[PATCH] gas: bpf: new test for MOV with C-like numbers ll suffix

[Jan Beulich]

On 31.10.2023 11:12, Jose E. Marchesi wrote:
> 
>> On 30.10.2023 18:34, Jose E. Marchesi wrote:
>>>
>>>> On 30.10.2023 16:02, Jose E. Marchesi wrote:
>>>>> --- a/gas/testsuite/gas/bpf/alu-pseudoc.d
>>>>> +++ b/gas/testsuite/gas/bpf/alu-pseudoc.d
>>>>> @@ -69,3 +69,4 @@ Disassembly of section .text:
>>>>>   1d0:	d7 01 00 00 10 00 00 00 	r1 = bswap16 r1
>>>>>   1d8:	d7 02 00 00 20 00 00 00 	r2 = bswap32 r2
>>>>>   1e0:	d7 03 00 00 40 00 00 00 	r3 = bswap64 r3
>>>>> + 1e8:	b7 02 00 00 9a 02 00 00 	r2=0x29a
>>>>> \ No newline at end of file
>>>>
>>>> Just as a remark - issues like this would be really nice to be taken
>>>> care of as files are touched. Imo much better than waiting for a
>>>> separate cleanup patch.
>>>
>>> Your point being?
>>
>> "No newline at end of file" is something that shouldn't happen in any
>> source files. Therefore when changing adjacent code it would be nice
>> if the missing newline was inserted at the same time.
> 
> Yes, agreed.  In all candour, in this case I didn't even notice it
> before pushing (was in a rush) or i would have done it.
> 
> Would you say the same applies to typos and other trivial fixes?
> Usually I refrain from fixing these that I happen to notice while
> working on something else, unless they are very coupled with the change
> (like a typo in a comment describing the code I am changing) even if it
> is very tempting to do otherwise.  I guess it is a question of degree.

My rule of thumb is that anything that's touched anyway can also have
cosmetic corrections alongside the main intended adjustment(s).

Jan