0.2.8_beta4: release process and bugs

Help test release candidates for the next release
Post Reply
User avatar
Lucifer
Project Developer
Posts: 8640
Joined: Sun Aug 15, 2004 3:32 pm
Location: Republic of Texas
Contact:

Post by Lucifer »

That link knows more about manifest files than I do. All I know is that NSIS says you should have them, so I've always let it include them in the installers. So far as I know it's a default action, and unless anybody's specifically disabled it, then it's there. You guys did use NSIS for these, right?

I have also suffered source downloads from sourceforge being incomplete. I'm trying to remember which project, it might have been this one. So the checksum is more likely to yield an answer, I think. ;)
Image

Be the devil's own, Lucifer's my name.
- Iron Maiden
User avatar
andi75
On Lightcycle Grid
Posts: 44
Joined: Mon Dec 19, 2005 4:57 pm
Contact:

Post by andi75 »

For the sake of completeness:

$ md5sum armagetronad-0.2.8_beta4.2.vc8.win32.exe
a55b819b5d10dc28bd754571f4520aa9 *armagetronad-0.2.8_beta4.2.vc8.win32.exe
User avatar
Z-Man
God & Project Admin
Posts: 11585
Joined: Sun Jan 23, 2005 6:01 pm
Location: Cologne
Contact:

Post by Z-Man »

The checksums for the installer and executable are identical.
The .nsi does not say anything about manifest files, so I assume one is generated. And the same script is used for the other two builds, and they work well.
I could set up a development environment on WinXP to see what's wrong, but that would consume an entire day probably. We should just dump the VC8 builds instead for now and wait for time/someone with WinXP programming competence to solve the problem further down the road. According to the benchmarks, the build is the slowest one anyway and almost as big as the GCC build due to the inclusion of the runtime libraries.

Which reminds me of another clue: The runtime dlls are not required for the dedicated server as it is statically linked with them (can't do that on the client, or SDL and AA would not be able to share state), and the dedicated VC8 build works in XP.
User avatar
joda.bot
Match Winner
Posts: 421
Joined: Sun Jun 20, 2004 11:00 am
Location: Germany
Contact:

Post by joda.bot »

Bug:
- WIN_ZONE_INITIAL_SIZE is not respected.
If WIN_ZONE_EXPANSION is 0 it's ok, if it's 1 or bigger the zone really
starts really big :(.
Seems like the zone size is calculated from round start to current time.
User avatar
Z-Man
God & Project Admin
Posts: 11585
Joined: Sun Jan 23, 2005 6:01 pm
Location: Cologne
Contact:

Post by Z-Man »

joda.bot wrote:Seems like the zone size is calculated from round start to current time.
Correct. The reference time was set to the round start. It's a server bug and should be fixed now.
User avatar
joda.bot
Match Winner
Posts: 421
Joined: Sun Jun 20, 2004 11:00 am
Location: Germany
Contact:

Post by joda.bot »

@z-man: your fix broke displaying zones on the fortress server or at least delayed their start up.
User avatar
Z-Man
God & Project Admin
Posts: 11585
Joined: Sun Jan 23, 2005 6:01 pm
Location: Cologne
Contact:

Post by Z-Man »

Hehe. Well seen :) Yes, the fix was the reason. Apparently, at the time the fortress zones are created, the current game time is the end time of the last round. Should be back to normal now.
wd-40
On Lightcycle Grid
Posts: 28
Joined: Fri Dec 30, 2005 6:18 pm

Issue's I've had

Post by wd-40 »

I haven't got anything recorded for proof, but I've had the following problems:

- Camera sometimes suddenly moves away from the cycle to another player as if you had died then moves back.

- It seems harder to do 180s with beta4 than it is with 2.7.1 When double binding the cycle doesn't turn back directly on the side of the wall, its always away from the wall, so then another 2 double binds is necessary to get it done.

- Harder to get behind certain walls when trying to escape from a box this could be a server setting but I don't think it is. I don't have the problem when using 2.7.1

I think that's it. I hope these haven't been posted yet before but I didn't want to read through 10 pages of bugs
User avatar
Phytotron
Formerly Oscilloscope
Posts: 5041
Joined: Thu Jun 09, 2005 10:06 pm
Location: A site or situation, especially considered in regard to its surroundings.
Contact:

Post by Phytotron »

The second two aren't bugs. :D
User avatar
Z-Man
God & Project Admin
Posts: 11585
Joined: Sun Jan 23, 2005 6:01 pm
Location: Cologne
Contact:

Re: Issue's I've had

Post by Z-Man »

wd-40 wrote:- Camera sometimes suddenly moves away from the cycle to another player as if you had died then moves back.
Hmm, this should be fixed in beta4, possibly already beta3. It was caused by the sound processing (running in parallel to other stuff) messing with data structures read by the camera. Or the other way round.

