> What instance should be returned from
HostServicesDelegate#getInstance() by default if javafx.deploy is not
loaded?
It will need to be a new instance that implements the methods of
HostServices in a way that makes sense for a standalone app. The deploy
code already does this, so I can likely reuse that. In any event, I'll
do this when I fix the bug. It won't happen in the next few weeks,
though; until then please use the workaround or just ignore the warning
message.
-- Kevin
Yasumasa Suenaga wrote:
Hi Kevin,
Currently, ctor of HostServices will not be initialized because it
will recieve ClassNotFoundException. However, the application seems
to be run.
What instance should be returned from HostServicesDelegate#getInstance()
by default if javafx.deploy is not loaded?
I guess we can use ModuleFinder#find() to refactoring.
Thanks,
Yasumasa
On 2016/06/30 0:09, Kevin Rushforth wrote:
Hi Yasumasa,
No, this is not a correct solution as it just masks the problem. A
likely solution will be to refactor the code to provide an
alternative implementation of HostServicesDelegate in case the
javafx.deploy module is not loaded (either in the case of a purely
open build of OpenJFX or in the case where the deploy module doesn't
get loaded by default).
-- Kevin
Yasumasa Suenaga wrote:
Hi Kevin,
java -addmods javafx.deploy ....
We will need a fix that doesn't require this.
I guess you will fix to add exports directive to module-info.java in
closed source.
However, I think we should fix as below:
----------------
diff -r 3597495167c7
modules/graphics/src/main/java/com/sun/javafx/application/HostServicesDelegate.java
---
a/modules/graphics/src/main/java/com/sun/javafx/application/HostServicesDelegate.java
Tue Jun 28 01:48:07 2016 +0530
+++
b/modules/graphics/src/main/java/com/sun/javafx/application/HostServicesDelegate.java
Wed Jun 29 23:53:07 2016 +0900
@@ -55,7 +55,6 @@
getInstanceMeth =
factoryClass.getMethod(
"getInstance",
Application.class);
} catch (Exception ex) {
- ex.printStackTrace();
return null;
}
}
----------------
Otherwise, the user who uses pure OpenJFX will see this stack trace.
I think this exception can be ignored.
So it should not be printed.
Thanks,
Yasumasa
On 2016/06/29 0:44, Kevin Rushforth wrote:
A workaround is:
java -addmods javafx.deploy ....
We will need a fix that doesn't require this.
-- Kevin
Kevin Rushforth wrote:
Thanks for the pointer. That was enough for me to reproduce the
bug. I filed the following issue to track it:
https://bugs.openjdk.java.net/browse/JDK-8160464
-- Kevin
Yasumasa Suenaga wrote:
Hi Kevin,
I use JDK 9 EA b124 for Windows x64 from java.net <http://java.net>.
https://jdk9.java.net/download/
If you want to full stack trace, please download from here:
http://icedtea.classpath.org/bugzilla/attachment.cgi?id=1540&action=edit
<http://icedtea.classpath.org/bugzilla/attachment.cgi?id=1540&action=edit>
This exception does not occur on OracleJDK 8. So I guess this
exception is from Java FX 9.
Thanks,
Yasumasa
2016/06/28 23:52 "Kevin Rushforth" <kevin.rushfo...@oracle.com
<mailto:kevin.rushfo...@oracle.com>>:
Hi Yasumasa,
I assume this is a download of the Oracle JDK from java.net
<http://java.net>? We haven't seen any problems like this, so it
might be a problem with HeapStats analyzer. The class in
question
is a closed-source class that lives in the javafx.deploy module.
What exactly are you running?
-- Kevin
Yasumasa Suenaga wrote:
Hi all,
When I ran HeapStats Analyzer [1] on JDK 9 EA b124, I
encountered ClassNotFoundException as below:
------------
java.lang.ClassNotFoundException:
com.sun.deploy.uitoolkit.impl.fx.HostServicesFactory
at
jdk.internal.loader.BuiltinClassLoader.loadClass(java.base@9-ea/BuiltinClassLoader.java:366)
at
java.lang.ClassLoader.loadClass(java.base@9-ea/ClassLoader.java:419)
at java.lang.Class.forName0(java.base@9-ea/Native
Method)
at
java.lang.Class.forName(java.base@9-ea/Class.java:378)
at
com.sun.javafx.application.HostServicesDelegate.lambda$getInstance$0(javafx.graphics@9-ea/HostServicesDelegate.java:52)
:
------------
This message does not appear on JDK 8u92.
I think this message is printed at [2].
I think this exception message do not need to be printed.
At least, it is debug level.
Does someone work for it?
If not so, I want to contribute patch.
Thanks,
Yasumasa
[1] http://icedtea.classpath.org/wiki/HeapStats
[2]
http://hg.openjdk.java.net/openjfx/9-dev/rt/file/3597495167c7/modules/graphics/src/main/java/com/sun/javafx/application/HostServicesDelegate.java#l58
