Closed Bug 88870 Opened 24 years ago Closed 23 years ago

Java support needed for carbon Mac OS X build

Categories

(Core Graveyard :: Plug-ins, defect, P1)

Product:
Core Graveyard
Component:
Plug-ins
Platform:
PowerPC
macOS
Type:
defect

Tracking

(Not tracked)

Status:
VERIFIED FIXED

People

(Reporter: chrispetersen, Assigned: beard)

References

Details

(Keywords: meta)

Attachments

(2 files)

Patch for building the MRJ plugin for Mac OS X.
1.04 KB, patch
Details | Diff | Splinter Review
crash log from trying to use java on .9.9+ and 1.b1
13.44 KB, text/plain
Details
Build: 2001061314 Platform: Mac OS X Expected Results: Java support should be functional What I got: Applets are not loaded. This report has been mentioned in http://bugzilla.mozilla.org/show_bug.cgi?id=81659. However, that solution will resolved this particular issue with our MRJ plug-in. To track this problem, I have created a new bug report. Steps to reproduce: 1) With MRJ Plugin installed, go to any page that contains applets. Sample test case:(http://mozilla.org/quality/browser/standards/html/applet_align_top.html) 2) Applets are not rendered.
Changing Severity
Severity: normal → major
I think Patrick Beard is working on this.
Assignee: av → beard
The following statement should have said: "However, that solution will not resolved this particular issue with our MRJ plug-in." in regards to bug 81659. If this problem can't be resolved in the near future, I would see this problem mentioned in the Read me for the OS X build.
Adding myself to the CC list. I'm anxious to see this move forward; With 0.9.2, this (and the resolved plugins issue) seem to be the biggest impediments to mozilla being the best browser on the platform. (MSIE/MacOSX is still pretty flaky).
*** Bug 81458 has been marked as a duplicate of this bug. ***
Set a dependency between this bug & bug 81659, since it is not fixed (I had to reopen). This obviously won't be looked at until that one is complete. - Adam
Depends on: 81659
*** Bug 91645 has been marked as a duplicate of this bug. ***
*** Bug 93707 has been marked as a duplicate of this bug. ***
Keywords: nsenterprise+
Priority: -- → P1
Target Milestone: --- → mozilla0.9.4
I will start examining what it will take to write a new Java plugin for Mac OS X.
Status: NEW → ASSIGNED
Target Milestone: mozilla0.9.4 → mozilla0.9.5
Changed milestone to future, as CPD currently does not have resources working on this feature. Patrick: Are there any progress since we last talk about this bug?
Target Milestone: mozilla0.9.5 → Future
I have been making some progress on the implementation of an OJI plugin for Mac OS X. I have been in contact with Apple engineers who have assisted me in building a simple test application that successfully loads a Java applet. Converting this to an OJI plugin isn't a huge leap.
Target Milestone: Future → mozilla0.9.6
Nominating for PDT+ - -> Adding nsbranch+.
Keywords: nsbranch+
QA Contact: shrir → pmac
nsenterprise-; not a stopper for enterprise
Keywords: nsenterprise+ → nsenterprise-
Can you give me a timeframe on when you'll have the plugin? Need it asap if we want this in 0.9.4.
Keywords: nsenterprise- → nsenterprise+
This plugin will not be ready in time for 0.9.4, but that's probably OK, as it is a standalone piece of code.
This is not on Tian Tian Must Fix list for this release. Should we consider minusing it, and changing the priority for this release?
This is not on Tian Tian Must Fix list for this release, and is targeted for 0.9.6. Should we consider minusing it, and changing the priority for this release?
Not a stop ship
Whiteboard: [OSX-]
Keywords: nsenterprise+
marking nsenterprise- to get off the enterprise stop ship list.
Keywords: nsenterprise-
adding PDT- as per all comments above
Whiteboard: [OSX-] → [OSX-], PDT-
Blocks: 102998
Blocks: 102989
Blocks: 107066
Keywords: nsbranch+
There have been no updates in over a month.
Yes, it a bit of a shame! I bought OS X 10.1 hoping to get trading/push capacity via Java 2, but neither Greatest Mozilla nor little Explorer 5.1.2 are able to do so. :-(
I am currently working on some LiveConnect issues that have to do with the order of ProxyJNI and SecurityManager creation.
Patrick, We see all the sub-bugs you've been posting recently related to this, and it looks like things are coming together; I was just wondering, are you developing the OJI plugin for the CFM build, or the Mach-0 build (or will it be compatible with either build)? Thanks!
It's currently a CFM plugin. Our mach-O build will have to work with CFM plugins, so this shouldn't be a problem.
The latest development version of the plugin is available here: http://homepage.mac.com/pcbeard/MRJPluginCarbon-1.0d6.sit
