coroutine has a @main@ function, which takes a reference to a coroutine
 object and returns @void@.
 %[numbers=left] Why numbers on this one?
 \begin{cfa}
+\begin{cfa}[numbers=left,numberstyle=\scriptsize\sf]
 void main(CountUp & this) {
         for (unsigned int next = 0 ; true ; ++next) {
                 this.next = next;
+        for (unsigned int up = 0;; ++up) {
+                this.next = up;
                 suspend;$\label{suspend}$
+        }
 …
 \end{cfa}
 In this function, or functions called by this function (helper functions), the
 @suspend@ statement is used to return execution to the coroutine's caller
 without terminating the coroutine's function.
+@suspend@ statement is used to return execution to the coroutine's resumer
+without terminating the coroutine's function(s).
 A coroutine is resumed by calling the @resume@ function, \eg @resume(countup)@.
 The first resume calls the @main@ function at the top. Thereafter, resume calls
 continue a coroutine in the last suspended function after the @suspend@
+statement. In this case there is only one and, hence, the difference between
+subsequent calls is the state of variables inside the function and the
+coroutine object.
+The return value of @resume@ is a reference to the coroutine, to make it
+convent to access fields of the coroutine in the same expression.
+Here is a simple example in a helper function:
+\begin{cfa}
+unsigned int get_next(CountUp & this) {
+        return resume(this).next;
+}
+\end{cfa}
+When the main function returns the coroutine halts and can no longer be
+resumed.
+statement, in this case @main@ line~\ref{suspend}.  The @resume@ function takes
+a reference to the coroutine structure and returns the same reference. The
+return value allows easy access to communication variables defined in the
+coroutine object. For example, the @next@ value for coroutine object @countup@
+is both generated and collected in the single expression:
+@resume(countup).next@.
 \subsection{Monitor and Mutex Parameter}
 …
 exclusion on a monitor object by qualifying an object reference parameter with
 @mutex@.
 \begin{cfa}
 void example(MonitorA & mutex argA, MonitorB & mutex argB);
 \end{cfa}
+\begin{lstlisting}[language=CFA,{moredelim=**[is][\color{red}]{@}{@}}]
+void example(MonitorA & @mutex@ argA, MonitorB & @mutex@ argB);
+\end{lstlisting}
 When the function is called, it implicitly acquires the monitor lock for all of
 the mutex parameters without deadlock.  This semantics means all functions with
 …
+{
         StringWorker stringworker; // fork thread running in "main"
 } // Implicit call to join(stringworker), waits for completion.
+} // implicitly join with thread / wait for completion
 \end{cfa}
 The thread main is where a new thread starts execution after a fork operation

