User-Agent: Mozilla/5.0 (Macintosh; U; PPC Mac OS X; en-US; rv:1.0.0) Gecko/20020818 Build Identifier: Mozilla/5.0 (Macintosh; U; PPC Mac OS X; en-US; rv:1.0.0) Gecko/20020818 When I try to write text into the textfields on this page, nothing happens. I even cannot mark the text, that is already in the fields. Reproducible: Always Steps to Reproduce: 1. Go to www.lohn1.de/lobn.htm 2. Try to mark the text in the textfields. 3. Try to write anything into the textfields. Actual Results: Nothing. Expected Results: Show the typed text.

Reporter, are you referring to the html form in the lower left of the page, or the java applet on the right, or both? fwiw both work for me in mozilla (trunk cvs 2002-08-19)

I mean the java applet. The HTML form works for me. And I do not use Mozilla, but Chimera.

->plugins

.

WorksForMe using Chimera/2002081605.

Well, I'm using Chimera/2002081805. If you can tell, that it works there, then I have to look for another solution.

Using the Chimera 2002-08-21-05 NB, I can reproduce the problem when attempting to type in the any of the text fields located in the applet. Checkboxes, radio buttons, pull down menus all work in the applet though. Tested under OS X 10.1.5.

*** Bug 166653 has been marked as a duplicate of this bug. ***

*** Bug 169381 has been marked as a duplicate of this bug. ***

*** Bug 171450 has been marked as a duplicate of this bug. ***

I am using OSX 10.2 Jaguar and have these problems.....I really need to use these chat rooms........

*** Bug 173202 has been marked as a duplicate of this bug. ***

New York Times online crossword puzzles also are refusing text entry, with build 2002101504. http://www.nytimes.com/pages/crosswords/index.html then click on any of the main puzzle links.

On http://www.hwattmba.dk/password.htm is an applet login form (no, it is not secure), however, I cannot type in the text fields (typing does not appear). Works in Mozilla 1.2b Build ID: 2002102104

->peterl

another URL that seems to have this issue https://www.cusa-hfs.com/cgi-bin/hfssvc/ppecu/hfs.html

*** Bug 175457 has been marked as a duplicate of this bug. ***

*** Bug 181207 has been marked as a duplicate of this bug. ***

I'm working with Apple on this issue. We're doing all the right things in the MRJPlugin, but the Apple JavaControl doesn't like the Carbon Events we send it, for some reason.

*** Bug 182582 has been marked as a duplicate of this bug. ***

mouseDragged events are also broken -- they only fire once when the drag is started. Here's an example: http://www.pubbawup.net/dand/home/programs/whiteboard/ This applet works with Omni and IE. Daniel

Blocker for some people.

I take it that we're still waiting for feedback from Apple on this. Could we ping them again to follow up?

*** Bug 186165 has been marked as a duplicate of this bug. ***

In my test, applets received mouse events correctly in popup windows (running OS X Jaguar (10.2.1) and the Chimera December 20 nightly build). For a test case, go to http://www.jars.com/jars_search.html

That's probably a Swing applet. Key events do work in Swing, probably because it uses some native controls.

I visited http://www.jars.com/jars_search.html (in a new tab), opened java-search, typed in a text and closed the search (without searching). after reopening the java-search again, I couldn't change the text, I couldn't close the tab, I had to restart chimera (Build 2002122804).

*** Bug 177380 has been marked as a duplicate of this bug. ***

*** Bug 173802 has been marked as a duplicate of this bug. ***

Would you please ping Apple again, this is a blocker for me. I need to use GoToMyPC for work. Every other Mac browser works!

That bug is the only one that makes Chimera not recommandable for our ActiveMath learning environment. On Macs, it means, the normal Mozilla is the only choice... very very sad. Paul

With the development of Apple's Safari web browser (www.apple.com/safari), I seriously doubt that we're going to get any response at all from Apple.

> With the development of Apple's Safari web browser (www.apple.com/safari), I > seriously doubt that we're going to get any response at all from Apple. That is not at all the case. I'm still in communication with Apple over this issue. In addition, Safari shares a number of Java bugs with Chimera.

You may want to check out Apple's DP9, just in case it includes any new plugins. Not, of course, that I can say whether it does or not, being under NDA.