Glad to see it, Patrick! I've gotten it and tried it out already, with our product which relies heavily on Java Applets. First thing I noticed is that, for one of our applets which loads in a frame, it does _NOT_ suffer from the bug in MSIE where the applet draws itself relative to a whole window, not it's proper frame. Also noticed the following: a) One of our applets didn't completely initialize (we'll look into this further), b) Occasionally, I've found our chat applet wasn't accepting keyboard input, and then it would "flash", and start accepting it, c) I've crashed a couple of times, but not sure if that's a Mozilla+Java issue, or a Mozilla+QT issue. I'll try to get an externally accessable and relatively unchanging external testcase for these, and I'll attach the crashreporter log from the crashing I've gotten here. Are you ready for any bugreports on this, or should we hold off for a little while?
I've been using the plug-in, and it works well, rendering Applets accurately, but now the browser refuses to quit normally whenver an applet has loaded in the session. This us under OS X 10.1.
*** Bug 109757 has been marked as a duplicate of this bug. ***
Keywords: meta
Very nice, but it doesn't seem to like having two codebases specified - eg one for the .cab specified in the <object> tag, as well as the <param name="codebase">. It then takes the .cab as the applet's codebase, where it use use the <param>. Should I file a new bug for that?
*** Bug 110130 has been marked as a duplicate of this bug. ***
This is a meta tracking bug, so it has no target milestone.
Target Milestone: mozilla0.9.6 → ---
Mozilla build ID 2001 11 15 06 Platform: MacOSX 10.0.4 version of the plugin is: http://homepage.mac.com/pcbeard/MRJPluginCarbon-1.0d6.sit Menu:Help:About Plugins: States that MRJPlugin is loaded & operational. However, following the steps to reproduce error still fails. I still get the 'click here to get the plug-in message' with all javaApplet pages I try. Am I missing something?
Nagano, if I remember correctly, the plug-in requires Mac OS X 10.1 or later.
the plug-in I had to remove after upgrading to 10.1.1
Yes, the Mac OS X MRJ plugin uses APIs that are only available in Mac OS X version 10.1 and later. There is code in the plugin which checks, and explicitly refuses to load on earlier systems (otherwise it would crash).
The plug-in should be truthful and indicate that it doesn't work with 10.04. Plug-in Lies to User. MenuBar:Help:About Plug-ins: Taunts the 10.04 user. Lists functions of the plug-in as enabled and ready. (in reality they're not even loaded, skipped over?)
It works for me in mozilla 0.9.6 and macosx 10.1 on some java apps but try this one as a test: http://neo.jpl.nasa.gov/cgi-bin/db?200317 The framework of the applet shows up but the central area which contains the orbits does not, it just appears blank white.
Excellent work! I was very happy to see that even LiveConnect is being implemented. However, Java to JavaScript communication does not seem to work in 1.0d6. getWindow(Applet) always returns null. Is that work in progress?
On OS X 10.1.1, Mozilla build 2001112020, Java plug-in 1.0d6, the mouse-over applets on the following page do not work. Instead, Mozilla indicates a new plug-in is required so it's not recognizing the applets as being Java I guess. http://www.pstnet.com/E-Prime/e-prime.htm It will be a great relief to have Java on the Mac again! Joe
I found a couple of issues and filed bug 113499 and bug 113500.
*** Bug 115426 has been marked as a duplicate of this bug. ***
*** Bug 115509 has been marked as a duplicate of this bug. ***
The Java plugin http://homepage.mac.com/pcbeard/MRJPluginCarbon- 1.0d6.sit is very good...But what puzzles me, is I can enter Yahoo's http:// chat.yahoo.com chat rooms with it using Netscape 6.2.1 for Mac OS X, but not in Mozilla 0.9.6 latest nightly builds for X. Other Java applets seem to work though in Mozilla 0.9.6 latest nightly build. Might note though, this Java plugin does not work with any other browser for X than Mozilla and Netscape, and the Netscape 6.2.1 ability to view Java chat on Yahoo is limited to some extent because instead of just embedding the applet, it shrinks and then embeds the applet. The shrinkage is about 1/4 off the size as appears normally in Classic browsers. Thank you for making that plugin available.
I might add, I am using Mac OS X 10.1.1 5M28, 192MB of RAM, Powerbook G3/233 with 512k backside.
Why is the MRJPluginCarbon not part of the daily builds for Mac OS X. It works quite well and is definitely better than nothing.
Unfortunately, the interfaces used to build it were hand crafted out of those available in /System/Library/Frameworks/JavaEmbedding.framework. The changes to build it are small, but nonetheless have to be made. I'll attach a diff to the bug, and let's see what progress we can make in getting this automated.
