AWT problem on linux

gelderk at gelderk at
Mon Sep 7 00:14:29 PDT 1998

> 	The problem is that in kaffe the invalidate method does not
> seem to work: if I call invalidate() and then I query the state
> through isValid() right away, it returns true (when it should return
> false since I invalidated the component). So in kaffe it seems that
> invalidate() does not work as it should and therefore it is
> impossible to invalidate something and then validate it to force a
> new layout to be done.

You just reported one of the bugs I wanted to post today :-)

If you look in the source-code ( !), you will find:

  public void validate() {}
  public void invalidate() {}
  public boolean isValid() { return true; }

which is plainly non-compliant to the standard.
validate() should call doLayout() and propagate downwards (through
all childs), invalidate() should propagate upwards (through the parent).

While there is no fix for this, call doLayout() yourself, but document
that thoroughly, so anybody knows what to change when.


+--- Kero ------------------ kero at ---+
| I ain't changed,                          |
| but I know I ain't the same               |
+--- M38C --- ---+

More information about the kaffe mailing list