doc/theses/andrew_beach_MMath/features.tex

-              reaeca5f
+              r1d402be
 \paragraph{Raise}
 The raise is the starting point for exception handling,
+The raise is the starting point for exception handling
 by raising an exception, which passes it to
 the EHM.
 …
 \paragraph{Handle}
 The primary purpose of an EHM is to run some user code to handle a raised
+exception. This code is given, along with some other information,
+in a handler.
+exception. This code is given, with some other information, in a handler.
 A handler has three common features: the previously mentioned user code, a
 region of code it guards and an exception label/condition that matches
 against the raised exception.
+region of code it guards, and an exception label/condition that matches
+the raised exception.
 Only raises inside the guarded region and raising exceptions that match the
 label can be handled by a given handler.
 …
 The @try@ statements of \Cpp, Java and Python are common examples. All three
+also show another common feature of handlers, they are grouped by the guarded
+region.
+show the common features of guarded region, raise, matching and handler.
+\begin{cfa}
+try {                           // guarded region
+        ...
+        throw exception;        // raise
+        ...
+} catch( exception ) {  // matching condition, with exception label
+        ...                             // handler code
+}
+\end{cfa}
 \subsection{Propagation}
 After an exception is raised comes what is usually the biggest step for the
+EHM: finding and setting up the handler for execution.
+The propagation from raise to
+EHM: finding and setting up the handler for execution. The propagation from raise to
 handler can be broken up into three different tasks: searching for a handler,
 matching against the handler and installing the handler.
 …
 \paragraph{Searching}
 The EHM begins by searching for handlers that might be used to handle
 the exception.
 The search will find handlers that have the raise site in their guarded
+the exception. The search is restricted to
+handlers that have the raise site in their guarded
 region.
 The search includes handlers in the current function, as well as any in
 …
 \paragraph{Matching}
 Each handler found is with the raised exception. The exception
+Each handler found is matched with the raised exception. The exception
 label defines a condition that is used with the exception and decides if
 there is a match or not.
+%
 In languages where the first match is used, this step is intertwined with
 searching; a match check is performed immediately after the search finds
 …
 different course of action for this case.
 This situation only occurs with unchecked exceptions as checked exceptions
 (such as in Java) can make the guarantee.
+(such as in Java) are guaranteed to find a matching handler.
 The unhandled action is usually very general, such as aborting the program.
 …
 A handler labeled with any given exception can handle exceptions of that
 type or any child type of that exception. The root of the exception hierarchy
 (here \code{C}{exception}) acts as a catch-all, leaf types catch single types
+(here \code{C}{exception}) acts as a catch-all, leaf types catch single types,
 and the exceptions in the middle can be used to catch different groups of
 related exceptions.
 This system has some notable advantages, such as multiple levels of grouping,
 the ability for libraries to add new exception types and the isolation
+the ability for libraries to add new exception types, and the isolation
 between different sub-hierarchies.
 This design is used in \CFA even though it is not a object-orientated
 …
 For effective exception handling, additional information is often passed
 from the raise to the handler and back again.
+So far, only communication of the exceptions' identity is covered.
+A common communication method for adding information to an exception
+is putting fields into the exception instance
+So far, only communication of the exception's identity is covered.
+A common communication method for passing more information is putting fields into the exception instance
 and giving the handler access to them.
+% You can either have pointers/references in the exception, or have p/rs to
+% the exception when it doesn't have to be copied.
+Passing references or pointers allows data at the raise location to be
+updated, passing information in both directions.
+Using reference fields pointing to data at the raise location allows data to be
+passed in both directions.
 \section{Virtuals}
 \label{s:virtuals}
+\label{s:Virtuals}
 Virtual types and casts are not part of \CFA's EHM nor are they required for
 an EHM.
 However, one of the best ways to support an exception hierarchy
 is via a virtual hierarchy and dispatch system.
 Ideally, the virtual system would have been part of \CFA before the work
+Ideally, the virtual system should have been part of \CFA before the work
 on exception handling began, but unfortunately it was not.
 Hence, only the features and framework needed for the EHM were
+designed and implemented for this thesis.
+Other features were considered to ensure that
+designed and implemented for this thesis. Other features were considered to ensure that
 the structure could accommodate other desirable features in the future
 but are not implemented.
 The rest of this section only discusses the implemented subset of the
 virtual system design.
+virtual-system design.
 The virtual system supports multiple ``trees" of types. Each tree is
 …
 number of children.
 Any type that belongs to any of these trees is called a virtual type.
+For example, the following hypothetical syntax creates two virtual-type trees.
+\begin{flushleft}
+\lstDeleteShortInline@
+\begin{tabular}{@{\hspace{20pt}}l@{\hspace{20pt}}l}
+\begin{cfa}
+vtype V0, V1(V0), V2(V0);
+vtype W0, W1(W0), W2(W1);
+\end{cfa}
+&
+\raisebox{-0.6\totalheight}{\input{vtable}}
+\end{tabular}
+\lstMakeShortInline@
+\end{flushleft}
 % A type's ancestors are its parent and its parent's ancestors.
 % The root type has no ancestors.
 % A type's descendants are its children and its children's descendants.
+For the purposes of illistration, a proposed -- but unimplemented syntax --
+will be used. Each virtual type is repersented by a trait with an annotation
+that makes it a virtual type. This annotation is empty for a root type, which
+creates a new tree:
+\begin{cfa}
+trait root_type(T) virtual() {}
+\end{cfa}
+The annotation may also refer to any existing virtual type to make this new
+type a child of that type and part of the same tree. The parent may itself
+be a child or a root type and may have any number of existing children.
+\begin{cfa}
+trait child_a(T) virtual(root_type) {}
+trait grandchild(T) virtual(child_a) {}
+trait child_b(T) virtual(root_type) {}
+\end{cfa}
+\todo{Update the diagram in vtable.fig to show the new type tree.}
+Every virtual type also has a list of virtual members and a unique id,
+both are stored in a virtual table.
+Every instance of a virtual type also has a pointer to a virtual table stored
+in it, although there is no per-type virtual table as in many other languages.
+The list of virtual members is built up down the tree. Every virtual type
+inherits the list of virtual members from its parent and may add more
+virtual members to the end of the list which are passed on to its children.
+Again, using the unimplemented syntax this might look like:
+\begin{cfa}
+trait root_type(T) virtual() {
+        const char * to_string(T const & this);
+        unsigned int size;
+}
+trait child_type(T) virtual(root_type) {
+        char * irrelevant_function(int, char);
+}
+\end{cfa}
+% Consider adding a diagram, but we might be good with the explanation.
+As @child_type@ is a child of @root_type@ it has the virtual members of
+@root_type@ (@to_string@ and @size@) as well as the one it declared
+(@irrelivant_function@).
+It is important to note that these are virtual members, and may contain
+arbitrary fields, functions or otherwise.
+The names ``size" and ``align" are reserved for the size and alignment of the
+virtual type, and are always automatically initialized as such.
+The other special case are uses of the trait's polymorphic argument
+(@T@ in the example), which are always updated to refer to the current
+virtual type. This allows functions that refer to to polymorphic argument
+to act as traditional virtual methods (@to_string@ in the example), as the
+object can always be passed to a virtual method in its virtual table.
+Every virtual type (tree node) has a pointer to a virtual table with a unique
+@Id@ and a list of virtual members (see \autoref{s:VirtualSystem} for
+details). Children inherit their parent's list of virtual members but may add
+and/or replace members.  For example,
+\begin{cfa}
+vtable W0 | { int ?<?( int, int ); int ?+?( int, int ); }
+vtable W1 | { int ?+?( int, int ); int w, int ?-?( int, int ); }
+\end{cfa}
+creates a virtual table for @W0@ initialized with the matching @<@ and @+@
+operations visible at this declaration context.  Similarly, @W1@ is initialized
+with @<@ from inheritance with @W0@, @+@ is replaced, and @-@ is added, where
+both operations are matched at this declaration context. It is important to
+note that these are virtual members, not virtual methods of object-orientated
+programming, and can be of any type. Finally, trait names can be used to
+specify the list of virtual members.
+\PAB{Need to look at these when done.
+\CFA still supports virtual methods as a special case of virtual members.
+Function pointers that take a pointer to the virtual type are modified
+with each level of inheritance so that refers to the new type.
+This means an object can always be passed to a function in its virtual table
+as if it were a method.
+\todo{Clarify (with an example) virtual methods.}
+}%
 Up until this point the virtual system is similar to ones found in
+object-oriented languages but this is where \CFA diverges.
+Objects encapsulate a single set of methods in each type,
+universally across the entire program,
+and indeed all programs that use that type definition.
+The only way to change any method is to inherit and define a new type with
+its own universal implementation. In this sense,
+these object-oriented types are ``closed" and cannot be altered.
+% Because really they are class oriented.
+In \CFA, types do not encapsulate any code.
+Whether or not satisfies any given assertion, and hence any trait, is
+context sensitive. Types can begin to satisfy a trait, stop satisfying it or
+satisfy the same trait at any lexical location in the program.
+In this sense, an type's implementation in the set of functions and variables
+that allow it to satisfy a trait is ``open" and can change
+throughout the program.
+object-orientated languages but this is where \CFA diverges. Objects encapsulate a
+single set of methods in each type, universally across the entire program,
+and indeed all programs that use that type definition. Even if a type inherits and adds methods, it still encapsulate a
+single set of methods. In this sense,
+object-oriented types are ``closed" and cannot be altered.
+In \CFA, types do not encapsulate any code. Traits are local for each function and
+types can satisfy a local trait, stop satisfying it or, satisfy the same
+trait in a different way at any lexical location in the program where a function is call.
+In this sense, the set of functions/variables that satisfy a trait for a type is ``open" as the set can change at every call site.
 This capability means it is impossible to pick a single set of functions
 that represent a type's implementation across a program.
 …
 type. A user can define virtual tables that are filled in at their
 declaration and given a name. Anywhere that name is visible, even if it is
 defined locally inside a function (although in this case the user must ensure
 it outlives any objects that use it), it can be used.
+defined locally inside a function \PAB{What does this mean? (although that means it does not have a
+static lifetime)}, it can be used.
 Specifically, a virtual type is ``bound" to a virtual table that
 sets the virtual members for that object. The virtual members can be accessed
 through the object.
-This means virtual tables are declared and named in \CFA.
-They are declared as variables, using the type
-@vtable(VIRTUAL_TYPE)@ and any valid name. For example:
-\begin{cfa}
-vtable(virtual_type_name) table_name;
-\end{cfa}
-Like any variable they may be forward declared with the @extern@ keyword.
-Forward declaring virtual tables is relatively common.
-Many virtual types have an ``obvious" implementation that works in most
-cases.
-A pattern that has appeared in the early work using virtuals is to
-implement a virtual table with the the obvious definition and place a forward
-declaration of it in the header beside the definition of the virtual type.
-Even on the full declaration, no initializer should be used.
-Initialization is automatic.
-The type id and special virtual members ``size" and ``align" only depend on
-the virtual type, which is fixed given the type of the virtual table and
-so the compiler fills in a fixed value.
-The other virtual members are resolved, using the best match to the member's
-name and type, in the same context as the virtual table is declared using
-\CFA's normal resolution rules.
 While much of the virtual infrastructure is created, it is currently only used
 …
 @EXPRESSION@ object, otherwise it returns @0p@ (null pointer).
+\section{Exceptions}
+The syntax for declaring an exception is the same as declaring a structure
+except the keyword that is swapped out:
+\begin{cfa}
+exception TYPE_NAME {
+        FIELDS
+};
+\end{cfa}
+Fields are filled in the same way as a structure as well. However an extra
+field is added, this field contains the pointer to the virtual table.
+It must be explicitly initialised by the user when the exception is
+constructed.
+Here is an example of declaring an exception type along with a virtual table,
+assuming the exception has an ``obvious" implementation and a default
+virtual table makes sense.
+\begin{minipage}[t]{0.4\textwidth}
+Header:
+\begin{cfa}
+exception Example {
+        int data;
+};
+extern vtable(Example)
+        example_base_vtable;
+\end{cfa}
+\end{minipage}
+\begin{minipage}[t]{0.6\textwidth}
+Source:
+\begin{cfa}
+vtable(Example) example_base_vtable
+\end{cfa}
+\vfil
+\end{minipage}
+%\subsection{Exception Details}
+If one is only raising and handling exceptions, that is the only interface
+that is needed. However it is actually a short hand for a more complex
+trait based interface.
+The language views exceptions through a series of traits,
+\section{Exception}
+% Leaving until later, hopefully it can talk about actual syntax instead
+% of my many strange macros. Syntax aside I will also have to talk about the
+% features all exceptions support.
+Exceptions are defined by the trait system; there are a series of traits, and
 if a type satisfies them, then it can be used as an exception. The following
 is the base trait all exceptions need to match.
 …
 completing the virtual system). The imaginary assertions would probably come
 from a trait defined by the virtual system, and state that the exception type
 is a virtual type, is a descendant of @exception_t@ (the base exception type)
 and allow the user to find the virtual table type.
+is a virtual type, is a descendant of @exception_t@ (the base exception type),
+and note its virtual table type.
 % I did have a note about how it is the programmer's responsibility to make
 …
 \end{cfa}
 Both traits ensure a pair of types are an exception type, its virtual table
 type
+type,
 and defines one of the two default handlers. The default handlers are used
 as fallbacks and are discussed in detail in \vref{s:ExceptionHandling}.
 …
 facing way. So these three macros are provided to wrap these traits to
 simplify referring to the names:
 @IS_EXCEPTION@, @IS_TERMINATION_EXCEPTION@ and @IS_RESUMPTION_EXCEPTION@.
+@IS_EXCEPTION@, @IS_TERMINATION_EXCEPTION@, and @IS_RESUMPTION_EXCEPTION@.
 All three take one or two arguments. The first argument is the name of the
 …
 These twin operations are the core of \CFA's exception handling mechanism.
 This section covers the general patterns shared by the two operations and
 then goes on to cover the details each individual operation.
+then goes on to cover the details of each individual operation.
 Both operations follow the same set of steps.
 First, a user raises an exception.
 Second, the exception propagates up the stack, searching for a handler.
+Second, the exception propagates up the stack.
 Third, if a handler is found, the exception is caught and the handler is run.
 After that control continues at a raise-dependent location.
+As an alternate to the third step,
+if a handler is not found, a default handler is run and, if it returns,
+then control
+Fourth, if a handler is not found, a default handler is run and, if it returns, then control
 continues after the raise.
+The differences between the two operations include how propagation is
+performed, where excecution after an exception is handler
 and which default handler is run.
+%This general description covers what the two kinds have in common.
+The differences in the two operations include how propagation is performed, where execution continues
+after an exception is caught and handled, and which default handler is run.
 \subsection{Termination}
 \label{s:Termination}
+Termination handling is the familiar kind of handling
+and used in most programming
+Termination handling is the familiar EHM and used in most programming
 languages with exception handling.
 It is a dynamic, non-local goto. If the raised exception is matched and
 …
 Then propagation starts with the search. \CFA uses a ``first match" rule so
 matching is performed with the copied exception as the search key.
 It starts from the raise site and proceeds towards base of the stack,
+It starts from the raise in the throwing function and proceeds towards the base of the stack,
 from callee to caller.
 At each stack frame, a check is made for termination handlers defined by the
 …
 \end{cfa}
 When viewed on its own, a try statement simply executes the statements
 in the \snake{GUARDED_BLOCK} and when those are finished,
+in the \snake{GUARDED_BLOCK}, and when those are finished,
 the try statement finishes.
 …
 termination exception types.
 The global default termination handler performs a cancellation
+(as described in \vref{s:Cancellation})
+on the current stack with the copied exception.
+Since it is so general, a more specific handler can be defined,
+overriding the default behaviour for the specific exception types.
+(see \vref{s:Cancellation} for the justification) on the current stack with the copied exception.
+Since it is so general, a more specific handler is usually
+defined, possibly with a detailed message, and used for specific exception type, effectively overriding the default handler.
 \subsection{Resumption}
 \label{s:Resumption}
+Resumption exception handling is less familar form of exception handling,
+but is
+Resumption exception handling is the less familar EHM, but is
 just as old~\cite{Goodenough75} and is simpler in many ways.
 It is a dynamic, non-local function call. If the raised exception is
 …
 function once the error is corrected, and
 ignorable events, such as logging where nothing needs to happen and control
+should always continue from the raise site.
+Except for the changes to fit into that pattern, resumption exception
+handling is symmetric with termination exception handling, by design
+(see \autoref{s:Termination}).
+should always continue from the raise point.
 A resumption raise is started with the @throwResume@ statement:
 …
 \end{cfa}
