It will definitely be built at Java 7 level then, which means Java 6 will throw an exception if you try to use it.
@rimadoma BTW:
$ mvn -v
Apache Maven 3.3.9 (bb52d8502b132ec0a5a3f4c09453c07478323dc5; 2015-11-10T10:41:47-06:00)
Maven home: /usr/local/Cellar/maven/3.3.9/libexec
Java version: 1.8.0_73, vendor: Oracle Corporation
Java home: /Library/Java/JavaVirtualMachines/jdk1.8.0_73.jdk/Contents/Home/jre
Default locale: en_US, platform encoding: UTF-8
OS name: "mac os x", version: "10.11.4", arch: "x86_64", family: "mac"
You could try a different Maven core version to see if that makes a difference with your local enforcer error.
@ctrueden Ah, well I guess it's again related to the problems with legacy code, Java3d, Java 8, that forum thread (http://forum.imagej.net/t/java3d-issue-bonej-with-latest-fiji-version-problem-solved/1151/29). I made a post there earlier today, so let's take that jvm version discussion there, shall we?
@ctrueden Sure, I'll try a different core version, and report back later
@ctrueden Interestingly the newest maven from apt-get was 3.0.5, so I had to add a repo (https://launchpad.net/~andrei-pozolotin/+archive/ubuntu/maven3) to apt-get to install 3.3.9
@ctrueden Also when I googled for the exception I was getting I ran into this old conversation: https://groups.google.com/forum/#!topic/fiji-devel/k_6X3DY_cPY
If 3.0.5 really does not work, we should update the minimum requirement in pom-scijava accordingly.
@ctrueden ImageJ doesn’t currently run on 32-bit Windows 10 due to the 3/4 of available memory approximation for maximum heap space. This evaluates to 1.5G (=3/4*2G) on two different systems, which leads to the JVM telling me
Could not reserve enough space for 1572864KB object heap.. Can anyone confirm that issue? Also fails for 32-bit ImageJ/Fiji on a 64-bit Windows 10..
I have Win systems where 32-bit Java cannot start with as little as 1.1g. And others where it will start fine with 1.4g.
My statements above were mostly Windows XP, IIRC. Maybe some Windows 7.
@stelfrich The patch would be a super easy one-liner; see this line.
All I ask is that you clone imagej-launcher locally, build locally, and actually test that it fixes your problem when you lower that value.
And do some testing to determine exactly how low you need to put the bar.
Then, you can file a PR with that new number.
@/all New version of ImageJ released: 2.0.0-rc-47. Lots of new stuff, including a revamped (but still buggy/alpha-level!) Script Interpreter. Also fixes the long-standing inability to use "choices" in a parameterized script, as well as a recent BeanShell bug. And tons of other things I don't have time to to mention at this moment.
Version
2.0b6 is already on Maven Central, and we ship it with ImageJ...
Note that the groupId changed though. See http://maven.imagej.net/index.html#nexus-search;gav~~bsh~~~
See also beanshell/beanshell#17
My problem was that I was searching for "beanshell".
Which only turns up the 2.0b4 jar.
Thanks!
I'm not entirely clear on exactly what's going on, but micro-manager/micro-manager#170 details the issue.
Basically, if you define a function in a REPL (read-eval-print loop), then you can't later redefine it; the name you assign to that function is permanent AFAICT.
(In 2.0b4 the same was true of functions defined in scripts, which was a major problem for script development, but 2.0b6 fixes that...for some unknown reason)
So you guys use ..
What? A bsh-specific repl?
Do your scripts suffer from bsh's "null vs void" problem?
All I ask is that you clone imagej-launcher locally, build locally, and actually test that it fixes your problem when you lower that value.
@ctrueden: I was about to do that and I am struggling with the compilation on Windows. Is there any source of information on the topic?
The 32-bit portion of the build script is:
# build 32-bit
UNAME="$(uname -s)"
case "$UNAME" in
Linux)
(export JAVA_HOME=/home/dscho/fiji/java/linux/jdk1.6.0_24 &&
export PATH=$JAVA_HOME/bin:$PATH &&
mvn -P i386-Linux,copy-launchers -P !amd64-Linux -X install)
;;
MINGW*)
(export JAVA_HOME=$(cd /src/fiji/java/win32/jdk1.6.0_24 && pwd -W) &&
export PATH=/c/msysgit/mingw/bin:/c/msysgit/bin:$JAVA_HOME/bin:$PATH &&
mvn -Pcopy-launchers -U -Dos.arch=x86 -X -e install)
esac
So, it uses msysgit.
It will be really nice not to maintain the launcher anymore...
@stelfrich If you like, another thing you could test is the JavaFX launcher, to see whether that resolves the issue.