Re: GCC 3.4 regression reproduceable under Linux PR 12548

[Mohan Embar <gnustuff at thisiscool dot com>]

Hi Ãyvind and Rutger,

>This email is perhaps a bit superfluous since the bug is marked as
>critical, but this compilation error is in the way of a half a dozen
>other tests that I'd like to run against GCC 3.4.
>
>I verified that the problem is reproduceable under Linux using a
>Linux->Windows cross compiler.
>
>http://gcc.gnu.org/bugzilla/show_bug.cgi?id=12548

Can you do the following?:

- Get and install my latest build:

  http://www.thisiscool.net/gcc34-20031120.zip

  (This is the UNICOWS build, but forget that for now.)

- Get the following file:

  http://www.thisiscool.net/jc1R_20031120.zip

- Unzip the contents of jc1R_20031120.zip to the thisiscool-gcc/gcc-3.4
  directory. This should replace the previous jc1.exe at
  thisiscool-gcc/gcc-3.4/libexec/gcc/i686-pc-mingw32/3.4.

- Verify BOTH of the following:

  - that the compile works instead of failing
  - that resources in a jarfile are handled properly. (I've never
    done compiled resources yet so I don't know how to test these.)

(Can someone send me a tiny sample program I can add to my examples
which illustrates compiling resources into the executable and
accessing them?)

If you verify both of the above, I'll post my findings so as not
to lead people down a blind alley.

-- Mohan
http://www.thisiscool.com/
http://www.animalsong.org/