Closed Bug 53907 Opened 25 years ago Closed 25 years ago

New Linux Installer bits for Java 2 support

Categories

(SeaMonkey :: Installer, defect, P1)

x86
Linux
defect

Tracking

(Not tracked)

VERIFIED FIXED

People

(Reporter: drapeau, Assigned: samir_bugzilla)

References

Details

(Whiteboard: [nsbeta3++][rtm+] Reviewed and approved fix in hand)

Attachments

(1 file)

Linux Java 2 bits are ready for PR 3. They are delivered in a similar form to PR 2, with small modifications: * This deliverable is the JRE only; the PR 2 deliverable handed to Netscape was a full JDK. This deliverable is smaller. * This deliverable must not be modified (i.e., remove files from the JRE distribution). Java Software informs that this will make it not a real JRE. Instructions for installation: * Take the binary tar.gz ball and make it into a .xpi. * When the installer installs it, installer may unpack the thing anywhere it wants to. * After unpacking the JRE, browser installer must make a symlink from $MOZILLA/dist/bin/plugins/libjavaplugin_oji.so to $JRE/lib/i386/libjavaplugin_oji.so (i.e., actual location of OJI plugin). The OJI plugin can figure out where the rest of the JRE is from there. No other installation necessary. No environment variables, etc. Questions? Please send email to drapeau@eng.sun.com. I hope this is clear; please forgive me if not, and I'm happy to try and clarify.
Nominating for NSBeta3 and RTM, assigning P1 priority to the bug. Java will not work on Linux without applying this patch with these instructions, for both default install and for auto-download from the Plugin Finder page.
Priority: P3 → P1
Whiteboard: [nsbeta3] [rtm]
Assignee: ssu → sgehani
Keywords: nsbeta3, rtm
Whiteboard: [nsbeta3] [rtm] → [nsbeta3+]
Fixing nominations and approving nsbeta3+. Checkins to the Netscape branch require nsbeta double-plus from the PDT so we have to wait for that.
Status: NEW → ASSIGNED
George, So am I correct in assuming that mozilla will look for libjavaplugin_oji.so in "<mozilla-bin_INSTALL_DIRECTORY>/plugins"?
Samir: yes, that is correct. To verify, you can try the Linux bits that you already have for PR 2; same behavior. On Linux Mozilla and Netscape 6, the browser will look in the plugins/ subdirectory for plugins; if the libjavaplugin_oji.so is there, OJI can be activated, and the JRE will be correctly launched. The alternative would have been to put libjavaplugin_oji.so in the Mozilla/N6 components/ subdirectory, but libjavaplugin_oji.so is not an XPCOM component.
Hello? PDT? This should most definately be nsbeta3++, otherwise there won't be any linux java support (again!)
Whiteboard: [nsbeta3+] → [nsbeta3+] Awaiting bits from leaf
It would be a shame to ship PR3 without Java on Linux again. Where are the bits and the patch? Will minus tomorrow if the pdt doesn't beat me to it.
Whiteboard: [nsbeta3+] Awaiting bits from leaf → [nsbeta3+][rtm+] Awaiting bits from leaf
Target Milestone: --- → M18
Leaf? Do you have the bits?
Blocks: 49451
I can bring Linux bits in person on CD, if necessary, but I believe leaf already has them. Those guys are busy. That's probably the reason they haven't closed every bug in the universe yet. I'm happy to help in any way I can.
Whiteboard: [nsbeta3+][rtm+] Awaiting bits from leaf → [nsbeta3++][rtm+] Awaiting bits from leaf
Marking nsbeta3++ if we've already got the bits. It's great to see this landing.
*** Bug 49451 has been marked as a duplicate of this bug. ***
Bringing Linux and Win32 bits in person on CD on Wednesday morning. Go, Netscape RE!
Have you fixed the problems with Java & *nix/X windows detailed on this page? http://www.mozilla.org/unix/using-x-with-mozilla-threads.html
OS=Windows2000 sounds wrong, doesn't it?
Yep
OS: Windows 2000 → Linux
we have the bits. I've placed them at http://blues/users/granrose/publish/j2se_1.3.0_01_linux.tar.gz for easier access.
XPInstall does not have sym linking functionality. So, I have worked out a hack: in the jre.xpi we will include a shell script wrapper to "ln" and call it using Install.execute() with args having the fully resolved path names (from the toString() form returned by getFolder("Plugins") + {<srcfile>,<linkname>}). I just tested this on a recent build and it works. Open to alternate suggestions. As far as I can tell this is lower risk than implementing the functionality in the XPInstall engine. (The hooks for implementing the File.unixLink() API are in the XPInstall engine but it appears that nsLocalFileUnix doesn't have symlinking functionality which would need fleshing out.)
Whiteboard: [nsbeta3++][rtm+] Awaiting bits from leaf → [nsbeta3++][rtm+]
Hey, I'm no expert, but this seems like a pretty safe workaround that gets the job done.
George, There is no libjavaplugin_oji.so in the lib/i386 directory in the current drop. However, there is a libjavaplugin_jni.so. Do you want me to create a symlink to this latter named lib (with *jni*, not *oji*)? If so, I presume the symlink should be identical to the libname: i.e. does Netscape 6 look for libjavaplugin_jni.so or libjavaplugin_oji.so? (Both moz and ns lxr didn't help -- couldn't find any "libjavaplugin" string). This is urgent so please respond at the earliest. Today is the last day to get this change in. Thanks. The lib/i386 directory in your current drop to granrose looks like this: {orb} /u/sgehani/dev/linux/jre/jre-image-i386/lib/i386 > ls -la total 5752 drwxr-xr-x 5 sgehani wheel 4096 Sep 23 05:55 ./ drwxr-xr-x 10 sgehani wheel 4096 Sep 23 05:55 ../ drwxr-xr-x 2 sgehani wheel 4096 Sep 23 05:33 client/ lrwxrwxrwx 1 sgehani wheel 6 Sep 27 19:34 hotspot -> client/ -rwxr-xr-x 1 sgehani wheel 48004 Sep 4 23:36 libJdbcOdbc.so* -rwxr-xr-x 1 sgehani wheel 3385161 Sep 4 23:18 libawt.so* -rwxr-xr-x 1 sgehani wheel 286878 Sep 4 23:30 libcmm.so* -rwxr-xr-x 1 sgehani wheel 177572 Sep 4 23:08 libdcpr.so* -rwxr-xr-x 1 sgehani wheel 517611 Sep 4 23:27 libfontmanager.so* -rwxr-xr-x 1 sgehani wheel 72424 Sep 4 22:35 libhprof.so* -rwxr-xr-x 1 sgehani wheel 19021 Sep 4 23:33 libioser12.so* -rwxr-xr-x 1 sgehani wheel 224952 Sep 5 05:32 libjava.so* -rwxr-xr-x 1 sgehani wheel 180528 Sep 23 05:04 libjavaplugin_jni.so* -rwxr-xr-x 1 sgehani wheel 5486 Sep 4 23:39 libjawt.so* -rwxr-xr-x 1 sgehani wheel 51547 Sep 4 22:35 libjcov.so* -rwxr-xr-x 1 sgehani wheel 162977 Sep 4 23:29 libjpeg.so* -rwxr-xr-x 1 sgehani wheel 246735 Sep 4 22:58 libjsound.so* -rwxr-xr-x 1 sgehani wheel 214414 Sep 4 23:08 libmlib_image.so* -rwxr-xr-x 1 sgehani wheel 34591 Sep 4 22:48 libnet.so* -rwxr-xr-x 1 sgehani wheel 79103 Sep 4 21:45 libverify.so* -rwxr-xr-x 1 sgehani wheel 73913 Sep 5 05:32 libzip.so* drwxr-xr-x 2 sgehani wheel 4096 Sep 23 05:33 native_threads/ drwxr-xr-x 2 sgehani wheel 4096 Sep 23 05:33 server/ {orb} /u/sgehani/dev/linux/jre/jre-image-i386/lib/i386 >
Whiteboard: [nsbeta3++][rtm+] → [nsbeta3++][rtm+] [need info from Java folks]
If I make a symlink named $MOZILLA/plugins/libjavaplugin_jni.so -> $JRE/lib/i386/libjavaplugin_jni.so ^^^ then Netscape 6 does not launch. Further, if I make a symlink named $MOZILLA/plugins/libjavaplugin_oji.so -> $JRE/lib/i386/libjavaplugin_jni.so ^^^ then Netscape 6 still does not launch. George, Can you put me in touch with the folks who have tested these JRE bits against Linux Netscape 6? Thanks.
Working on it now, Samir. I'll also contact the manager of the Java Plug-In group at Sun, since I believe he's the one who created the bits.
Arrrggghh! My bad! Poor instructions on my part. Here's the correction: To find the file "libjavaplugin_oji.so", look here: $JRE/plugin/i386/libjavaplugin_oji.so NOT $JRE/lib/i386/ My fault. This should have been obvious to me when I read the message. Please look at the Linux Java bits you have. When unpacked, you should have a "plugin/" directory at the top level; if not, then I believe I gave you bad bits. This should do it. Please feel free to contact me via AIM (gdeweydrapeau) or Yahoo! (georgedrapeau) if you wish to chat.
George, Thanks for the update. That worked. Java works in a recent Netscape 6 linux build.
Whiteboard: [nsbeta3++][rtm+] [need info from Java folks] → [nsbeta3++][rtm+]
sgehani, I believe the plugin doesn't work with debug builds (in xpi?). (see comments in 49451). I don't know if this is your problem.
ssu, Please review the attached patch. Thanks. vishy, Please super review the attached patch. Thanks.
Whiteboard: [nsbeta3++][rtm+] → [nsbeta3++][rtm+] Fix in hand
r=ssu
Whiteboard: [nsbeta3++][rtm+] Fix in hand → [nsbeta3++][rtm+] Reviewed fix in hand
sr=a=vishy.
It's been reviewed and approved. Please get it in.
Whiteboard: [nsbeta3++][rtm+] Reviewed fix in hand → [nsbeta3++][rtm+] Reviewed and approved fix in hand
Fix checked into branch and trunk.
Status: ASSIGNED → RESOLVED
Closed: 25 years ago
Resolution: --- → FIXED
Thank you. :-) (sorry if my earlier comments appeared rude :-( )
Blocks: 54594
Java did install per the installation UI but once in N6, when I tried to access www.java.sun.com I got the following error in console: Create instance called in JavaPluginFactory5!!!!!!!! There was an error trying to initialize the HPI library. Please check your installation, HotSpot does not work correctly when installed in the JDK 1.2 Linux Production Release, or with any JDK 1.1.x release. Could not startup JVM properly! java_vm process: could not start Java VM INTERNAL ERROR on Browser End: Could not read ack from browser System error?:: Resource temporarily unavailable not sure if this should be new bug?
I could play the Java Hangman applet just fine the 2000-09-29-09-MN6 Linux installed build. Sounds like java is working.
When I download and untar the .tar.gz for today's branch build, I don't get a build with java working. Is there something special I have to do to see this?
I logged bug 54763 to deal with crash when trying to access www.java.sun.com to test applets It did install and sym links are set. I believe this bug is complete- bits are installed
vfy build 2000092909
Status: RESOLVED → VERIFIED
oops, added vtrunk
Keywords: vtrunk
Reopening to Resolve as Fixed. Bugs stay resolved until verified on both the branch and the trunk.
Status: VERIFIED → REOPENED
Resolution: FIXED → ---
Resolving as Fixed. vtrunk keyword is in place requesting trunk verification before this bug changes state to Verified. When this is verified on the trunk vtrunk keyword should be removed and the bug Status should be set to Verified. Sorry for the spam, just trying to make sure we cover all of our bases here.
Status: REOPENED → RESOLVED
Closed: 25 years ago25 years ago
Resolution: --- → FIXED
sgehani, The file jre.jst in the patch is under the NPL/MPL, too, right?
mozilla@bucksch.org, Actually jre.jst is in the ns source tree (non-public). This bug should have really been in bugscape.
Verified on the 11/14 Linux trunk build.
Status: RESOLVED → VERIFIED
Keywords: vtrunk
Product: Browser → Seamonkey
Component: Installer: XPI Packages → Installer
QA Contact: agracebush → general
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: