Hi,
I’ve recently customized Spark for evaluation in company use.
Currently only windows, modified look and feel and: with** integrated JRE** folder (WinRAR SFX)
The most people use Java 7. Spark doesn’t use wordwrap in chatrooms, thats why I packed the latest JRE 6
I’ve asked an employee, “Hey, could you test it” and Spark didn’t even start, Why? -> Java x64
So I ran Spark with procmon and was wondering why:
-
Spark finds its JRE folder, scans it
-
Spark is searching afterwards in registry
-
Last result of found JRE is taken (Registry)
So I thought checking out Spark, with all those weird tools (yes, I’m not very familiar with java/eclipse/install4j/izpack), building it could be an option, so I would have been removing the “Registry search” from it. But even with Spark trunk checkout I was not able to build the very same exe as shipped from igniterealtime download page. Build went fine but: Unstable, crashing at logout, strange.
I’ve ended up building an EXE with AutoIT and thats using the JRE I’ve packed with it.
ShellExecute(@ScriptDir & “\jre\bin\javaw.exe”, “-Dappdir=. -cp lib/jdom.jar;lib/log4j.jar;lib/lti-civil.jar;lib/fmj.jar;lib/startup.jar;lib/wi ndows/jdic.jar;resources;lib/windows; -Djava.library.path=”“lib/windows”" org.jivesoftware.launcher.Startup", @ScriptDir)
This is something that should be fixed.
We have many devs here which are using different versions of Java, and IF a JRE folder EXISTS, then this version should also be used… not?
Many thanks for this great Project!