Page 1 of 1

Alt-Tab bug in 0.70b3 is still here

Posted: 22 Mar 2006, 17:28
by ginekolog
Just so u know. I am also willing to help my best at repairing this.

I know good deal about windows, networking and some deal about java, c++ etc. Contact me and i can help debuging it.

Posted: 22 Mar 2006, 17:43
by LathanStanley
I can alt-tab just fine.. :roll:

Posted: 22 Mar 2006, 17:52
by ginekolog
i know some can, but many cant. We could back in 67b , but not since .70

Posted: 22 Mar 2006, 18:16
by mother
Yeah, and no easily detectable rhyme or reason why some can/can't.

The behavior is currently improved for me- Spring no longer seems to closed immediately when I alt-tab out.

It does however close after being out for a few seconds.

Re: Alt-Tab bug in 0.70b3 is still here

Posted: 22 Mar 2006, 22:29
by Tobi
ginekolog wrote:Just so u know. I am also willing to help my best at repairing this.

I know good deal about windows, networking and some deal about java, c++ etc. Contact me and i can help debuging it.
It would definitely help if you could figure out what goes wrong exactly. Anything which can point us in some direction can help. E.g. does it ever happen when not hosting, or only on hosting?

I myself am suspecting somehow the server thread stops or the ticks reported by SDL_GetTicks() isn't increased (fast enough) when switched out.

A fix could maybe be as easy as handling some switch in / switch out events properly. Heck, I might even try to read some docs about that myself soonish.

Posted: 23 Mar 2006, 09:43
by Zenka
The alt-tab bug still in it for me. Though things did improve, It can stay minimised for about 5 sec. And I think I never saw it crash when watching a replay. Can't say if it happens if I host yes or no, since I'm to carefull with alt-tabbing then.

Posted: 23 Mar 2006, 10:06
by ginekolog
My first research showed that when i host there are no problems with tabbing out. If i play single player no problems also.

Problem is ONLY when i am client. I think tobi is on correct track. I also suspect that spring.exe stops sending "im a alive" packets to host(server) - so it just closes as it suspects that server is down.

Heck, i will fire on traffic sniffer today to see if that is the problem.

Posted: 23 Mar 2006, 19:52
by ginekolog
I did some more research, see inluded traffic log.

This bug only happens when i am not host. In this case i was tabed out for 3 seconds before spring closed. In log eveerything is quite normal, client just stop sendig data to host. It can take up to 10 seconds before this happens, somtimes it can even work fine (rarely)

I have winxp2sp, very low ping internet and no programs runing in test.



Host has ip 1.1.1.1, my ip as client is 66.66.66.66.

http://home.amis.net/radtest/taspring/temp.txt

Check log, see if it helps tobi.

Posted: 03 Apr 2006, 12:01
by Tobi
I don't see anything abnormal, but it (the log) doesn't contain really much information either. Still no clue abotu the bug (except maybe handling of activation events, but I still didn't test anything concerning that -- been too busy with other things...)