 \todo{Decide on a final set of keywords and use them everywhere.}
+It works much the same way as the termination raise, except the
+type must satisfy the \snake{is_resumption_exception} that uses the
+default handler: \defaultResumptionHandler.
+This can be specialized for particular exception types.
+At run-time, no exception copy is made. Since
+It works much the same way as the termination throw.
+The expression must return a reference to a resumption exception,
+where the resumption exception is any type that satisfies the trait
+@is_resumption_exception@ at the call site.
+The assertions from this trait are available to
+the exception system while handling the exception.
+At run-time, no exception copy is made, since
 resumption does not unwind the stack nor otherwise remove values from the
 current scope, there is no need to manage memory to keep the exception
+allocated.
 Then propagation starts with the search,
 following the same search path as termination,
+from the raise site to the base of stack and top of try statement to bottom.
 However, the handlers on try statements are defined by @catchResume@ clauses.
+current scope, so there is no need to manage memory to keep the exception in scope.
+Then propagation starts with the search. It starts from the raise in the
+resuming function and proceeds towards the base of the stack,
+from callee to caller.
+At each stack frame, a check is made for resumption handlers defined by the
+@catchResume@ clauses of a @try@ statement.
 \begin{cfa}
 try {
 …
+}
 \end{cfa}
+Note that termination handlers and resumption handlers may be used together
+% PAB, you say this above.
+% When a try statement is executed, it simply executes the statements in the
+% @GUARDED_BLOCK@ and then finishes.
+%
+% However, while the guarded statements are being executed, including any
+% invoked functions, all the handlers in these statements are included in the
+% search path.
+% Hence, if a resumption exception is raised, these handlers may be matched
+% against the exception and may handle it.
+%
+% Exception matching checks the handler in each catch clause in the order
+% they appear, top to bottom. If the representation of the raised exception type
+% is the same or a descendant of @EXCEPTION_TYPE@$_i$, then @NAME@$_i$
+% (if provided) is bound to a pointer to the exception and the statements in
+% @HANDLER_BLOCK@$_i$ are executed.
+% If control reaches the end of the handler, execution continues after the
+% the raise statement that raised the handled exception.
+%
+% Like termination, if no resumption handler is found during the search,
+% then the default handler (\defaultResumptionHandler) visible at the raise
+% statement is called. It will use the best match at the raise sight according
+% to \CFA's overloading rules. The default handler is
+% passed the exception given to the raise. When the default handler finishes
+% execution continues after the raise statement.
+%
+% There is a global @defaultResumptionHandler{} is polymorphic over all
+% resumption exceptions and performs a termination throw on the exception.
+% The \defaultTerminationHandler{} can be overridden by providing a new
+% function that is a better match.
+The @GUARDED_BLOCK@ and its associated nested guarded statements work the same
+for resumption as for termination, as does exception matching at each
+@catchResume@. Similarly, if no resumption handler is found during the search,
+then the currently visible default handler (\defaultResumptionHandler) is
+called and control continues after the raise statement if it returns. Finally,
+there is also a global @defaultResumptionHandler@, which can be overridden,
+that is polymorphic over all resumption exceptions but performs a termination
+throw on the exception rather than a cancellation.
+Throwing the exception in @defaultResumptionHandler@ has the positive effect of
+walking the stack a second time for a recovery handler. Hence, a programmer has
+two chances for help with a problem, fixup or recovery, should either kind of
+handler appear on the stack. However, this dual stack walk leads to following
+apparent anomaly:
+\begin{cfa}
+try {
+        throwResume E;
+} catch (E) {
+        // this handler runs
+}
+\end{cfa}
+because the @catch@ appears to handle a @throwResume@, but a @throwResume@ only
+matches with @catchResume@. The anomaly results because the unmatched
+@catchResuem@, calls @defaultResumptionHandler@, which in turn throws @E@.
+% I wonder if there would be some good central place for this.
+Note, termination and resumption handlers may be used together
 in a single try statement, intermixing @catch@ and @catchResume@ freely.
 Each type of handler only interacts with exceptions from the matching
 kind of raise.
-Like @catch@ clauses, @catchResume@ clauses have no effect if an exception
-is not raised.
-The matching rules are exactly the same as well.
-The first major difference here is that after
-@EXCEPTION_TYPE@$_i$ is matched and @NAME@$_i$ is bound to the exception,
-@HANDLER_BLOCK@$_i$ is executed right away without first unwinding the stack.
-After the block has finished running control jumps to the raise site, where
-the just handled exception came from, and continues executing after it,
-not after the try statement.
 \subsubsection{Resumption Marking}
 …
 and run, its try block (the guarded statements) and every try statement
 searched before it are still on the stack. There presence can lead to
+the recursive resumption problem.
+\todo{Is there a citation for the recursive resumption problem?}
+the \emph{recursive resumption problem}.
 The recursive resumption problem is any situation where a resumption handler
 …
 When this code is executed, the guarded @throwResume@ starts a
 search and matches the handler in the @catchResume@ clause. This
+call is placed on the stack above the try-block.
+Now the second raise in the handler searches the same try block,
+matches again and then puts another instance of the
+call is placed on the stack above the try-block. Now the second raise in the handler
+searches the same try block, matches, and puts another instance of the
 same handler on the stack leading to infinite recursion.
+While this situation is trivial and easy to avoid, much more complex cycles
+can form with multiple handlers and different exception types.
+While this situation is trivial and easy to avoid, much more complex cycles can
+form with multiple handlers and different exception types.  The key point is
+that the programmer's intuition expects every raise in a handler to start
+searching \emph{below} the @try@ statement, making it difficult to understand
+and fix the problem.
 To prevent all of these cases, each try statement is ``marked" from the
 time the exception search reaches it to either when a handler completes
 handling that exception or when the search reaches the base
+time the exception search reaches it to either when a matching handler
+completes or when the search reaches the base
 of the stack.
 While a try statement is marked, its handlers are never matched, effectively
 …
 for instance, marking just the handlers that caught the exception,
 would also prevent recursive resumption.
 However, the rules selected mirrors what happens with termination,
 so this reduces the amount of rules and patterns a programmer has to know.
 The marked try statements are the ones that would be removed from
+However, the rule selected mirrors what happens with termination,
+and hence, matches programmer intuition that a raise searches below a try.
+In detail, the marked try statements are the ones that would be removed from
 the stack for a termination exception, \ie those on the stack
 between the handler and the raise statement.
 …
 \subsection{Comparison with Reraising}
+In languages without conditional catch, that is no ability to match an
+exception based on something other than its type, it can be mimicked
+by matching all exceptions of the right type, checking any additional
+conditions inside the handler and re-raising the exception if it does not
+match those.
+Here is a minimal example comparing both patterns, using @throw;@
+(no argument) to start a re-raise.
+Without conditional catch, the only approach to match in more detail is to reraise
+the exception after it has been caught, if it could not be handled.
 \begin{center}
 \begin{tabular}{l r}
 \begin{cfa}
 try {
     do_work_may_throw();
 } catch(exception_t * exc ;
+                can_handle(exc)) {
     handle(exc);
+}
+\begin{tabular}{l|l}
+\begin{cfa}
+try {
+        do_work_may_throw();
+} catch(excep_t * ex; can_handle(ex)) {
+        handle(ex);
+}
 \end{cfa}
+&
 \begin{cfa}
 try {
+    do_work_may_throw();
+} catch(exception_t * exc) {
+    if (can_handle(exc)) {
+        handle(exc);
+    } else {
+        throw;
+    }
+}
+\end{cfa}
+\end{tabular}
+\end{center}
+At first glance catch-and-reraise may appear to just be a quality of life
+feature, but there are some significant differences between the two
+stratagies.
+A simple difference that is more important for \CFA than many other languages
+is that the raise site changes, with a re-raise but does not with a
+conditional catch.
+This is important in \CFA because control returns to the raise site to run
+the per-site default handler. Because of this only a conditional catch can
+allow the original raise to continue.
+The more complex issue comes from the difference in how conditional
+catches and re-raises handle multiple handlers attached to a single try
+statement. A conditional catch will continue checking later handlers while
+a re-raise will skip them.
+If the different handlers could handle some of the same exceptions,
+translating a try statement that uses one to use the other can quickly
+become non-trivial:
+\noindent
+Original, with conditional catch:
+\begin{cfa}
+...
+} catch (an_exception * e ; check_a(e)) {
+        handle_a(e);
+} catch (exception_t * e ; check_b(e)) {
+        handle_b(e);
+}
+\end{cfa}
+Translated, with re-raise:
+\begin{cfa}
+...
+} catch (exception_t * e) {
+        an_exception * an_e = (virtual an_exception *)e;
+        if (an_e && check_a(an_e)) {
+                handle_a(an_e);
+        } else if (check_b(e)) {
+                handle_b(e);
+        do_work_may_throw();
+} catch(excep_t * ex) {
+        if (can_handle(ex)) {
+                handle(ex);
         } else {
                 throw;
 …
+}
 \end{cfa}
+(There is a simpler solution if @handle_a@ never raises exceptions,
+using nested try statements.)
+% } catch (an_exception * e ; check_a(e)) {
+%     handle_a(e);
+% } catch (exception_t * e ; !(virtual an_exception *)e && check_b(e)) {
+%     handle_b(e);
+% }
+\end{tabular}
+\end{center}
+Notice catch-and-reraise increases complexity by adding additional data and
+code to the exception process. Nevertheless, catch-and-reraise can simulate
+conditional catch straightforwardly, when exceptions are disjoint, \ie no
+inheritance.
+However, catch-and-reraise simulation becomes unusable for exception inheritance.
+\begin{flushleft}
+\begin{cfa}[xleftmargin=6pt]
+exception E1;
+exception E2(E1); // inheritance
+\end{cfa}
+\begin{tabular}{l|l}
+\begin{cfa}
+try {
+        ... foo(); ... // raise E1/E2
+        ... bar(); ... // raise E1/E2
+} catch( E2 e; e.rtn == foo ) {
+        ...
+} catch( E1 e; e.rtn == foo ) {
+        ...
+} catch( E1 e; e.rtn == bar ) {
+        ...
+}
+\end{cfa}
+&
+\begin{cfa}
+try {
+        ... foo(); ...
+        ... bar(); ...
+} catch( E2 e ) {
+        if ( e.rtn == foo ) { ...
+        } else throw; // reraise
+} catch( E1 e ) {
+        if (e.rtn == foo) { ...
+        } else if (e.rtn == bar) { ...
+        else throw; // reraise
+}
+\end{cfa}
+\end{tabular}
+\end{flushleft}
+The derived exception @E2@ must be ordered first in the catch list, otherwise
+the base exception @E1@ catches both exceptions. In the catch-and-reraise code
+(right), the @E2@ handler catches exceptions from both @foo@ and
+@bar@. However, the reraise misses the following catch clause. To fix this
+problem, an enclosing @try@ statement is need to catch @E2@ for @bar@ from the
+reraise, and its handler must duplicate the inner handler code for @bar@. To
+generalize, this fix for any amount of inheritance and complexity of try
+statement requires a technique called \emph{try-block
+splitting}~\cite{Krischer02}, which is not discussed in this thesis. It is
+sufficient to state that conditional catch is more expressive than
+catch-and-reraise in terms of complexity.
+\begin{comment}
+That is, they have the same behaviour in isolation.
+Two things can expose differences between these cases.
+One is the existence of multiple handlers on a single try statement.
+A reraise skips all later handlers for a try statement but a conditional
+catch does not.
+% Hence, if an earlier handler contains a reraise later handlers are
+% implicitly skipped, with a conditional catch they are not.
+Still, they are equivalently powerful,
+both can be used two mimic the behaviour of the other,
+as reraise can pack arbitrary code in the handler and conditional catches
+can put arbitrary code in the predicate.
+% I was struggling with a long explanation about some simple solutions,
+% like repeating a condition on later handlers, and the general solution of
+% merging everything together. I don't think it is useful though unless its
+% for a proof.
+% https://en.cppreference.com/w/cpp/language/throw
+The question then becomes ``Which is a better default?"
+We believe that not skipping possibly useful handlers is a better default.
+If a handler can handle an exception it should and if the handler can not
+handle the exception then it is probably safer to have that explicitly
+described in the handler itself instead of implicitly described by its
+ordering with other handlers.
+% Or you could just alter the semantics of the throw statement. The handler
+% index is in the exception so you could use it to know where to start
+% searching from in the current try statement.
+% No place for the `goto else;` metaphor.
+The other issue is all of the discussion above assumes that the only
+way to tell apart two raises is the exception being raised and the remaining
+search path.
+This is not true generally, the current state of the stack can matter in
+a number of cases, even only for a stack trace after an program abort.
+But \CFA has a much more significant need of the rest of the stack, the
+default handlers for both termination and resumption.
+% For resumption it turns out it is possible continue a raise after the
+% exception has been caught, as if it hadn't been caught in the first place.
+This becomes a problem combined with the stack unwinding used in termination
+exception handling.
+The stack is unwound before the handler is installed, and hence before any
+reraises can run. So if a reraise happens the previous stack is gone,
+the place on the stack where the default handler was supposed to run is gone,
+if the default handler was a local function it may have been unwound too.
+There is no reasonable way to restore that information, so the reraise has
+to be considered as a new raise.
+This is the strongest advantage conditional catches have over reraising,
+they happen before stack unwinding and avoid this problem.
+% The one possible disadvantage of conditional catch is that it runs user
+% code during the exception search. While this is a new place that user code
+% can be run destructors and finally clauses are already run during the stack
+% unwinding.
+%
+% } catch (an_exception * e)
+%   if (check_a(e)) {
+%     handle_a(e);
+%   } else throw;
+% } catch (exception_t * e)
+%   if (check_b(e)) {
+%     handle_b(e);
+%   } else throw;
+% }
+In similar simple examples translating from re-raise to conditional catch
+takes less code but it does not have a general trivial solution either.
+So, given that the two patterns do not trivially translate into each other,
+it becomes a matter of which on should be encouraged and made the default.
+From the premise that if a handler that could handle an exception then it
+should, it follows that checking as many handlers as possible is preferred.
+So conditional catch and checking later handlers is a good default.
+% https://www.cplusplus.com/reference/exception/current_exception/
+%   `exception_ptr current_exception() noexcept;`
+% https://www.python.org/dev/peps/pep-0343/
+\end{comment}
 \section{Finally Clauses}
 …
 The @FINALLY_BLOCK@ is executed when the try statement is removed from the
 stack, including when the @GUARDED_BLOCK@ finishes, any termination handler
 finishes or during an unwind.
+finishes, or during an unwind.
 The only time the block is not executed is if the program is exited before
 the stack is unwound.
 …
 they have their own strengths, similar to top-level function and lambda
 functions with closures.
 Destructors take more work to create, but if there is clean-up code
+Destructors take more work for their creation, but if there is clean-up code
 that needs to be run every time a type is used, they are much easier
 to set-up for each use. % It's automatic.
+to set-up.
 On the other hand finally clauses capture the local context, so is easy to
 use when the clean-up is not dependent on the type of a variable or requires
 …
 raise, this exception is not used in matching only to pass information about
 the cause of the cancellation.
 Finally, as no handler is provided, there is no default handler.
+Finaly, since a cancellation only unwinds and forwards, there is no default handler.
 After @cancel_stack@ is called the exception is copied into the EHM's memory
 …
 After the main stack is unwound there is a program-level abort.
 The first reason for this behaviour is for sequential programs where there
 is only one stack, and hence to stack to pass information to.
+Second, even in concurrent programs, the main stack has no dependency
 on another stack and no reliable way to find another living stack.
+Finally, keeping the same behaviour in both sequential and concurrent
 programs is simple and easy to understand.
+The reasons for this semantics in a sequential program is that there is no more code to execute.
+This semantics also applies to concurrent programs, too, even if threads are running.
+That is, if any threads starts a cancellation, it implies all threads terminate.
+Keeping the same behaviour in sequential and concurrent programs is simple.
+Also, even in concurrent programs there may not currently be any other stacks
+and even if other stacks do exist, main has no way to know where they are.
 \paragraph{Thread Stack}
 …
 With explicit join and a default handler that triggers a cancellation, it is
+possible to cascade an error across any number of threads,
+alternating between the resumption (possibly termination) and cancellation,
+cleaning up each
+possible to cascade an error across any number of threads, cleaning up each
 in turn, until the error is handled or the main thread is reached.
 …
 caller's context and passes it to the internal report.
+A coroutine only knows of two other coroutines,
+its starter and its last resumer.
+A coroutine only knows of two other coroutines, its starter and its last resumer.
 The starter has a much more distant connection, while the last resumer just
 (in terms of coroutine state) called resume on this coroutine, so the message
 …
 With a default handler that triggers a cancellation, it is possible to
+cascade an error across any number of coroutines,
+alternating between the resumption (possibly termination) and cancellation,
+cleaning up each in turn,
+cascade an error across any number of coroutines, cleaning up each in turn,
 until the error is handled or a thread stack is reached.
+\PAB{Part of this I do not understand. A cancellation cannot be caught. But you
+talk about handling a cancellation in the last sentence. Which is correct?}

