Posts: 1,049
Threads: 55
Joined: Nov 2010
Well yet again AC servers seem to be having a demo problem. I was in my server doing a match and at the end I simply wanted to get the demo. I received ac_shine just like I wanted yet it played something completely different. From this, I noticed that my other server had players in it and that I had somehow received the demo from that server. I'm not sure how this happens but I do know the circumstances it takes for it.
Ok say you have two servers, one on the default port (28763) and one not on default (28765). If you are playing on the 28765 non default and the default 28763 server is in current use and had games switch in it, then you will get a demo from the 28763 when you ask for a demo from the 28765. Why does this happen? The default 28763 server has to be in use and maybe rotated a few games in it before the 28765 server decides to mess the demos up. Can someone knowledgeable with servers please describe to me why this happens and if it can be fixed? All help appreciated :)
Posts: 140
Threads: 9
Joined: Jun 2010
I have had this same problem many times in 1.1.0.4 I no longer am able to download any demos correctly.
Posts: 3,462
Threads: 72
Joined: Jun 2010
This is the first I have heard of the relation to ports. This may help the Devs debug this. If you can trouble shoot this any more, that would probably be appreciated.
Although, I cannot imagine why there would be a port mixup. AC works in very mysterious ways.
Posts: 2,136
Threads: 50
Joined: Jun 2010
27 Feb 11, 02:56PM
(This post was last modified: 27 Feb 11, 02:56PM by jamz.)
I'm pretty sure it's to do with the way Windows stores demo data temporarily. Individual servers are indiscriminate about how they access 'demorecord' in the demos folder. Linux (and probably Mac) doesn't have this file.
Posts: 1,436
Threads: 7
Joined: Jun 2010
27 Feb 11, 07:51PM
(This post was last modified: 27 Feb 11, 07:56PM by tempest.)
You're absolutely right, Jamz. Someone else already reported that problem, and I came to the same conclusion and posted a possible way to fix it. Of course, nothing happened. Just don't post bug reports, it's useless. Not like the devs would fix a bug, y'know.
Btw: Linux and Mac use tmpfile(), but of course Windows doesn't have that, because why should they abide by standards?
See here for another example. Crash bug, two days old, exactly described, fixable through a single line of code.
Posts: 533
Threads: 33
Joined: Jun 2010
tempest: like the revision 6111? ;)
Posts: 586
Threads: 24
Joined: Jun 2010
27 Feb 11, 11:51PM
(This post was last modified: 27 Feb 11, 11:52PM by flowtron.)
Yeah, what an incredible wait - 2 days - it's like we weren't being paid to work on this 24/7!
Actually, I'm inclined to turn off the sf.net bug-tracker completely. 99.1% of the "bugs" reported there aren't any .. and the 0.9% that are only have contact data in 50% (=0.45% of all reports) - I usually only scan it every 3 months or so; so he's actually lucky it was only 2 days :-P
Posts: 1,436
Threads: 7
Joined: Jun 2010
(27 Feb 11, 11:26PM)ärkefiende Wrote: tempest: like the revision 6111? ;)
Oh dear, looks like I'll have to pull back :P
@flowtron: turn it off then, and point people to the forum.