This patch changes the files in /System/Library/Frameworks/JavaEmbedding.framework/Headers so they can be used with CodeWarrior Pro 7.
Keywords: nsenterprise-
Whiteboard: [OSX-], PDT-
Keywords: nsbeta1
It's been a while since the last update. What's the status on this one?
Depends on: 113499
Depends on: 113500
agree with the last comment. dagley/beard, what are the chances this is gonna make Mozilla 1.0?
It should make Moz 1.0 - bnesse and beard are actively working on the plugin. We expect a new build to be posted later this week. We want to get it into the Mozilla build RSN but we're talking to Apple about some headers we'll need before that can land.
*** Bug 123577 has been marked as a duplicate of this bug. ***
Plugin 1.0d7 no longer functions for Javascript on some URL's with the 0.9.8 build. Example: http://www.wrh.noaa.gov/satellite/wr4kmiranim.html The plugin worked with the 0.9.7 build for this URL and with similar Java UI controlled animations.
I can't seem to get Yahoo Games to work with this. Going to the game rooms (such as http://games.yahoo.com/games/login?game=Go) and clicking on one of the room links lets me bring up a game window, and the "Starting Plugin for type application/x-java-vm" message appears in the status bar, but the game applet does not load. It works fine under IE. This is with plugin version 1.0d6 and 1.0d7.
Please also try PlaySite: http://www.playsite.com/ PlaySite's games are very well done, and fairly complex Java. PlaySite almost works in IE. There's a bug which prevents you from seeing the browser that lets you join/watch a game. It would be nice if Mozilla could do this while IE can not.
No longer blocks: 107066
Blocks: 107066
No longer blocks: 107066
The plugin seems to work alright, however, the CPU usage is completely ridiculous. Using a G4 550 with 512MB of RAM and OSX 10.1.2, it uses between 60% and 75% of the cpu just displaying a page such as java.sun.com or any of the games at games.yahoo.com. I could sort of understand a game of sorts, but not a standard page with an applet on it such as java.sun.com. IE only takes 14 to 15% which I still think is bad for a program just sitting stagnant and not being used, but the improvement over previous Netscape 6 and Mozilla builds in getting rid of idle CPU usage (which I'm very pleased by because I can actually use it in the real world now) gets negated by this plugin on pages with applets. Has this been seen on other systems, and I hope that efficiency will be something looked at before releasing this, and hopefully in a build coming soon, because I can't use a plugin that takes over 3/4's of my system just to display some moving text.
In my previous post I talked about how Mozilla 0.9.8 uses over 70% of the CPU showing java applets such as at java.sun.com. Interestingly, using Netscape 6.2 with the same plugin instead of Mozilla, the CPU usage is less than half for Netscape than it is for Mozilla. Anybody doing the builds understand why this is and what can be done to make this better? and also consistent?
Under Mac OS X, we shouldn't have to give plugins nearly as much idle time as we currently do, because if they really need time to animate, etc. they can surely spin up their own threads. This is especially true for the Java plugin. We should really consider implementing the feature where we ask the plugin the period at which it would like to receive idle messages. We could implement this feature using Carbon timers, and then stop using idle events altogether. The Java plugin would likely return a value indicating that it doesn't need any idle time.
Sounds like what NPPVpluginTimerInterval may be good for: http://lxr.mozilla.org/mozilla/source/include/npapi.h#316 ..but it doesn't look like it's implemented in the browser yet...
*** Bug 126733 has been marked as a duplicate of this bug. ***
Depends on: 127085
Since this is a META bug, we ought to "plus" the bug to make MRJ part of the OSX builds. Patrick, when you identify that bug, please plus it (nsbeta1+)!
Keywords: nsbeta1
What is the estimated landing date for this? Could you please assign a milestone, 0.9.9 or mozilla 1.0?
We're waiting on Apple to make some headers publicly available so the actual build of the MRJPluginCarbon can be landed in the tree. Work will continue from that point. As for when it'll be finished/declared a release build, the Moz 1.0 timeframe sounds about right.
I think the "acid test" should be using ICQ Lite. With Moz 0.9.8 (Build 2002020516) and MRJ Plug-In (Carbon) 1.0d7, I only get a gray window when I click on the link at http://www.icqlite.com/. Since ICQ and Netscape are part of AOL, I thought ICQ would work with Moz's MRJ Plug-in, but apparently not. I'm running Moz on Mac OS X 10.1.3.
