From the java-dev Apple list, I can see that there is a problem in the latest Java update for Leopard and Snow Leopard.
It will probably exist in Servoy too with fields of type HTML_AREA or TEXT_AREA which, I believe are based on the problematic class.
Here is the bug report that was just submitted:
Problem ID: 8013559
1.6.0_20: JFormattedTextField backspace deletes two characters for each keypress
Summary: JFormattedTextField backspace deletes two characters for each keypress.
Steps to Reproduce: Run any sample containing a JFormattedTextField; press backspace. For example, pick one of the Sun samples from here: http://java.sun.com/docs/books/tutorial … field.html Expected Results: Backspace should delete either the current selection, or the single character before the cursor.
Actual Results: DefaultEditorKit.deletePrevCharAction is apparently being invoked twice for a single keypress. Regression: This issue was introduced in: * Java for Mac OS X 10.5 Update 7 (1.6.0_20 + 1.5.0_24) * Java for Mac OS X 10.6 Update 2 (1.6.0_20)
Notes: http://lists.apple.com/archives/java-de … 00082.html
Yes, I noticed this today when working. This happens frequently when you hit delete on a mac os x 10…5.8 and the latest version of Java.
It must be said that if the circus continues that has been the last 4-6 weeks with java continue then the ability of this operating system to deliver robust solutions to serious industries will and must be bought into serious questions.
We have 150 doctors and the cost of having to do team viewer sessions to empty caches, update software, downgrade java to 18, then updgrade it again and then to have to say sorry to mac users who cannot delete, is seriously eroding all faith in the underlying OS. Then have to put web shortcuts on their desktops as webstart then cannot find the cached files when we restart the servoy service
We are continually are told that our competitors who use .net solutions just put an icon on the desktop and work when clicked and update when an update comes out!
Are Oracle in a secret deal with Microsoft to destroy this plaftform?
dpearce:
We are continually are told that our competitors who use .net solutions just put an icon on the desktop and work when clicked and update when an update comes out!
Are Oracle in a secret deal with Microsoft to destroy this plaftform?
David
Well David, I guess you are not on the .Net forum to see how much bugs there are still and introduced with every windows update, do you?
I’m not on them either, but I’m pretty sure that they have their issues too!
And knowing Microsoft, I guess they are plenty!
Now, I’m sure that your Mac users will agree that having a system that runs on their platform is better than having a .Net solution and having to run a Windows emulation or buy a PC, don’t you think?
I agree that Oracle’s latest bad moves can shake your client’s confidence in the platform but I can’t think of any environment that has no shortcomings.
And in that particular case, the problem is more likely to come from Apple than Oracle, really.
ptalbot:
1.6.0_20: JFormattedTextField backspace deletes two characters for each keypress
these are not TEXT_AREA or HTML_AREA
But our normal TEXT_FIELD (and password field) those are build on the JFormattedTextField
So if that bug in in _20 you should be able to very quickly see it because i guess 75%+ of the fields in servoy are of that type…
dpearce:
We have 150 doctors and the cost of having to do team viewer sessions to empty caches, update software, downgrade java to 18, then updgrade it again and then to have to say sorry to mac users who cannot delete, is seriously eroding all faith in the underlying OS. Then have to put web shortcuts on their desktops as webstart then cannot find the cached files when we restart the servoy service
that caching bug should be fixed in _20 that i think was a bug of 17/18…
and U20 with the latest servoy code also shouldnt force you to clean any cache anymore… Because we will make sure that any change in any jnlp or jar file will be seen webstart on the client.
ptalbot:
1.6.0_20: JFormattedTextField backspace deletes two characters for each keypress
these are not TEXT_AREA or HTML_AREA
But our normal TEXT_FIELD (and password field) those are build on the JFormattedTextField
So if that bug in in _20 you should be able to very quickly see it because i guess 75%+ of the fields in servoy are of that type…
Just to confirm: elements of type ‘TEXT_FIELD’ indeed have this problem.
Mac implemented however a wonderful shortcut for a single backspace: [SPACE]-[BACKSPACE]
(But really, it’s bullsh*t - that part, I’m not kidding about).
The real question is - how, or when or if it is going to be fixed.
My dilema is that now that I actually HAVE Mac customers - that I want to (or have convinced to use) Servoy - space + backspace isn’t really an option for stuff that’s in production…
How is everyone else on a Mac getting around this problem? Can we downgrade the Java version - or ??
A month later, any hope for a fix in Java on this? Do you guys at Servoy have a developer preview of the next java release yet? Crossing my fingers, because if I have to tell I.T. one more time that a Java upgrade caused a fatal (must go workstation to workstation to downgrade) bug, it’s not going to go well.
After repeated questions by Patick Talbot on the java-dev mailinglist Apple gave a workaround for it on that same mailinglist and Servoy added it in the 5.2 code:
After repeated questions by Patick Talbot on the java-dev mailinglist Apple gave a workaround for it on that same mailinglist and Servoy added it in the 5.2 code:
Cool. Can we get this patch in the 5.1.4 branch? As mentioned by others elsewhere (servoy clients not getting discarded thread), we’re barely getting our live sites stable at 5.1.4. Moving to 5.2 is going to entail a lot of testing before we dare move live sites over. And this java bug is definitely smacking us around.