doc/theses/andrew_beach_MMath/future.tex

-              reaeca5f
+              r1d402be
 \label{c:future}
+The following discussion covers both possible interesting research
+that could follow from this work as long as simple implementation
+improvements.
+The following discussion covers both missing language features that affected my
+work and research based improvements.
 \section{Language Improvements}
 …
 \CFA is a developing programming language. As such, there are partially or
 unimplemented features (including several broken components)
+that I had to workaround while building the EHM largely in
+the \CFA language (some C components). Below are a few of these issues
+and how implementing/fixing them would affect the EHM.
+In addition there are some simple improvements that had no interesting
+research attached to them but would make using the language easier.
+that I had to workaround while building an EHM largely in
+the \CFA language (some C components).  The following are a few of these
+issues, and once implemented/fixed, how they would affect the exception system.
 \begin{itemize}
+\item
+The implementation of termination is not portable because it includes
+hand-crafted assembly statements for each architecture, where the
+ARM processor was just added.
+% The existing compilers cannot translate that for other platforms and those
+% sections must be ported by hand to
+Supporting more hardware architectures in a general way is important.
 \item
 Due to a type-system problem, the catch clause cannot bind the exception to a
 …
 @return@, \etc. The reason is that current code generation hoists a handler
 into a nested function for convenience (versus assemble-code generation at the
+try statement). Hence, when the handler runs, it can still access local
+variables in the lexical scope of the try statement. Still, it does mean
+that seemingly local control flow is not in fact local and crosses a function
+boundary.
+Making the termination handlers code within the surrounding
+function would remove this limitation.
+% Try blocks are much more difficult to do practically (requires our own
+% assembly) and resumption handlers have some theoretical complexity.
+@try@ statement). Hence, when the handler runs, its can access local variable
+in the lexical scope of the @try@ statement, but the closure does not capture
+local control-flow points so it cannot perform non-local transfers in the
+hoisted function.
 \item
 There is no detection of colliding unwinds. It is possible for clean-up code
 run during an unwind to trigger another unwind that escapes the clean-up code
 itself; such as a termination exception caught further down the stack or a
 cancellation. There do exist ways to handle this case, but currently there is
 no detection and the first unwind will simply be forgotten, often leaving
+cancellation. There do exist ways to handle this case, but currently there is no
+detection and the first unwind is simply forgotten, often leaving
 it in a bad state.
 \item
 Finally, the exception system has not had a lot of programmer testing.
+Finally, the exception system has not have a lot programmer testing.
 More time with encouraged usage will reveal new
 quality of life upgrades that can be made.
 …
 project, but was thrust upon it to do exception inheritance; hence, only
 minimal work is done. A draft for a complete virtual system is available but
 not finalized. A future \CFA project is to complete that work and then
+not finalized.  A future \CFA project is to complete that work and then
 update the exception system that uses the current version.
 …
 types to allow traits to refer to types not listed in their header. This
 feature allows exception traits to not refer to the virtual-table type
+explicitly, removing the need for the current interface macros,
+such as @EHM_IS_EXCEPTION@.
+explicitly. %, removing the need for the current interface macros.
 \section{Additional Raises}
 …
 Non-local/concurrent raise requires more
 coordination between the concurrency system
 and the exception system. Many of the interesting design decisions center
+and the exception system. Many of the interesting design decisions centre
 around masking, \ie controlling which exceptions may be thrown at a stack. It
 would likely require more of the virtual system and would also effect how
 …
 exception signature. An exception signature must declare all checked
 exceptions that could propagate from the function, either because they were
 raised inside the function or came from a sub-function. This improves safety
+raised inside the function or a call to a sub-function. This improves safety
 by making sure every checked exception is either handled or consciously
 passed on.
 …
 Workarounds are possible but awkward. Ideally an extension to libunwind could
 be made, but that would either require separate maintenance or gaining enough
 support to have it folded into the official library itself.
+support to have it folded into the code base.
 Also new techniques to skip previously searched parts of the stack need to be

