Re: gdbus branch review

[Tomas Bzatek <tbzatek redhat com>]

On Tue, 2012-04-24 at 12:03 +0200, Alexander Larsson wrote:
> On Mon, 2012-04-23 at 11:05 -0400, David Zeuthen wrote:
> 
> > I see this too and there are some inconsistencies too, for example
> > DriveStop() vs. lookupMount().
> > 
> > GNOME code generally uses CamelCase (not initialLowerCase) for
> > methods, signals and properties and lower_case for parameter names. At
> > least that's what gdbus-codegen(1) is optimized insofar that it
> > transforms names for use in C structs and C functions.
> > 
> > Of course, if you change lookupMount() to LookupMount() you
> > effectively change (break) the ABI and maybe you don't want that
> > (changing parameter names is not a break though). But then again maybe
> > you do, I mean, IIRC we never promised ABI stability on any of the
> > internals of GVfs. Then again, it's nice to not break ABI if you can
> > avoid it.
> 
> Well, the method call to signal change means we're breaking ABI anyway.
> Might as well get the rest right then.

Yes, trying not to break the ABI was the reason I kept original names. I
did break it on a few places though for code simplicity. But it's not
much a concern since GVfs never had a stable public ABI.

The naming style would surely need some cleanup, even the core daemon
interface is a bit mess sometimes. I'll do that with other changes while
addressing the review issues.

-- 
Tomas Bzatek <tbzatek redhat com>