Comment on design

[-]

I have a method

public void setGap(Gap gap) {
...
doLayout();
}


After calling the method I then do a static call to Gap.setXXX(..).
This means that I have to call setGap(gap) again because I want
"doLayout()" to be called since my setXXX(..) cannot call doLayout() due
to Gap not containing a reference to the class that has doLayout().

Is this good design?

 

[Sonal]

Why dont you define setGap as:

public void setGap(Gap gap, xx XXX){
....
gap.setXXX(XXX);
doLayout();

}

 

[JScoobyCed]

I have a method

public void setGap(Gap gap) {
...
doLayout();
}


After calling the method I then do a static call to Gap.setXXX(..). This
means that I have to call setGap(gap) again because I want "doLayout()"
to be called since my setXXX(..) cannot call doLayout() due to Gap not
containing a reference to the class that has doLayout().

Is this good design?

What about:
public interface GapListener {
public void updateGap();
}

public class GapManager implements GapListener {

public void setGap(Gap gap) {
// ...
gap.setGapListener(this);
doLayout();
}

public void updateGap() {
doLayout();
}
}

public class NewGap extends Gap {
// or maybe if you own the Gap object simply:
// public class Gap {

private static GapListener gapListener = null;

public static void setGapListener(GapListener gapListener) {
this.gapListener = gapListener;
}

public static void setXXX(XXX xxx) {
// do XXX stuff here
if(gapListener != null) {
gapListener.updateGap();
}
}
}

 

[-]

What about:
public interface GapListener {
public void updateGap();
}

public class GapManager implements GapListener {

public void setGap(Gap gap) {
// ...
gap.setGapListener(this);
doLayout();
}

public void updateGap() {
doLayout();
}
}

public class NewGap extends Gap {
// or maybe if you own the Gap object simply:
// public class Gap {

private static GapListener gapListener = null;

public static void setGapListener(GapListener gapListener) {
this.gapListener = gapListener;
}

public static void setXXX(XXX xxx) {
// do XXX stuff here
if(gapListener != null) {
gapListener.updateGap();
}
}
}

In my case, since Gap is my class, I can pass to its constructor, or in
your case the NewGap's constructor, a reference of the class that has
the doLayout() method.

I'm trying to avoid doing that because it looks out of place if you do a
new Gap(someClass, x, y) rather than a new Gap(x, y);

I'm wondering whether in the real world, the way I am going about doing
it is accepted - that is to leave it up to the user to call setGap(Gap
gap) again after he has set any new values of gap.

 

[Wibble]

In my case, since Gap is my class, I can pass to its constructor, or in
your case the NewGap's constructor, a reference of the class that has
the doLayout() method.

I'm trying to avoid doing that because it looks out of place if you do a
new Gap(someClass, x, y) rather than a new Gap(x, y);

I'm wondering whether in the real world, the way I am going about doing
it is accepted - that is to leave it up to the user to call setGap(Gap
gap) again after he has set any new values of gap.

I think the doLayout doesnt belong in setGap. What if you want to
change a few parameters before doLayout? You want to batch the
modifications and then doLayout.

Its unnatural to call setGap at the end to doLayout, just call doLayout.

 

[Jesper Nordenberg]

I have a method

public void setGap(Gap gap) {
...
doLayout();
}


After calling the method I then do a static call to Gap.setXXX(..).
This means that I have to call setGap(gap) again because I want
"doLayout()" to be called since my setXXX(..) cannot call doLayout() due
to Gap not containing a reference to the class that has doLayout().

Is this good design?

No.

/JN

 

[-]

No.

Do share with me the proper way of accomplishing it.