[PATCH,V2 2/4] gas: sframe: reword diagnostic to address ambiguity

Jan Beulich jbeulich@suse.com
Sun May 18 08:09:51 GMT 2025
On 16.05.2025 17:43, Indu Bhagat wrote:
> On 5/16/25 4:36 AM, Jan Beulich wrote:
>>> --- a/gas/testsuite/gas/cfi-sframe/cfi-sframe-aarch64-4.d
>>> +++ b/gas/testsuite/gas/cfi-sframe/cfi-sframe-aarch64-4.d
>>> @@ -1,7 +1,7 @@
>>>   #as: --gsframe
>>>   #objdump: --sframe=.sframe
>>>   #name: SFrame cfi_negate_ra_state_with_pc test
>>> -#warning: Warning: skipping SFrame FDE; \.cfi_negate_ra_state_with_pc
>>> +#warning: Warning: no SFrame FDE emitted; \.cfi_negate_ra_state_with_pc
>> The specific wording isn't relevant for this kind of test, so please
>> consider further replacing the part of the text you already replace
>> by .* (the parts after the semicolons probably want to stay, as that's
>> what disambiguates the individual diagnostics).
> 
> I find that .* is unnecessary.  I have removed the "no SFrame FDE 
> emitted" from all warning texts in all testcases now.  Simply the 
> substring which distinguishes the individual diagnostics.  So for the 
> above test, its like:

If this works - even better. My prior okay remains in place then.

Jan

> --- a/gas/testsuite/gas/cfi-sframe/cfi-sframe-aarch64-4.d
> +++ b/gas/testsuite/gas/cfi-sframe/cfi-sframe-aarch64-4.d
> @@ -1,7 +1,7 @@
>   #as: --gsframe
>   #objdump: --sframe=.sframe
>   #name: SFrame cfi_negate_ra_state_with_pc test
> -#warning: Warning: skipping SFrame FDE; \.cfi_negate_ra_state_with_pc
> +#warning: \.cfi_negate_ra_state_with_pc
> 
> Similarly,
> 
> --- a/gas/testsuite/gas/cfi-sframe/cfi-sframe-x86_64-empty-1.d
> +++ b/gas/testsuite/gas/cfi-sframe/cfi-sframe-x86_64-empty-1.d
> @@ -1,5 +1,5 @@
>   #as: --gsframe
> -#warning: skipping SFrame FDE; \.cfi_escape DW\_CFA\_expression with SP 
> reg 7
> +#warning: \.cfi_escape DW\_CFA\_expression with SP reg 7
>   #objdump: --sframe=.sframe
>   #name: CFI_escape with register of significance to SFrame
>   #...
> 
> etc..



More information about the Binutils mailing list