\begin{describe}{mac}
{with-gensyms (@{ @<var> @! (@<var> @[@<name>@]) @}^*) \\ \ind
@<declaration>^* \\
- @<form>^*}
+ @<form>^*
+ \-\nlret @<value>^*}
Bind each @<var> (a symbol, not evaluated) to a freshly made gensym whose
name is based on the corresponding @<name> (a string, evaluated), and
evaluate the @<form>s as an implicit @|progn| in the resulting environment.
@{ @<var> @! (@<var> @[@<value-form>@]) @}^*) \\ \ind
@<declaration>^* \\
@<form>^*
- \nlret @<result-form>}
+ \-\nlret @<result-form>}
This is a helper to ensure that macro expansions evaluate their arguments
exactly once each, in the correct order.
@{ @<var> @! (@{ @<var> @!
(@<var> @[@<locative>@]) @}^*) @} \\ \ind
@<declaration>^* \\
- @<form>^*}
+ @<form>^*
+ \-\nlret @<values>^*}
This is a macro which hides the use of locatives from its caller using
symbol-macros.
An anaphoric macro implicitly binds a well-known name to a value of interest,
in the course of doing something else. The concept was popularized by Paul
-Graham \cite[FIXME:OnLisp].
+Graham \cite{graham-1993:on-lisp}.
The macros described here all bind the variable @|it|.
The symbol @|it| is exported by the @|sod-utilities| package.
\end{describe}
-\begin{describe}{mac}{aif @<condition> @<consequent> @[@<alt>@]}
+\begin{describe}{mac}{aif @<condition> @<consequent> @[@<alt>@] @> @<value>^*}
Evaluate the @<condition>. If @<condition> is non-nil, then bind @|it| to
the resulting value and evaluate the @<consequent>, returning all of its
values. Otherwise, evaluate @<alt>, returning all of its values.
\end{describe}
-\begin{describe}{mac}{aand @<form>^*}
+\begin{describe}{mac}{aand @<form>^* @> @<value>^*}
Evaluate each @<form> in turn. If any @<form> evaluates to nil, then stop
and return nil. Each form except the first is evaluated with @|it| bound
to the (necessarily non-nil) value of the previous form. If all but the
(No @|aor| is provided, since @|it| would necessarily be bound to nil.)
-\begin{describe}{mac}{awhen @<condition> @<form>^*}
+\begin{describe}{mac}{awhen @<condition> @<form>^* @> nil}
If @<condition> evaluates to a non-nil value, bind @|it| to that value, and
evaluate the @<form>s as an implicit @|progn|. Otherwise, return nil.
\end{describe}
-\begin{describe}{mac}{acond @{ (@<condition> @<form>^*) @}^*}
+\begin{describe}{mac}{acond @{ (@<condition> @<form>^*) @}^* @> @<value>^*}
Evaluate each @<condition> in turn, until one of them produces a non-nil
value. If the @<condition> is followed by one or more @<form>s, then bind
@|it| to the non-nil value of the @<condition> and evaluate the @<form>s as
\end{describe}
\begin{describe*}
- {\dhead{mac}
- {acase @<scrutinee> @{ (@{ @<case> @! (@<case>^*) @} @<form>^*) @}^*}
- \dhead{mac}
- {aecase @<scrutinee> @{ (@{ @<case> @! (@<case>^*) @} @<form>^*) @}^*}
- \dhead{mac}{atypecase @<scrutinee> @{ (@<type> @<form>^*) @}^*}
- \dhead{mac}{aetypecase @<scrutinee> @{ (@<type> @<form>^*) @}^*}}
+ {\dhead{mac}{acase @<scrutinee>
+ @{ (@{ @<case> @! (@<case>^*) @} @<form>^*) @}^*
+ @> @<value>^*}
+ \dhead{mac}{aecase @<scrutinee>
+ @{ (@{ @<case> @! (@<case>^*) @} @<form>^*) @}^*
+ @> @<value>^*}
+ \dhead{mac}{atypecase @<scrutinee> @{ (@<type> @<form>^*) @}^*
+ @> @<value>^*}
+ \dhead{mac}{aetypecase @<scrutinee> @{ (@<type> @<form>^*) @}^*
+ @> @<value>^*}}
These are like the Common Lisp macros @|case|, @|ecase|, @|typecase|, and
@|etypecase|, except that @|it| is bound to the value of the @<scrutinee>
while evaluating the matching @<form>s.
