RE: Static array length access bug in gcj

["Anand Krishnaswamy" <anandkr at nortelnetworks dot com>]

Title: RE: Static array length access bug in gcj

Thanks for the workaround. It works fine for my needs. I did file the bug report and here's the information:
        Thank you very much for your problem report.
        It has the internal identification `gcj/129'.
        The individual assigned to look at your
        report is: apbianco.

        >Category:       gcj
        >Responsible:    apbianco
        >Synopsis:       Static array length access bug in gcj
        >Arrival-Date:   Fri Jan 07 12:20:00 PST 2000

Anand

-----Original Message-----
From:   Jeff Sturm [SMTP:jsturm@sigma6.com]
Sent:   Friday, January 07, 2000 2:51 PM
To:     Krishnaswamy, Anand [RICH2:2K25:EXCH]
Cc:     'java-discuss@sourceware.cygnus.com'
Subject:        Re: Static array length access bug in gcj

Anand Krishnaswamy wrote:
> public class B {
>    public static void main ( String[] args ) {
>         int length = A.strArr.length;

Gcj still has trouble parsing ambiguous expressions like above (it has
to decide if A is a class or package name).  It will compile if you
decompose the complex expression:

          String[] array = A.strArr;
          int length = array.length;

I reproduced this bug with a recent snapshot.  I still don't see it
mentioned in gnats however.  Can you file a bug report?

http://sourceware.cygnus.com/cgi-bin/gnatsweb.pl

--
Jeff Sturm
jsturm@sigma6.com