[PATCH v0 07/15] gas tests: add new tests for build attributes

[Richard Earnshaw]

On 11/03/2025 12:06, Jan Beulich wrote:
> On 11.03.2025 12:16, Matthieu Longo wrote:
>> On 2025-03-11 08:13, Jan Beulich wrote:
>>> On 10.03.2025 18:51, Matthieu Longo wrote:
>>>> --- /dev/null
>>>> +++ b/gas/testsuite/gas/aarch64/build-attributes/ba-failures-1.d
>>>> @@ -0,0 +1,14 @@
>>>> +# name: EABI build attributes: error cases
>>>> +# source: ba-failures-1.s
>>>> +# as:
>>>> +#error: \A[^\n]*\.s: Assembler messages:\n
>>>> +#error: [^\n]*\.s:[0-9]+: Error: A subsection has to be declared before declaring attributes\n
>>>> +#error: [^\n]*\.s:[0-9]+: Error: Attribute 1 cannot be redefined\n
>>>> +#error: [^\n]*\.s:[0-9]+: Error: Attribute 4 cannot be redefined\n
>>>> +#error: [^\n]*\.s:[0-9]+: Error: Invalid token for section name: \.\n
>>>> +#error: [^\n]*\.s:[0-9]+: Error: Expected <subsection_name>, <optional>, <format>\n
>>>> +#error: [^\n]*\.s:[0-9]+: Error: Recalled subsections must have the same parameters\n
>>>> +#error: [^\n]*\.s:[0-9]+: Error: Recalled subsections must have the same parameters\n
>>>> +#error: [^\n]*\.s:[0-9]+: Error: Recalled subsections must have the same parameters\n
>>>> +#error: [^\n]*\.s:[0-9]+: Error: Expected subsection values for <optional> are 0 or 1\n
>>>> +#error: [^\n]*\.s:[0-9]+: Error: Expected subsection values for <format> are 0 or 1
>>>> \ No newline at end of file
>>>
>>> Please can this be taken care of (throughout the series)?
>>
>> What do you mean by "being taken care of (throughout the series)" ?
> 
> Hmm, odd - I don't see what's unclear about my request, taking the
> context into account. I'm asking that - throughout the series - you
> make sure you don't introduce any file lacking a trailing newline.
> 
> Jan

I must admit that I didn't immediately grasp what you were asking for 
either.  If you'd trimmed the other lines from the test, so that it read

--- /dev/null
+++ b/gas/testsuite/gas/aarch64/build-attributes/ba-failures-1.d
@@ -0,0 +1,14 @@<diff header>
[...]
\ No newline at end of file

Then the comment would have been unambiguous.  As it was, you could be 
commenting on almost anything in that hunk.

R