org.aspectj/docs/pdGuideDB/pointcuts.xml

<chapter id="pointcuts" xreflabel="Debugging Pointcuts">
    <title>Debugging Pointcuts</title>
    <sect1 id="pointcuts-introduction">
    <title>Introduction</title>
    
    <para> 
        This section describes how to write and debug pointcuts
        using the usual approach of iteration and decomposition.
        New users are often stumped when their advice does not match.
        That means the pointcut doesn't match; they rewrite the
        pointcut and it still doesn't match, with no new information.
        This can be frustrating if each iteration involves building,
        deploying, and testing a complex application.  Learning to
        break it down, particularly into parts that can be checked
        at compile-time, can save a lot of time.
    </para>
    </sect1>    
    
    <sect1 id="pointcuts-debugging">
    <title>Debugging pointcuts</title>
    <para>
Go at it top-down and then bottom-up.  Top-down, draft significant
aspects by first writing the comments to specify responsibilities.
Advice responsibility usually takes the form, "When X, do Y."
Pointcut responsibility for "When X" often takes the form,
"When [join points] [in locations] [are ...]".  These []'s often
translate to named pointcuts (like `libraryCalls() &amp;&amp; within(Client)
&amp;&amp; args(Context)`) which form a semantic bridge to the plain-text
meaning in a comment (e.g., `// when the client passes only context into
the library`).        
This gets you to a point where you can debug the parts of the
pointcut independently.
    </para>
    <para>
Bottom up (to build each part), consider each primitive pointcut
designator (PCD), then the composition, and then any implicit
constraints:
    <orderedlist>
    <listitem><para>
What kinds of join points should it match? (constructor-call?
field-get?)?  This translates to using the kinded pointcuts
(`call(..)`, `get(..)`, etc.).
    </para></listitem>
    <listitem><para>
Are these restricted to being lexically within something?  This
translates to using `within{code}(..)`.  If this is true, it should
always be used, to speed up weaving.
    </para></listitem>
    <listitem><para>
What runtime constraints and context should be true and available at
each join point?  This translates to `this()`, `target()`, `args()`,
`cflow{below}()` and `if(..)`.
    </para></listitem>
    <listitem><para>
Are there any advice or implementation limitations at issue?  This
involves knowing the few constraints on AspectJ imposed by Java bytecode 
as listed in the AspectJ Programming Guide section on 
        <ulink url="../progguide/implementation.html">Implementation Notes</ulink>.
    </para></listitem>
    </orderedlist>
    </para>
    <para>
       It's much faster to iterate a pointcut at compile-time
        using declare warning (even better, some errors are identified
        at parse-time in the latest versions of AJDT).
        Start with the parts of the pointcut
        that are staticly-determinable (i.e., they do not involve
        the runtime PCD's listed above).  If compiles themselves
        take too long because of all the AspectJ weaving, then
        try to only include the debugging aspect with the prototype
        pointcut, and limit the scope using <literal>within(..)</literal>.
    </para>
    <para>
        Some mistakes in primitive pointcuts:
        <itemizedlist>
<listitem><para>
`this(Foo) &amp;&amp; execution(static * *(..))`: There is no `this` in a static
context, so `this()` or `target()` should not be used in a static
context or when targetting a static context (respectively).  This
happens most often when you want to say things like "all calls to Foo from Bar"
and you only pick out calls to instance methods of Foo
or you try to pick out calls from static methods of Bar.
</para></listitem>
<listitem><para>
`target(Foo) &amp;&amp;  call(new(..)`: This will never match.  In
constructor-call join points, there is no target because the object
has not been created yet. 
</para></listitem>
<listitem><para>
`call(* Foo.*(..))`: `Foo` refers to the compile-time type of the
invoking reference, not the implementing class.  In Java before 1.4,
the compile-time type was rendered as the defining type, not the
reference type; this was corrected in 1.4 (as shown when using ajc
with the -1.4 flag)  Most people should use `target(Foo) &amp;&amp;  call(...)`.
</para></listitem>
<listitem><para>
`execution(* Foo.bar(..))`: An execution join point for Foo is
always within Foo, so this won't pick out any overrides of bar(..).
Use `target(Foo) &amp;&amp;  execution(* bar(..))` for instance methods.
</para></listitem>
<listitem><para>
`within(Foo)`: anonymous types are not known at weave-time to be
within the lexically-enclosing type (a limitation of Java bytecode).
</para></listitem>
        </itemizedlist>
    </para>
    <para>
        Some mistakes in composition:
        <itemizedlist>
<listitem><para>
`call(* foo(Bar, Foo)) &amp;&amp; args(Foo)`: This will never match.
The parameters in `args(..)` are position-dependent, so `args(Foo)` only picks
 out join points where there is only one argument possible, of type Foo.
Use the indeterminate-arguments operator '..' as needed, e.g., `args(Foo, ..)`.
</para></listitem>
<listitem><para>
`call(* foo()) &amp;&amp; execution(* foo())`: This will never match.  Each
pointcut must be true at each join point matched.  For a union of different
kinds of join points (here, call or execution), use '||'.
E.g., to match both method-call and field-get join points, use 
    `call(* ...) || get(...)`.
</para></listitem>
</itemizedlist>
    </para>
    <para>
        Some mistakes in implicit advice constraints:
        <itemizedlist>
<listitem><para>
`after () returning (Foo foo) : ...`: after advice can bind the
returned object or exception thrown.  That effectively acts like
`target()`, `this()`, or `args()` in restricting when the advice
runs based on the runtime type of the bound object, even though it is
not explicitly part of the pointcut.
</para></listitem>
</itemizedlist>
    </para>
    <para>
        Some mistakes in implementation requirements:
        <itemizedlist>
<listitem><para>
`ajc` has to control the code for a join point in order to implement
the join point.  This translates to an implicit `within({code under
the control of the compiler})` for all join points, with additional
caveat for some join points.  Take exception handlers, for example:
there is no way to be sure from the bytecode where the original handler
ends, so `ajc` can't implement after advice on handler join points.
(Since these are on a per-join-point basis, they should be considered
for each corresponding primitive pointcut designator.)  Unlike the
mistakes with the primitive PCDs above, the compiler will emit an
error for these caveats.
</para></listitem>
<listitem><para>
`call(@SuperAnnotation Subclass.meth()`: Annotations are not inherited
by default, so e.g., if the pointcut specifies an annotation, then 
subclass implementations of that method will not be matched.
</para></listitem>
</itemizedlist>
    </para>


    </sect1>
</chapter>