doc/theses/andrew_beach_MMath/implement.tex

-              reaeca5f
+              r1d402be
 \label{s:VirtualSystem}
 % Virtual table rules. Virtual tables, the pointer to them and the cast.
+While the \CFA virtual system currently has only one public features, virtual
+cast and virtual tables,
+% ??? refs (see the virtual cast feature \vpageref{p:VirtualCast}),
+substantial structure is required to support them,
+While the \CFA virtual system currently has only one public feature, virtual
+cast (see the virtual cast feature \vpageref{p:VirtualCast}),
+substantial structure is required to support it,
 and provide features for exception handling and the standard library.
 \subsection{Virtual Type}
+A virtual type~(see \autoref{s:virtuals}) has a pointer to a virtual table,
+called the \emph{virtual-table pointer},
+which binds each instance of a virtual type to a virtual table.
+Internally, the field is called \snake{virtual_table}
+and is fixed after construction.
+This pointer is also the table's id and how the system accesses the
+virtual table and the virtual members there.
+It is always the first field in the
+structure so that its location is always known.
+% We have no special rules for these constructors.
+Virtual table pointers are passed to the constructors of virtual types
+as part of field-by-field construction.
+A virtual type~(see \autoref{s:Virtuals}) has a pointer to a virtual table,
+called the \emph{virtual-table pointer}, which binds an instance of a virtual
+type to a virtual table.  Internally, the field is called \snake{virtual_table}
+and is fixed after construction.  This pointer is also the table's id and how
+the system accesses the virtual table and the virtual members there. It is
+always the first field in the structure so that its location is always known.
+\todo{Talk about constructors for virtual types (after they are working).}
 \subsection{Type Id}
+Every virtual type has a unique id.
+These are used in type equality, to check if the representation of two values
+are the same, and to access the type's type information.
+This uniqueness means across a program composed of multiple translation
+units (TU), not uniqueness across all programs or even across multiple
+processes on the same machine.
+Our approach for program uniqueness is using a static declaration for each
+type id, where the run-time storage address of that variable is guaranteed to
+be unique during program execution.
+The type id storage can also be used for other purposes,
+and is used for type information.
+Every virtual type needs a unique id, so that type ids can be compared for
+equality, which checks if the types representation are the same, or used to
+access the type's type information.  Here, uniqueness means within a program
+composed of multiple translation units (TU), not uniqueness across all
+programs.
+One approach for program uniqueness is declaring a static declaration for each
+type id, where the runtime storage address of that variable is guaranteed to be
+unique during program execution. The type id storage can also be used for other
+purposes.
 The problem is that a type id may appear in multiple TUs that compose a
+program (see \autoref{ss:VirtualTable}); so the initial solution would seem
+to be make it external in each translation unit. Honever, the type id must
+have a declaration in (exactly) one of the TUs to create the storage.
+No other declaration related to the virtual type has this property, so doing
+this through standard C declarations would require the user to do it manually.
+Instead the linker is used to handle this problem.
+% I did not base anything off of C++17; they are solving the same problem.
+A new feature has been added to \CFA for this purpose, the special attribute
+\snake{cfa_linkonce}, which uses the special section @.gnu.linkonce@.
+When used as a prefix (\eg @.gnu.linkonce.example@) the linker does
+not combine these sections, but instead discards all but one with the same
+full name.
+So each type id must be given a unique section name with the linkonce
+prefix. Luckily \CFA already has a way to get unique names, the name mangler.
+For example, this could be written directly in \CFA:
+\begin{cfa}
+__attribute__((cfa_linkonce)) void f() {}
+\end{cfa}
+This is translated to:
+\begin{cfa}
+__attribute__((section(".gnu.linkonce._X1fFv___1"))) void _X1fFv___1() {}
+\end{cfa}
+This is done internally to access the name manglers.
+This attribute is useful for other purposes, any other place a unique
+instance required, and should eventually be made part of a public and
+stable feature in \CFA.
+\subsection{Type Information}
+There is data stored at the type id's declaration, the type information.
+program, see \autoref{ss:VirtualTable}; hence in each TU, it must be declared
+as external to prevent multiple definitions. However, the type id must actually
+be declared in one of the TUs so the linker creates the storage.  Hence, the
+problem becomes designating one TU to insert an actual type-id declaration. But
+the \CFA compiler does not know the set of the translation units that compose a
+program, because TUs can be compile separately, followed by a separate link
+step.
+The solution is to mimic a \CFA feature in \Cpp{17}, @inline@ variables and
+function:
+\begin{quote}
+There may be more than one definition of an inline function or variable (since
+\Cpp{17} in the program as long as each definition appears in a different
+translation unit and (for non-static inline functions and variables (since
+\Cpp{17})) all definitions are identical. For example, an inline function or an
+inline variable (since \Cpp{17}) may be defined in a header file that is
+@#include@'d in multiple source files.~\cite{C++17}
+\end{quote}
+The underlying mechanism to provide this capability is attribute
+\begin{cfa}
+section(".gnu.linkonce.NAME")
+\end{cfa}
+where @NAME@ is the variable/function name duplicated in each TU.  The linker than
+provides the service of generating a single declaration (instance) across all
+TUs, even if a program is linked incrementally.
+C does not support this feature for @inline@, and hence, neither does \CFA.
+Again, rather than implement a new @inline@ extension for \CFA, a temporary
+solution for the exception handling is to add the following in \CFA.
+\begin{lstlisting}[language=CFA,{moredelim=**[is][\color{red}]{@}{@}}]
+@__attribute__((cfa_linkonce))@ void f() {}
+\end{lstlisting}
+which becomes
+\begin{lstlisting}[language=CFA,{moredelim=**[is][\color{red}]{@}{@}}]
+__attribute__((section(".gnu.linkonce._X1fFv___1"))) void @_X1fFv___1@(){}
+\end{lstlisting}
+where @NAME@ from above is the \CFA mangled variable/function name.  Note,
+adding this feature is necessary because, when using macros, the mangled name
+is unavailable.  This attribute is useful for purposes other than exception
+handling, and should eventually be rolled into @inline@ processing in \CFA.
+Finally, a type id's data implements a pointers to the type's type information
+instance.  Dereferencing the pointer gets the type information.
+\subsection{Implementation}
 The type information currently is only the parent's type id or, if the
 type has no parent, the null pointer.
 …
 \end{cfa}
 Type information is constructed as follows:
+The type information is constructed as follows:
 \begin{enumerate}
 \item
 …
 \item
 \CFA's name mangler does its regular name mangling encoding the type of
+the declaration into the instance name.
+This process gives a completely unique name
+the declaration into the instance name. This process gives a program unique name
 including different instances of the same polymorphic type.
 \end{enumerate}
 \todo{The list is making me realize, some of this isn't ordered.}
-Writing that code manually, with helper macros for the early name mangling,
-would look like this:
-\begin{cfa}
-struct INFO_TYPE(TYPE) {
-        INFO_TYPE(PARENT) const * parent;
-};
-__attribute__((cfa_linkonce))
-INFO_TYPE(TYPE) const INFO_NAME(TYPE) = {
-        &INFO_NAME(PARENT),
-};
-\end{cfa}
 \begin{comment}
 …
 and the other is discarded.
 \end{comment}
 \subsection{Virtual Table}
 …
 The first and second sections together mean that every virtual table has a
 prefix that has the same layout and types as its parent virtual table.
 This, combined with the fixed offset to the virtual table pointer, means that
+This, combined with the fixed offset to the virtual-table pointer, means that
 for any virtual type, it is always safe to access its virtual table and,
 from there, it is safe to check the type id to identify the exact type of the
 …
 type's alignment, is set using an @alignof@ expression.
-Most of these tools are already inside the compiler. Using the is a simple
-code transformation early on in compilation allows most of that work to be
-handed off to the existing tools. \autoref{f:VirtualTableTransformation}
-shows an example transformation, this example shows an exception virtual table.
-It also shows the transformation on the full declaration,
-for a forward declaration the @extern@ keyword is preserved and the
-initializer is not added.
-\begin{figure}[htb]
-\begin{cfa}
-vtable(example_type) example_name;
-\end{cfa}
-\transformline
-% Check mangling.
-\begin{cfa}
-const struct example_type_vtable example_name = {
-        .__cfavir_typeid : &__cfatid_example_type,
-        .size : sizeof(example_type),
-        .copy : copy,
-        .^?{} : ^?{},
-        .msg : msg,
-};
-\end{cfa}
-\caption{Virtual Table Transformation}
-\label{f:VirtualTableTransformation}
-\end{figure}
 \subsection{Concurrency Integration}
 Coroutines and threads need instances of @CoroutineCancelled@ and
 …
 These transformations are shown through code re-writing in
 \autoref{f:CoroutineTypeTransformation} and
+\autoref{f:CoroutineMainTransformation}.
+Threads use the same pattern, with some names and types changed.
+In both cases, the original declaration is not modified,
+only new ones are added.
+\begin{figure}[htb]
+\autoref{f:CoroutineMainTransformation} for a coroutine and a thread is similar.
+In both cases, the original declaration is not modified, only new ones are
+added.
+\begin{figure}
 \begin{cfa}
 coroutine Example {
 …
 \caption{Coroutine Type Transformation}
 \label{f:CoroutineTypeTransformation}
+\end{figure}
+\begin{figure}[htb]
+%\end{figure}
+\bigskip
+%\begin{figure}
 \begin{cfa}
 void main(Example & this) {
 …
 \begin{cfa}
 void * __cfa__virtual_cast(
+        struct __cfavir_type_id * parent,
+        struct __cfavir_type_id * const * child );
+\end{cfa}
+The type id for the target type of the virtual cast is passed in as
+@parent@ and
+        struct __cfavir_type_td parent,
+        struct __cfavir_type_id const * child );
+\end{cfa}
+The type id for the target type of the virtual cast is passed in as @parent@ and
 the cast target is passed in as @child@.
 The generated C code wraps both arguments and the result with type casts.
 …
 \section{Exceptions}
+% The implementation of exception types.
+Creating exceptions can roughly divided into two parts,
+the exceptions themselves and the virtual system interactions.
+Creating an exception type is just a matter of preppending the field
+with the virtual table pointer to the list of the fields
+(see \autoref{f:ExceptionTypeTransformation}).
+\begin{figure}[htb]
+\begin{cfa}
+exception new_exception {
+        // EXISTING FIELDS
+};
+\end{cfa}
+\transformline
+\begin{cfa}
+struct new_exception {
+        struct new_exception_vtable const * virtual_table;
+        // EXISTING FIELDS
+};
+\end{cfa}
+\caption{Exception Type Transformation}
+\label{f:ExceptionTypeTransformation}
+\end{figure}
+The integration between exceptions and the virtual system is a bit more
+complex simply because of the nature of the virtual system prototype.
+The primary issue is that the virtual system has no way to detect when it
+should generate any of its internal types and data. This is handled by
+the exception code, which tells the virtual system when to generate
+its components.
+All types associated with a virtual type,
+the types of the virtual table and the type id,
+are generated when the virtual type (the exception) is first found.
+The type id (the instance) is generated with the exception if it is
+a monomorphic type.
+However if the exception is polymorphic then a different type id has to
+be generated for every instance. In this case generation is delayed
+until a virtual table is created.
+% There are actually some problems with this, which is why it is not used
+% for monomorphic types.
+When a virtual table is created and initialized two functions are created
+to fill in the list of virtual members.
+The first is a copy function which adapts the exception's copy constructor
+to work with pointers, avoiding some issues with the current copy constructor
+interface.
+Second is the msg function, which returns a C-string with the type's name,
+including any polymorphic parameters.
+\todo{Anything about exception construction.}
 \section{Unwinding}
 …
 stack. On function entry and return, unwinding is handled directly by the
 call/return code embedded in the function.
+% Discussing normal stack unwinding:
+\PAB{Meaning: In many cases, the position of the instruction pointer (relative to parameter
+and local declarations) is enough to know the current size of the stack
+frame.}
 Usually, the stack-frame size is known statically based on parameter and
 local variable declarations. Even for a dynamic stack-size, the information
 …
 bumping the hardware stack-pointer up or down as needed.
 Constructing/destructing values within a stack frame has
+a similar complexity but larger constants.
+% Discussing multiple frame stack unwinding:
+a similar complexity but larger constants, which takes longer.
 Unwinding across multiple stack frames is more complex because that
 information is no longer contained within the current function.
+With seperate compilation,
+a function does not know its callers nor their frame layout.
+Even using the return address, that information is encoded in terms of
+actions in code, intermixed with the actions required finish the function.
+Without changing the main code path it is impossible to select one of those
+two groups of actions at the return site.
+With separate compilation a function does not know its callers nor their frame size.
+In general, the caller's frame size is embedded only at the functions entry (push
+stack) and exit (pop stack).
+Without altering the main code path it is also hard to pass that work off
+to the caller.
 The traditional unwinding mechanism for C is implemented by saving a snap-shot
 …
 many languages define clean-up actions that must be taken when certain
 sections of the stack are removed. Such as when the storage for a variable
+is removed from the stack, possibly requiring a destructor call,
+or when a try statement with a finally clause is
+is removed from the stack (destructor call) or when a try statement with a finally clause is
 (conceptually) popped from the stack.
 None of these cases should be handled by the user --- that would contradict the
 …
 In plain C (which \CFA currently compiles down to) this
 flag only handles the cleanup attribute:
-%\label{code:cleanup}
 \begin{cfa}
 void clean_up( int * var ) { ... }
 …
 To get full unwinding support, all of these features must be handled directly
 in assembly and assembler directives; partiularly the cfi directives
+in assembly and assembler directives; particularly the cfi directives
 \snake{.cfi_lsda} and \snake{.cfi_personality}.
 …
 needs its own exception context.
 The current exception context should be retrieved by calling the function
+An exception context is retrieved by calling the function
 \snake{this_exception_context}.
 For sequential execution, this function is defined as
 …
 function. The LSDA in particular is hard to mimic in generated C code.
 The workaround is a function called \snake{__cfaehm_try_terminate} in the
+standard \CFA library. The contents of a try block and the termination
+handlers are converted into nested functions. These are then passed to the
 try terminate function and it calls them, appropriately.
+The workaround is a function called @__cfaehm_try_terminate@ in the standard
+\CFA library. The contents of a try block and the termination handlers are converted
+into nested functions. These are then passed to the try terminate function and it
+calls them, appropriately.
 Because this function is known and fixed (and not an arbitrary function that
 happens to contain a try statement), its LSDA can be generated ahead
 of time.
+Both the LSDA and the personality function for \snake{__cfaehm_try_terminate}
+are set ahead of time using
+Both the LSDA and the personality function for @__cfaehm_try_terminate@ are set ahead of time using
 embedded assembly. This assembly code is handcrafted using C @asm@ statements
 and contains
 enough information for the single try statement the function represents.
+enough information for a single try statement the function represents.
 The three functions passed to try terminate are:
 …
 decides if a catch clause matches the termination exception. It is constructed
 from the conditional part of each handler and runs each check, top to bottom,
+in turn, to see if the exception matches this handler.
+The match is performed in two steps, first a virtual cast is used to check
+if the raised exception is an instance of the declared exception type or
+one of its descendant types, and then the condition is evaluated, if
+present.
+The match function takes a pointer to the exception and returns 0 if the
+in turn, first checking to see if the exception type matches.
+The match is performed in two steps, first a virtual cast is used to see
+if the raised exception is an instance of the declared exception or one of
+its descendant type, and then is the condition true, if present.
+It takes a pointer to the exception and returns 0 if the
 exception is not handled here. Otherwise the return value is the id of the
 handler that matches the exception.
 …
 All three functions are created with GCC nested functions. GCC nested functions
 can be used to create closures,
+in other words,
+functions that can refer to variables in their lexical scope even
+those variables are part of a different function.
+This approach allows the functions to refer to all the
+in other words, functions that can refer to their lexical scope in other
+functions on the stack when called. This approach allows the functions to refer to all the
 variables in scope for the function containing the @try@ statement. These
 nested functions and all other functions besides @__cfaehm_try_terminate@ in
 …
 the operation finishes, otherwise the search continues to the next node.
 If the search reaches the end of the list without finding a try statement
+with a handler clause
+that can handle the exception, the default handler is executed.
+If the default handler returns, control continues after the raise statement.
+that can handle the exception, the default handler is executed and the
+operation finishes, unless it throws an exception.
 Each node has a handler function that does most of the work.
 …
 If no match is found the function returns false.
 The match is performed in two steps, first a virtual cast is used to see
+if the raised exception is an instance of the declared exception type or one
+of its descendant types, if so then it is passed to the custom predicate
+if one is defined.
+% You need to make sure the type is correct before running the predicate
+% because the predicate can depend on that.
+if the raised exception is an instance of the declared exception or one of
+its descendant type, and then is the condition true, if present.
+\PAB{I don't understand this sentence.
+This ordering gives the type guarantee used in the predicate.}
 \autoref{f:ResumptionTransformation} shows the pattern used to transform
 a \CFA try statement with catch clauses into the approprate C functions.
+a \CFA try statement with catch clauses into the appropriate C functions.
 \todo{Explain the Resumption Transformation figure.}
 …
 (see \vpageref{s:ResumptionMarking}), which ignores parts of
 the stack
+already examined, and is accomplished by updating the front of the list as
+the search continues.
+Before the handler is called at a matching node, the head of the list
+already examined, and is accomplished by updating the front of the list as the
+search continues. Before the handler is called at a matching node, the head of the list
 is updated to the next node of the current node. After the search is complete,
 successful or not, the head of the list is reset.
 …
 \section{Finally}
 % Uses destructors and GCC nested functions.
+%\autoref{code:cleanup}
+A finally clause is handled by converting it into a once-off destructor.
+The code inside the clause is placed into GCC nested-function
+with a unique name, and no arguments or return values.
+This nested function is
+\autoref{f:FinallyTransformation} shows the pattern used to transform a \CFA
+try statement with finally clause into the appropriate C functions.
+The finally clause is placed into a GCC nested-function
+with a unique name, and no arguments or return values.  This nested function is
 then set as the cleanup function of an empty object that is declared at the
 beginning of a block placed around the context of the associated try
 statement (see \autoref{f:FinallyTransformation}).
+beginning of a block placed around the context of the associated @try@
+statement.
 \begin{figure}
 …
                 // TRY BLOCK
+        }
+}
 \end{cfa}
 …
 \end{figure}
+The rest is handled by GCC.
+The TRY BLOCK
+contains the try block itself as well as all code generated for handlers.
+Once that code has completed,
+control exits the block and the empty object is cleaned
+The rest is handled by GCC. The try block and all handlers are inside this
+block. At completion, control exits the block and the empty object is cleaned
 up, which runs the function that contains the finally code.
 …
 The first step of cancellation is to find the cancelled stack and its type:
 coroutine, thread or main thread.
+coroutine, thread, or main thread.
 In \CFA, a thread (the construct the user works with) is a user-level thread
 (point of execution) paired with a coroutine, the thread's main coroutine.
 The thread library also stores pointers to the main thread and the current
 thread.
+coroutine.
 If the current thread's main and current coroutines are the same then the
 current stack is a thread stack, otherwise it is a coroutine stack.

