[RFC] .fill does not accept forward labels

[H.J. Lu]

On 10/9/17, Andreas Krebbel <krebbel@linux.vnet.ibm.com> wrote:
> On 10/09/2017 11:42 AM, Maciej W. Rozycki wrote:
>> On Mon, 9 Oct 2017, Andreas Krebbel wrote:
>>
>>> diff --git a/gas/testsuite/gas/all/fill-1.s
>>> b/gas/testsuite/gas/all/fill-1.s
>>> new file mode 100644
>>> index 0000000..92d495a
>>> --- /dev/null
>>> +++ b/gas/testsuite/gas/all/fill-1.s
>>> @@ -0,0 +1,5 @@
>>> +	.text
>>> +	.fill (2f-1f), 1, 0x90
>>> +1:
>>> +        nop
>>> +2:
>>
>>  Have you verified your new test case? -- I think not all targets have a
>> `nop' instruction mnemonic.  I reckon this is already handled in other
>> tests that need to emit a piece of code across all targets, although ISTM
>>
>> here you could have just used a portable data directive instead, such as
>> `.word', as it doesn't really matter for the objective of this test what
>> is there between the two labels.
>>
>>   Maciej
>>
> Good point. I've replaced it with .word as you suggested. Thanks!
>

Please use "objdump" to verify that contents are filled properly.

Thanks.

-- 
H.J.