On 12/04/2010 02:06 AM, Richard Earnshaw wrote:
> 
> On Fri, 2010-12-03 at 18:05 +0000, Joseph S. Myers wrote:
>> On Fri, 3 Dec 2010, Richard Earnshaw wrote:
>>
>>> It's probably better to make this a specific test for strongarm* and
>>> arm* rather than a very wide-ranging glob.
>>
>> We eliminated support for strongarm*-*-*, ep9312*-*-*, xscale*-*-* 
>> canonical target triplets in GCC 4.3, advising users to use arm*-*-* 
>> instead.  Perhaps config.sub should be changed to make those triplets 
>> noncanonical (mapping them to arm*-*-* canonical forms) so that no 
>> software using config.sub needs to deal with them?
>>
> 
> makes sense to me.
> 

Most of these are "dwarf2 is supported" test, instead of "object format
is ELF" test.  These common code is extracted to
lib/dwarf.exp:dwarf2_support, and test cases are refactored accordingly.
 We can add dwarf3_support and dwrf4_support easily in the future.

In this new patch, xscale-*-* is removed from two test cases.

Only two cases (py-section-script.exp and dup-sect.exp) are about
"object format is ELF" test, so I don't create a common function in
gdb/testsuite/lib for this purpose.

OK for mainline and 7.2 branch?

-- 
Yao (齐尧)