doc/theses/andrew_beach_MMath/intro.tex

-              reaeca5f
+              r1d402be
 % Now take a step back and explain what exceptions are generally.
+A language's EHM is a combination of language syntax and run-time
+components that are used to construct, raise, and handle exceptions,
+including all control flow.
+Exceptions are an active mechanism for replacing passive error/return codes and return unions (Go and Rust).
 Exception handling provides dynamic inter-function control flow.
-A language's EHM is a combination of language syntax and run-time
-components that construct, raise, propagate and handle exceptions,
-to provide all of that control flow.
 There are two forms of exception handling covered in this thesis:
 termination, which acts as a multi-level return,
 and resumption, which is a dynamic function call.
+% About other works:
+Often, when this separation is not made, termination exceptions are assumed
+as they are more common and may be the only form of handling provided in
+a language.
+All types of exception handling link a raise with a handler.
+Both operations are usually language primitives, although raises can be
+treated as a primitive function that takes an exception argument.
+Handlers are more complex as they are added to and removed from the stack
+during execution, must specify what they can handle and give the code to
+handle the exception.
+Exceptions work with different execution models but for the descriptions
+that follow a simple call stack, with functions added and removed in a
+first-in-last-out order, is assumed.
+Termination exception handling searches the stack for the handler, then
+unwinds the stack to where the handler was found before calling it.
+The handler is run inside the function that defined it and when it finishes
+it returns control to that function.
+% PAB: Maybe this sentence was suppose to be deleted?
+Termination handling is much more common,
+to the extent that it is often seen as the only form of handling.
+% PAB: I like this sentence better than the next sentence.
+% This separation is uncommon because termination exception handling is so
+% much more common that it is often assumed.
+% WHY: Mention other forms of continuation and \cite{CommonLisp} here?
+Exception handling relies on the concept of nested functions to create handlers that deal with exceptions.
 \begin{center}
+\input{callreturn}
+\begin{tabular}[t]{ll}
+\begin{lstlisting}[aboveskip=0pt,belowskip=0pt,language=CFA,{moredelim=**[is][\color{red}]{@}{@}}]
+void f( void (*hp)() ) {
+        hp();
+}
+void g( void (*hp)() ) {
+        f( hp );
+}
+void h( int @i@, void (*hp)() ) {
+        void @handler@() { // nested
+                printf( "%d\n", @i@ );
+        }
+        if ( i == 1 ) hp = handler;
+        if ( i > 0 ) h( i - 1, hp );
+        else g( hp );
+}
+h( 2, 0 );
+\end{lstlisting}
+&
+\raisebox{-0.5\totalheight}{\input{handler}}
+\end{tabular}
 \end{center}
+Resumption exception handling searches the stack for a handler and then calls
+it without removing any other stack frames.
+The handler is run on top of the existing stack, often as a new function or
+closure capturing the context in which the handler was defined.
+After the handler has finished running it returns control to the function
+that preformed the raise, usually starting after the raise.
+The nested function @handler@ in the second stack frame is explicitly passed to function @f@.
+When this handler is called in @f@, it uses the parameter @i@ in the second stack frame, which is accessible by an implicit lexical-link pointer.
+Setting @hp@ in @h@ at different points in the recursion, results in invoking a different handler.
+Exception handling extends this idea by eliminating explicit handler passing, and instead, performing a stack search for a handler that matches some criteria (conditional dynamic call), and calls the handler at the top of the stack.
+It is the runtime search $O(N)$ that differentiates an EHM call (raise) from normal dynamic call $O(1)$ via a function or virtual-member pointer.
+Termination exception handling searches the stack for a handler, unwinds the stack to the frame containing the matching handler, and calling the handler at the top of the stack.
+\begin{center}
+\input{termination}
+\end{center}
+Note, since the handler can reference variables in @h@, @h@ must remain on the stack for the handler call.
+After the handler returns, control continues after the lexical location of the handler in @h@ (static return)~\cite[p.~108]{Tennent77}.
+Unwinding allows recover to any previous
+function on the stack, skipping any functions between it and the
+function containing the matching handler.
+Resumption exception handling searches the stack for a handler, does \emph{not} unwind the stack to the frame containing the matching handler, and calls the handler at the top of the stack.
 \begin{center}
 \input{resumption}
 \end{center}
+After the handler returns, control continues after the resume in @f@ (dynamic return).
+Not unwinding allows fix up of the problem in @f@ by any previous function on the stack, without disrupting the current set of stack frames.
 Although a powerful feature, exception handling tends to be complex to set up
 and expensive to use
 so it is often limited to unusual or ``exceptional" cases.
 The classic example is error handling, exceptions can be used to
 remove error handling logic from the main execution path, and pay
+The classic example is error handling, where exceptions are used to
+remove error handling logic from the main execution path, while paying
 most of the cost only when the error actually occurs.
 …
 some of the underlying tools used to implement and express exception handling
 in other languages are absent in \CFA.
 Still the resulting syntax resembles that of other languages:
 \begin{cfa}
 try {
+Still the resulting basic syntax resembles that of other languages:
+\begin{lstlisting}[language=CFA,{moredelim=**[is][\color{red}]{@}{@}}]
+@try@ {
         ...
         T * object = malloc(request_size);
         if (!object) {
                 throw OutOfMemory{fixed_allocation, request_size};
+                @throw@ OutOfMemory{fixed_allocation, request_size};
+        }
         ...
 } catch (OutOfMemory * error) {
+} @catch@ (OutOfMemory * error) {
         ...
+}
 \end{cfa}
+\end{lstlisting}
 % A note that yes, that was a very fast overview.
 The design and implementation of all of \CFA's EHM's features are
 …
 % The current state of the project and what it contributes.
+All of these features have been implemented in \CFA,
+covering both changes to the compiler and the run-time.
+In addition, a suite of test cases and performance benchmarks were created
+along side the implementation.
+The implementation techniques are generally applicable in other programming
+The majority of the \CFA EHM is implemented in \CFA, except for a small amount of assembler code.
+In addition,
+a suite of tests and performance benchmarks were created as part of this project.
+The \CFA implementation techniques are generally applicable in other programming
 languages and much of the design is as well.
+Some parts of the EHM use other features unique to \CFA and would be
+harder to replicate in other programming languages.
+Some parts of the EHM use features unique to \CFA, and hence,
+are harder to replicate in other programming languages.
+% Talk about other programming languages.
+Three well known programming languages with EHMs, %/exception handling
+C++, Java and Python are examined in the performance work. However, these languages focus on termination
+exceptions, so there is no comparison with resumption.
 The contributions of this work are:
 \begin{enumerate}
 \item Designing \CFA's exception handling mechanism, adapting designs from
 other programming languages and creating new features.
 \item Implementing stack unwinding and the \CFA EHM, including updating
 the \CFA compiler and the run-time environment.
 \item Designed and implemented a prototype virtual system.
+other programming languages, and creating new features.
+\item Implementing stack unwinding for the \CFA EHM, including updating
+the \CFA compiler and run-time environment to generate and execute the EHM code.
+\item Designing and implementing a prototype virtual system.
 % I think the virtual system and per-call site default handlers are the only
 % "new" features, everything else is a matter of implementation.
+\item Creating tests to check the behaviour of the EHM.
+\item Creating benchmarks to check the performances of the EHM,
+as compared to other languages.
+\item Creating tests and performance benchmarks to compare with EHM's in other languages.
 \end{enumerate}
+The rest of this thesis is organized as follows.
 The current state of exceptions is covered in \autoref{s:background}.
 The existing state of \CFA is also covered in \autoref{c:existing}.
 New EHM features are introduced in \autoref{c:features},
+%\todo{I can't figure out a good lead-in to the roadmap.}
+The thesis is organization as follows.
+The next section and parts of \autoref{c:existing} cover existing EHMs.
+New \CFA EHM features are introduced in \autoref{c:features},
 covering their usage and design.
 That is followed by the implementation of these features in
 \autoref{c:implement}.
+Performance results are examined in \autoref{c:performance}.
+Possibilities to extend this project are discussed in \autoref{c:future}.
+Finally, the project is summarized in \autoref{c:conclusion}.
+Performance results are presented in \autoref{c:performance}.
+Summing up and possibilities for extending this project are discussed in \autoref{c:future}.
 \section{Background}
 \label{s:background}
 Exception handling has been examined before in programming languages,
 with papers on the subject dating back 70s.\cite{Goodenough75}
+Exception handling is a well examined area in programming languages,
+with papers on the subject dating back the 70s~\cite{Goodenough75}.
 Early exceptions were often treated as signals, which carried no information
 except their identity. Ada still uses this system.\todo{cite Ada}
+except their identity. Ada~\cite{Ada} still uses this system.
 The modern flag-ship for termination exceptions is \Cpp,
 which added them in its first major wave of non-object-orientated features
 in 1990.
+\todo{cite https://en.cppreference.com/w/cpp/language/history}
+Many EHMs have special exception types,
+however \Cpp has the ability to use any type as an exception.
+These were found to be not very useful and have been pushed aside for classes
+inheriting from
+% https://en.cppreference.com/w/cpp/language/history
+While many EHMs have special exception types,
+\Cpp has the ability to use any type as an exception.
+However, this generality is not particularly useful, and has been pushed aside for classes, with a convention of inheriting from
 \code{C++}{std::exception}.
+Although there is a special catch-all syntax (@catch(...)@) there are no
+operations that can be performed on the caught value, not even type inspection.
+Instead the base exception-type \code{C++}{std::exception} defines common
+functionality (such as
+the ability to describe the reason the exception was raised) and all
+While \Cpp has a special catch-all syntax @catch(...)@, there is no way to discriminate its exception type, so nothing can
+be done with the caught value because nothing is known about it.
+Instead the base exception-type \code{C++}{std::exception} is defined with common functionality (such as
+the ability to print a message when the exception is raised but not caught) and all
 exceptions have this functionality.
+That trade-off, restricting usable types to gain guaranteed functionality,
+is almost universal now, as without some common functionality it is almost
+impossible to actually handle any errors.
+Java was the next popular language to use exceptions. \todo{cite Java}
+Its exception system largely reflects that of \Cpp, except that requires
+you throw a child type of \code{Java}{java.lang.Throwable}
+Having a root exception-type seems to be the standard now, as the guaranteed functionality is worth
+any lost in flexibility from limiting exceptions types to classes.
+Java~\cite{Java} was the next popular language to use exceptions.
+Its exception system largely reflects that of \Cpp, except it requires
+exceptions to be a subtype of \code{Java}{java.lang.Throwable}
 and it uses checked exceptions.
+Checked exceptions are part of a function's interface,
+the exception signature of the function.
+Every function that could be raised from a function, either directly or
+because it is not handled from a called function, is given.
+Using this information, it is possible to statically verify if any given
+exception is handled and guarantee that no exception will go unhandled.
+Making exception information explicit improves clarity and safety,
+but can slow down or restrict programming.
+For example, programming high-order functions becomes much more complex
+if the argument functions could raise exceptions.
+However, as odd it may seem, the worst problems are rooted in the simple
+inconvenience of writing and updating exception signatures.
+This has caused Java programmers to develop multiple programming ``hacks''
+to circumvent checked exceptions, negating their advantages.
+One particularly problematic example is the ``catch-and-ignore'' pattern,
+where an empty handler is used to handle an exception without doing any
+recovery or repair. In theory that could be good enough to properly handle
+the exception, but more often is used to ignore an exception that the
+programmer does not feel is worth the effort of handling it, for instance if
+they do not believe it will ever be raised.
+If they are incorrect the exception will be silenced, while in a similar
+situation with unchecked exceptions the exception would at least activate
+the language's unhandled exception code (usually program abort with an
+error message).
+Checked exceptions are part of a function's interface defining all exceptions it or its called functions raise.
+Using this information, it is possible to statically verify if a handler exists for all raised exception, \ie no uncaught exceptions.
+Making exception information explicit, improves clarity and
+safety, but can slow down programming.
+For example, programming complexity increases when dealing with high-order methods or an overly specified
+throws clause. However some of the issues are more
+programming annoyances, such as writing/updating many exception signatures after adding or remove calls.
+Java programmers have developed multiple programming ``hacks'' to circumvent checked exceptions negating the robustness it is suppose to provide.
+For example, the ``catch-and-ignore" pattern, where the handler is empty because the exception does not appear relevant to the programmer versus
+repairing or recovering from the exception.
 %\subsection
 Resumption exceptions are less popular,
+although resumption is as old as termination; hence, few
+although resumption is as old as termination;
+hence, few
 programming languages have implemented them.
 % http://bitsavers.informatik.uni-stuttgart.de/pdf/xerox/parc/techReports/
 %   CSL-79-3_Mesa_Language_Manual_Version_5.0.pdf
 Mesa is one programming language that did.\todo{cite Mesa} Experience with Mesa
 is quoted as being one of the reasons resumptions were not
+Mesa~\cite{Mesa} is one programming languages that did. Experience with Mesa
+is quoted as being one of the reasons resumptions are not
 included in the \Cpp standard.
 % https://en.wikipedia.org/wiki/Exception_handling
+Since then resumptions have been ignored in main-stream programming languages.
+However, resumption is being revisited in the context of decades of other
+developments in programming languages.
+While rejecting resumption may have been the right decision in the past,
+the situation has changed since then.
+Some developments, such as the function programming equivalent to resumptions,
+algebraic effects\cite{Zhang19}, are enjoying success.
+A complete reexamination of resumptions is beyond this thesis,
+but there reemergence is enough to try them in \CFA.
+As a result, resumption has ignored in main-stream programming languages.
+However, ``what goes around comes around'' and resumption is being revisited now (like user-level threading).
+While rejecting resumption might have been the right decision in the past, there are decades
+of developments in computer science that have changed the situation.
+Some of these developments, such as functional programming's resumption
+equivalent, algebraic effects\cite{Zhang19}, are enjoying significant success.
+A complete reexamination of resumptions is beyond this thesis, but their re-emergence is
+enough to try them in \CFA.
 % Especially considering how much easier they are to implement than
 % termination exceptions and how much Peter likes them.
 %\subsection
 Functional languages tend to use other solutions for their primary error
 handling mechanism, but exception-like constructs still appear.
 Termination appears in the error construct, which marks the result of an
 expression as an error; then the result of any expression that tries to use
 it also results in an error, and so on until an appropriate handler is reached.
+% termination exceptions.
+%\subsection
+Functional languages tend to use other solutions for their primary EHM,
+but exception-like constructs still appear.
+Termination appears in error construct, which marks the result of an
+expression as an error; thereafter, the result of any expression that tries to use it is also an
+error, and so on until an appropriate handler is reached.
 Resumption appears in algebraic effects, where a function dispatches its
 side-effects to its caller for handling.
 %\subsection
+More recently exceptions seem to be vanishing from newer programming
 languages, replaced by ``panic".
 In Rust, a panic is just a program level abort that may be implemented by
 unwinding the stack like in termination exception handling.\todo{cite Rust}
+Some programming languages have moved to a restricted kind of EHM
+called ``panic".
+In Rust~\cite{Rust}, a panic is just a program level abort that may be implemented by
+unwinding the stack like in termination exception handling.
 % https://doc.rust-lang.org/std/panic/fn.catch_unwind.html
 Go's panic through is very similar to a termination, except it only supports
+In Go~\cite{Go}, a panic is very similar to a termination, except it only supports
 a catch-all by calling \code{Go}{recover()}, simplifying the interface at
 the cost of flexibility.\todo{cite Go}
+the cost of flexibility.
 %\subsection
 While exception handling's most common use cases are in error handling,
 here are some other ways to handle errors with comparisons with exceptions.
+here are other ways to handle errors with comparisons to exceptions.
 \begin{itemize}
 \item\emph{Error Codes}:
+This pattern has a function return an enumeration (or just a set of fixed
+values) to indicate if an error has occurred and possibly which error it was.
+Error codes mix exceptional/error and normal values, enlarging the range of
+possible return values. This can be addressed with multiple return values
+(or a tuple) or a tagged union.
+However, the main issue with error codes is forgetting to check them,
+This pattern has a function return an enumeration (or just a set of fixed values) to indicate
+if an error occurred and possibly which error it was.
+Error codes mix exceptional and normal values, artificially enlarging the type and/or value range.
+Some languages address this issue by returning multiple values or a tuple, separating the error code from the function result.
+However, the main issue with error codes is forgetting to checking them,
 which leads to an error being quietly and implicitly ignored.
+Some new languages and tools will try to issue warnings when an error code
+is discarded to avoid this problem.
+Checking error codes also bloats the main execution path,
+especially if the error is not handled immediately hand has to be passed
+through multiple functions before it is addressed.
+Some new languages have tools that issue warnings, if the error code is
+discarded to avoid this problem.
+Checking error codes also results in bloating the main execution path, especially if an error is not dealt with locally and has to be cascaded down the call stack to a higher-level function..
 \item\emph{Special Return with Global Store}:
+Similar to the error codes pattern but the function itself only returns
+that there was an error
+and store the reason for the error in a fixed global location.
+For example many routines in the C standard library will only return some
+error value (such as -1 or a null pointer) and the error code is written into
+the standard variable @errno@.
+This approach avoids the multiple results issue encountered with straight
+error codes but otherwise has the same disadvantages and more.
+Every function that reads or writes to the global store must agree on all
+possible errors and managing it becomes more complex with concurrency.
+Some functions only return a boolean indicating success or failure
+and store the exact reason for the error in a fixed global location.
+For example, many C routines return non-zero or -1, indicating success or failure,
+and write error details into the C standard variable @errno@.
+This approach avoids the multiple results issue encountered with straight error codes
+but otherwise has many (if not more) of the disadvantages.
+For example, everything that uses the global location must agree on all possible errors and global variable are unsafe with concurrency.
 \item\emph{Return Union}:
 …
 so that one type can be used everywhere in error handling code.
 This pattern is very popular in any functional or semi-functional language
 with primitive support for tagged unions (or algebraic data types).
 % We need listing Rust/rust to format code snippets from it.
+This pattern is very popular in functional or any semi-functional language with
+primitive support for tagged unions (or algebraic data types).
+% We need listing Rust/rust to format code snipits from it.
 % Rust's \code{rust}{Result<T, E>}
 The main advantage is that an arbitrary object can be used to represent an
 error so it can include a lot more information than a simple error code.
+The disadvantages include that the it does have to be checked along the main
+execution and if there aren't primitive tagged unions proper usage can be
 hard to enforce.
+The main advantage is providing for more information about an
+error, other than one of a fix-set of ids.
+While some languages use checked union access to force error-code checking,
+it is still possible to bypass the checking.
+The main disadvantage is again significant error code on the main execution path and cascading through called functions.
 \item\emph{Handler Functions}:
 This pattern associates errors with functions.
 On error, the function that produced the error calls another function to
+This pattern implicitly associates functions with errors.
+On error, the function that produced the error implicitly calls another function to
 handle it.
 The handler function can be provided locally (passed in as an argument,
 either directly as as a field of a structure/object) or globally (a global
 variable).
+C++ uses this approach as its fallback system if exception handling fails,
+such as \snake{std::terminate_handler} and, for a time,
+\snake{std::unexpected_handler}.
+Handler functions work a lot like resumption exceptions,
+but without the dynamic search for a handler.
+Since setting up the handler can be more complex/expensive,
+especially when the handler has to be passed through multiple layers of
+function calls, but cheaper (constant time) to call,
+they are more suited to more frequent (less exceptional) situations.
+C++ uses this approach as its fallback system if exception handling fails, \eg
+\snake{std::terminate_handler} and for a time \snake{std::unexpected_handler}
+Handler functions work a lot like resumption exceptions, without the dynamic handler search.
+Therefore, setting setting up the handler can be more complex/expensive, especially if the handle must be passed through multiple function calls, but cheaper to call $O(1)$, and hence,
+are more suited to frequent exceptional situations.
+% The exception being global handlers if they are rarely change as the time
+% in both cases shrinks towards zero.
 \end{itemize}
 %\subsection
 Because of their cost, exceptions are rarely used for hot paths of execution.
+Hence, there is an element of self-fulfilling prophecy as implementation
+techniques have been focused on making them cheap to set-up,
+happily making them expensive to use in exchange.
+This difference is less important in higher-level scripting languages,
+where using exception for other tasks is more common.
+An iconic example is Python's \code{Python}{StopIteration} exception that
+is thrown by an iterator to indicate that it is exhausted.
+When paired with Python's iterator-based for-loop this will be thrown every
+time the end of the loop is reached.
+\todo{Cite Python StopIteration and for-each loop.}
+Therefore, there is an element of self-fulfilling prophecy for implementation
+techniques to make exceptions cheap to set-up at the cost
+of expensive usage.
+This cost differential is less important in higher-level scripting languages, where use of exceptions for other tasks is more common.
+An iconic example is Python's @StopIteration@ exception that is thrown by
+an iterator to indicate that it is exhausted, especially when combined with Python's heavy
+use of the iterator-based for-loop.
 % https://docs.python.org/3/library/exceptions.html#StopIteration