YMMV, but it seems for some people now with Mozilla 0.9.8 for Mac OS X they are having no trouble with the new plugin 1.0d8 http://www.versiontracker.com/moreinfo.fcgi?id=13225&db=mac , and Apple's Java 1.3.1 Update 1 Article 122000 when connecting to chat.yahoo.com. Though I tried the same thing on an iMac G4 800 with yesterday's nightly build (2/27), and it wasn't able to load the chatroom in Yahoo. It was able to load nbc4.com's calendar on the right side though which it has always been able to do. When I tried it, the applet loaded on chat.yahoo.com, but no chatroom showed up. And after about 3 minutes Mozilla unexpectedly quit in Mac OS X 10.1.3.
It seems like new releases (such as 1.0d8) should be mentioned here and/or bug 120222 instead of sending people to VersionTracker. Agree/disagree?
If it's popular enough to warrant attention on VersionTracker, it's probably good enough to be part of the standard build, then people don't have to worry about tracking it. Sure, it's not perfect yet, but Mozilla isn't perfect yet, so that's OK. In my experience, the stability problems are gone, leaving only compatibility problems, and applet users are used to that. Dev. versions can continue to be mentioned on the newsgroup and hosted on mac.com, where people who should be trying out every new version can find them.
Regarding the MRJCarbon plugin becoming part of the standard build - see my earlier comment <http://bugzilla.mozilla.org/show_bug.cgi?id=88870#c63>
Java applets served up via iPlanet/Netscape Proxy servers still don't work. Using the latest MRJ Plugin (1.0d8) in both Netscape 6.2.1 and the latest nightly build of Mozilla for OSX I get the same error. Care of Java Console Logging: java.lang.ClassFormatError: NetMail (Bad magic number) at java.lang.ClassLoader.defineClass0(Native Method) at java.lang.ClassLoader.defineClass(ClassLoader.java:486) at java.security.SecureClassLoader.defineClass(SecureClassLoader.java:111) at sun.applet.AppletClassLoader.findClass(AppletClassLoader.java:142) at sun.plugin.security.PluginClassLoader.findClass(PluginClassLoader.java:213) at java.lang.ClassLoader.loadClass(ClassLoader.java:297) at sun.applet.AppletClassLoader.loadClass(AppletClassLoader.java:108) at java.lang.ClassLoader.loadClass(ClassLoader.java:253) at sun.applet.AppletClassLoader.loadCode(AppletClassLoader.java:373) at sun.applet.AppletPanel.createApplet(AppletPanel.java:598) at sun.applet.AppletPanel.runLoader(AppletPanel.java:534) at sun.applet.AppletPanel.run(AppletPanel.java:308) at java.lang.Thread.run(Thread.java:496) The browser itself reports "Start: applet not initialized" Other, simple Java pages (http://java.sun.com) seem to work without difficulty on this system.
I got the lastest build and plugin and got an error while trying to chat on chat.yahoo.com on several instances. The good news is that I finally managed to chat. Here is my console log: http://www.index-site.com/console.log
Jon, can you give a link that fails? Abrody, you ran into the JS exception recursive death problem. I filed bug 127136 on this issue on 2/21 and the fix was landed over the weekend. I haven't verified the final patch yet, but earlier versions fixed the problem.
*** Bug 130259 has been marked as a duplicate of this bug. ***
Added myself to the CC list.
Brian, the fix is yet to work on Build 2002031005. Same problem. I open Mozilla, enter a Chat room on Yahoo, and browser quits and Console comes up. Date/Time: 2002-02-27 22:22:51 -0500 OS Version: 10.1.3 (Build 5Q45) Host: localhost Command: Mozilla PID: 347 Exception: EXC_BAD_ACCESS (0x0001) Codes: KERN_PROTECTION_FAILURE (0x0002) at 0xbff82ff0 with lots after this.
I went to chat.yahoo.com using a 2002030608 Commercial build... worked flawlessly for me. Are you having problems with any particular item, or just in general?
Entering the chat room, the browser quits. It tries to load the applet, and when the applet tries to render, the browser quits. Running Mac OS X 10.1.3, MRJ Carbon Plugin 1.0d8. If there is a newer version of the plugin, please let me know where I can download it?
Strange you say commercial build, I downloaded what appeared to be a commercial build as well. 0.9.9. See my build number is even higher than yours. Are we using the same OS?
Sorry, commercial == Netscape 6. latest plugin is at: http://homepage.mac.com/pcbeard/MRJPluginCarbon-1.0b1.sit
b1 is later than b8?
Yes. b(eta)1 is newer than d(evelopment)8.
I repeat: please Please PLEASE announce new versions of MRJ Plugin right here in this bug. How was anyone supposed to know about 1.0b1? It's mentioned on neither netscape.public.mozilla.macosx nor VersionTracker (which are also nice places to announce, AFTER here).
