RE: Problem in Class.java

[John Keiser]

> -----Original Message-----
> From: John Leuner [mailto:address@hidden
> Sent: Wednesday, April 25, 2001 7:11 PM
> To: John Keiser
> Cc: address@hidden
> Subject: Re: Problem in Class.java
>
>
> > Very good catch.  Definitely needs a test, too. The only thing
> it'll work
> > for right now is nested arrays.
> >
> > I think it would be a lot easier (and more efficient) to parse in
> > Class.java.  How's this?
> >
> > (Source:
> >
> http://java.sun.com/products/jdk/1.1/docs/guide/jni/spec/types.doc
> .html#1643
> > 2)
> >
> > This is not even compiled but It Should Work (TM).
>
> I don't mean to criticise, but writing code without compiling it
> is extremely error prone.
>

Sure is.  That's why I sent it here instead of checking it in :)  I didn't
have a copy of Classpath handy and I don't feel right spending more than a
few minutes on it at work.

The problem my patch had, though, wouldn't have been caught by the compiler.
It was a good deal more insidious.

> The following patch works with my JVM (and I've run some tests
> against the IBM JVM), I hope I haven't missed anything.
>
> BTW, is this the correct way to generate and submit a patch?
>

I believe so.

The patch looks good (much better than my attempt); the only thing I would
do is put the '[' and 'L' case into the switch ... it's slightly more
efficient that way, less function calls and such.

I'll try and get a cvs tree down tomorrow.

--John Keiser