vnc2swf-users

[Vnc2swf-users] Bug in vnclog recordings

[<lfk...@ya...>]

Hi,=0A=0AI found the reason for the timing problem in vnclog recordings.=0A=
When recording using type shape or video, every framebuffer update (screen =
update) creates the appropriate number of frames using the time it finished=
 processing the update (finish_update function of RFBProxy class) as the ti=
mestamp of the update.=0ABut when recording with type vnc, the timestamp us=
ed is the time of the request for update (request_update function of RFBNet=
workClientForRecording class).  But that's not a good timestamp, since we d=
on't know how long the server will take to answer the request.=0AActually, =
both are taking the wrong timestamp.=0AIn my opinion the correct timestamp =
should be the moment the client starts receiving the framebuffer update bec=
ause that's the closest to the moment that the server actually checked the =
screen.=0AThe problem is most noticeable when there are none or just a few =
interactions (keyboard and mouse movements). When the mouse moves alot the =
server is constantly sending an update so the difference in timestamp is no=
t noticeable.=0AAnother problem lies in function update_screen of RFBMovieC=
onverter class (rfb.py).  Given a timestamp, it should first create empty f=
rames to catchup the elapsed time and only after that it should insert a fr=
ame with the framebuffer update. Currently it's in the other way around.  T=
his is used in vnclog playback.   SWF recordings use a different class whic=
h is currently working correctly.=0A=0AYusuki, can you check this out for u=
s?=0A=0AThank you.=0A=0ALuis Fernando=0A=0A----- Mensagem original ----=0AD=
e: Tom Panning <lu...@gm...>=0APara: General discussion about vnc2swf=
 <vnc...@li...>=0AEnviadas: Quinta-feira, 22 de Mar=
=E7o de 2007 10:10:18=0AAssunto: Re: [Vnc2swf-users] Res: VNCLog playback i=
s nonlinear=0A=0A=0A=0AOn 3/22/07, lfk...@ya... <lfkpoa-69@yahoo.=
com.br> wrote:=0A=0ATom,=0A=0AI'll try to check why the vnclog playback is =
not working properly, but you'll have to wait until next week. =0A=0A=0A=0A=
=0AThanks, I appreciate it.=0A=0A=0A=0ATom=0A=0A=0A=0A=0A=0A---------------=
----------------------------------------------------------=0ATake Surveys. =
Earn Cash. Influence the Future of IT=0AJoin SourceForge.net's Techsay pane=
l and you'll get the chance to share your=0Aopinions on IT & business topic=
s through brief surveys-and earn cash=0Ahttp://www.techsay.com/default.php?=
page=3Djoin.php&p=3Dsourceforge&CID=3DDEVDEV=0A____________________________=
___________________=0AVnc2swf-users mailing list=0AV...@li...ur=
ceforge.net=0Ahttps://lists.sourceforge.net/lists/listinfo/vnc2swf-users=0A=
=0A=0A=0A=0A=0A=0A__________________________________________________=0AFale=
 com seus amigos  de gra=E7a com o novo Yahoo! Messenger =0Ahttp://br.messe=
nger.yahoo.com/ 

Re: [Vnc2swf-users] Bug in vnclog recordings

[Yusuke S. <yu...@cs...>]

Hi Luis,

On Thu, 29 Mar 2007 06:44:55 -0700 (PDT), lfk...@ya... wrote:
> In my opinion the correct timestamp should be the moment the
> client starts receiving the framebuffer update because that's
> the closest to the moment that the server actually checked the
> screen.  The problem is most noticeable when there are none or
> just a few interactions (keyboard and mouse movements). When the
> mouse moves alot the server is constantly sending an update so
> the difference in timestamp is not noticeable.  Another problem
> lies in function update_screen of RFBMovieConverter class
> (rfb.py).  Given a timestamp, it should first create empty
> frames to catchup the elapsed time and only after that it should
> insert a frame with the framebuffer update. Currently it's in
> the other way around.  This is used in vnclog playback.  SWF
> recordings use a different class which is currently working
> correctly.

Hmm... you seem to be right. I haven't thought much about
this timing issue, but experienced weird phenomena 
(screen updates are sometime too off timing) a couple of times.
I'll take a look into this.

Sorry for the late response.

Yusuke

Re: [Vnc2swf-users] Bug in vnclog recordings

[Tom P. <lu...@gm...>]

On 3/30/07, Yusuke Shinyama <yu...@cs...> wrote:
>
> Hi Luis,
>
> On Thu, 29 Mar 2007 06:44:55 -0700 (PDT), lfk...@ya... wrote:
> > In my opinion the correct timestamp should be the moment the
> > client starts receiving the framebuffer update because that's
> > the closest to the moment that the server actually checked the
> > screen.  The problem is most noticeable when there are none or
> > just a few interactions (keyboard and mouse movements). When the
> > mouse moves alot the server is constantly sending an update so
> > the difference in timestamp is not noticeable.  Another problem
> > lies in function update_screen of RFBMovieConverter class
> > (rfb.py).  Given a timestamp, it should first create empty
> > frames to catchup the elapsed time and only after that it should
> > insert a frame with the framebuffer update. Currently it's in
> > the other way around.  This is used in vnclog playback.  SWF
> > recordings use a different class which is currently working
> > correctly.
>
> Hmm... you seem to be right. I haven't thought much about
> this timing issue, but experienced weird phenomena
> (screen updates are sometime too off timing) a couple of times.
> I'll take a look into this.


Yusuke,
Have you had a chance to look at this problem? Is there anything that I can
do to help?

Thanks,
Tom