Oh, that's true, although it wouldn't work before 1.4 is delivered...Two comments:-> the AWT of the Java 1.4.1 for OSX draws in Cocoa, I presume this should help Chimera-> Chimera would be the first that would be compatible with this plugin...Paul

On the other hand, JDK 1.4.1 requires 10.2 It would be a real shame to tie Chimera needlessly to 10.2, since the fact that it works in 10.1 is a real distinguishing feature from Safari at the moment.

Hi Robbie, 10.2 is so much better than any previous OS X versions that I'd never go back. I understand what you say, but I'd also not stick with too old systems when the one we've today is better. We gotta move forward!

True, for those unix geeks bound to ypbind, there's no choice except refusing 10.2... but honestly it should be worth attempting, possibly as an option. Paul

Let's clarify matters here: 1. We don't use apple's plugin, we use our own (MRJ) since theirs doesn't do LiveConnect. 2. We can't use the 1.4VM because the embedding APIs have changed. 3. We need to continue to run on 10.1 so we just have to get a fix for the current issue.

Now, no-one can be against a bug-fix.However I do believe there are good hopes, hence bug 190830 has been filed.

Perhaps it would be possible to make a runtime check for 10.2, and if it's found, don't load the MRJPlugin. That would allow the Apple plugin to take over, since it's known to work OK, that fixes this bug on 10.2. On 10.1 people will still be able to use java applets, without key events until this bug is fixed.

Time to ping Apple again? Can they still help, since Safari is using a different plug-in?

I don't understand why you would stay wit 10.1.....although some of you say the java scripts worked for you...they certainly didn't for me.....however Safari works with these issues....has a few other little quirks...but at least I can use my chat windows properly with it.

I have to say that it is still broken for Java 1.4.1. However, it accepts Key input sometimes. Of course, 9 out of 10 doesn't. It affects more than text field and combo box. Java games that utilize keyboard input fail as well. :( However, Mozilla and Phoenix work. Could Camino share the code from them?

It would seem that nothing has progressed for this bug since late March. Any thoughts as to when typing into Java applets will work for Camino? They do for Firebird and Safari, interestingly.

Except Safari doesn't do the java-javascript connection... Paul

I personlly have no Java at all in both nightly builds of Camino and last Mozilla 1.4rc1. I have MacOs X 10.2.6 and Java 1.4.1 installed. Java used to work in Camino 0.7 and Mozilla < 1.3 if I remember well (until Mozilla was build as MachO I guess). If I am not the only case like this, it may be a 1.4 blocker ?

I don't think anyone's mentioned this yet, but when I try to provide some key input into a java applet, nothing is typed into the applet but instead type-ahead find receives the key input. This can be seen in the status bar. Here's a link to the applet I was testing on. http://www.shodor.org/interactivate/activities/piechart/index.html

Still true with 2003090102 on 10.2.6

I've upgraded my Java engine with the lastest release from a few days ago. Problem still occurs ...

*** Bug 212914 has been marked as a duplicate of this bug. ***

*** Bug 193895 has been marked as a duplicate of this bug. ***

sfraser: what code would you look at for this bug? Is there a standard plugins API or is it java specific? Once we dump 10.1 support how will this change?

This is Apple's bug, not ours. There's nothing we can do to fix it.

ok. you wrote in comment 39: > 1. We don't use apple's plugin, we use our own (MRJ) since theirs doesn't do > LiveConnect. I think that key events are more important than LiveConnect > 2. We can't use the 1.4VM because the embedding APIs have changed. explain? > 3. We need to continue to run on 10.1 so we can switch after 0.8

sfraser wrote: > I'm working with Apple on this issue. We're doing all the right things in the > MRJPlugin, but the Apple JavaControl doesn't like the Carbon Events we send it, > for some reason. So we have a plugin that runs on top of MRJ. MRJ's JavaControl has some kind of bug where the carbon events for key events are being processed correctly. Thoughts: a. Apple's not doing any more work on MRJ I doubt. b. If mozilla ALSO runs MRJPlugin on top of MRJ (and it works) ... then what's different in Camino? Is this something to do with QuickDrawView ? c. How hard would it be to port MRJPlugin to Apple's Java 1.4 ?

