Re: Software evolution around scripts for the semantic patch langugae

[Cornelia Huck <cornelia.huck@de.ibm.com>]

On Thu, 25 Aug 2016 14:23:35 -0400 (EDT)
Julia Lawall <julia.lawall@lip6.fr> wrote:

> On Thu, 25 Aug 2016, SF Markus Elfring wrote:
> 
> > > Or some close variant.
> >
> > I have got more script variants evolving in my software collection.
> >
> > There are further approaches available from various contributors,
> > aren't there?
> 
> What she is asking for is a concise and precise decription of what you
> have done.  If you have some other variants, eg controlling where the
> sizeof argument is (left or right of *), you don't necessarily have to
> include it in the patch, if such a rule was not used for the specific
> patch anyway.

*nod*

If I see a patch that says "I've run the following cocchinelle patch to
perform $TRANSFORMATION, and here's the result", I can be reasonably
sure that the result will be what is intended to be changed in the
first place (and I can assess whether the change makes sense at all.)
If I see only the resulting patch, I won't know whether you have
performed the changes manually (and possibly introduced bugs, as
happens to all of us.)

Moreover, a good semantic patch is useful to others as well and might
even be reused in other contexts that have similar requirements. You
really lose value if you don't publish them.