Regression for csl-arm on arm-none-elf

[Phil Edwards]

On Sun, Sep 19, 2004 at 04:14:06PM -0700, Mark Mitchell wrote:
> Also, for arm-none-elf,  do we have the arm-none-elf binutils in the
> PATH when we're running tests?  I've discovered that DejaGNU wants to be
> able to run "objdump" directly; if it can't, these tests might fail
> spuriously.

Do you recall where DejaGNU tries to do this?  I'll look closer.

For clarification:  Does it try to run objdump, or arm-none-elf-objdump?
The former won't be available even if $prefix/$target/bin leads the PATH.


> I cannot reproduce these failures on arm-none-elf using recent 
> binutils.  Phil, what binutils are you using?  If you have a recent 
> binutils, and it is in your PATH, I'd like to understand where these 
> failures are coming from.

CVS binutils, HEAD, updated weekly (on Sunday mornings).

The failures here are all coming from the compiler; none of them ever
make it to running a binutils tool.  Where does binutils come into play
for the visibility tests, and which tool?


Phil