\end{describe*}
-\begin{describe}{mac}{asetf @{ @<place> @<value> @}^*}
+\begin{describe}{mac}{asetf @{ @<place> @<value> @}^* @> @<value>^*}
For each @<place> and @<value> in turn: bind @|it| to the current value of
the @<place>, evaluate the @<value> expression, and store the resulting
- value back in the @<place>.
+ value back in the @<place>. Return the @<value>(s) stored by the final
+ pair: there may be more than one value, e.g., if @<place> is a @|values|
+ form.
For example, @|(asetf @<place> (1+ it))| is almost equivalent to @|(incf
@<place>)|, even if evaluating @<place> has side-effects.
The following utilities make use of the introspection features of the CLOS
metaobject protocol.
-\begin{describe}{gf}{instance-initargs @<instance>}
+\begin{describe}{gf}{instance-initargs @<instance> @> @<initargs-list>}
Return a fresh list of plausible initargs for the given @<instance>.
This is done by digging through the instance's class's slot definitions and
\dhead{gf}{merge-error-candidates @<error> @> @<list>}
\dhead{gf}{merge-error-present-function @<error> @> @<function>}}
The @|inconsistent-merge-error| condition class used to represent a failure
- of the \descref{merge-lists}[function]{fun}.
+ of the \descref{fun}{merge-lists}[function].
The @<candidates> are a list of offending items from the input lists, in
some order: the error is reporting that the function has failed because it
the partial order.
\end{describe}
+\begin{describe}{fun}{cross-product \&rest @<pieces>}
+ Return the cross product of the @<pieces>.
+
+ Each arguments may be a list, or a (non-nil) atom, which is equivalent to a
+ singleton list containing just that atom. Return a list of all possible
+ lists which can be constructed by taking one item from each argument list
+ in turn, in an arbitrary order.
+\end{describe}
+
\begin{describe}{fun}
{find-duplicates @<report> @<sequence> \&key :key :test}
Call @<report> on each pair of duplicate items in a @<sequence>.
and $y$ are considered equal if and only if @|(funcall @<test> (funcall
@<key> $x$) (funcall @<key> $y$))| returns non-nil.
+ The @<report> function is called as @|(funcall @<report> @<duplicate>
+ @<previous>)|. Duplicates are reported in order; the @<previous> item is
+ always the first matching item in the sequence.
+
This function will work for arbitrary @<test> functions, but it will run
- much more efficiently if @<test> is @|eq|, @|eql|, @|equal|, or @|equalp|
- (because it can use hash-tables).
+ much more efficiently if @<test> is @|eq|, @|eql|, @|equal|, or @|equalp|,
+ because it can use hash-tables. (The generic implementation for lists is
+ especially inefficient.)
\end{describe}
Assume that we are currently at a particular @<line> and @<column> in a
file, and wish to \emph{unread} @<character>: return an @<old-line> and
@<old-column> at which we might plausibly re-read the character, so that
- the next call to \descref{update-position}{fun} will return us to @<line>
+ the next call to \descref{fun}{update-position} will return us to @<line>
and @<column>. (Specifically, the @<old-column> will likely be wrong if
@<character> is a horizontal tab. It is expected that this won't matter:
the purpose of this function is to set things up so that the
The main purpose for this is to be able to access features of the
underlying stream which a pretty-printing stream can't proxy. Most
notably, this is used by C fragment output, which takes advantage of an
- underlying \descref{position-aware-output-stream}{cls} to print @|\#line|
+ underlying \descref{cls}{position-aware-output-stream} to print @|\#line|
directives, so that a C~compiler will blame the original fragment in the
Sod module source rather than the generated C code.
\end{describe}
\end{describe}
\begin{describe}{fun}{cerror* @<datum> \&rest @<arguments>}
- A simplified version of \descref{promiscuous-cerror}{fun} which uses the
+ A simplified version of \descref{fun}{promiscuous-cerror} which uses the
hardcoded string @|Continue| for the restart. This makes calling the
function more similar to other condition-signalling functions, at the
expense of some usability in environments which don't continue after
\begin{describe}{fun}
{frob-identifier @<string> \&key :swap-case :swap-hyphen
@> @<frobbed-string>}
+ Return a `frobbed' version of the identifier @<string>. Two different
+ transformations can be applied.
+
+ \begin{itemize}
+
+ \item If @<swap-case> is non-nil (the default), and the letters in
+ @<string> are either all uppercase or all lowercase, then switch the case
+ of all of the letters.
+
+ \item If @<swap-hyphen> is non-nil (the default), and @<string> contains
+ either hyphens @`--' or underscores @`_', but not both, then replace the
+ hyphens by underscores or \emph{vice-versa}.
