Re: Some code changes wanted in Rygel

[Kjartan Maraas <kmaraas broadpark no>]

ma., 27.09.2010 kl. 17.29 +0200, skrev Vincent Untz:
> Le lundi 27 septembre 2010, à 16:27 +0300, Zeeshan Ali (Khattak) a écrit :
> > Hi,
> > 
> > On Mon, Sep 27, 2010 at 2:30 PM, Vincent Untz <vuntz gnome org> wrote:
> > > What are
> > >> the risks involved in approving this patch?
> > 
> >   None that I can imagine for the changes in 8c0391d..14071c0.
> > 
> > >  - there's at least one unrelated fix ("media-export: Add null checks to
> > >   standard XDG folders")
> > 
> >   There was a bug about this on debian bugzilla. Now that you mention
> > it, this shouldn't matter to a normal GNOME desktop but rygel is used
> > by people on headless machines where this issue will realize.
> 
> I'm sorry if this sounds like a lot of paperwork, but this is another
> issue, so it's better for us to track this in a separate thread -- it's
> already a bit confusing right now, so adding something else on top of
> that won't help. And there is a question to start this thread: what does
> it do with NULL? Does it crash?
> 
> > > My understanding is that the first three commits go together. How
> > > critical are those changes?
> > 
> >   The first one (206f019) fixes a leak introduced in previous release.
> > 8c0391d..14071c0 are the fixes for IOP.
> 
> Ah, I hadn't see this one due to the way git log works (see below).
> 
> I'm confused: the commit says "Don't set description if already set",
> but you always do a set_content(). I don't know if this is what you
> wanted?
> 
> (Else, I guess the old code was adding a second description child to the
> XML; that part is okay)
> 
> > > How many devices will fail to work with
> > > rygel?
> > 
> >   AFAICT 3 different commercial products, however I can only be 100%
> > certain about one of them since the other two bug reporters haven't
> > managed to verify the bugs yet.
> > 
> >   So I modify my request to only include the following commits:
> > 
> > 206f019..f6bfed4, 42a0f7b and f0b0127.
> 
> So, for people using git log, you want to do git log on
> 206f019^..f6bfed4, I guess.
> 
> For 206f019, see above.
> 
> For the rest of 206f019..f6bfed4, I think I'm okay with it, but the
> renames make it hard to see if there was any change in the xml files.
> Even with git diff -M :/ So I'll give a first approval.
> 
2/2

Cheers
Kjartan