Re: Deja Vu: multiple definition of `java::lang::String::length()'?

[Tom Tromey <tromey at redhat dot com>]

>>>>> "Adam" == Adam Megacz <gcj@lists.megacz.com> writes:

Adam> Okay, this works, but I'm at a loss for how to add it to gcjh -- it
Adam> assumes that it doesn't need to know if the method is 'decompilable'
Adam> until after the function name and arg list have been printed to the
Adam> output file -- at which point it's too late to add an __attribute__.

Adam> Is there any way for me to add this to gcjh without major
Adam> restructuring?

You could do it by making the code uglier without major
restructuring.  This isn't a real problem since the code is already
extremely ugly.

The idea would be to have the decompilation function take a flag
indicating whether it should do work or not.  Then have it return a
value indicating whether the function will be decompiled.  (I haven't
looked at the code recently.  There may be some other reason this
can't work.)

gcjh needs a major overhaul.  Unfortunately it is very low priority.

Tom