Home » Archived » Visual Editor (VE) » VE Release 0.5.0 is final and available.
| VE Release 0.5.0 is final and available. [message #8998] |
Mon, 08 December 2003 17:18  |
Eclipse User |
|
|
|
Originally posted by: richkulp.NO.SPAM.us.ibm.com
The final release of VE 0.5.0 is now available from the download site.
At this time we've started working on 1.0.0.
--
Thanks, Rich Kulp
|
|
|
| Re: VE Release 0.5.0 is final and available. [message #9019 is a reply to message #8998] |
Mon, 08 December 2003 21:20  |
| Eclipse User |
|
|
|
Originally posted by: robmv.promca.coma
Rich Kulp wrote:
> The final release of VE 0.5.0 is now available from the download site.
> At this time we've started working on 1.0.0.
I have tried this release and i have been unable to migrate Netbeans
generated code. The method initComponents is listed as an initializer, but
all widget hierarchy that make the entire frame, dialog, or panel is
messed up.
I have noticed a slow performance on Linux GTK, (I haven't tried another
release), maybe the current hack to support Swing inside SWT components is
the reason to this problem. I hope the cooperation between SWT-SWING can
be improved without the use of this hack ;-)
Another problem found is that my Eclipse crashed, but the "hidden" Swing
window remains open, is it running on a different JVM?, maybe making it
run on the same JVM will improve the performance
The only java development tool that I use outside of Eclipse is the
NetBeans Swing GUI builder, that works like a charm, and it clearly
separates the automatically generated code from the hand written regions
(VisualAge4J used the same concept), I do not like the current incarnation
of VE because there is no separation like Netbeans/VisualAge, I will like
a preference to control if develpopers can change generated code, this
will make my code more manageable, and enforce a common way to write GUI
code across the development team
|
|
|
| Re: VE Release 0.5.0 is final and available. [message #9040 is a reply to message #9019] |
Mon, 08 December 2003 21:37 |
| Eclipse User |
|
|
|
Originally posted by: richkulp.NO.SPAM.us.ibm.com
Robert wrote:
> Rich Kulp wrote:
>
>
>>The final release of VE 0.5.0 is now available from the download site.
>>At this time we've started working on 1.0.0.
>
>
> I have tried this release and i have been unable to migrate Netbeans
> generated code. The method initComponents is listed as an initializer, but
> all widget hierarchy that make the entire frame, dialog, or panel is
> messed up.
We don't directly support NetBeans (we look for the initialize() method.
Maybe someone else can point out how to get NetBeans to work, if it can).
>
> I have noticed a slow performance on Linux GTK, (I haven't tried another
> release), maybe the current hack to support Swing inside SWT components is
> the reason to this problem. I hope the cooperation between SWT-SWING can
> be improved without the use of this hack ;-)
>
Swing within SWT is not involved at all. Swing is not used within SWT,
Swing is used as Swing with no reference to SWT. The Eclipse SWT
developers know about this performance problem with GTK/SWT.
> Another problem found is that my Eclipse crashed, but the "hidden" Swing
> window remains open, is it running on a different JVM?, maybe making it
> run on the same JVM will improve the performance
>
We purposely do not run the Swing windows on the IDE vm. That could
actually cause performance problems. We run on a separate VM because
user code is executing and we don't want to take the chance that user
code could contaminate the IDE's VM, plus it allows us to run a
different level of jdk, and other reasons that it is better to be in a
separate vm.
Also, if the main IDE crashes, there is a timer in the other VM's that
if they don't hear from the main IDE after five minutes they will shut
themselves down. If they don't, then this is a bug that needs to be
addressed.
--
Thanks, Rich Kulp
|
|
|
| Re: VE Release 0.5.0 is final and available. [message #9060 is a reply to message #9040] |
Mon, 08 December 2003 22:21 |
| Eclipse User |
|
|
|
Originally posted by: robmv.NO***SPAM.promca.com
Rich Kulp wrote:
> Robert wrote:
> > Rich Kulp wrote:
> >
> >
> >>The final release of VE 0.5.0 is now available from the download site.
> >>At this time we've started working on 1.0.0.
> >
> >
> > I have tried this release and i have been unable to migrate Netbeans
> > generated code. The method initComponents is listed as an initializer, but
> > all widget hierarchy that make the entire frame, dialog, or panel is
> > messed up.
> We don't directly support NetBeans (we look for the initialize() method.
> Maybe someone else can point out how to get NetBeans to work, if it can).
