Re: SAF-* comment at the end of the line

[Julien Grall <julien@xxxxxxx>]

Hi,

On 06/12/2023 02:42, Stefano Stabellini wrote:
> On Tue, 5 Dec 2023, Luca Fancellu wrote:
> > Hi all,
> >
> > I’m writing this mail to collect thoughts about the need to improve the SAF-* 
> > comments.
> >
> > I think we reached a point where we need to use deviations for some violation 
> > that we want
> > to keep in the code with a proper justification and an issue was raised when 
> > the comment
> > cannot be put on a line on its own.
> >
> > e.g.
> >
> > If ( condition-1 &&
> >       condition-2 &&
> >        [...] )
> > {
> >      ...
> > }
> >
> > For example in the code above, if the violation is in the second condition, 
> > breaking the conditions
> > to have an empty line between them for the SAF-* comment is not ideal, so we 
> > could maybe
> > improve the in-code comment to be used at the end of the line:
> >
> > e.g.
> >
> > If ( condition-1 &&
> >       condition-2 && /* SAF-*-safe [...] */
> >        [...] )
> > {
> >      ...
> > }
> >
> > This might require also a deviation on the coding style to allow the comment to 
> > overcome the line length.
> >
> > Bertrand, from his experience with safety certifications, feels that adding 
> > this feature could be enough
> > to cover the majority of the cases where we need to deviate a violation in the 
> > code.
> >
> > Using it consistently in the code base as the only way to deviate a violation 
> > can also help the adoption
> > of the project to people who might want to fix them instead of deviating them, 
> > the only thing they would need
> > to do is to grep for SAF-* to have a rough idea of how many justified violation 
> > are in the code.
> >
> > Please let me know your thoughts before I start to implement the feature.
> I think we need this feature and in fact we have already been adding it
> in an ad-hoc way with /* octal-ok */
>
> It would like to remove octal-ok and use a generic way (SAF) to do the
> same.
+1.

Cheers,

--
Julien Grall