[Spice-devel] [PATCH spice-gtk] NEWS: prepare v0.36 release

[Frediano Ziglio]

> 
> Hi
> 
> On Sat, Jan 12, 2019 at 1:41 PM Frediano Ziglio <fziglio at redhat.com> wrote:
> >
> > >
> > > From: Marc-André Lureau <marcandre.lureau at redhat.com>
> > >
> > > Signed-off-by: Marc-André Lureau <marcandre.lureau at redhat.com>
> > > ---
> > >  NEWS | 26 ++++++++++++++++++++++++++
> > >  1 file changed, 26 insertions(+)
> > >
> > > This is a preliminary patch for the v0.36 release notes, which should
> > > happen after the spice-gtk patch queue on the ML is flushed.
> > >
> > > diff --git a/NEWS b/NEWS
> > > index 3f54f77..83ace29 100644
> > > --- a/NEWS
> > > +++ b/NEWS
> > > @@ -1,3 +1,29 @@
> > > +v0.36
> > > +=====
> > > +
> > > +- Add meson build: autotools will be removed in a future release
> > > +- Deprecate PulseAudio backend: it will be removed in a future
> > > +  release, please use GStreamer instead and report issues
> > > +- Add h265 stream support (requires GStreamer support)
> >
> > small, H265 (capital)
> 
> ok
> 
> >
> > > +- Add SpiceQmpPort helper to interact with QEMU monitor over a Spice
> > > port
> > > +- Display a message if EGL support is required (with dmabuf local
> > > rendering)
> > > +- Many GstVideoOverlay improvements
> > > +- Smooth-scrolling improvements
> > > +- Fix accumulating output frames with GStreamer
> >
> > I think you mean Snir patches. They are targeting a specific case.
> 
> commit 8835e757922cb0342b6236997f5ef2ab61bac1a9
> Author: Frediano Ziglio <fziglio at redhat.com>
> Date:   Fri Apr 13 21:14:56 2018 +0100
> 
>     channel-display-gst: Prevent accumulating output queue
> 
> 
> Do you have a better wording to suggest?
> 

Oh, completely forgot about this. Yes, fixes the output part of
the issue. Maybe

"Fix accumulating decoded frames with GStreamer" ?

is not clear what "output" means in the general case, the patch
was more specific about a "queue" and the code, in a general NEWS
file it became not much clear.

> > > +- Fix reconnection handling
> > > +- Fix memory leaks due to bad refcount of images in the cache
> >
> > I would say "Fix small memory leaks" ...
> 
> ok, I am not sure how bad it is tbh, probably minor since it was
> unnoticed. I guess REPLACE_ME is not being used that often by the
> driver/server.
> 
> >
> > > +- Fix build for newer LibreSSL
> > > +- win32: fix crash with libusb 1.0.22 and UsbDk is not installed
> >
> > is -> if ??
> 
> no,  commit 3e23fff91ea15421a404fb8fc0cb5d86bfbeaa95
> 
> Do you have a better wording?
> 

No, but is not clear what's the subject. Maybe:
"Fix a crash on Windows with libusb 1.0.22 when UsbDk is not installed"

> > It seems a copy&paste from a commit title, also all other
> > points start with capital.
> >
> > > +- win32: do not try to connect to org.gnome.SessionManager
> >
> > Isn't this too technical? Also it affected Mac.
> > Something like "Do not attempt Gnome desktop integration on Windows and
> > Mac" ?
> 
> I can drop it from the NEWS, it can be considered quite minor indeed.
> 

As you like, rewording is fine for me.

> > > +- Require Gtk >= 3.22
> > > +- Require GStreamer >= 1.0
> > > +- Require spice-protocol >= 1.12.15
> > > +- Require usbredir >= 0.5
> > > +- Require libusb >= 1.0.16
> > > +- Require libcacard >= 2.5.1
> > > +- Require lz4 >= 1.7.3
> >
> > I don't think this was changed for this release, the commit
> > in spice-common is from November 2016, I cannot find nothing
> > in spice-gtk.
> 
> commit f494335bc34773f27bd9c519b8b383cc011e7528
> 
> We discussed with Christophe whether it make sense to support the old
> version scheme (~129 etc), and decided it was ok to depend on 1.7.3
> instead based on repology
> 

Missed that. Maybe it would be good to update spice-common m4 macro
accordingly or do we want to keep supporting old lz4 versions
for spice-server ?

Frediano