OK, I got the new beta 1 build installed, before restarting I tried running the same Mozilla build as above, and got http://www.macmaps.com/console.log error. After restarting the Mac and trying the same chat applet over again I got an error saying the applet did not initialize. It seems plugins do not even begin to work without a restart. And even so, chat.yahoo.com chat applet which is strangely supported on Netscape 6.2.1 for X is not under Mozilla 0.9.9. Somewhere along the lines some Java applets appear to be disabled by later Mozilla builds.
i know this bug is abt osx...On win NT, with jre 1.3.1_02,I too cannot see the java chat applet working ( using nscp 621)
This is from my attempt to log into yahoo chat with the 1.b1 java plugin and mozilla build 2002031308... just incase someone wants to look thru
abrody, do a Cmd-K on your log...it's getting tough to find the good stuff ;) Riscky's attachment shows the current crash... this is almost certainly related to bug 129250, the fix for which did not land on the 0.9.9 branch
change to all platforms since the applets not loading on Mac os 9.2 as well (commercial netscape build: 2002-03-11-08-trunk). 1. Type this url: http://chat.yahoo.com/ 2. Make sure you log in your name and password (if you don't have one, you need to creat it, it's free). 3. Click on "Enter Chat Room" Notice the applet is not loading.
OS: MacOS X → All
Hardware: Macintosh → All
No. Please open a specific bug for this. This is a META bug for the OSX MRJ plugin. This will (does) work under OS X with the b2 plugin which we will post as soon as we possibly can.
OS: All → MacOS X
Hardware: All → Macintosh
Nominating for topembed
Keywords: nsbeta1, topembed
Keywords: topembed → topembed+
b2 is available at the new and improved page: http://homepage.mac.com/pcbeard/MRJPlugin/ This should work better with Mozilla 0.9.9, but we have been seeing some issues when used with Netscape 6.2.1.
Comment on attachment 73973 [details] crash log from trying to use java on .9.9+ and 1.b1 This crash is caused by the fact of an incompatible interface change made to nsIURI. This is why we had to respin v1.0b2.
Depends on: 126578
Depends on: 130740
Performance of 1.0b2 with http://goproblems.com/ is still ****. It works just fine with the Windows Java plug-in under W2K. But with the Carbon Java plug-in, sometimes things don't draw, sometimes things don't draw until after the mouse cursor is moved, and sometimes the wrong hit coordinates result from a mouse click.
*** Bug 130554 has been marked as a duplicate of this bug. ***
*** Bug 120222 has been marked as a duplicate of this bug. ***
After using the b2 plugin on .9.9 (OS 10.1.3) I noticed it was a little quicker to load the java chat client (of the Digi-Chat variety), but there are still three major weaknesses I noticed: 1. After a minute or two, typing into the text field becomes agonizingly slow. As in, type 15 words, wait 45 seconds for it to fill in. Major text lag. 2. If you collapse a Window using WindowShade X, the contents of said window will not fill in when you expand it again. You have to manually drag the resize handle to get things to fill in properly. 3. You cannot launch a new Navigator window while a chat session is active (or if you open a private chat window to talk to one person). The File menu just stays highlighted after you let go of the mouse, but nothing happens. All plugin files were installed directly into the internet plugins folder, rather than using an alias.
In re my comment 91, updating to Java 1.3.1 Update 1 via Software Update has completely fixed my trouble with http://goproblems.com/ .
Depends on: 132564
Depends on: 132325
Blocks: 136063
No longer blocks: 136063
Depends on: 136063
Depends on: 135474
I just installed the new MRJPlugin but still cannot look at my aquarium-page at URL http://www.macadam.dyndns.org/html/aquarium.html under MacOS X using Netscape 6.2.2 All other browsers including a very old Netscape under MacOS 9 show this page just fine. Kind regards, Adam@macadam.dyndns.org
The question is, does that aquarium page load in the latest Mozilla build with the plugin. Netscape builds are actually always older than Mozilla builds based on the same code as Mozilla. If Mozilla can load the page and no major changes are made to Mozilla, then Netscape 6.2.3 should be able to load it when it comes out.
The applet loads for me, with 0.9.9 and version 1.0b2 of the plugin.
Looks like the plugin has reached final candidate 1 status: http://homepage.mac.com/pcbeard/MRJPlugin/ Now we have another opportunity to see if the new plugin actually improves Mac OS X java loading in Mozilla and Netscape.
Looks like the plugin has reached final candidate 1 status: http://homepage.mac.com/pcbeard/MRJPlugin/ Now we have another opportunity to see if the new plugin actually improves Mac OS X java loading in Mozilla and Netscape.