Silly is right about the other two. Server admins can revert to the old behavior by setting all RUBBER_MINDISTANCE config items to 0, but that's unsupported. It can lead to phasing. Sorry, technical necessity. I'll see what happens if I make the code defaults zero, they are what is in effect when you join a 0.2.7.x server with a 0.2.8 client.
User avatar
joda.bot
Match Winner
Posts: 421
Joined: Sun Jun 20, 2004 11:00 am
Location: Germany
Contact:

Post by joda.bot »

Camera bug... or packet loss ?

The render engine just stops in the middle of the round and after I die it returns. Really a strange bug.

Fast Forward to timestamp: 1100 ...
Beta 4.2 client

I'm not really happy with the map downloading ... the console messages can't be reviewed and the error message is just not telling the actual problem. Message is like "Can't download map" ... I'd like to atleast have the chance to retry downloading the map. Perhaps clients can start downloading the map before joining the server as visible player ? ... for example in spectator mode ?
You do not have the required permissions to view the files attached to this post.
User avatar
joda.bot
Match Winner
Posts: 421
Joined: Sun Jun 20, 2004 11:00 am
Location: Germany
Contact:

Post by joda.bot »

Current CVS is not able to download map for racer server.

Recording is attached...

Windows CVS 2006.01.04 compiled with VC7 (VC++ .NET 2003).

This also shows the missing details in the error messages.
You do not have the required permissions to view the files attached to this post.
User avatar
joda.bot
Match Winner
Posts: 421
Joined: Sun Jun 20, 2004 11:00 am
Location: Germany
Contact:

Post by joda.bot »

Suicide without any wall...

After 2 or 3 grinds on a rim wall, I'm killed without any obstracle wall ahead.

TimeStamp 90 (Round start)
TimeStamp 110 (after doing 1 pass along the wall)

Windows CVS 2006.01.04 compiled with VC7 (VC++ .NET 2003).

I'm still trying to record the teleport bug and get a proof for a rubber loss but the last thing might just exist in my imagination.
You do not have the required permissions to view the files attached to this post.
User avatar
Z-Man
God & Project Admin
Posts: 11585
Joined: Sun Jan 23, 2005 6:01 pm
Location: Cologne
Contact:

Post by Z-Man »

The camera recording does not play back for me right now. I'll try later in Windows. But I'm afraid it won't play back all alone because wrtl seems to delete old maps from the download server. So don't delete the resource cache. Was it the smart camera that was buggy? It isn't exactly made for many small obstacles...

I'm not really sure what problem you have with the error message. Well, that may be because I get a regular 404 map not found error now... Philippe or Luke, do you know how to help?

The strange death would be really useful as a serverside recording, the client does not know why you died either. But it looks like you are pretty close to the rim before the last 180 already, so I'd bet you had zero space left anyway, give or take floating point inaccuracies. Usually, the movement code does not dump you into walls that are almost parallel to your movement direction, but perhaps an inpassable bottleneck was created there on your last run.
User avatar
joda.bot
Match Winner
Posts: 421
Joined: Sun Jun 20, 2004 11:00 am
Location: Germany
Contact:

Post by joda.bot »

@z-man: race server maps

oh, right I'll try to identify the corret map.
Oh, that's useless you'll need all maps up to that point (I guess).

The camera used was the (server) custom camera.
Perhaps the server side cam was changed at that point ? no idea.
I tar-bzipped the all maps from wrtlprnft. (Attached)

Might be good to add a 3 week timeout for maps ?
What happens if wrtlprnft's server reuses a map filename and the client has a different copy cached locally ? Will it check for modification times or another version information ?

about the error messages:

I don't like the error message as it does not give the exact error details.
And accessing the ingame console is pretty difficult without entering a game (which scolls the messages).

Suggestion for a better error description:
Error 404 http://.... could not be read. Can't load DTD XYZ for map ZYX.

Suggestion for a better error description:
Error 404 http://.... could not be read. Can't load map ZYX.

EDIT EDIT EDIT

about the "bounce like bug":

I just don't know how to do a serverside recording. I guess it could be the bounce bug, as I definitly saw a real bounce on the 0.2.8 nixda settings (about one grid cell wide). I saw it just once that's my problem with that while other problems like the camera & cycle teleportation issue happend 3-4 times (which have not been able to reproduce). The situation in the recording is reproducable though.

As the more severe problems occured on my 0.2.8 client with older clients connected, too, I guess the 0.2.6 or 0.2.7 compatible mode might be more buggy ?
You do not have the required permissions to view the files attached to this post.
Post Reply