+
+ \end{itemize}
+
+ (These are the `obvious' transformations to convert a C identifier into a
+ Lisp symbol.)
+
+ Some examples:
+ \begin{itemize}
+ \item @|(frob-identifier "foo")| $\Longrightarrow$ @|"FOO"|
+ \item @|(frob-identifier "FOO")| $\Longrightarrow$ @|"foo"|
+ \item @|(frob-identifier "FooBar")| $\Longrightarrow$ @|"FooBar"|
+ \item @|(frob-identifier "Foo-Bar")| $\Longrightarrow$ @|"Foo_Bar"|
+ \item @|(frob-identifier "Foo_Bar")| $\Longrightarrow$ @|"Foo-Bar"|
+ \item @|(frob-identifier "foo_bar")| $\Longrightarrow$ @|"FOO-BAR"|
+ \item @|(frob-identifier "foo_bar" :swap-hyphen nil)| $\Longrightarrow$
+ @|"FOO_BAR"|
+ \item @|(frob-identifier "foo_bar" :swap-case nil)| $\Longrightarrow$
+ @|"foo-bar"|
+ \item @|(frob-identifier "foo_bar" :swap-case nil :swap-hyphen nil)|
+ $\Longrightarrow$ @|"foo_bar"|
+ \end{itemize}
\end{describe}
\begin{describe}{fun}
- {compose @<function> \&rest @<more-functions> @> @<function>}
+ {compose @<functions> @> @<function>}
+ Return the left-to-right composition zero or more @<functions>.
+
+ Let $f_1$, $f_2$, \ldots, $f_n$ be functions, and let $g = @|(compose $f_1$
+ $f_2$ $\cdots$ $f_n$)|$ is their composition. If $g$ is applied to
+ arguments, the effect is as follows: first, $f_1$ is applied to the
+ arguments, yielding some value; $f_2$ is applied to this value, yielding a
+ second value; and so on, until finally the value yielded by $f_n$ is
+ returned as the result of $g$. Note that this is the reverse of the usual
+ mathematician's convention, but the author finds this ordering
+ significantly easier to work with:
+ \[ g = f_n \circ \cdots \circ f_2 \circ f_1 \]
+
+ If any of the input functions return multiple values then \emph{all} of the
+ values are passed on to the next function in the list. (If the last
+ function returns multiple values then all of the values are returned from
+ the composition.
+
+ The result of composing no functions is a function which simply returns all
+ of its arguments as values; essentially, $@|(compose)| \equiv
+ @|\#'values|$.
\end{describe}
\begin{describe}{mac}{defvar-unbound @<name> @<documentation> @> @<name>}
+ Define a variable called @<name>, with a @<documentation> string.
+
+ The Common Lisp @|defvar| macro accepts both an initial value and a
+ doc-string as optional arguments, in that order, with the result that it's
+ not possible to define a variable and establish a documentation string for
+ it without also giving it an initial value. The @|defvar-unbound| macro,
+ on the other hand, never changes the symbol's variable-value.
\end{describe}
\begin{describe}{mac}
{dosequence (@<var> @<sequence>
@[[ :start @<start> @! :end @<end> @!
- :indexvar @<var> @]]) \\ \ind
+ :indexvar @<index-var> @]]) \\ \ind
@<declaration>^* \\
@{ @<tag> @! @<statement> @}^*}
+ Iterate over a @<sequence>. Common Lisp has a rich collection of iteration
+ primitives, and a rich collection of functions for working with sequences,
+ but no macro for iterating over the items of a sequence.
+
+ First, the @<sequence> is evaluated. If @<start> and/or @<end> are
+ provided, they are also evaluated (in that order), which should produce
+ integers; @<end> may be also be nil. If not provided, or nil (in the case
+ of @<end>), @<start> and @<end> default respectively to zero and the length
+ of the @<sequence>. For each item in the sequence between the @<start> and
+ @<end> positions (i.e., each item in @|(subseq @<sequence> @<start>
+ @<end>)|, in order, the body is evaluated as an implicit @|tagbody|, with
+ @<var> bound to the item and, if provided, @<index-var> bound to the item's
+ index. It is not specified whether the @<var> and @<index-var> are
+ let-bound or mutated in each iteration.
+
+ Unlike other Common Lisp @|do|\dots\ forms, there is no `result' form.