I played with the VE preferences, and found initComponents (used by
NetBeans) listed as initializer so i thought that NetBeans migration was a
feature (sorry), well NetBeans form files are xml so a tool can be made to
generate a java file compatible with VE
> >
> > I have noticed a slow performance on Linux GTK, (I haven't tried another
> > release), maybe the current hack to support Swing inside SWT components is
> > the reason to this problem. I hope the cooperation between SWT-SWING can
> > be improved without the use of this hack ;-)
> >
> Swing within SWT is not involved at all. Swing is not used within SWT,
> Swing is used as Swing with no reference to SWT. The Eclipse SWT
> developers know about this performance problem with GTK/SWT.
If I understand correctly the workaround to be able to create a Swing
Visual Editor inside an SWT container, is based in the creation of an
offscreen image of the real swing windows, and it is drawn on a SWT
component, sorry but this is what i call a hack, maybe someday it will not
be needed, but if works i don't have problem to use it. I'm happy to know
that the performance problem is known. Is it planned to add a feature to
ease migration from NetBeans forms?, i can help the VE team to test it, i
have a lot of them ;-) and don't want to start NetBeans only to edit a
Swing component
> > Another problem found is that my Eclipse crashed, but the "hidden" Swing
> > window remains open, is it running on a different JVM?, maybe making it
> > run on the same JVM will improve the performance
> >
> We purposely do not run the Swing windows on the IDE vm. That could
> actually cause performance problems. We run on a separate VM because
> user code is executing and we don't want to take the chance that user
> code could contaminate the IDE's VM, plus it allows us to run a
> different level of jdk, and other reasons that it is better to be in a
> separate vm.
> Also, if the main IDE crashes, there is a timer in the other VM's that
> if they don't hear from the main IDE after five minutes they will shut
> themselves down. If they don't, then this is a bug that needs to be
> addressed.
Thanks for your input, i killed it without waiting for the timer that you
talk about, i will try it.
|
|
|
| Re: VE Release 0.5.0 is final and available. [message #9081 is a reply to message #9060] |
Mon, 08 December 2003 23:40 |
| Eclipse User |
|
|
|
Originally posted by: richkulp.NO.SPAM.us.ibm.com
The getting of the image from an off-screen area is not a hack. Is it a
very important funtion.
It allows:
1) To have the IDE isolated from user code. Very important because user
code can kill an IDE.
2) Run the user's code on a different JDK than the IDE. Very important
to allow different JDKs to run.
3) If it was put directly onto an SWT control we would have to put in
handlers and setting all over the place to try to prevent user
interaction with the live control. This must be prevented from
happening. Since the live control could be user controls that they've
created, we may not even know all of the possible interaction points, or
the user may not have put the necessary hooks in for us to disable
interaction.
4) Using an image allows us to draw things over top of the image.
Otherwise we can't draw lines or outlines over the control because the
Swing control refreshing would wipe out our lines.
5) It allows to actually change pieces of the image to highlight parts
of the image.
--
Thanks, Rich Kulp
|
|
| |
| Re: VE Release 0.5.0 is final and available. [message #9226 is a reply to message #9081] |
Tue, 09 December 2003 18:13 |
| Eclipse User |
|
|
|
Originally posted by: robmv.No****Spam********.promca.com
Rich Kulp wrote:
> The getting of the image from an off-screen area is not a hack. Is it a
> very important funtion.
> It allows:
> 1) To have the IDE isolated from user code. Very important because user
> code can kill an IDE.
> 2) Run the user's code on a different JDK than the IDE. Very important
> to allow different JDKs to run.
> 3) If it was put directly onto an SWT control we would have to put in
> handlers and setting all over the place to try to prevent user
> interaction with the live control. This must be prevented from
> happening. Since the live control could be user controls that they've
> created, we may not even know all of the possible interaction points, or
> the user may not have put the necessary hooks in for us to disable
> interaction.
> 4) Using an image allows us to draw things over top of the image.
> Otherwise we can't draw lines or outlines over the control because the
> Swing control refreshing would wipe out our lines.
> 5) It allows to actually change pieces of the image to highlight parts
> of the image.
I think 2 is the real advantage of this approach, all the others are
problems that arises by the usage of live components (i have used
VisualAge Smalltalk and know a lot about the problems of live components
;-) ) and the need of a round trip VE (visual editor to code, code to
visual editor).