spinning beachball occurs now with 200204172 build of Mozilla for X and release candidate 1 of the carbon Java plugin when visiting http://chat.yahoo.com/?room= Macintosh::1600326592 after logging into my mail.yahoo.com mail account. I can't login to http://chat.yahoo.com directly because now Mozilla doesn't render Javascript and stylesheets on that page like it did before this new build. In any case, the applet loads, and I am stuck with a spinning beachball and forced to force quit the browser.
Is there a correlation between this bug and why the homepage.mac.com web page list does not render properly. I'm on a Ti G400 running OS 10.1.4. When I login to the "Homepage" start page, the list of web pages appears below the square which the list should occupy. I am unable to select any of the pages on my web page. I am also unable to select anything below the floating list of Homepage pages. Are there any other iTools users who can duplicate this problem? Is this related to WebObjects?
edwin, no, this bug has nothing to do with the homepage.mac.com problem. That is a problem with the HTML/CSS being served by mac.com. People should complain to Apple that they need to make their pages compatible with Mozilla (I believe the _supported_ browsers for mac.com are Netscape 4.x and IE 5.x). There is an evangelism bug logged for this I believe but the more people Apple hear it from the more likely it will get fixed. They have updated the slide show pages so they work in Mozilla for example.
Edwin, This is the bug your are mentioning. http://bugzilla.mozilla.org/show_bug.cgi?id=105791
abrody, re: chat.yahoo.com, you might try a newer build. I got the same results with a 20020419 branch build, but the builds around the weekend were pretty flakey. Using a debug build from yesterday, I got into the chat rooms with no problems.
removing topembed as this is a tracking bug
Keywords: topembed+
There is a typo in the install doc for MRJPluginCarbon-1.0fc2 The path "/Library/Internet Plugins." should be "/Library/Internet Plug-Ins", or at least that's what it is on my system. (Note the hyphen.) It might also be a good idea to move the period outside of the quotation marks, but that's a clarity vs. correctness issue I'll leave to others.
Still doesn't consistantly fix the bug for http://chat.yahoo.com/ I thought a latest nightly build did that, but then installed the new plugin and once again the chat client on Java doesn't work. Anyone know who to contact at Yahoo to ensure their Java works with your plugin?
I might add, I sometimes get a spinning beachball, sometimes am able to enter the Java chat just fine at http://chat.yahoo.com/ There is no consistancy. I'm always trying to enter the Macintosh:1 chat room.
No longer depends on: 141176
Still problematic with http://chat.yahoo.com and build 2002051005 and MRJ Carbon plugin 1.0fc2. Using Mac OS X 10.1.4 on a Flat Panel iMac. Get the dreaded spinning beachball, though I am able to force quit the program and reload it.
*** Bug 144142 has been marked as a duplicate of this bug. ***
Current trunk builds have a plugin module originated regression caused by the checkin for bug 120875. This is, unfortunately, causing major problems for the MRJ plug-in right now. The branch builds (1.0.0rc builds) should not exhibit these problems.
Build 2002051908 still doesn't solve the bug with the 1.0fc2 plugin. As a matter of fact it only gets worse. Now when I try to chat, Mozilla unexpectedly crashes and generates this crash log: http://www.macmaps.com/Mozillacrash.log
According to the bug report, the fix wasn't checked in until last night... it should be in this morning's build.
On the site www.dominategame.org (an online risk type game) Mozilla is actually the only browser I can find for OSX that will play this game... at least to some extent. There is however a few problems, first of all processor utilization during gameplay is absolutely insane... it sits at around 100% on average. I'm wondering if this is normal? (I have an 800Mhz TiBook) Secondly, and the more annoying problem is a constant and random disconnect/reconnect problem. It seems to randomly disconnect and reconnect to the game, rendering a game finished. Especially annoying after spending an hour playing :) On a warm note, good work with the plugin, and being the ONLY OSX browser that works with this site!! Mark
I don't know if this will help anyone, but I had a problem where the Java Plugin wouldn't work. It wasn't even showing up in About Plugins. Finally, I tried getting info on the plugin file, and the default association was with Netscape 4.x in OS 9 instead of Mozilla in OS X. I changed the association to Mozilla and the plugin started working.
Always the last to know... as mentioned on VersionTracker, MRJ Plugin Carbon 1.0 final is available now.