\end{describe}
\begin{describe}{mac}
{define-access-wrapper @<from> @<to>
@[[ :read-only @<read-only-flag> @]]}
+ Define @<from> as a function of one argument, so that @|(@<from> @<thing>)|
+ is equivalent to @|(@<to> @<thing>)|. If @<read-only-flag> is nil (the
+ default), then also define @|(setf @<from>)| so that @|(setf (@<from>
+ @<thing>) @<value>)| is equivalent to @|(setf (@<to> @<thing>) @<value>)|.
+
+ In a @|defstruct| form, the accessor function names are constructed based
+ on the structure name and slot names. The structure name and accessor
+ names are part of the exported interface, but the slot names ideally
+ shouldn't be. This causes a problem when the slot name which will lead to
+ the right accessor is already an external symbol in some package. You can
+ solve this problem by choosing an internal name for the symbol, and then
+ using this macro to define an accessor function with the name that you
+ want, in terms of the accessor that @|defstruct| made.
\end{describe}
\begin{describe}{fun}
{distinguished-point-shortest-paths @<root> @<neighbours-func>
@> @<list>}
+ Calculate the shortest path from the @<root> to each node reachable from it
+ in a directed graph. The nodes of the graph can be any kind of object;
+ they will be compared using @|eql|.
+
+ The @<neighbours-func> should be a function which, given a node~$v$ as its
+ only argument, returns a list of cons cells @|($v'$ . $c'$)|, one for each
+ node~$v'$ adjacent to $v$, indicating the cost $c'$ of traversing the arc
+ from $v$ to $v'$.
+
+ The return value is a list of cons cells @|($c$ . $p$)|, where $p$ is list
+ of nodes, in reverse order, along a path from the @<root> to some other
+ node, and $c$ is the total cost of traversing this path. (Therefore @|(car
+ $p$)| is the destination node, and @|(car (last $p$))| is always the
+ @<root> itself.)
+
+ The function runs in $O(n^2)$ time, where $n$ is the number of nodes
+ reachable from the @<root>. Currently, it uses an algorithm due to Edsger
+ Dijkstra.
+\end{describe}
+
+
+\subsection{Other exported symbols}
+
+\begin{describe}{sym}{int}
+ The symbol @|int| is exported by the @|sod-utilities| package, without
+ giving it any particular meaning. This is done because it's given
+ non-conflicting meanings by two different packages, and it's more
+ convenient for user code not to have to deal with an unnecessary symbol
+ conflict. Specifically, the @|sod| package wants to define it as a C type
+ specifier, see \descref{cls}{simple-c-type}; and @|optparse| wants to
+ define it as an option handler, see \descref{opt}{int}.
\end{describe}
%%%--------------------------------------------------------------------------
\section{Option parser} \label{sec:misc.optparse}
-These symbols are defined in the @|optparse| package.
+Most of these symbols are defined in the @|optparse| package.
\begin{describe}{fun}{exit \&optional (@<code> 0) \&key :abrupt}
\end{describe}
\dhead{fun}{setf (opt-negated-tag @<option>) @<tag>}
\dhead{fun}{opt-arg-name @<option> @> @<string-or-null>}
\dhead{fun}{setf (opt-arg-name @<option>) @<string-or-null>}
- \dhead{fun}{opt-optional-p @<option> @> @<generalized-boolean>}
- \dhead{fun}{setf (opt-optional-p @<option>) @<generalized-boolean>}
+ \dhead{fun}{opt-arg-optional-p @<option> @> @<generalized-boolean>}
+ \dhead{fun}{setf (opt-arg-optional-p @<option>) @<generalized-boolean>}
\dhead{fun}{opt-documentation @<option> @> @<string-or-null>}
\dhead{fun}{setf (opt-documentation @<option>) @<string-or-null>}}
\end{describe*}
\&key :format-control :format-arguments}
\end{describe}
+\begin{describe}{fun}{option-parse-error @<msg> \&optional @<args>}
+\end{describe}
+
\begin{describe}{fun}{option-parse-remainder @<option-parser>}
\end{describe}
@<form>^*) @}^*}
\end{describe}
+\begin{describe}{fun}{sod-frontend:augment-options @<options-list>}
+\end{describe}
+
%%%--------------------------------------------------------------------------
\section{Property sets} \label{sec:misc.pset}
\begin{describe*}
{\dhead{fun}{p-name @<property> @> @<name>}
+ \dhead{meth}{property}{file-location (@<property> property) @> @<floc>}
\dhead{fun}{p-value @<property> @> @<value>}
\dhead{fun}{p-type @<property> @> @<type>}
\dhead{fun}{p-key @<property> @> @<symbol>}