[Freeciv-Dev] Re: (PR#3664) Server doesn't react while send_all_info(&pc
[Top] [All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index] [Thread Index]
On Mon, Mar 10, 2003 at 04:06:12PM -0800, Vasco Alexandre da Silva Costa wrote:
>
> [ue80@xxxxxxxxxxxxxxxxxxxxx - Mon Mar 10 19:21:45 2003]:
>
> > Tried it out with my connection 5 secs for a 200x100 map. (No units)
>
> Cities (especially on players using ICS tactics) and units take a lot of
> space as well.
>
> What is taking 5 secs for you can take a minute for some poor loser in
> an analog modem. I get 4 KBps with luck and a favourable wind in my
> modem. Some people get 2-3 KBps.
(AI on Big maps playing alone ~16000 freights)
And > 1000 cities .. but in that case the gui starts to be to slow ..
> > Can't the buffer send the data without blocking the other players?
>
> If you allow the other players to continue playing they will cause the
> game state to change and trigger even more data sends to the player in
> question. The guy on the slow connection is having enough problems
> without being spammed even more. You have to *stop* the data pileup at
> some point.
Ok we during a reconnect we have things that can be changed and things
which can't. I think a bigger part isn't updating. Why don't we send that
first with a nonblocking state, and the rest after that information is
in the client with blocking?
> > With the current solution some players need > 1 minute to reconnect
> > and the usuall responsetime for an action in normal game is < 1 sec.
>
> Sending data takes time and if the other guy is on a slow connection...
> Since the Freeciv protocol is synchronous either you compress the data
> more (up to a certain limit) or insert waitstates.
> You can't drop data like in a video streaming server where losing some
> frames is of little consequence.
I know that after the delta-patch all sendet data is important, so its
ok that every packet is sent.
> > People don't like to wait + they want to know what happens. And they
> > want to see if they are the only people who have the problem or all.
>
> When flush_packets causes slowdowns *everyone* gets slowed down. So it
> is fair if annoying.
I don't like things that are fair like this. In quakeworld people with
bad pings only lowered there own fun, but not the fun from other
people (ok in 4on4 the fun of their hole team) I think the effect for
other players should be minimized but it also should be possible for
people with crap connections to play.
> > Think at least the chatline should work during the long reconnects
> > from some players.
>
> This is more special casing. I don't like it.
Ok.
> The server could send a message to the players each time flush_packets
> is issued. Do you think this would be useful?
When a connect needs more than 10 seconds people want to have
information. Don't know which is the best place to send that to the
players.
> If you don't want players on slow connections cramping your style just:
> /set netwait 0
>
> Just don't complain if noone is fast enough to re-connect. ;-)
Think that is only a hack, and i think that i'll try it out. But i
really dislike to lose a player in the game because his connection is to
poor and we don't like to wait.
Thomas
--
Thomas Strub *** eMail ue80@xxxxxxxxxxxxxxxxxxxxx
Wenn Du nicht programmieren kannst und Dir für Arbeit zu schade bist:
Werde Berater, Analyst oder organisiere Kongresse.
[Freeciv-Dev] (PR#3664) Server doesn't react while send_all_info(&pconn->self);, Vasco Alexandre da Silva Costa, 2003/03/10
- Message not available
- [Freeciv-Dev] Re: (PR#3664) Server doesn't react while send_all_info(&pconn->self);,
ue80@xxxxxxxxxxxxxxxxxxxxx <=
Message not available
[Freeciv-Dev] (PR#3664) Server doesn't react while send_all_info(&pconn->self);, Vasco Alexandre da Silva Costa, 2003/03/10
[Freeciv-Dev] (PR#3664) Server doesn't react while send_all_info(&pconn->self);, Vasco Alexandre da Silva Costa, 2003/03/10
[Freeciv-Dev] (PR#3664) Server doesn't react while send_all_info(&pconn->self);, Vasco Alexandre da Silva Costa, 2003/03/11
[Freeciv-Dev] (PR#3664) Server doesn't react while send_all_info(&pconn->self);, Vasco Alexandre da Silva Costa, 2003/03/11
|
|