Re: [PATCH] libnm-glib: add errors to device classes v2
- From: Jirka Klimes <jklimes redhat com>
- To: Dan Williams <dcbw redhat com>
- Cc: NetworkManager mailing list <networkmanager-list gnome org>
- Subject: Re: [PATCH] libnm-glib: add errors to device classes v2
- Date: Fri, 16 Mar 2012 10:55:50 +0100
On Thursday 15 of March 2012 15:25:17 Dan Williams wrote:
> On Thu, 2012-03-15 at 17:06 +0100, Jirka Klimes wrote:
> > On Thursday 15 of March 2012 08:35:07 Dan Williams wrote:
> > > On Wed, 2012-03-14 at 15:24 +0100, Jirka Klimes wrote:
> > > > The patch adds errors to libnm-glib's device classes.
> > > > The errors are returned in
> > > > nm-device_connection_valid(dev,con,error)
> > > > when
> > > > the device and connection don't match.
> > > >
> > > > Currently, the error is not used. But it could be useful for
> > > > nmcli and elsewere.
> > >
> > > It is an ABI break though since we're changing a function signature,
> > > right? We should probably make another function for this and then
> > > have
> > > nm_device_connection_valid() call that. Maybe name the new one
> > > nm_device_connection_compatible() :)
> > >
> > > Dan
> >
> > Oh, you're right that breaks ABI. Here is the respin.
> > Or do you also prefer to rename connection_valid() class functions to
> > connection_compatible() in order to be consistent?
>
> Yeah, we probably should do that to reduce confusion. To be completely
> fair that's an ABI break too, in case somebody is subclassing NMDevice
> outside of libnm-glib, but that seems incredibly unlikely as there's no
> way to get libnm-glib to create your subclass instead of the built-in
> types. Low enough risk of that though, so we should just rename the
> class function.
>
Renamed and pushed.
a11067f5846e3effe2f653b8babdfbabf7d2254d
Thanks.
Jirka
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]