You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Previously we had restored the whitespace for *.aj files in 6888a6f
to avoid a but in aspectj.
We have updated to the latest version of apsectj and restored the
changes in commit 6888a6f which included a significant cleanup of
whitespace.
Issue: SPR-10208
Copy file name to clipboardExpand all lines: spring-aspects/src/main/java/org/springframework/beans/factory/aspectj/AbstractDependencyInjectionAspect.aj
Copy file name to clipboardExpand all lines: spring-aspects/src/main/java/org/springframework/beans/factory/aspectj/AbstractInterfaceDrivenDependencyInjectionAspect.aj
+33-33Lines changed: 33 additions & 33 deletions
Original file line number
Diff line number
Diff line change
@@ -26,94 +26,94 @@ import java.io.Serializable;
26
26
* upon deserialization. Subaspects need to simply provide definition for the configureBean() method. This
27
27
* method may be implemented without relying on Spring container if so desired.
28
28
* </p>
29
-
* <p>
29
+
* <p>
30
30
* There are two cases that needs to be handled:
31
31
* <ol>
32
-
* <li>Normal object creation via the '<code>new</code>' operator: this is
33
-
* taken care of by advising <code>initialization()</code> join points.</li>
32
+
* <li>Normal object creation via the '{@codenew}' operator: this is
33
+
* taken care of by advising {@codeinitialization()} join points.</li>
34
34
* <li>Object creation through deserialization: since no constructor is
35
35
* invoked during deserialization, the aspect needs to advise a method that a
36
36
* deserialization mechanism is going to invoke. Ideally, we should not
37
37
* require user classes to implement any specific method. This implies that
38
38
* we need to <i>introduce</i> the chosen method. We should also handle the cases
39
39
* where the chosen method is already implemented in classes (in which case,
40
-
* the user's implementation for that method should take precedence over the
40
+
* the user's implementation for that method should take precedence over the
41
41
* introduced implementation). There are a few choices for the chosen method:
42
42
* <ul>
43
43
* <li>readObject(ObjectOutputStream): Java requires that the method must be
44
-
* <code>private</p>. Since aspects cannot introduce a private member,
44
+
* {@codeprivate}</p>. Since aspects cannot introduce a private member,
45
45
* while preserving its name, this option is ruled out.</li>
46
-
* <li>readResolve(): Java doesn't pose any restriction on an access specifier.
47
-
* Problem solved! There is one (minor) limitation of this approach in
48
-
* that if a user class already has this method, that method must be
49
-
* <code>public</code>. However, this shouldn't be a big burden, since
50
-
* use cases that need classes to implement readResolve() (custom enums,
46
+
* <li>readResolve(): Java doesn't pose any restriction on an access specifier.
47
+
* Problem solved! There is one (minor) limitation of this approach in
48
+
* that if a user class already has this method, that method must be
49
+
* {@codepublic}. However, this shouldn't be a big burden, since
50
+
* use cases that need classes to implement readResolve() (custom enums,
51
51
* for example) are unlikely to be marked as @Configurable, and
52
-
* in any case asking to make that method <code>public</code> should not
52
+
* in any case asking to make that method {@codepublic} should not
53
53
* pose any undue burden.</li>
54
54
* </ul>
55
-
* The minor collaboration needed by user classes (i.e., that the
56
-
* implementation of <code>readResolve()</code>, if any, must be
57
-
* <code>public</code>) can be lifted as well if we were to use an
58
-
* experimental feature in AspectJ - the <code>hasmethod()</code> PCD.</li>
55
+
* The minor collaboration needed by user classes (i.e., that the
56
+
* implementation of {@codereadResolve()}, if any, must be
57
+
* {@codepublic}) can be lifted as well if we were to use an
58
+
* experimental feature in AspectJ - the {@codehasmethod()} PCD.</li>
59
59
* </ol>
60
60
61
61
* <p>
62
62
* While having type implement the {@link ConfigurableObject} interface is certainly a valid choice, an alternative
63
63
* is to use a 'declare parents' statement another aspect (a subaspect of this aspect would be a logical choice)
64
64
* that declares the classes that need to be configured by supplying the {@link ConfigurableObject} interface.
Copy file name to clipboardExpand all lines: spring-aspects/src/main/java/org/springframework/beans/factory/aspectj/GenericInterfaceDrivenDependencyInjectionAspect.aj
0 commit comments