doc/theses/andrew_beach_MMath/performance.tex

-              reaeca5f
+              r1d402be
 Tests were run in \CFA, C++, Java and Python.
 In addition there are two sets of tests for \CFA,
 one for termination and once with resumption.
+one for termination and one for resumption exceptions.
 C++ is the most comparable language because both it and \CFA use the same
 framework, libunwind.
 In fact, the comparison is almost entirely in quality of implementation.
+In fact, the comparison is almost entirely a quality of implementation.
 Specifically, \CFA's EHM has had significantly less time to be optimized and
 does not generate its own assembly. It does have a slight advantage in that
 \Cpp has to do some extra bookkeeping to support its utility functions,
+there are some features it handles directly instead of through utility functions,
 but otherwise \Cpp should have a significant advantage.
 Java a popular language with similar termination semantics, but
+Java is a popular language with similar termination semantics, but
 it is implemented in a very different environment, a virtual machine with
 garbage collection.
 It also implements the finally clause on try blocks allowing for a direct
+It also implements the @finally@ clause on @try@ blocks allowing for a direct
 feature-to-feature comparison.
 As with \Cpp, Java's implementation is mature, has more optimizations
 and extra features as compared to \CFA.
+As with \Cpp, Java's implementation is mature, optimized
+and has extra features.
 Python is used as an alternative comparison because of the \CFA EHM's
 current performance goals, which is to not be prohibitively slow while the
+current performance goals, which is not to be prohibitively slow while the
 features are designed and examined. Python has similar performance goals for
 creating quick scripts and its wide use suggests it has achieved those goals.
 …
 resumption exceptions. Even the older programming languages with resumption
 seem to be notable only for having resumption.
+Instead, resumption is compared to its simulation in other programming
+languages: fixup functions that are explicity passed into a function.
+So instead, resumption is compared to its simulation in other programming
+languages using fixup functions that are explicitly passed for correction or
+logging purposes.
+% So instead, resumption is compared to a less similar but much more familiar
+%feature, termination exceptions.
 All tests are run inside a main loop that repeatedly performs a test.
 This approach avoids start-up or tear-down time from
 affecting the timing results.
+The number of times the loop is run is configurable from the command line,
+the number used in the timing runs is given with the results per test.
+Tests ran their main loop a million times.
+Each test is run a N times (configurable from the command line).
 The Java tests runs the main loop 1000 times before
 beginning the actual test to ``warm-up" the JVM.
-% All other languages are precompiled or interpreted.
 Timing is done internally, with time measured immediately before and
 …
 The exceptions used in these tests are always based off of
+the base exception for the language.
+This requirement minimizes performance differences based
+a base exception. This requirement minimizes performance differences based
 on the object model used to represent the exception.
 …
 For example, empty inline assembly blocks are used in \CFA and \Cpp to
 prevent excessive optimizations while adding no actual work.
+Each test was run eleven times. The top three and bottom three results were
+discarded and the remaining five values are averaged.
+The tests are compiled with gcc-10 for \CFA and g++-10 for \Cpp. Java is
+compiled with version 11.0.11. Python with version 3.8. The tests were run on:
+\begin{itemize}[nosep]
+\item
+ARM 2280 Kunpeng 920 48-core 2$\times$socket \lstinline{@} 2.6 GHz running Linux v5.11.0-25
+\item
+AMD 6380 Abu Dhabi 16-core 4$\times$socket \lstinline{@} 2.5 GHz running Linux v5.11.0-25
+\end{itemize}
+Two kinds of hardware architecture allows discriminating any implementation and
+architectural effects.
 % We don't use catch-alls but if we did:
 % Catch-alls are done by catching the root exception type (not using \Cpp's
 % \code{C++}{catch(...)}).
-When collecting data each test is run eleven times. The top three and bottom
-three results are discarded and the remaining five values are averaged.
-The test are run with the latest (still pre-release) \CFA compiler was used,
-using gcc-10 as a backend.
-g++-10 is used for \Cpp.
-Java tests are complied and run with version 11.0.11.
-Python used version 3.8.
-The machines used to run the tests are:
-\todo{Get patch versions for python, gcc and g++.}
-\begin{itemize}[nosep]
-\item ARM 2280 Kunpeng 920 48-core 2$\times$socket
-      \lstinline{@} 2.6 GHz running Linux v5.11.0-25
-\item AMD 6380 Abu Dhabi 16-core 4$\times$socket
-      \lstinline{@} 2.5 GHz running Linux v5.11.0-25
-\end{itemize}
-Representing the two major families of hardware architecture.
 \section{Tests}
 …
 They should provide a guide as to where the EHM's costs are found.
+\paragraph{Stack Traversal}
+This group measures the cost of traversing the stack,
+(and in termination, unwinding it).
+Inside the main loop is a call to a recursive function.
+This function calls itself F times before raising an exception.
+F is configurable from the command line, but is usually 100.
+This builds up many stack frames, and any contents they may have,
+before the raise.
+The exception is always handled at the base of the stack.
+For example the Empty test for \CFA resumption looks like:
+\begin{cfa}
+\paragraph{Raise and Handle}
+This group measures the cost of a try statement when exceptions are raised and
+the stack is unwound (termination) or not unwound (resumption).  Each test has
+has a repeating function like the following
+\begin{lstlisting}[language=CFA,{moredelim=**[is][\color{red}]{@}{@}}]
 void unwind_empty(unsigned int frames) {
         if (frames) {
+                unwind_empty(frames - 1);
+                @unwind_empty(frames - 1);@ // AUGMENTED IN OTHER EXPERIMENTS
+        } else throw (empty_exception){&empty_vt};
+}
+\end{lstlisting}
+which is called N times, where each call recurses to a depth of R (configurable from the command line), an
+exception is raised, the stack is a unwound, and the exception caught.
+\begin{itemize}[nosep]
+\item Empty:
+For termination, this test measures the cost of raising (stack walking) an
+exception through empty stack frames from the bottom of the recursion to an
+empty handler, and unwinding the stack. (see above code)
+\medskip
+For resumption, this test measures the same raising cost but does not unwind
+the stack. For languages without resumption, a fixup function is to the bottom
+of the recursion and called to simulate a fixup operation at that point.
+\begin{cfa}
+void nounwind_fixup(unsigned int frames, void (*raised_rtn)(int &)) {
+        if (frames) {
+                nounwind_fixup(frames - 1, raised_rtn);
         } else {
+                throwResume (empty_exception){&empty_vt};
+                int fixup = 17;
+                raised_rtn(fixup);
+        }
+}
 \end{cfa}
+Other test cases have additional code around the recursive call add
+something besides simple stack frames to the stack.
+Note that both termination and resumption will have to traverse over
+the stack but only termination has to unwind it.
+\begin{itemize}[nosep]
+% \item None:
+% Reuses the empty test code (see below) except that the number of frames
+% is set to 0 (this is the only test for which the number of frames is not
+% 100). This isolates the start-up and shut-down time of a throw.
+\item Empty:
+The repeating function is empty except for the necessary control code.
+As other traversal tests add to this, so it is the baseline for the group
+as the cost comes from traversing over and unwinding a stack frame
+that has no other interactions with the exception system.
+where the passed fixup function is:
+\begin{cfa}
+void raised(int & fixup) {
+        fixup = 42;
+}
+\end{cfa}
+For comparison, a \CFA version passing a function is also included.
 \item Destructor:
+The repeating function creates an object with a destructor before calling
+itself.
+Comparing this to the empty test gives the time to traverse over and/or
+unwind a destructor.
+This test measures the cost of raising an exception through non-empty frames,
+where each frame has an object requiring destruction, from the bottom of the
+recursion to an empty handler. Hence, there are N destructor calls during
+unwinding.
+\medskip
+This test is not meaningful for resumption because the stack is only unwound as
+the recursion returns.
+\begin{cfa}
+        WithDestructor object;
+        unwind_destructor(frames - 1);
+\end{cfa}
 \item Finally:
+The repeating function calls itself inside a try block with a finally clause
+attached.
+Comparing this to the empty test gives the time to traverse over and/or
+unwind a finally clause.
+This test measures the cost of establishing a try block with an empty finally
+clause on the front side of the recursion and running the empty finally clauses
+during stack unwinding from the bottom of the recursion to an empty handler.
+\begin{cfa}
+        try {
+                unwind_finally(frames - 1);
+        } finally {}
+\end{cfa}
+\medskip
+This test is not meaningful for resumption because the stack is only unwound as
+the recursion returns.
 \item Other Handler:
+The repeating function calls itself inside a try block with a handler that
+will not match the raised exception, but is of the same kind of handler.
+This means that the EHM will have to check each handler, but will continue
+over all of the until it reaches the base of the stack.
+Comparing this to the empty test gives the time to traverse over and/or
+unwind a handler.
+For termination, this test is like the finally test but the try block has a
+catch clause for an exception that is not raised, so catch matching is executed
+during stack unwinding but the match never successes until the catch at the
+bottom of the recursion.
+\begin{cfa}
+        try {
+                unwind_other(frames - 1);
+        } catch (not_raised_exception *) {}
+\end{cfa}
+\medskip
+For resumption, this test measures the same raising cost but does not unwind
+the stack. For languages without resumption, the same fixup function is passed
+and called.
 \end{itemize}
+\paragraph{Cross Try Statement}
+This group of tests measures the cost setting up exception handling if it is
+not used (because the exceptional case did not occur).
+Tests repeatedly cross (enter and leave, execute) a try statement but never
+preform a raise.
+\paragraph{Try/Handle/Finally Statement}
+This group measures just the cost of executing a try statement so
+\emph{there is no stack unwinding}.  Hence, the program main loops N times
+around:
+\begin{cfa}
+try {
+} catch (not_raised_exception *) {}
+\end{cfa}
 \begin{itemize}[nosep]
 \item Handler:
 The try statement has a handler (of the appropriate kind).
+The try statement has a handler (catch/resume).
 \item Finally:
 The try statement has a finally clause.
 …
 This group measures the cost of conditional matching.
 Only \CFA implements the language level conditional match,
 the other languages mimic it with an ``unconditional" match (it still
 checks the exception's type) and conditional re-raise if it is not supposed
+the other languages mimic with an ``unconditional" match (it still
+checks the exception's type) and conditional re-raise if it is not suppose
 to handle that exception.
+There is the pattern shown in \CFA and \Cpp. Java and Python use the same
+pattern as \Cpp, but with their own syntax.
+\begin{minipage}{0.45\textwidth}
+\begin{center}
+\begin{tabular}{ll}
+\multicolumn{1}{c}{\CFA} & \multicolumn{1}{c}{\Cpp, Java, Python} \\
 \begin{cfa}
 try {
+        ...
+} catch (exception_t * e ;
+                should_catch(e)) {
+        ...
+        throw_exception();
+} catch (empty_exception * exc;
+                 should_catch) {
+}
 \end{cfa}
+\end{minipage}
+\begin{minipage}{0.55\textwidth}
+\begin{lstlisting}[language=C++]
+&
+\begin{cfa}
 try {
+        ...
+} catch (std::exception & e) {
+        if (!should_catch(e)) throw;
+        ...
+        throw_exception();
+} catch (EmptyException & exc) {
+        if (!should_catch) throw;
+}
+\end{lstlisting}
+\end{minipage}
+\end{cfa}
+\end{tabular}
+\end{center}
 \begin{itemize}[nosep]
 \item Match All:
 …
 \end{itemize}
+\paragraph{Resumption Simulation}
+A slightly altered version of the Empty Traversal test is used when comparing
+resumption to fix-up routines.
+The handler, the actual resumption handler or the fix-up routine,
+always captures a variable at the base of the loop,
+and receives a reference to a variable at the raise site, either as a
+field on the exception or an argument to the fix-up routine.
+% I don't actually know why that is here but not anywhere else.
+\medskip
+\noindent
+All omitted test code for other languages is functionally identical to the \CFA
+tests or simulated, and available online~\cite{CforallExceptionBenchmarks}.
 %\section{Cost in Size}
 …
 \section{Results}
+% First, introduce the tables.
+\autoref{t:PerformanceTermination},
+\autoref{t:PerformanceResumption}
+and~\autoref{t:PerformanceFixupRoutines}
+show the test results.
+In cases where a feature is not supported by a language, the test is skipped
+for that language and the result is marked N/A.
+There are also cases where the feature is supported but measuring its
+cost is impossible. This happened with Java, which uses a JIT that optimize
+away the tests and it cannot be stopped.\cite{Dice21}
+These tests are marked N/C.
+To get results in a consistent range (1 second to 1 minute is ideal,
+going higher is better than going low) N, the number of iterations of the
+main loop in each test, is varied between tests. It is also given in the
+results and usually have a value in the millions.
+An anomaly in some results came from \CFA's use of gcc nested functions.
+These nested functions are used to create closures that can access stack
+variables in their lexical scope.
+However, if they do so then they can cause the benchmark's run-time to
+increase by an order of magnitude.
+The simplest solution is to make those values global variables instead
+of function local variables.
+% Do we know if editing a global inside nested function is a problem?
+Tests that had to be modified to avoid this problem have been marked
+with a ``*'' in the results.
+% Now come the tables themselves:
+% You might need a wider window for this.
+\begin{table}[htb]
+One result not directly related to \CFA but important to keep in
+mind is that, for exceptions, the standard intuition about which languages
+should go faster often does not hold. For example, there are a few cases where Python out-performs
+\CFA, \Cpp and Java. The most likely explanation is that, since exceptions are
+rarely considered to be the common case, the more optimized languages
+make that case expense. In addition, languages with high-level
+representations have a much easier time scanning the stack as there is less
+to decode.
+Tables~\ref{t:PerformanceTermination} and~\ref{t:PerformanceResumption} show
+the test results for termination and resumption, respectively.  In cases where
+a feature is not supported by a language, the test is skipped for that language
+(marked N/A).  For some Java experiments it was impossible to measure certain
+effects because the JIT corrupted the test (marked N/C). No workaround was
+possible~\cite{Dice21}.  To get experiments in the range of 1--100 seconds, the
+number of times an experiment is run (N) is varied (N is marked beside each
+experiment, e.g., 1M $\Rightarrow$ 1 million test iterations).
+An anomaly exists with gcc nested functions used as thunks for implementing
+much of the \CFA EHM. If a nested-function closure captures local variables in
+its lexical scope, performance dropped by a factor of 10.  Specifically, in try
+statements of the form:
+\begin{cfa}
+        try {
+                unwind_other(frames - 1);
+        } catch (not_raised_exception *) {}
+\end{cfa}
+the try block is hoisted into a nested function and the variable @frames@ is
+the local parameter to the recursive function, which triggers the anomaly. The
+workaround is to remove the recursion parameter and make it a global variable
+that is explicitly decremented outside of the try block (marked with a ``*''):
+\begin{cfa}
+        frames -= 1;
+        try {
+                unwind_other();
+        } catch (not_raised_exception *) {}
+\end{cfa}
+To make comparisons fair, a dummy parameter is added and the dummy value passed
+in the recursion. Note, nested functions in gcc are rarely used (if not
+completely unknown) and must follow the C calling convention, unlike \Cpp
+lambdas, so it is not surprising if there are performance issues efficiently
+capturing closures.
+% Similarly, if a test does not change between resumption
+% and termination in \CFA, then only one test is written and the result
+% was put into the termination column.
+% Raw Data:
+% run-algol-a.sat
+% ---------------
+% Raise Empty   &  82687046678 &  291616256 &   3252824847 & 15422937623 & 14736271114 \\
+% Raise D'tor   & 219933199603 &  297897792 & 223602799362 &         N/A &         N/A \\
+% Raise Finally & 219703078448 &  298391745 &          N/A &         ... & 18923060958 \\
+% Raise Other   & 296744104920 & 2854342084 & 112981255103 & 15475924808 & 21293137454 \\
+% Cross Handler &      9256648 &   13518430 &       769328 &     3486252 &    31790804 \\
+% Cross Finally &       769319 &        N/A &          N/A &     2272831 &    37491962 \\
+% Match All     &   3654278402 &   47518560 &   3218907794 &  1296748192 &   624071886 \\
+% Match None    &   4788861754 &   58418952 &   9458936430 &  1318065020 &   625200906 \\
+%
+% run-algol-thr-c
+% ---------------
+% Raise Empty   &   3757606400 &   36472972 &   3257803337 & 15439375452 & 14717808642 \\
+% Raise D'tor   &  64546302019 &  102148375 & 223648121635 &         N/A &         N/A \\
+% Raise Finally &  64671359172 &  103285005 &          N/A & 15442729458 & 18927008844 \\
+% Raise Other   & 294143497130 & 2630130385 & 112969055576 & 15448220154 & 21279953424 \\
+% Cross Handler &      9646462 &   11955668 &       769328 &     3453707 &    31864074 \\
+% Cross Finally &       773412 &        N/A &          N/A &     2253825 &    37266476 \\
+% Match All     &   3719462155 &   43294042 &   3223004977 &  1286054154 &   623887874 \\
+% Match None    &   4971630929 &   55311709 &   9481225467 &  1310251289 &   623752624 \\
+%
+% run-algol-04-a
+% --------------
+% Raise Empty   & 0.0 & 0.0 &  3250260945 & 0.0 & 0.0 \\
+% Raise D'tor   & 0.0 & 0.0 & 29017675113 & N/A & N/A \\
+% Raise Finally & 0.0 & 0.0 &         N/A & 0.0 & 0.0 \\
+% Raise Other   & 0.0 & 0.0 & 24411823773 & 0.0 & 0.0 \\
+% Cross Handler & 0.0 & 0.0 &      769334 & 0.0 & 0.0 \\
+% Cross Finally & 0.0 & N/A &         N/A & 0.0 & 0.0 \\
+% Match All     & 0.0 & 0.0 &  3254283504 & 0.0 & 0.0 \\
+% Match None    & 0.0 & 0.0 &  9476060146 & 0.0 & 0.0 \\
+% run-plg7a-a.sat
+% ---------------
+% Raise Empty   &  57169011329 &  296612564 &   2788557155 & 17511466039 & 23324548496 \\
+% Raise D'tor   & 150599858014 &  318443709 & 149651693682 &         N/A &         N/A \\
+% Raise Finally & 148223145000 &  373325807 &          N/A &         ... & 29074552998 \\
+% Raise Other   & 189463708732 & 3017109322 &  85819281694 & 17584295487 & 32602686679 \\
+% Cross Handler &      8001654 &   13584858 &      1555995 &     6626775 &    41927358 \\
+% Cross Finally &      1002473 &        N/A &          N/A &     4554344 &    51114381 \\
+% Match All     &   3162460860 &   37315018 &   2649464591 &  1523205769 &   742374509 \\
+% Match None    &   4054773797 &   47052659 &   7759229131 &  1555373654 &   744656403 \\
+%
+% run-plg7a-thr-a
+% ---------------
+% Raise Empty   &   3604235388 &   29829965 &   2786931833 & 17576506385 & 23352975105 \\
+% Raise D'tor   &  46552380948 &  178709605 & 149834207219 &         N/A &         N/A \\
+% Raise Finally &  46265157775 &  177906320 &          N/A & 17493045092 & 29170962959 \\
+% Raise Other   & 195659245764 & 2376968982 &  86070431924 & 17552979675 & 32501882918 \\
+% Cross Handler &    397031776 &   12503552 &      1451225 &     6658628 &    42304965 \\
+% Cross Finally &      1136746 &        N/A &          N/A &     4468799 &    46155817 \\
+% Match All     &   3189512499 &   39124453 &   2667795989 &  1525889031 &   733785613 \\
+% Match None    &   4094675477 &   48749857 &   7850618572 &  1566713577 &   733478963 \\
+%
+% run-plg7a-04-a
+% --------------
+% 0.0 are unfilled.
+% Raise Empty   & 0.0 & 0.0 &  2770781479 & 0.0 & 0.0 \\
+% Raise D'tor   & 0.0 & 0.0 & 23530084907 & N/A & N/A \\
+% Raise Finally & 0.0 & 0.0 &         N/A & 0.0 & 0.0 \\
+% Raise Other   & 0.0 & 0.0 & 23816827982 & 0.0 & 0.0 \\
+% Cross Handler & 0.0 & 0.0 &     1422188 & 0.0 & 0.0 \\
+% Cross Finally & 0.0 & N/A &         N/A & 0.0 & 0.0 \\
+% Match All     & 0.0 & 0.0 &  2671989778 & 0.0 & 0.0 \\
+% Match None    & 0.0 & 0.0 &  7829059869 & 0.0 & 0.0 \\
+\begin{table}
 \centering
