[libvirt] [PATCH v3 04/35] cfg.mk: single variable declaration per line when using cleanup macro

[Erik Skultety]

On Wed, Jul 11, 2018 at 08:36:59PM +0530, Sukrit Bhatnagar wrote:
> The problem is not that it is initialized to a non-NULL value.
> If we were to detect multiple declarations in a line. we would
> search for a comma (separator), right? In the case I mentioned,
> the comma inside the function has to be avoided by the rule.

Syntax-wise, our macro signatures follow the ones of a function, i.e. you also
have parentheses. If we're ever going to create a syntax-check rule for that
it won't be as simple as matching commas, you'd use more context for such a
regex, for the reasons you've mentioned. Therefore, the example you provided
would never be affected by such a rule.

Erik

> On Wed, 11 Jul 2018 at 15:57, Erik Skultety <eskultet at redhat.com> wrote:
> >
> > On Wed, Jul 11, 2018 at 10:35:25AM +0200, Pavel Hrdina wrote:
> > > On Wed, Jul 11, 2018 at 12:42:43AM +0530, Sukrit Bhatnagar wrote:
> > > > On Tue, 10 Jul 2018 at 16:24, Erik Skultety <eskultet at redhat.com> wrote:
> > > > >
> > > > > On Sat, Jun 30, 2018 at 02:30:08PM +0530, Sukrit Bhatnagar wrote:
> > > > > > Add rule to ensure that each variable declaration made using
> > > > > > a cleanup macro is in its own separate line.
> > > > > >
> > > > > > Sometimes a variable might be initialized from a value returned
> > > > > > by a macro or a function, which may take on more than one
> > > > > > parameter, thereby introducing a comma, which might be mistaken
> > > > > > for multiple declarations in a line. This rule takes care of
> > > > > > that too.
> > > > >
> > > > > I can't think of an example or I'm just not seeing it, can you please give me
> > > > > an example where you actually need the rule below? Because right now I don't
> > > > > see a need for it.
> > > >
> > > > In src/util/virfile.c in virFileAbsPath function:
> > > > ...
> > > > VIR_AUTOFREE(char *) buf = getcwd(NULL, 0);
> > > > ...
> > >
> > > I don't see anything wrong with it, it is properly initialized to some
> > > value, it doesn't have to be only NULL.
> > >
> > > Pavel
> >
> > Agreed,
> >
> > Erik
> >
> >