Jar to exe is *really fragging simple* - read on...

Question

Step 1) compile this code to xjar.bin
int main(int argc, char *argv) {
    return system("javaw -jar xjar.exe");
}step 2) cat xjar.bin yourapp.jar > xjar.exe
viola, it's an exe, it can have a pretty icon, a user can double click to run, no .bat files, AND it's cross platform! because you can legally prepend data to a zip file, the exe is merely a header, for other platforms you can run the normal .bat / .jar way. (you may need to do zip -F to fix though, the kaffe VM doesn't like headered zips, sun's is ok with it though)
===
My knowledge of C was really basic a few months ago, and has rotted since, methinkst it should only take another couple of lines to do
"javaw -jar %s", args[0]
to allow renaming of the exe, but my many attempts all segfault, can someone here add to it?

Answer 1

still need a JVMWherever you go you still need libraries - and compare
the size of 8MB JRE download to all the gigabytes of
stuff in /usr/lib
The most common reason is because they want the appto be faster, this solution,
at least on start up is slightly slower.It starts slightly slower, and after that JIT takes
care of the rest. With the exception of array use
(with enforced bounds checking each access) and the
first time a each class is loaded, JITted programs can
run faster than CI didn't claim that it would run faster if actually compiled. I did however say that is what people think when they are looking for a compiled solution. I doubt that they will think that your solution meets that need.
>
>
The second most common reason is to "protect" thecode. This doesn't help
with that either.A) most people wouldn't even attempt to decompile
any .exe file, unless they know its a jar in disguise
B) Obfuscation helps a lot, even if it isn't perfect
C) C binaries can be decompiled to C source, just like
java. I can't remember the program (dcc?), it's still
in alpha and the author got a job and moved on before
finishing it, but it still works for small simple
apps.
D) Writing / editing parts in jasmin tends to screw up
decompilers, and they either crash or resort to
spewing bytecode mnemonics
Anywho, the point is is solves some of the problems,
without creating any new ones (unless you count the
nanoseconds for a 5KB C prog to start as a
problem...)...Again I didn't say that I wanted to do this. I argue against using any methods designed to protect code. But other people do. The slight hiding that this achieves might be what they want. Or not.