The final plugin doesn't appear to be working on http://chat.yahoo.com/ yet with Mozilla final candidate 3. Obviously it isn't final yet.
The plugin is 1.0. The Yahoo chat pages work for most (as well as for internal testing). Try bugging the folks at Yahoo.
Steve, I've tried to find a place to contact a real person at Yahoo but it seems impossible. But perhaps coming from someone at Netscape maybe they'll listen.
Yahoo Java chat WorksForMe using FizzillaCFM/2002052305 (RC3) and MRJ Plug-in 1.0.
Please see my new bug 148471 about a problem between Java applets and tabbed browsing (now blocking this bug).
Depends on: 148471
The links are not broken persay... apple has started to impose bandwidth control on iDisk users... no one knows exactly how much bandwidth you get. However I would think it would be a good idea to place the the plugin on the mozilla.org site someplace... BTW Kevin Gerich pinstripe theme had the same problem.
Blocks: 148450
Not only http://chat.yahoo.com/ but http://www.instantchess.com/ as well won't work. What happens in both instances with the latest release candidate 1.00rc3 of Mozilla for Mac OS X along with the MRJ Carbon plugin is that it gets stuck trying to load the applet, leaving me with a spinning beachball. It is possible to force quit the browser, but that should not have to be done. If a browser finds an error in code, it should generate an error report.
Keep in mind the browser simply uses the plugin to pass any Java applet to the JVM built into Mac OS X. No error checking of the code is possible. When you encounter the technicolor beachball of death with an applet use the Sampler app to grab a log of what is happening in the Mozilla process (a 5 second sample should be plenty). Please log a seperate bug and attach the Sampler output if you do that
> use the Sampler app What is the sampler app? From where can it be obtained, and how is it used (or where is its usage documented)? -matt
The Sampler app is part of the Mac OS X Developer Tools installation. If the tools installer isn't one of the CDs that came with your OS X package or pre-installed on your Mac the tools image can be downloaded from the developer.apple.com site. There are some docs installed with the dev tools but they're somewhat sparse when it comes to Sampler. You might find more info somewhere on the developer.apple.com website but personally I found Sampler pretty clear w/o docs.
WFM at chat.yahoo.com with RC3 and MRJ 1.0. I did reproduce the hang at instantchess.com. Filed a general "MRJ hangs" bug (bug 148829) to cover this and other "beachball" state bugs.
Depends on: 148829
Developers of the Java plugin may be interested in this webpage which Apple has posted on their server: http://developer.apple.com/internet/java/enterprisejava.html Hopefully this will help someone develop better Java support in Mozilla for Mac OS X.
Developers of the Java plugin may be interested in this webpage which Apple has posted on their server: http://developer.apple.com/internet/java/enterprisejava.html Hopefully this will help someone develop better Java support in Mozilla for Mac OS X.
This bug is all but finished. The remaining bugs are features we would like to have, but aren't critical.
Keywords: nsbeta1 → nsbeta1-
It is by no means finished...the start of this bug's issue hasn't been resolved by the latest browser or plugin. This bug won't be finished until we Mac users can go to any Java site supported by any web browser and use Mozilla to view the same Java.
Java is not working in most of the applications I try. For instance, the Compaq Remote Insight Manager, Lights Out card's applets fail to initialize on the Mac with the latest Moz stable and the final MRJ plugin.
Do the applets that don't work in Mozilla with the 1.0 release plugin work in any browser on OS X? Keep in mind that some applets aren't going to work with Apple's JVM and you'll need to evangelize the developer of the applet to fix them.
Lavalife java instant messenger works fine in IE 5.2 on OSX 10.1.5, however Mozilla build 2002062503 with MRJ plugin 1.0 locks up with the spinning disk. www.lavalife.com - (unfortunately, you need to be a memeber to get at the java instant messenger- the guest login only gives a 'lookalike' frame that isn't functional.)
Please do not add comments about sites with applets not working to this bug. Please file a separate bug for each site. This bug was logged as a meta bug to track the issues that blocked the 1.0 release of the MRJCarbon plugin. Now that we've shipped that this bug is likely to be closed.
I refuse to let this bug be closed until the bug is fixed. It hasn't been fixed, the 1.0 plugin does not address the issue, we are waiting for a new plugin to address the issue. I'll just reopen the bug if you close it.
