This is the mail archive of the java@gcc.gnu.org mailing list for the Java project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: two darwin issues


Hi Geoff,


Geoff Strom wrote:


I've built recent cvs versions of 3.4 on both Jaguar and Panther and can happily report that both JNI and threads seem to work fine.

However I am seeing two specific failures that I'm not able to adequately debug.

The first is running SWT programs. I can compile a SWT based program fine. When running it certainly looks like it is linking to the .dylib correctly, but it always dies after some delay with a "Bus Error" (condensed crash log #1 below).

The second issue seems like a regression. I can build and run Mozilla's Rhino interpreter and it works great for pure javascript. Any time I attempt to use LiveConnect and load Java objects, it dies immediately (condensed crash log #2 below). I swear this worked for me before when I was working with GCC 3.3.1 (FSF).

Could either of these be affected by gc? Should I try updating to 6.3alpha4?

How can I help find the real issues here?

Just a fyi, darwin is missing an important part to handle NPE correctly. I sent a preliminary patch around the 30.1 to gcc-patches, it implements the MD_FALLBACK_FRAMESTATE_FOR macro which is neccessary to unwind in case of a NPE. (Null Pointer Exception).


Also a patch to libjava signal handling is necessary.

And second, between 3.3.x and 3.4.x fsf gcc on darwin got the dylib implementation. So it's no longer a static library .....

These things maybe related to your failures you see.

We hope to get these things sorted out soon.

I was able to run some awt gtk stuff, yes some windows without swing.



Andreas



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]