Bugzilla Bug 200550

dev-java/java-gnome-4.0.5 doesn't compile with -jN with N > 1; the compilation
fails at the beginning with the following error:

make: *** No rule to make target `tmp/gtk-4.0.jar', needed by `build-native'. 
Stop.
make: *** Waiting for unfinished jobs....
JAVA    BindingsGenerator

It then continues to compile libgtkjni-4.0.so, and fails when it ends.

Reproducible: Always

Steps to Reproduce:
1. Set MAKEOPTS=jN with N > 1
2. Try emerge =dev-java/java-gnome-4.0.5
3. The compilation fails

Actual Results:  
The compilation fails because the package cannot be compiled in parallel.

Expected Results:  
The package should compile.

Using "emake -j1" instead of "emake" solves the problem.

That's interesting. I have MAKEOPTS=-j3 here and the java-gnome ebuild installs
fine.

[As it happens, `make` doesn't actually do anything other than act as a
familiar front end to the script that actually does the build, (which,
regrettably and admittedly doesn't do parallel compiling, but that's not the
issue here) so the whole matter (beyond getting it to not splat for you) is
somewhat academic].

AfC

Weird; it fails for me in my amd64 desktop and in my x86 laptop, every time,
with the ebuild (which uses emake) and calling make by myself.

It fails with -j3 and -j2, works without any of them or with -j1. I'm using
stable packages for the most part: in particular nothing from Java (except
java-gnome) is keyworded ~x86 or ~amd64.

I'm using the Sun JDK 1.6, but it also fails with the 1.5 version.

No, I don't think it's anything to do with Java or your arch (stable or
otherwise) for that matter.... oh wait.

Target 'build-native'. Huh. Ok, that's the GCJ build path. It *is* still a
Makefile (and has been ... not deprecated, but somewhat ignored for quite a
while; with no standalone gcj ebuild, we haven't had GCJ installed on any of
our systems since we upgraded everyone a few months ago).

If you care to, join me in #java-gnome on gimpnet (irc.gnome.org) or in
#gentoo-java on Freenode (irc.freenode.net) and we can debug it further there,
but as far as Gentoo is concerned, yes, it would seem wise to either a) pass
-j1 regardless, or b) pass -j1 if GCJ is likely to be found.

AfC

(In reply to comment #3)
> 
> If you care to, join me in #java-gnome on gimpnet (irc.gnome.org) or in
> #gentoo-java on Freenode (irc.freenode.net) and we can debug it further there,
> but as far as Gentoo is concerned, yes, it would seem wise to either a) pass
> -j1 regardless, or b) pass -j1 if GCJ is likely to be found.
> 

Added -j1 so done on the Gentoo side.