Re: [patch] noinst_JAVA

[Tom Tromey]

>>>>> "Rusty" == Rusty Ballinger <address@hidden> writes:

Rusty> Hey, a thing that's bothered me for a while is that noinst_JAVA
Rusty> is broken.

Thanks, I checked in your patch.

Rusty> (The reason for this test is that the install rule in java.am
Rusty> installs *.class; in Todd's case, for example, any class files
Rusty> which happened to be built for his noinst_JAVA files would have
Rusty> been installed in javadir.)

Yeah.  I think it will be far more common to install .jar files
though.  Really for Java we should have something that makes it as
easy to make jar files as it is to make .a libraries.

Rusty> It seems like multiple _JAVA directories could be allowed if
Rusty> the install & uninstall rules transformed each _JAVA file name
Rusty> from Foo.java to "Foo.class Foo\$*.class"; are there some class
Rusty> files that would miss?

Yes.  Consider `bar.java':

    class foo { }
    public class bar { }

Here you could miss `foo.class'.

gcj has a tool to detect this (it will list all the classes in a
file), but of course we can't rely on the existence of this tool.
Most Java developers don't (yet :-) use gcj.

Tom