I really like to see a traditional VE, one that you open a file (for
example a XUL file or XML serialized swing hierarchy), and that allow you
to edit the layout of the components, then generate the code. It is
possible to build one using only the VE SDK?
|
|
|
| Re: VE Release 0.5.0 is final and available. [message #9726 is a reply to message #9226] |
Tue, 09 December 2003 21:06 |
| Eclipse User |
|
|
|
Originally posted by: richkulp.NO.SPAM.us.ibm.com
>
> I really like to see a traditional VE, one that you open a file (for
> example a XUL file or XML serialized swing hierarchy), and that allow you
> to edit the layout of the components, then generate the code. It is
> possible to build one using only the VE SDK?
>
>
Actually at the moment no you can't. But much of the basics are in there
to allow an XML master with java code gen attached to it. There just
isn't any code to hook the two together.
--
Thanks, Rich Kulp
|
|
| |
| Re: VE Release 0.5.0 is final and available. [message #9843 is a reply to message #9728] |
Wed, 10 December 2003 14:01 |
| Eclipse User |
|
|
|
Originally posted by: mendelgili.netscape.net
Francesc Rosés wrote:
> Hi,
>
> Is there a list of changes from RC2?
>
> Thanks,
>
> Francesc
>
> Rich Kulp wrote:
>
>
>>The final release of VE 0.5.0 is now available from the download site.
>>At this time we've started working on 1.0.0.
>
>
>
>
>
No new fixes were made verifying RC2 and onto the build for the released
version.
|
|
|
| Re: VE Release 0.5.0 is final and available. [message #572873 is a reply to message #8998] |
Mon, 08 December 2003 21:20 |
 Robert
Messages: 45
Registered: July 2009 |
Member |
|
|
Rich Kulp wrote:
> The final release of VE 0.5.0 is now available from the download site.
> At this time we've started working on 1.0.0.
I have tried this release and i have been unable to migrate Netbeans
generated code. The method initComponents is listed as an initializer, but
all widget hierarchy that make the entire frame, dialog, or panel is
messed up.
I have noticed a slow performance on Linux GTK, (I haven't tried another
release), maybe the current hack to support Swing inside SWT components is
the reason to this problem. I hope the cooperation between SWT-SWING can
be improved without the use of this hack ;-)
Another problem found is that my Eclipse crashed, but the "hidden" Swing
window remains open, is it running on a different JVM?, maybe making it
run on the same JVM will improve the performance
The only java development tool that I use outside of Eclipse is the
NetBeans Swing GUI builder, that works like a charm, and it clearly
separates the automatically generated code from the hand written regions
(VisualAge4J used the same concept), I do not like the current incarnation
of VE because there is no separation like Netbeans/VisualAge, I will like
a preference to control if develpopers can change generated code, this
will make my code more manageable, and enforce a common way to write GUI
code across the development team
|
|
|
| Re: VE Release 0.5.0 is final and available. [message #572900 is a reply to message #9019] |
Mon, 08 December 2003 21:37 |
| Eclipse User |
|
|
|
Originally posted by: richkulp.NO.SPAM.us.ibm.com
Robert wrote:
> Rich Kulp wrote:
>
>
>>The final release of VE 0.5.0 is now available from the download site.
>>At this time we've started working on 1.0.0.
>
>
> I have tried this release and i have been unable to migrate Netbeans
> generated code. The method initComponents is listed as an initializer, but
> all widget hierarchy that make the entire frame, dialog, or panel is
> messed up.
We don't directly support NetBeans (we look for the initialize() method.
Maybe someone else can point out how to get NetBeans to work, if it can).
>
> I have noticed a slow performance on Linux GTK, (I haven't tried another
> release), maybe the current hack to support Swing inside SWT components is
> the reason to this problem. I hope the cooperation between SWT-SWING can
> be improved without the use of this hack ;-)
>
Swing within SWT is not involved at all. Swing is not used within SWT,
Swing is used as Swing with no reference to SWT. The Eclipse SWT
developers know about this performance problem with GTK/SWT.
> Another problem found is that my Eclipse crashed, but the "hidden" Swing
> window remains open, is it running on a different JVM?, maybe making it
> run on the same JVM will improve the performance
>
We purposely do not run the Swing windows on the IDE vm. That could
actually cause performance problems. We run on a separate VM because
user code is executing and we don't want to take the chance that user
code could contaminate the IDE's VM, plus it allows us to run a
different level of jdk, and other reasons that it is better to be in a
separate vm.
Also, if the main IDE crashes, there is a timer in the other VM's that
if they don't hear from the main IDE after five minutes they will shut
themselves down. If they don't, then this is a bug that needs to be
addressed.
