Hi,

> On 31 May 2022, at 15:28, Jan Beulich <[email protected]> wrote:
> 
> On 31.05.2022 16:14, Bertrand Marquis wrote:
>>> On 31 May 2022, at 14:50, Jan Beulich <[email protected]> wrote:
>>> On 31.05.2022 15:30, Bertrand Marquis wrote:
>>>> +# Replace current by goal in the dependency to generate an analysis for 
>>>> all
>>>> +# rules we would like to respect.
>>>> +cppcheck-misra.json cppcheck-misra.txt: $(XEN_ROOT)/docs/misra/rules.rst
>>>> +  $(Q)$(srctree)/tools/convert_misra_doc.py -i $< -o cppcheck-misra.txt \
>>>> +          -j cppcheck-misra.json
>>>> +
>>>> +# Prevent parallel make issues as script is generating both files
>>>> +cppcheck-misra.json: cppcheck-misra.txt
>>> 
>>> With this dependency the earlier rule should not list multiple targets
>>> (and it generally should not, for not being a pattern rule). If I'm not
>>> mistaken the way you have it the Python script would be invoked twice,
>>> and all you prevent is that it is invoked twice in parallel. And then
>>> please use $@ inside the rule. Additionally, with the script being an
>>> in-tree one, the rule should also have a dependency on that script
>>> (such that the targets would be rebuilt if the script alone changes).
>> 
>> I am a bit lost on the $@, previous patch adding cppcheck was changed
>> to use $(Q) instead. So can you justify this request ?
> 
> I'm talking of $@ (a macro expansion), not @ as a prefix to suppress
> command echoing.
> 

That make a lot more sense.

Thanks
Bertrand

> Jan
> 
> 


Reply via email to