 \caption{Termination Performance Results (sec)}
+\caption{Performance Results Termination (sec)}
 \label{t:PerformanceTermination}
 \begin{tabular}{|r|*{2}{|r r r r|}}
 \hline
                        & \multicolumn{4}{c||}{AMD}         & \multicolumn{4}{c|}{ARM}  \\
+                        & \multicolumn{4}{c||}{AMD}             & \multicolumn{4}{c|}{ARM}      \\
 \cline{2-9}
 N\hspace{8pt}          & \multicolumn{1}{c}{\CFA} & \multicolumn{1}{c}{\Cpp} & \multicolumn{1}{c}{Java} & \multicolumn{1}{c||}{Python} &
                          \multicolumn{1}{c}{\CFA} & \multicolumn{1}{c}{\Cpp} & \multicolumn{1}{c}{Java} & \multicolumn{1}{c|}{Python} \\
 \hline
 Empty Traversal (1M)   & 3.4   & 2.8   & 18.3  & 23.4      & 3.7   & 3.2   & 15.5  & 14.8  \\
 D'tor Traversal (1M)   & 48.4  & 23.6  & N/A   & N/A       & 64.2  & 29.0  & N/A   & N/A   \\
 Finally Traversal (1M) & 3.4*  & N/A   & 17.9  & 29.0      & 4.1*  & N/A   & 15.6  & 19.0  \\
 Other Traversal (1M)   & 3.6*  & 23.2  & 18.2  & 32.7      & 4.0*  & 24.5  & 15.5  & 21.4  \\
 Cross Handler (100M)   & 6.0   & 0.9   & N/C   & 37.4      & 10.0  & 0.8   & N/C   & 32.2  \\
 Cross Finally (100M)   & 0.9   & N/A   & N/C   & 44.1      & 0.8   & N/A   & N/C   & 37.3  \\
 Match All (10M)        & 32.9  & 20.7  & 13.4  & 4.9       & 36.2  & 24.5  & 12.0  & 3.1   \\
 Match None (10M)       & 32.7  & 50.3  & 11.0  & 5.1       & 36.3  & 71.9  & 12.3  & 4.2   \\
+N\hspace{8pt}           & \multicolumn{1}{c}{\CFA} & \multicolumn{1}{c}{\Cpp} & \multicolumn{1}{c}{Java} & \multicolumn{1}{c||}{Python} &
+                          \multicolumn{1}{c}{\CFA} & \multicolumn{1}{c}{\Cpp} & \multicolumn{1}{c}{Java} & \multicolumn{1}{c|}{Python} \\
+\hline
+Throw Empty (1M)        & 3.4   & 2.8   & 18.3  & 23.4          & 3.7   & 3.2   & 15.5  & 14.8  \\
+Throw D'tor (1M)        & 48.4  & 23.6  & N/A   & N/A           & 64.2  & 29.0  & N/A   & N/A   \\
+Throw Finally (1M)      & 3.4*  & N/A   & 17.9  & 29.0          & 4.1*  & N/A   & 15.6  & 19.0  \\
+Throw Other (1M)        & 3.6*  & 23.2  & 18.2  & 32.7          & 4.0*  & 24.5  & 15.5  & 21.4  \\
+Try/Catch (100M)        & 6.0   & 0.9   & N/C   & 37.4          & 10.0  & 0.8   & N/C   & 32.2  \\
+Try/Finally (100M)      & 0.9   & N/A   & N/C   & 44.1          & 0.8   & N/A   & N/C   & 37.3  \\
+Match All (10M)         & 32.9  & 20.7  & 13.4  & 4.9           & 36.2  & 24.5  & 12.0  & 3.1   \\
+Match None (10M)        & 32.7  & 50.3  & 11.0  & 5.1           & 36.3  & 71.9  & 12.3  & 4.2   \\
 \hline
 \end{tabular}
 \end{table}
 \begin{table}[htb]
+\begin{table}
 \centering
+\caption{Resumption Performance Results (sec)}
+\small
+\caption{Performance Results Resumption (sec)}
 \label{t:PerformanceResumption}
+\begin{tabular}{|r||r||r|}
+\setlength{\tabcolsep}{5pt}
+\begin{tabular}{|r|*{2}{|r r r r|}}
 \hline
+N\hspace{8pt}
+                        & AMD     & ARM  \\
+\hline
 Empty Traversal (10M)   & 0.2     & 0.3  \\
+D'tor Traversal (10M)   & 1.8     & 1.0  \\
 Finally Traversal (10M) & 1.7     & 1.0  \\
 Other Traversal (10M)   & 22.6    & 25.9 \\
 Cross Handler (100M)    & 8.4     & 11.9 \\
 Match All (100M)        & 2.3     & 3.2  \\
 Match None (100M)       & 2.9     & 3.9  \\
+                        & \multicolumn{4}{c||}{AMD}             & \multicolumn{4}{c|}{ARM}      \\
+\cline{2-9}
+N\hspace{8pt}           & \multicolumn{1}{c}{\CFA (R/F)} & \multicolumn{1}{c}{\Cpp} & \multicolumn{1}{c}{Java} & \multicolumn{1}{c||}{Python} &
+                          \multicolumn{1}{c}{\CFA (R/F)} & \multicolumn{1}{c}{\Cpp} & \multicolumn{1}{c}{Java} & \multicolumn{1}{c|}{Python} \\
+\hline
+Resume Empty (10M)      & 3.8/3.5       & 14.7  & 2.3   & 176.1 & 0.3/0.1       & 8.9   & 1.2   & 119.9 \\
+Resume Other (10M)      & 4.0*/0.1*     & 21.9  & 6.2   & 381.0 & 0.3*/0.1*     & 13.2  & 5.0   & 290.7 \\
+Try/Resume (100M)       & 8.8           & N/A   & N/A   & N/A   & 12.3          & N/A   & N/A   & N/A   \\
+Match All (10M)         & 0.3           & N/A   & N/A   & N/A   & 0.3           & N/A   & N/A   & N/A   \\
+Match None (10M)        & 0.3           & N/A   & N/A   & N/A   & 0.4           & N/A   & N/A   & N/A   \\
 \hline
 \end{tabular}
 \end{table}
+\begin{table}[htb]
+\centering
+\small
+\caption{Resumption/Fixup Routine Comparison (sec)}
+\label{t:PerformanceFixupRoutines}
+\setlength{\tabcolsep}{5pt}
+\begin{tabular}{|r|*{2}{|r r r r r|}}
+\hline
+            & \multicolumn{5}{c||}{AMD}     & \multicolumn{5}{c|}{ARM}  \\
+\cline{2-11}
+N\hspace{8pt}       & \multicolumn{1}{c}{Raise} & \multicolumn{1}{c}{\CFA} & \multicolumn{1}{c}{\Cpp} & \multicolumn{1}{c}{Java} & \multicolumn{1}{c||}{Python} &
+              \multicolumn{1}{c}{Raise} & \multicolumn{1}{c}{\CFA} & \multicolumn{1}{c}{\Cpp} & \multicolumn{1}{c}{Java} & \multicolumn{1}{c|}{Python} \\
+\hline
+Resume Empty (10M)  & 3.8  & 3.5  & 14.7  & 2.3   & 176.1 & 0.3  & 0.1  & 8.9   & 1.2   & 119.9 \\
+%Resume Other (10M)  & 4.0* & 0.1* & 21.9  & 6.2   & 381.0 & 0.3* & 0.1* & 13.2  & 5.0   & 290.7 \\
+\hline
+\end{tabular}
+\end{table}
+% Now discuss the results in the tables.
+One result not directly related to \CFA but important to keep in mind is that,
+for exceptions the standard intuition about which languages should go
+faster often does not hold.
+For example, there are a few cases where Python out-performs
+\CFA, \Cpp and Java.
+The most likely explanation is that, since exceptions
+are rarely considered to be the common case, the more optimized languages
+make that case expensive to improve other cases.
+In addition, languages with high-level representations have a much
+easier time scanning the stack as there is less to decode.
+As stated,
+the performance tests are not attempting to show \CFA has a new competitive
+way of implementing exception handling.
+The only performance requirement is to insure the \CFA EHM has reasonable
+performance for prototyping.
+Although that may be hard to exactly quantify, we believe it has succeeded
+in that regard.
+Details on the different test cases follow.
+As stated, the performance tests are not attempting to compare exception
+handling across languages.  The only performance requirement is to ensure the
+\CFA EHM implementation runs in a reasonable amount of time, given its
+constraints. In general, the \CFA implement did very well. Each of the tests is
+analysed.
 \begin{description}
+\item[Empty Traversal]
+\CFA is slower than \Cpp, but is still faster than the other languages
+and closer to \Cpp than other languages.
+This is to be expected as \CFA is closer to \Cpp than the other languages.
+\item[D'tor Traversal]
+Running destructors causes huge slowdown in every language that supports
+them. \CFA has a higher proportionate slowdown but it is similar to \Cpp's.
+Considering the amount of work done in destructors is so low the cost
+likely comes from the change of context required to do that work.
+\item[Finally Traversal]
+Speed is similar to Empty Traversal in all languages that support finally
+clauses. Only Python seems to have a larger than random noise change in
+its run-time and it is still not large.
+Despite the similarity between finally clauses and destructors,
+finally clauses seem to avoid the spike in run-time destructors have.
+Possibly some optimization removes the cost of changing contexts.
+\todo{OK, I think the finally clause may have been optimized out.}
+\item[Other Traversal]
+For \Cpp, stopping to check if a handler applies seems to be about as
+expensive as stopping to run a destructor.
+This results in a significant jump.
+Other languages experiance a small increase in run-time.
+The small increase likely comes from running the checks,
+but they could avoid the spike by not having the same kind of overhead for
+switching to the check's context.
+\todo{Could revist Other Traversal, after Finally Traversal.}
+\item[Cross Handler]
+Here \CFA falls behind \Cpp by a much more significant margin.
+This is likely due to the fact \CFA has to insert two extra function
+calls while \Cpp doesn't have to do execute any other instructions.
+Python is much further behind.
+\item[Cross Finally]
+\CFA's performance now matches \Cpp's from Cross Handler.
+If the code from the finally clause is being inlined,
+which is just a asm comment, than there are no additional instructions
+to execute again when exiting the try statement normally.
+\item[Conditional Match]
+Both of the conditional matching tests can be considered on their own,
+however for evaluating the value of conditional matching itself the
+comparison of the two sets of results is useful.
+Consider the massive jump in run-time for \Cpp going from match all to match
+none, which none of the other languages have.
+Some strange interaction is causing run-time to more than double for doing
+twice as many raises.
+Java and Python avoid this problem and have similar run-time for both tests,
+possibly through resource reuse or their program representation.
+However \CFA is built like \Cpp and avoids the problem as well, this matches
+the pattern of the conditional match which makes the two execution paths
+much more similar.
+\item[Throw/Resume Empty]
+For termination, \CFA is close to \Cpp, where other languages have a higher cost.
+For resumption, \CFA is better than the fixup simulations in the other languages, except Java.
+The \CFA results on the ARM computer for both resumption and function simulation are particularly low;
+I have no explanation for this anomaly, except the optimizer has managed to remove part of the experiment.
+Python has a high cost for passing the lambda during the recursion.
+\item[Throw D'tor]
+For termination, \CFA is twice the cost of \Cpp.
+The higher cost for \CFA must be related to how destructors are handled.
+\item[Throw Finally]
+\CFA is better than the other languages with a @finally@ clause, which is the
+same for termination and resumption.
+\item[Throw/Resume Other]
+For termination, \CFA is better than the other languages.
+For resumption, \CFA is equal to or better the other languages.
+Again, the \CFA results on the ARM computer for both resumption and function simulation are particularly low.
+Python has a high cost for passing the lambda during the recursion.
+\item[Try/Catch/Resume]
+For termination, installing a try statement is more expressive than \Cpp
+because the try components are hoisted into local functions.  At runtime, these
+functions are than passed to libunwind functions to set up the try statement.
+\Cpp zero-cost try-entry accounts for its performance advantage.
+For resumption, there are similar costs to termination to set up the try
+statement but libunwind is not used.
+\item[Try/Finally]
+Setting up a try finally is less expensive in \CFA than setting up handlers,
+and is significantly less than other languages.
+\item[Throw/Resume Match All]
+For termination, \CFA is close to the other language simulations.
+For resumption, the stack unwinding is much faster because it does not use
+libunwind.  Instead resumption is just traversing a linked list with each node
+being the next stack frame with the try block.
+\item[Throw/Resume Match None]
+The same results as for Match All.
 \end{description}
+Moving on to resumption there is one general note,
+resumption is \textit{fast}, the only test where it fell
+behind termination is Cross Handler.
+In every other case, the number of iterations had to be increased by a
+factor of 10 to get the run-time in an approprate range
+and in some cases resumption still took less time.
+% I tried \paragraph and \subparagraph, maybe if I could adjust spacing
+% between paragraphs those would work.
+\begin{description}
+\item[Empty Traversal]
+See above for the general speed-up notes.
+This result is not surprising as resumption's link list approach
+means that traversing over stack frames without a resumption handler is
+$O(1)$.
+\item[D'tor Traversal]
+Resumption does have the same spike in run-time that termination has.
+The run-time is actually very similar to Finally Traversal.
+As resumption does not unwind the stack both destructors and finally
+clauses are run while walking down the stack normally.
+So it follows their performance is similar.
+\item[Finally Traversal]
+The increase in run-time fromm Empty Traversal (once adjusted for
+the number of iterations) roughly the same as for termination.
+This suggests that the
+\item[Other Traversal]
+Traversing across handlers reduces resumption's advantage as it actually
+has to stop and check each one.
+Resumption still came out ahead (adjusting for iterations) but by much less
+than the other cases.
+\item[Cross Handler]
+The only test case where resumption could not keep up with termination,
+although the difference is not as significant as many other cases.
+It is simply a matter of where the costs come from. Even if \CFA termination
+is not ``zero-cost" passing through an empty function still seems to be
+cheaper than updating global values.
+\item[Conditional Match]
+Resumption shows a slight slowdown if the exception is not matched
+by the first handler, which follows from the fact the second handler now has
+to be checked. However the difference is not large.
+\end{description}
+Finally are the results of the resumption/fixup routine comparison.
+These results are surprisingly varied, it is possible that creating a closure
+has more to do with performance than passing the argument through layers of
+calls.
+Even with 100 stack frames though, resumption is only about as fast as
+manually passing a fixup routine.
+So there is a cost for the additional power and flexibility exceptions
+provide.
+\begin{comment}
+This observation means that while \CFA does not actually keep up with Python in
+every case, it is usually no worse than roughly half the speed of \Cpp. This
+performance is good enough for the prototyping purposes of the project.
+The test case where \CFA falls short is Raise Other, the case where the
+stack is unwound including a bunch of non-matching handlers.
+This slowdown seems to come from missing optimizations.
+This suggests that the performance issue in Raise Other is just an
+optimization not being applied. Later versions of gcc may be able to
+optimize this case further, at least down to the half of \Cpp mark.
+A \CFA compiler that directly produced assembly could do even better as it
+would not have to work across some of \CFA's current abstractions, like
+the try terminate function.
+Resumption exception handling is also incredibly fast. Often an order of
+magnitude or two better than the best termination speed.
+There is a simple explanation for this; traversing a linked list is much
+faster than examining and unwinding the stack. When resumption does not do as
+well its when more try statements are used per raise. Updating the internal
+linked list is not very expensive but it does add up.
+The relative speed of the Match All and Match None tests (within each
+language) can also show the effectiveness conditional matching as compared
+to catch and rethrow.
+\begin{itemize}[nosep]
+\item
+Java and Python get similar values in both tests.
+Between the interpreted code, a higher level representation of the call
+stack and exception reuse it it is possible the cost for a second
+throw can be folded into the first.
+% Is this due to optimization?
+\item
+Both types of \CFA are slightly slower if there is not a match.
+For termination this likely comes from unwinding a bit more stack through
+libunwind instead of executing the code normally.
+For resumption there is extra work in traversing more of the list and running
+more checks for a matching exceptions.
+% Resumption is a bit high for that but this is my best theory.
+\item
+Then there is \Cpp, which takes 2--3 times longer to catch and rethrow vs.
+just the catch. This is very high, but it does have to repeat the same
+process of unwinding the stack and may have to parse the LSDA of the function
+with the catch and rethrow twice, once before the catch and once after the
+rethrow.
+% I spent a long time thinking of what could push it over twice, this is all
+% I have to explain it.
+\end{itemize}
+The difference in relative performance does show that there are savings to
+be made by performing the check without catching the exception.
+\end{comment}
+\begin{comment}
+From: Dave Dice <dave.dice@oracle.com>
+To: "Peter A. Buhr" <pabuhr@uwaterloo.ca>
+Subject: Re: [External] : JIT
+Date: Mon, 16 Aug 2021 01:21:56 +0000
+> On 2021-8-15, at 7:14 PM, Peter A. Buhr <pabuhr@uwaterloo.ca> wrote:
+>
+> My student is trying to measure the cost of installing a try block with a
+> finally clause in Java.
+>
+> We tried the random trick (see below). But if the try block is comment out, the
+> results are the same. So the program measures the calls to the random number
+> generator and there is no cost for installing the try block.
+>
+> Maybe there is no cost for a try block with an empty finally, i.e., the try is
+> optimized away from the get-go.
+There's quite a bit of optimization magic behind the HotSpot curtains for
+try-finally.  (I sound like the proverbial broken record (:>)).
+In many cases we can determine that the try block can't throw any exceptions,
+so we can elide all try-finally plumbing.  In other cases, we can convert the
+try-finally to normal if-then control flow, in the case where the exception is
+thrown into the same method.  This makes exceptions _almost cost-free.  If we
+actually need to "physically" rip down stacks, then things get expensive,
+impacting both the throw cost, and inhibiting other useful optimizations at the
+catch point.  Such "true" throws are not just expensive, they're _very
+expensive.  The extremely aggressive inlining used by the JIT helps, because we
+can convert cases where a heavy rip-down would normally needed back into simple
+control flow.
+Other quirks involve the thrown exception object.  If it's never accessed then
+we're apply a nice set of optimizations to avoid its construction.  If it's
+accessed but never escapes the catch frame (common) then we can also cheat.
+And if we find we're hitting lots of heavy rip-down cases, the JIT will
+consider recompilation - better inlining -- to see if we can merge the throw
+and catch into the same physical frame, and shift to simple branches.
+In your example below, System.out.print() can throw, I believe.  (I could be
+wrong, but most IO can throw).  Native calls that throw will "unwind" normally
+in C++ code until they hit the boundary where they reenter java emitted code,
+at which point the JIT-ed code checks for a potential pending exception.  So in
+a sense the throw point is implicitly after the call to the native method, so
+we can usually make those cases efficient.
+Also, when we're running in the interpreter and warming up, we'll notice that
+the == 42 case never occurs, and so when we start to JIT the code, we elide the
+call to System.out.print(), replacing it (and anything else which appears in
+that if x == 42 block) with a bit of code we call an "uncommon trap".  I'm
+presuming we encounter 42 rarely.  So if we ever hit the x == 42 case, control
+hits the trap, which triggers synchronous recompilation of the method, this
+time with the call to System.out.print() and, because of that, we now to adapt
+the new code to handle any traps thrown by print().  This is tricky stuff, as
+we may need to rebuild stack frames to reflect the newly emitted method.  And
+we have to construct a weird bit of "thunk" code that allows us to fall back
+directly into the newly emitted "if" block.  So there's a large one-time cost
+when we bump into the uncommon trap and recompile, and subsequent execution
+might get slightly slower as the exception could actually be generated, whereas
+before we hit the trap, we knew the exception could never be raised.
+Oh, and things also get expensive if we need to actually fill in the stack
+trace associated with the exception object.  Walking stacks is hellish.
+Quite a bit of effort was put into all this as some of the specjvm benchmarks
+showed significant benefit.
+It's hard to get sensible measurements as the JIT is working against you at
+every turn.  What's good for the normal user is awful for anybody trying to
+benchmark.  Also, all the magic results in fairly noisy and less reproducible
+results.
+Regards
+Dave
+p.s., I think I've mentioned this before, but throwing in C++ is grim as
+unrelated throws in different threads take common locks, so nothing scales as
+you might expect.
+\end{comment}

doc/theses/andrew_beach_MMath/uw-ethesis.tex

reaeca5f	r1d402be
210	210	\lstMakeShortInline@
211	211	\lstset{language=CFA,style=cfacommon,basicstyle=\linespread{0.9}\tt}
	212	% PAB causes problems with inline @=
	213	%\lstset{moredelim=**[is][\protect\color{red}]{@}{@}}
212	214	% Annotations from Peter:
213	215	\newcommand{\PAB}[1]{{\color{blue}PAB: #1}}

Context Navigation

Changes in / [eaeca5f:1d402be]

Legend:

doc/theses/andrew_beach_MMath/conclusion.tex

doc/theses/andrew_beach_MMath/existing.tex

doc/theses/andrew_beach_MMath/features.tex

doc/theses/andrew_beach_MMath/future.tex

doc/theses/andrew_beach_MMath/implement.tex

doc/theses/andrew_beach_MMath/intro.tex

doc/theses/andrew_beach_MMath/performance.tex

doc/theses/andrew_beach_MMath/uw-ethesis.tex

Download in other formats: