Re: [Sisc-users] Re: java stack trace

On Tue, Jul 01, 2003 at 11:43:04PM +0100, Matthias Radestock wrote:
> [email protected] wrote:
> >On Tue, Jul 01, 2003 at 04:56:05PM -0400, Narang wrote:
> >
> >>can you please provide an example. Thanks.
> >
> >
> >Sorry.
> >
> >(import s2j)
> >(import generic-procedures)
> >(define-java-class <java.lang.Exception>)
> >(define-generic-java-method print-stack-trace)
> >
> >(define (print-java-stack-trace erec)
> >  (cond [(assq 'java-exception erec) =>
> >         (lambda (je)
> >           (print-stack-trace (cdr je)))]
> >        [else (error 'print-java-stack-trace 
> >                     "No Java exception associated.")]))
> >
> 
> There is no need for s2j.
> 
> Stack traces for exceptions thrown from s2j can be printed using 
> print-exception. They are also printed automatically by the REPL (just 
> like the Scheme exception stack trace) if the stack-trace-on-error 
> parameter is #t.
> 
> Stack traces for the nested Java exceptions of ordinary primitives can 
> be displayed as follows:
> 
> #;> (java/class 'foo)
> Error in java/class: class foo not found
> ?:?:?: <indeterminate call>
> file:repl.scm:79:34: <from call to eval>
> file:repl.scm:104:20: <indeterminate call>
> file:repl.scm:92:9: <indeterminate call>
> #;> (display-java-stack-trace (cdr (assq 'java-exception 
> (get-last-exception))))
> java.lang.RuntimeException: class foo not found
> 	at sisc.modules.s2j.Reflection.doApply(Reflection.java:281)
> 	at sisc.nativefun.NativeProcedure.apply(NativeProcedure.java:26)
> 	at sisc.exprs.AppEval.eval(AppEval.java:29)
> 	at sisc.interpreter.Interpreter.next(Interpreter.java:128)
> 	at sisc.exprs.AppExp.eval(AppExp.java:33)
> 	at sisc.interpreter.Interpreter.interpret(Interpreter.java:112)
> 	at sisc.interpreter.Interpreter.interpret(Interpreter.java:101)
> 	at sisc.interpreter.Interpreter.eval(Interpreter.java:244)
> 	at sisc.data.SchemeThread.run(SchemeThread.java:62)
> 	at java.lang.Thread.run(Thread.java:536)
> #;>
> 
> 
> You shouldn't need this in normal use.

I believe the example he was referring to fit into neither category.  In 
cases where a RuntimeException is thrown from a Primitive, we did not 
record the Java exception previously.  Recording it will be necessary 
for the IO Conditions SRFI for example.

	Scott

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQE/Amnyr9IW4v3mHtQRAlIWAJwNWuY/K0xXJTwH48ujWTZpUv64/QCfWsgP
My8F7mBZC+GATBQcHHvMEGw=
=63xo
-----END PGP SIGNATURE-----