Re: [Spice-devel] [spice-gtk v1 06/11] channel-display: add spice_frame_new() helper

On Tue, Apr 17, 2018 at 07:10:43AM -0400, Frediano Ziglio wrote: > > > From: Victor Toso > > > > As it makes easier to track what is allocated in its own function and > > ultimately what needs to be freed later on. > > > > Signed-off-by: Victor Toso

Re: [Spice-devel] [spice-gtk v1 10/11] channel-display-gst: summarize number of frames dropped

Hi, On Tue, Apr 17, 2018 at 07:38:22AM -0400, Frediano Ziglio wrote: > > > > From: Victor Toso > > > > For example, this has produced 9 lines of debug below instead of 31. > > > > GSpice-DEBUG: channel-display-gst.c:247 the GStreamer pipeline dropped 4 > > frames > >

Re: [Spice-devel] [PATCH spice-gtk] spice-channel: Use atomic operations for SpiceMsgIn::refcount

Hi, On Tue, Apr 17, 2018 at 09:45:10AM -0400, Frediano Ziglio wrote: > > > > On Tue, Apr 17, 2018 at 01:32:53PM +0100, Frediano Ziglio wrote: > > > This structure is potentially used in multiple thread. > > > Currently in Gstreamer thread using streaming data and coroutine > > > thread. > > > >

Re: [Spice-devel] [spice-gtk v1 05/11] channel-display: add spice_frame_free() helper

Hi, Many thanks for the reviews. On Tue, Apr 17, 2018 at 07:08:13AM -0400, Frediano Ziglio wrote: > > > > From: Victor Toso > > > > The SpiceFrame is created in channel-display.c but it is currently > > freed at each decoders' end. A helper function can reduce some code >

Re: [Spice-devel] [PATCH spice-gtk] build: Generate correct version when spice-gtk is a submodule

Hey, This script comes from https://git.savannah.gnu.org/gitweb/?p=gnulib.git;a=blob;f=build-aux/git-version-gen;h=6d073fcaddd827a396af4c52f1bf00bdd84a9f66;hb=HEAD where this issue might already be fixed, the line that you changed seems to be replaced by: elif test "`git log -1 --pretty=format:x

Re: [Spice-devel] [PATCH 1/2] Ensure that plugins cannot bypass version check

On Tue, Apr 17, 2018 at 06:39:05PM +0200, Christophe de Dinechin wrote: > When does this kind of scenario happen? Imagine we added support for > loading/unloading plugins. Version 13 adds a new interface to unload > plugins, and a new “unloadable plugin” entry point. Therefore, > starting with

[Spice-devel] [PATCH spice-gtk] build: Generate correct version when spice-gtk is a submodule

From: Christophe de Dinechin When spice-gtk is a submodule, .git is not a directory but a file. Testing for a file avoids an "UNKNOWN" version Signed-off-by: Christophe de Dinechin --- build-aux/git-version-gen | 4 ++-- 1 file changed, 2

Re: [Spice-devel] [spice-gtk v1 09/11] channel-display: don't debug latency for each frame

On Tue, Apr 17, 2018 at 07:46:42AM -0400, Frediano Ziglio wrote: > > > > From: Victor Toso > > > > Becomes quite hard to find meaning on something that is printed every > > time. Only print latency value if it is a new min/max or if average > > latency is 10% bigger/lower

[Spice-devel] [PATCH spice-streaming-agent 3/3] Add a macro to deal with the boilerplate of writing a streaming agent plugin

From: Christophe de Dinechin Signed-off-by: Christophe de Dinechin --- include/spice-streaming-agent/plugin.hpp | 8 1 file changed, 8 insertions(+) diff --git a/include/spice-streaming-agent/plugin.hpp

[Spice-devel] [PATCH spice-streaming-agent 0/3] Split plugin version check patch

This patchset just split previous series to be able to discuss separate changes. Christophe de Dinechin (3): Ensure that plugins cannot bypass version check Change numbering schema Add a macro to deal with the boilerplate of writing a streaming agent plugin

[Spice-devel] [PATCH spice-streaming-agent 1/3] Ensure that plugins cannot bypass version check

From: Christophe de Dinechin This change addresses two issues related to plugin version checking: 1. It is possible for plugins to bypass version checking or do it wrong (as a matter of fact, the mjpeg fallback sets a bad example) 2. The current plugin version check

[Spice-devel] [PATCH spice-streaming-agent 2/3] Change numbering schema

From: Christophe de Dinechin A major.minor numbering scheme is not ideal for ABI checks. In particular, it makes it difficult to react to an incompatibility that was detected post release. [More info] The major.minor numbering scheme initially selected makes it harder to

Re: [Spice-devel] [PATCH spice-server 2/3] red-stream: Implements flush using TCP_CORK

> > On 04/17/2018 08:50 PM, Frediano Ziglio wrote: > >> > >> On 04/13/2018 03:25 PM, Frediano Ziglio wrote: > >>> Cork is a system interface implemented by Linux and some *BSD systems to > >>> tell the system that other data are expected to be written to a socket. > >>> This allows the system to

Re: [Spice-devel] [PATCH 1/2] Ensure that plugins cannot bypass version check

> On 17 Apr 2018, at 20:11, Frediano Ziglio wrote: > >>> >>> On 17 Apr 2018, at 16:12, Frediano Ziglio wrote: >>> From: Christophe de Dinechin This change addresses three issues related to plugin version

Re: [Spice-devel] [PATCH 1/2] Ensure that plugins cannot bypass version check

On Tue, 2018-04-17 at 10:12 -0400, Frediano Ziglio wrote: > > > > From: Christophe de Dinechin > > > > This change addresses three issues related to plugin version checking: > > > > 1. It is possible for plugins to bypass version checking or do it wrong > >(as a matter

Re: [Spice-devel] [PATCH spice-server 2/3] red-stream: Implements flush using TCP_CORK

On 04/17/2018 08:50 PM, Frediano Ziglio wrote: On 04/13/2018 03:25 PM, Frediano Ziglio wrote: Cork is a system interface implemented by Linux and some *BSD systems to tell the system that other data are expected to be written to a socket. This allows the system to reduce network fragmentation

[Spice-devel] [spice-gtk v1] channel-display: always name GHashFunc and GEqualFunc

From: Victor Toso g_hash_table_new() allows to pass NULL for GHashFunc and GEqualFunc which defaults to g_direct_hash() and g_direct_equal(). This is the only place in our code that uses NULL while in channel-main and channel-smartcard and spice-file-transfer-task we name