Aleph One 0.21.1 / 2008-12-17

[treellama]

A bug fix release is available from http://source.bungie.org/get or http://marathon.sourceforge.net/

It is compatible with 0.21 but it is recommended that all players update to 0.21.1

Changes:Ensure Lua scripts work on all platforms (*cough* Windows)Work around a weird Windows crash when metaserver login is deniedFix crashes when gathering net gamesDouble UPnP discover and connect timeoutsInclude updated Lua and MML documentation

[Packard20]

Thanks again Treellama! Keep up the good work.

[treellama]

RPMs are uploaded

[chinkeeyong]

Lol @ Windows. Keep up the good work!

[acks45]

A bug fix release is available from http://source.bungie.org/get or http://marathon.sourceforge.net/

It is compatible with 0.21 but it is recommended that all players update to 0.21.1

Changes:Ensure Lua scripts work on all platforms (*cough* Windows)Work around a weird Windows crash when metaserver login is deniedFix crashes when gathering net gamesDouble UPnP discover and connect timeoutsInclude updated Lua and MML documentation

i still get crashes whenever i switch levels with windows users in (chicken)coop situations. try playing bighouse coop with a script, its a three minute fist fight then a crash.

[treellama]

i still get crashes

I don't suppose you filed a bug?

Do you expect bugs that nobody but you knows about to be fixed?

Edit: wait, is this with the co-op script that everyone knows crashes?

[RyokoTK]

I would think so.

[interion]

I don't suppose you filed a bug?

Do you expect bugs that nobody but you knows about to be fixed?

Edit: wait, is this with the co-op script that everyone knows crashes?

This may be a new bug. The 2.0 Co-Op scripts have always crashed on level change when a player goes OOS, and the 1.0 script crashes some Windows users in-game ever since the Lua API changed (in 0.20?).

But I just tried to jump to a new level with 0.21.1 and got the following assertion: network.cpp line: 1689 Expression: netState!=netUninitialized&&netState!=netDown&&netState!netJoining

[treellama]

But I just tried to jump to a new level with 0.21.1 and got the following assertion: network.cpp line: 1689 Expression: netState!=netUninitialized&&netState!=netDown&&netState!netJoining

That's with the control-shift level selector?

[interion]

That's with the control-shift level selector?

Sorry I wasn't clear. It's happening during the standard teleport to next level in a network game with the co-op 2.0 (original) script and one player.

[treellama]

Alright, if you get a chance, can you try it without the script? I suspect that will crash too

[interion]

Alright, if you get a chance, can you try it without the script? I suspect that will crash too

It does.

[treellama]

https://sourceforge.net/tracker2/?func= ... tid=101997

[interion]

Any idea what caused this assertion?

File: network_dialog_widgets_sdl.cpp Line 927 Expression: inMaxValue < 0

Happened when the PCR was supposed to display.

Game was a lua script type netscript with custom scoring.

[interion]

I'm happy to report I hosted a bunch of games tonight and didn't see a single person go OOS. Thousands of "discarding packet" events are still being logged, so I don't think my Ethernet card magically healed.

Thanks.

[Wrkncacnter]

No, it was your RAM that healed.

[interion]

No, it was your RAM that healed.

I'd probably find that more amusing if anyone could have suggested what hardware (or software) might be responsible for corrupting UDP packets or allowing them to pass through with invalid checksums, but you couldn't. And the RAM, where I assume all those packets are buffered until they pass through the Transport Layer was as good a guess as any.

Certainly a more likely scenario than those video drivers from Intel that caused random file corruption on the hard drive. Come to think of it, my Ethernet is Intel too.

[treellama]

File: network_dialog_widgets_sdl.cpp Line 927 Expression: inMaxValue < 0

Tried to draw a bar with negative length...could it have been a negative time? If you saved the film and script, that would help fix it.

I'd probably find that more amusing if anyone could have suggested what hardware (or software) might be responsible for corrupting UDP packets or allowing them to pass through with invalid checksums, but you couldn't. And the RAM, where I assume all those packets are buffered until they pass through the Transport Layer was as good a guess as any.

Certainly a more likely scenario than those video drivers from Intel that caused random file corruption on the hard drive. Come to think of it, my Ethernet is Intel too.

It's hard to know, but it's good that the checks actually being enforced on both sides (doh!) are helping. Supposedly some drivers will deliver packets that fail the UDP checksum anyway, with a warning--perhaps SDL_net doesn't pass those warnings on. And some hardware just crams 0's in there, so, end-to-end CRC is the best approach.

So, it looks like this release should have had some RCs, but I was worried that they wouldn't get tested enough to find these bugs anyway. What do you guys think? Would people have tested incompatible RCs on the metaserver enough to find all these issues?

[herecomethej2000]

Probably not, In fact haven't we been down that road before? Its all SEMANTlCS any way so whatever works best.

Edit: haha funny, but what if you use the word correctly. The word "meaning" is not exactly what I meant to say.
Oh the Irony

[treellama]

Yeah, it seems like the better approach is just to make metaserver users suffer through a couple "You need to update" messages.

[acks45]

a couple "You need to update" messages.

down with that.

[Iritscen]

I come back to A1 after a break, and find 0.21.x is out. Good for Windows guys, apparently, but it breaks hosting for me, 100% of the time instead of 66% of the time. And Tl even increased the UPnP timeout (for me? aww, you shouldn't have ), but to no avail, it just fails UPnP setup with -799 every time now.

I already know the solution, so don't bother telling me: buy new router.

[interion]

I come back to A1 after a break, and find 0.21.x is out. Good for Windows guys, apparently, but it breaks hosting for me, 100% of the time instead of 66% of the time. And Tl even increased the UPnP timeout (for me? aww, you shouldn't have ), but to no avail, it just fails UPnP setup with -799 every time now.

I already know the solution, so don't bother telling me: buy new router.

What router are you using again?

I know there are a number of poorly designed consumer routers out there, but I'm still convinced your problem is with the configuration. The fact that you're using UPnP is a bad sign.

[Iritscen]

Hmm, guess what? It was human error this time, heh heh heh... *sigh*

But, yeah, I do have to have UPnP turned on even though I also need port 4226 UDP/TCP opened too. When I do both, it works, so I guess the situation is no worse than before. Sorry for the false bug report of sorts. And it's a Netopia 2210, and I'm sure it is actually cheap crap and needs to be replaced (came from ISP). Once I have some money, I will be buying a good router so I can get a home network going with wireless and such, and this nonsense should all be ancient history (I will play wired, though, the wireless will be for other stuff).

[treellama]

Is it possible you have two routers, and don't know? You shouldn't both manually configure and use UPnP to get games to work--one or the other is the correct configuration.