Decimal128, printcmds.exp, dfp-exprs.exp

Decimal128, printcmds.exp, dfp-exprs.exp

[Christophe LYON]

Hi all,

On one of our internal targets, I observe some failures in the GDB 
testsuite in printcmds.exp and dfp-exprs.exp tests.
The failures in printcmds.exp occur only when the target is configured 
as big-endian, while endianness has no influence on the failure in 
dfp-exprs.exp.

More precisely,
* in dfp-exprs.exp:
p (_Decimal128) 3.7df
$89 = 2.00000000000000000000000800002422E+257
(gdb) FAIL: gdb.base/dfp-exprs.exp: p (_Decimal128) 3.7df

* in printcmds.exp:
printf "%DDf\n",1.234567890123456789012345678901234dl
1.234567890123454789012345678901234
(gdb) FAIL: gdb.base/printcmds.exp: printf 
"%DDf\n",1.234567890123456789012345678901234dl

My understanding is that none of these tests uses the underlying target, 
so what can make them fail?

Thanks,

Christophe.

Re: Decimal128, printcmds.exp, dfp-exprs.exp

[Joseph S. Myers]

On Fri, 20 Mar 2009, Christophe LYON wrote:

> Hi all,
> 
> On one of our internal targets, I observe some failures in the GDB testsuite
> in printcmds.exp and dfp-exprs.exp tests.
> The failures in printcmds.exp occur only when the target is configured as
> big-endian, while endianness has no influence on the failure in dfp-exprs.exp.

Are you sure you are using current GDB sources?  I fixed such failures 
with:

2008-06-27  Joseph Myers  <joseph@codesourcery.com>

        * dfp.c (decimal_convert): Call match_endianness before and after
        conversion.

If you're using current sources, maybe something has broken them again.

-- 
Joseph S. Myers
joseph@codesourcery.com

Re: Decimal128, printcmds.exp, dfp-exprs.exp

[Christophe LYON]

On 20.03.2009 16:35, Joseph S. Myers wrote:
> On Fri, 20 Mar 2009, Christophe LYON wrote:
> 
>> Hi all,
>>
>> On one of our internal targets, I observe some failures in the GDB testsuite
>> in printcmds.exp and dfp-exprs.exp tests.
>> The failures in printcmds.exp occur only when the target is configured as
>> big-endian, while endianness has no influence on the failure in dfp-exprs.exp.
> 
> Are you sure you are using current GDB sources?  I fixed such failures 
> with:
> 
No, I am based on GDB-6.8.
Thanks for the hint.

I have applied your patch and my problem has disappeared.

Thanks and sorry for the noise.

Christophe.