Yes, I read through the C++ code, and it seems strange that it's not caught. I'll put together that gdb stack trace for you.
Bill > > On Thu, 17 Jul 2008, Bill Janssen wrote: > > > Thanks, Andi. > > > > Unfortunately, I've already fixed the Java code so that this is no > > longer thrown. However, maybe I can put together a little test case > > for you. > > > > Remember that IllegalArgumentException is an "unchecked" exception, > > that can be thrown at any time by any code. > > Yes, I know, but it's an exception nonetheless that JNI's > ExceptionOccurred() method is telling me about and that fact triggers me to, > in turn, throw a C++ exception. The problem is not with your error but with > the fact that the C++ exception I throw because of it is not caught. > > Andi.. > > > > > Bill > > > >> > >> On Thu, 17 Jul 2008, Bill Janssen wrote: > >> > >>>>> Exception in thread "main" java.lang.IllegalArgumentException: Unknown > >>>>> image type 0 > >>> > >>>> And no stacktrace ? > >>> > >>> Yes, I have a stacktrace. (Or had -- I think I've deleted it.) > >>> What's interesting in it? I'm trying to create a BufferedImage from > >>> another BufferedImage, read in using ImageIO.read(). The constructor > >>> for BufferedImage is raising this exception. > >>> > >>>> Exception in thread "main" java.lang.IllegalArgumentException: Unknown > >>>> image type 0 > >>>> at java.awt.image.BufferedImage.<init>(BufferedImage.java:490) > >>> > >>> The code that it's objecting to is this: > >>> > >>> BufferedImage newImg = new BufferedImage((int)(scl*width), > >>> (int)(scl*height), oldImg.getType()); > >> > >> Ok, so we've determined that the original error is a user error. > >> Now, it'd be interesting to see why the JCC-generated code is not catching > >> the exception as expected and reporting it. > >> > >> Every generated call into Java is supposed to be surrounded by an > >> INT_CALL() > >> or OBJ_CALL() macro (defined in macros.h) that expands into a try/catch C++ > >> construct converting the C++ exception into a Python one. > >> > >> A gdb stacktrace (with symbols) should prove most useful. > >> > >> Andi.. > > > > > > _______________________________________________ pylucene-dev mailing list pylucene-dev@osafoundation.org http://lists.osafoundation.org/mailman/listinfo/pylucene-dev