> sfraser wrote: > > I'm working with Apple on this issue. We're doing all the right > > things in the MRJPlugin, but the Apple JavaControl doesn't like > > the Carbon Events we send it, for some reason. > > So we have a plugin that runs on top of MRJ. MRJ's JavaControl has > some kind of bug where the carbon events for key events are being > processed correctly. MRJ is "ours" (i.e. was developed at Netscape). It uses the Java Control embedding APIs in Apple's 1.3 version of the JVM, which allow third parties to embed Java in Carbon apps using a ControlRef-based API. Those APIs are no longer present in Java 1.4. When I've asked how to embed a Java applet in your own app, the only answer I've had back is to do it via WebKit. MRJ supports LiveConnect (the Java-J avaScript bridge). Apple's Java plugin does not. I belive that the key handling issues were reproducible in both our MRJ plugin, and the Apple plugin. > Thoughts: a. Apple's not doing any more work on MRJ I doubt. b. If > mozilla ALSO runs MRJPlugin on top of MRJ (and it works) ... then > what's different in Camino? Is this something to do with > QuickDrawView ? c. How hard would it be to port MRJPlugin to > Apple's Java 1.4 ? Hard, as I mentioned above. There appears to be no way to embed Java in another app other than to use WebKit. I don't belive there's even an NSView-based solution.

/me trying not to sound completely ignorant Simon, would someone at Sun possibly be of more help here, or is this completely Apple due to Camino being written in Cocoa?

This is all Apple-level stuff.

Just in case, allow me to point out that MRJplugin used to be precious because it was the only Applet environment on Macintosh that ever supported LiveConnect... But LiveConnect isn't supported by the plugin anymore. At least since the move to Mach-O...

bug 164712 it would seem indicates that LiveConnect is NOT working in Mozilla.

Correct. LiveConnect broke when we changed from CFM to Mach-O, because for plugins to be XPCOM-scriptable, they have to have the same ABI as the host (i.e. the plugin has to be Mach-O too). That's bug 97613. That one at least is fixable, and that would get LiveConnect working in Mozilla builds (but you'd still have key event issues in Camino). However, that was never completed because Apple said that they were going to implement LiveConnect in their plugin, which never happened. We were trying hard to push off MRJ support to Apple, but now everyone has dropped the ball. It's just a nasty mess, and I don't want to be on the receiving end, so reassigning.

I presume I shall not be the first to state such a thing but this state of things means, for our projects, MacOS is dropped for advanced features of our ActiveMath learning environment.

No, but it was, and I can promise you it was an agreable moment for me. Try Mozilla 1.2 or 1.3 and you'll see an applet with working LiveConnect. paul

Maybe we could use WebKit somehow...

The testcase URL appears to work with the JEP being developed on bug 197813, so adding dependency.

Using Camino 20041130 NB on 10.3.6 and the java embed plugin this page works correctly. It take key strokes and displays the text as I type. Only issue I see is that when a key is pressed the java app repaints which is very annoying.

Tried today's Java security update for Mac OS 10.3 on the site indicated in comment #48. The odd thing is that the applet slider works fine under Mac OS 10.3.8, but I cannot type any input into the textboxes using Camino nightly build 2005022208. I'm not getting type-ahead find at all. I'm getting zero text input. So this bug is still here, and it looks like it will be here for some time.

Try the Java Embedding plugin: <http://javaplugin.sourceforge.net/> This will likely be our Java plugin solution at some point.

I just did, Simon. You're right! It works! Freaky. Sorry for the spam, folks.

*** Bug 270203 has been marked as a duplicate of this bug. ***

I can write in java-textfields now too, with the plugin. (Still cant login with my home-banking, but thats a whole other issue probably due to my ****-ass-bank..)

*** Bug 288562 has been marked as a duplicate of this bug. ***

Has anyone contacted Steven Michaud on distributing his plugin with Camino? I know he wanted to go several weeks without bug reports before releasing it to the general audience. Now he hasn't had new reports for over 30 days, but there are a few outstanding bugs. None of these seem to detract from the improvement over the standard Java support (to me, at least), so I'd say: go for it. But it is Steven's call.

Packaging the java embedding plug-in seems like a no-brainer, since it is demonstrably better than Camino's current java, but pink needs to decide if the timing is right, or release 0.9 first. Including the JEP in Camino by default might be JEP's largest distribution channel, and could bring a lot of bugs out the the woodwork prior to 0.9.

We are working with Steven. Stay tuned.

Fixed by bundling the Java Embedding Plugin with Camino (nightly builds 20050729 and later), bug 301076.