--
Thanks, Rich Kulp
|
|
|
| Re: VE Release 0.5.0 is final and available. [message #572938 is a reply to message #9040] |
Mon, 08 December 2003 22:21 |
Robert Marcano
Messages: 2
Registered: July 2009 |
Junior Member |
|
|
Rich Kulp wrote:
> Robert wrote:
> > Rich Kulp wrote:
> >
> >
> >>The final release of VE 0.5.0 is now available from the download site.
> >>At this time we've started working on 1.0.0.
> >
> >
> > I have tried this release and i have been unable to migrate Netbeans
> > generated code. The method initComponents is listed as an initializer, but
> > all widget hierarchy that make the entire frame, dialog, or panel is
> > messed up.
> We don't directly support NetBeans (we look for the initialize() method.
> Maybe someone else can point out how to get NetBeans to work, if it can).
I played with the VE preferences, and found initComponents (used by
NetBeans) listed as initializer so i thought that NetBeans migration was a
feature (sorry), well NetBeans form files are xml so a tool can be made to
generate a java file compatible with VE
> >
> > I have noticed a slow performance on Linux GTK, (I haven't tried another
> > release), maybe the current hack to support Swing inside SWT components is
> > the reason to this problem. I hope the cooperation between SWT-SWING can
> > be improved without the use of this hack ;-)
> >
> Swing within SWT is not involved at all. Swing is not used within SWT,
> Swing is used as Swing with no reference to SWT. The Eclipse SWT
> developers know about this performance problem with GTK/SWT.
If I understand correctly the workaround to be able to create a Swing
Visual Editor inside an SWT container, is based in the creation of an
offscreen image of the real swing windows, and it is drawn on a SWT
component, sorry but this is what i call a hack, maybe someday it will not
be needed, but if works i don't have problem to use it. I'm happy to know
that the performance problem is known. Is it planned to add a feature to
ease migration from NetBeans forms?, i can help the VE team to test it, i
have a lot of them ;-) and don't want to start NetBeans only to edit a
Swing component
> > Another problem found is that my Eclipse crashed, but the "hidden" Swing
> > window remains open, is it running on a different JVM?, maybe making it
> > run on the same JVM will improve the performance
> >
> We purposely do not run the Swing windows on the IDE vm. That could
> actually cause performance problems. We run on a separate VM because
> user code is executing and we don't want to take the chance that user
> code could contaminate the IDE's VM, plus it allows us to run a
> different level of jdk, and other reasons that it is better to be in a
> separate vm.
> Also, if the main IDE crashes, there is a timer in the other VM's that
> if they don't hear from the main IDE after five minutes they will shut
> themselves down. If they don't, then this is a bug that needs to be
> addressed.
Thanks for your input, i killed it without waiting for the timer that you
talk about, i will try it.
|
|
|
| Re: VE Release 0.5.0 is final and available. [message #572972 is a reply to message #9060] |
Mon, 08 December 2003 23:40 |
| Eclipse User |
|
|
|
Originally posted by: richkulp.NO.SPAM.us.ibm.com
The getting of the image from an off-screen area is not a hack. Is it a
very important funtion.
It allows:
1) To have the IDE isolated from user code. Very important because user
code can kill an IDE.
2) Run the user's code on a different JDK than the IDE. Very important
to allow different JDKs to run.
3) If it was put directly onto an SWT control we would have to put in
handlers and setting all over the place to try to prevent user
interaction with the live control. This must be prevented from
happening. Since the live control could be user controls that they've
created, we may not even know all of the possible interaction points, or
the user may not have put the necessary hooks in for us to disable
interaction.
4) Using an image allows us to draw things over top of the image.
Otherwise we can't draw lines or outlines over the control because the
Swing control refreshing would wipe out our lines.
5) It allows to actually change pieces of the image to highlight parts
of the image.
--
Thanks, Rich Kulp
|
|
| | |
| Re: VE Release 0.5.0 is final and available. [message #573305 is a reply to message #9226] |
Tue, 09 December 2003 21:06 |
| Eclipse User |
|
|
|
Originally posted by: richkulp.NO.SPAM.us.ibm.com
>
> I really like to see a traditional VE, one that you open a file (for
> example a XUL file or XML serialized swing hierarchy), and that allow you
> to edit the layout of the components, then generate the code. It is
> possible to build one using only the VE SDK?
>
>
Actually at the moment no you can't. But much of the basics are in there
to allow an XML master with java code gen attached to it. There just
isn't any code to hook the two together.
--
Thanks, Rich Kulp
|
|
| | |
Goto Forum:
Current Time: Sun Oct 06 11:16:56 GMT 2024
Powered by FUDForum. Page generated in 0.05812 seconds |
] 
Search
Help
Register
Login
Home
