[Spice-devel] [PATCH v2 00/13] red_channel introducing refactoring, part 2

[Alon Levy]

On Fri, Feb 11, 2011 at 01:07:42PM +0100, Marc-André Lureau wrote:
> Hi
> 
> So ACK for everything except 11/13, this rename:
> 
>     s/hold_pipe_item_proc/channel_pipe_item_hold_proc/
> 
> It gives this inconsistency:
> 
> -typedef void (*channel_hold_pipe_item_proc)(PipeItem *item);
> +typedef void (*channel_pipe_item_hold_proc)(PipeItem *item);
>  typedef void (*channel_release_pipe_item_proc)(RedChannel *channel,
>                                                 PipeItem *item, int
> item_pushed);
> 
> Even if channel is not used by channel_pipe_item_hold_proc, it could,
> I mean, it's a callback with different possible implementations, I
> think it would be better wihtout rename:
> 
> typedef void (*channel_hold_pipe_item_proc)(RedChannel *channel,
> PipeItem *item);
> typedef void (*channel_release_pipe_item_proc)(RedChannel *channel,
>                                                 PipeItem *item, int
> item_pushed);
> 

Okay.

> On Thu, Feb 10, 2011 at 5:09 PM, Alon Levy <alevy at redhat.com> wrote:
> > v2: (Marc-Andre review)
> >  * squashed renames
> >  * squashed CommonChannel introduction and free cb addition
> >  * small fix from gcc 4.6.0
> >  * not addressed:
> >  * left CONTAINEROF for now
> >
> > This is a part of a much larger series so I don't have a good concise
> > description. Ultimately this all leads to a reuse of RedChannel across
> > most of the channels (sound remains an odd out because of not using a
> > pipe abstraction, but a bitmap of outstanding packets, you can have 0
> > or one of each type waiting to be sent, this makes sense for audio.
> > Probably for video too actually).
> >
> > Anyway this patchset touches almost only red_worker (and it's helper
> > cache files, split out because they are used for a number of different
> > caches via macros before inclusion, that usual trick). It is mostly about
> > renaming, making a CommonChannel take the place of the internal RedChannel
> > to allow future use of the external RedChannel defined in red_channel.h
> >
> > Alon Levy (13):
> >  server/red_worker: change hold_item sig, drop the void*
> >  server/red_worker: use ack_data struct
> >  server/red_worker: introduce CommonChannel
> >  server/red_worker: shorten some lines using alias variables
> >  server/red_worker: use red_channel pipe add versions
> >  server/red_worker: extract common_release_pipe_item from
> >    red_pipe_clear
> >  server/red_worker: split display_channel_send_item
> >  server/red_worker: add red_channel_init_send_data
> >  server/red_worker: use red_channel begin_send_message
> >  server/red_worker: split cursor_channel_send_item
> >  server/red_worker: renames to add channel_ prefix
> >  server/red_worker: introduce an outgoing struct around
> >    out_bytes_counter
> >  server/red_worker: match channel_release_pipe_item_proc to
> >    red_channel
> >
> >  server/red_channel.c             |    4 +-
> >  server/red_channel.h             |    8 +-
> >  server/red_client_cache.h        |    2 +-
> >  server/red_client_shared_cache.h |   20 +-
> >  server/red_worker.c              | 1022 ++++++++++++++++++++------------------
> >  5 files changed, 561 insertions(+), 495 deletions(-)
> >
> > --
> > 1.7.4
> >
> > _______________________________________________
> > Spice-devel mailing list
> > Spice-devel at lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/spice-devel
> >
> 
> 
> 
> -- 
> Marc-André Lureau