Please do not let this bug be closed until either Yahoo fixes their chat client to work with the release version of Mozilla for Mac OS X, or the Java plugin is able to understand Yahoo's chat in the release version of Mozilla for Mac OS X. And while some pre 1.0 builds of Mozilla and Mac OS X appeared to have fixed the problem, the release builds of Mozilla for Mac OS X and the plugin have the issue reocurring. Thank you.
Steve, On 2002-06-25 you asked if any applets that didn't work in Mozilla worked with any other browser. My comments about Lavalife were not to introduce new complaints, but to demonstrate that yes indeed, applets that don't work in Mozilla do work in other browsers on OSX.
I'm sorry to spam this bug [s'more], but I completely agree with Steve in comment #137 about what this bug refers to. If you read the summary, this bug tracks whether or not "support" is included for Java on Mac OS X. It doesn't say anywhere that the support should be flawless, or even "as functional as other browsers". While I agree that all of that is necessary; it is not necessary for *this* bug. Hopefully, if a single other bug is opened for tracking groups of java applets that *don't* work with the MRJ 1.0 plugin, a reference can be logged here.
Abrody, if this bug is closed and you re-open it your bugzilla access will be 'reconsidered'. Have you filed a separate bug on the Yahoo problem as was asked of you several months ago? If you have it's not listed as a blocker for this bug nor is it one of the 23 open bugzilla bugs with 'Yahoo' in the summary. For the last time I'll repeat that this bug was to get the version 1.0 of MRJCarbon plugin out the door and that is done. Anthony, telling us when a site works in other browsers but not Mozilla definitely helps but they need to be logged as seperate bugs. Closing this bug doesn't mean we're not going to work on the MRJCarbon plugin, it just means the 1.0 version is done.
No I haven't submitted a separate bug since I never received the request several months ago. As you'll see in posts: 44, 65, 70, 74, 76, 82, 101, 125 support at first seemed to come back, only to later fizzle. And the fact remains the reason for my continued posting on this bug is that support in the Carbon build is not complete. We need to work with Sun, Microsoft, and Yahoo to make sure their Java code is progressive and interchangeable, allowing people to use older versions of Java on their website. It obviously isn't progressive as Netscape 4.7.9 and Explorer 5.0 for Mac OS 9 have no trouble with this applet when they have difficulty in Mac OS X which has a higher version of Sun's Java. I would prefer it to be done in Mozilla since Mozilla has more standards support than any other web browser I've used to date in Mac OS X. Java was supposed to be crossplatform from the beginning. It looks like it never fullfilled that promise. Keeping this bug open hopefully will open the doors to helping it do that.
Oh, for heaven's sakes people, you can't have one bug that's used for every problem java page any more than you can have one bug that's used for every problem html page. That's what the bugzilla database is for. Maybe there should be a java plug-in component, but that's another issue. I filed a new bug for the yahoo chat problem, bug 154183 and set it to block bug 148829, the MRJ hangs bug. I cc'd the folks with recent activity here; please add or delete your cc: as per your interest. As for this bug, Java Support for Fizilla was needed, and Java Support for Fizilla was added. That's how I see it. Thanks for the hard work, guys, I really appreciate it.
Thank you Bill...just so long as we aren't going to let development of the Java Carbon plugin stall at 1.0. That's my concern, which hopefully is unfounded, and another reason why I want this bug to persist to help development of the next MRJ Carbon plugin.
abrody, if Mozilla HTML rendering support stalled as soon as they got a basic non-quirks parser going, your point would be well received. But the Mozilla guys go to great lengths to make everything work that can possibly work, so there's no reason to fear that the MRJ plug-in will rot. Bug 154183 has 0 votes right now - incrementing that number would be a start to showing some community interest in Yahoo! chat working on OSX.
Maybe a meta bug should exist for Mac OS X MRJ plug-in issues... just an idea.
Chris Petersen and I also agreed this bug should verify and close it since the MRJ v1.0 is included in the package of the OSX daily build.
Status: ASSIGNED → RESOLVED
Closed: 23 years ago
Resolution: --- → FIXED
Verified on Mac OS 10.1.5 (branch build: 2002-06-27-05-1.0.0)
Status: RESOLVED → VERIFIED
No longer depends on: 148471
Blocks: majorbugs
No longer blocks: majorbugs
Product: Core → Core Graveyard
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: