GeezerVison is showing github.com — Superfluous Java update? · Issue #285 · scijava/scijava
Superfluous Java update? #285Description
- I started from a "Fiji Latest" instance
- Turned on the SciJava Ops site and downloaded everything
- Restarted
- When updating again I was asked to upgrade my Java (which was already at 21)
- Restarted and checked. Asked to upgrade Java again.
- Clicked OK. Seemed like something was downloading
- Ended up with still just the one Java 21?
- No more update java requests
Anyway, if there's still any java upgrade logic in Ops we should remove it in favor of the app-launcher's. Reacted by Gabriel Selzer Activityelevans commented on Feb 16, 2026 @hinerm
This is still a problem and actually breaks Fiji (at least on Linux and macOS). I've had a couple Scijava Ops workflow fail because my users are updating Fiji and breaking it. Here's more details on what this looks like. Just like you if you install SciJava Ops from the update site and restart you will get prompted to upgrade your Java:
After clicking Ok you can see that it starts to download Java:
Then it crashes. Here's my terminal output:
[INFO] Reading available sites from https://imagej.net/
#
# A fatal error has been detected by the Java Runtime Environment:
#
# SIGBUS (0x7) at pc=0x000075bd6d88ee80, pid=24673, tid=25523
#
# JRE version: OpenJDK Runtime Environment Zulu21.42+19-CA (21.0.7+6) (build 21.0.7+6-LTS)
# Java VM: OpenJDK 64-Bit Server VM Zulu21.42+19-CA (21.0.7+6-LTS, mixed mode, sharing, tiered, compressed oops, compressed class ptrs, g1 gc, linux-amd64)
# Problematic frame:
# J 232 java.util.zip.Inflater.inflateBytesBytes(J[BII[BII)J [email protected] (0 bytes) @ 0x000075bd5dbae32c [0x000075bd5dbae220+0x000000000000010c]
#
# Core dump will be written. Default location: Core dumps may be processed with "/usr/share/apport/apport -p%p -s%s -c%c -d%d -P%P -u%u -g%g -F%F -- %E" (or dumping to /home/edward/Documents/software/fiji_latest/core.24673)
#
# An error report file with more information is saved as:
# /home/edward/Documents/software/fiji_latest/hs_err_pid24673.log
#
# If you would like to submit a bug report, please visit:
# http://www.azul.com/support/
# The crash happened outside the Java Virtual Machine in native code.
# See problematic frame for where to report the bug.
#
Aborted (core dumped)
Trying to start Fiji again after this crash results in this terminal output:
OpenJDK 64-Bit Server VM warning: You have loaded library /home/edward/Documents/software/fiji_latest/java/linux64/zulu21.42.19-ca-jdk21.0.7-linux_x64/lib/libzip.so which might have disabled stack guard. The VM will try to fix the stack guard now.
It's highly recommended that you fix the library with 'execstack -c <libfile>', or link it with '-z noexecstack'.
Error occurred during initialization of boot layer
java.lang.UnsatisfiedLinkError: /home/edward/Documents/software/fiji_latest/java/linux64/zulu21.42.19-ca-jdk21.0.7-linux_x64/lib/libzip.so: /home/edward/Documents/software/fiji_latest/java/linux64/zulu21.42.19-ca-jdk21.0.7-linux_x64/lib/libzip.so: file too short
What I think is happening here is that there is still some update logic that pulls down a Java 21, but then it seg faults before downloading completely resulting in a file thats too small. This failure path is reproducible for me. It always seg faults and produces this file too short error. hinerm commented on Feb 16, 2026 @elevans can you clarify where people are starting from? Presumably fiji-latest? Are there other update sites involved?
There is a note but non-issue in the Fiji fit project board to remove the scijava-ops update site. I am not sure what, if any, obstacles there are to integrating SciJava Ops into mainline Fiji at this point. @ctrueden do you remember? Is this documented anywhere?
In any case the short-term fix is:
- Remove the Java update logic from SciJava Ops update site
The true fix is:
- Migrate SciJava Ops content to Fiji-latest site and disable/archive the SciJava Ops update site
Reacted by Gabriel Selzer gselzer commented on Feb 16, 2026 Yeah, would be good to have a MVCE @elevans, I can reproduce @hinerm's process but I don't get crashes while doing so. elevans commented on Feb 16, 2026
@elevans can you clarify where people are starting from? Presumably fiji-latest? Are there other update sites involved?
Yes this is starting from fiji-latest. There are no other update sites involved. This happens with a fresh fiji-latest.
Yeah, would be good to have a MVCE @elevans,
Agreed! I just performed these steps minutes ago:
1). Download fiji-latest.
2). Unzip and start Fiji (I did it from the terminal).
3). Go to updates and select the Scijava Ops update site. Download and install all the .jars.
4). Restart Fiji.
5). Run Update again. This triggers the update prompt on my system that leads to the crash. gselzer commented on Feb 16, 2026
@elevans can you clarify where people are starting from? Presumably fiji-latest? Are there other update sites involved?
Yes this is starting from fiji-latest. There are no other update sites involved. This happens with a fresh fiji-latest.
Yeah, would be good to have a MVCE @elevans,
Agreed! I just performed these steps minutes ago:
1). Download fiji-latest. 2). Unzip and start Fiji (I did it from the terminal). 3). Go to updates and select the Scijava Ops update site. Download and install all the .jars. 4). Restart Fiji. 5). Run Update again. This triggers the update prompt on my system that leads to the crash.
Huh, interesting! I cannot replicate on Windows. It is possible though that the fix will avoid your error 🤞 ctrueden commented on Feb 16, 2026 I uploaded the current SciJava Ops to Fiji-Latest. Please test on your favorite platforms, and if all is well, we can update the ops.scijava.org docs, retire the SciJava Ops update site and then close this. gselzer commented on Feb 16, 2026 elevans commented on Feb 18, 2026 Thanks @ctrueden! Moving SciJava Ops to Fiji-latest fixes things for me!
MetadataMetadataDevelopmentNo branches or pull requestsIssue actions
|
