0.2.8_beta4: release process and bugs

Help test release candidates for the next release
Post Reply
User avatar
dlh
Formerly That OS X Guy
Posts: 2035
Joined: Fri Jan 02, 2004 12:05 am
Contact:

Post by dlh »

Did you change anything related to the messaging code? I keep getting the /team messages for older clients:

name: ( name --> Teammates ) message

Edit: This is in Fortress, so that thread probably would have been a better place for this.
User avatar
Z-Man
God & Project Admin
Posts: 11585
Joined: Sun Jan 23, 2005 6:01 pm
Location: Cologne
Contact:

Post by Z-Man »

No, no change.
Are you by chance quick to reconnect to the server when it restarts? And the login attempt takes a while, but then succeeds? Then the server will think you're using something earlier than 0.2.5.2. Yeah, strange problem. I know what causes it, but have not yet found a way to work around it, unless we destroy custom connect compatibility with versions earlier than that.
User avatar
dlh
Formerly That OS X Guy
Posts: 2035
Joined: Fri Jan 02, 2004 12:05 am
Contact:

Post by dlh »

z-man wrote:No, no change.
Are you by chance quick to reconnect to the server when it restarts? And the login attempt takes a while, but then succeeds? Then the server will think you're using something earlier than 0.2.5.2. Yeah, strange problem. I know what causes it, but have not yet found a way to work around it, unless we destroy custom connect compatibility with versions earlier than that.
Yes, I reconnected right after the restart.
User avatar
dlh
Formerly That OS X Guy
Posts: 2035
Joined: Fri Jan 02, 2004 12:05 am
Contact:

Post by dlh »

Start up Armagetron and connect to an internet server. Disconnect and go to a local game. Some of the settings do not reset, most noticeable are cycle_rubber and cycle_delay.
User avatar
Z-Man
God & Project Admin
Posts: 11585
Joined: Sun Jan 23, 2005 6:01 pm
Location: Cologne
Contact:

Post by Z-Man »

Yes, this has always been so. That does not make it not a bug, of course. I added a database entry.
User avatar
dlh
Formerly That OS X Guy
Posts: 2035
Joined: Fri Jan 02, 2004 12:05 am
Contact:

Post by dlh »

More bugs:

(1) On the Fortress Lobster Cage server, cycle_brake_deplete got stuck at 0. On my client it was set to the correct value (0.5). Joe said this has happened before (in 0.2.7.1?).

(2) During the intermission between rounds, I had my chat input active. I hit t and it inserted the instant chat text (/team).
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 network settings thing should now be fixed, as well as the login problem. I just added new information to the old style login message that new servers can use. I wonder why I have not done this earlier.
nemostultae wrote:(1) On the Fortress Lobster Cage server, cycle_brake_deplete got stuck at 0. On my client it was set to the correct value (0.5). Joe said this has happened before (in 0.2.7.1?).
Just so I understand this correctly: The BOTLC server uses CYCLE_BRAKE_DEPLETE 0.5 and CYCLE_BRAKE_REFILL 0. You played on it. Your client showed CYCLE_BRAKE_DEPLETE 0.5, but the server was using CYCLE_BRAKE_DEPLETE 0, so you had infinite lobster boost.
I think I know where this comes from. Some earlier version did not support brake depletion. Since todays settings-reverting code did not yet exist back when I introduced finite breaks, something different was done: the server falls back to unlimited brakes whenever an old client connects, without the server admin being able to do anything against it. I'll just change that to use the new mechanisms, then the admin can decide whether the setting should revert or whether old clients should be locked out.
nemostultae wrote:(2) During the intermission between rounds, I had my chat input active. I hit t and it inserted the instant chat text (/team).
That's probably a mix of two effects. First, whenever you press <Modifier except shift>+t, the chat input can't insert a valid character for it, so it takes the base keypress, the t, and looks for an instant chat and inserts that. But you probably were not really pressing a modifier together with t, right? But the input system, I think this is SDL itself, has the known problem that modifiers don't have an effect at the time of the keypress, but at the time the key is polled. And polling is a bit slower between rounds. So hitting t, then a modifier, would produce the observed result.
I'll see what I can do about the modifier input problem, maybe it is my fault after all. The last time I thought about it was before we had recordings, and it was impossible to debug back then.
User avatar
Z-Man
God & Project Admin
Posts: 11585
Joined: Sun Jan 23, 2005 6:01 pm
Location: Cologne
Contact:

Post by Z-Man »

About joda's double free glibc waring: Valgrind does not seem to find anything in that direction. Is the debugging glibc also available on Fedora? If so, I'd install it on a virtual machine and test there. If not, I probably can't do anything. I'll check the autopackage build directly, so far I worked with a regular build, and autopackage uses a compilter wrapper. Perhaps that messes things up.
User avatar
dlh
Formerly That OS X Guy
Posts: 2035
Joined: Fri Jan 02, 2004 12:05 am
Contact:

Post by dlh »

I saw you committed a change for the `date -r' error, but that did not fix the error.

The usage of my `date' requires the -r option to be before the format. This is the real reason I was getting the error.

The test you do to check for -r option actually exits with status 0.

Code: Select all

date -r this_will_exit_with_status_0
Wed Dec 31 19:00:00 EST 1969
however this will exit with an error

Code: Select all

date +"%Y" -r error
date: illegal time format
usage: date [-nu] [-r seconds] [+format]
       date [[[[[cc]yy]mm]dd]hh]mm[.ss]
but this will not

Code: Select all

date -r no_error +"%Y"  
1969
Very odd...

Code: Select all

STANDARDS
     The date utility supports the Version 3 of the Single UNIX Specification
     (``SUSv3'').  It is also expected to be compatible with IEEE Std 1003.2
     (``POSIX.2'').
User avatar
Z-Man
God & Project Admin
Posts: 11585
Joined: Sun Jan 23, 2005 6:01 pm
Location: Cologne
Contact:

Post by Z-Man »

Bizarre. I changed the test. Feel free to take over if the new version fails again.
User avatar
andi75
On Lightcycle Grid
Posts: 44
Joined: Mon Dec 19, 2005 4:57 pm
Contact:

VC8 version doesn't even start.

Post by andi75 »

VC6 version works fine. The VC8 executable seems corrupt. The md5sum after installation is:

$ md5sum armagetronad.exe
0153d2fed81a09af29e1a8823b360af7 *armagetronad.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 »

Using Windows XP? It works fine on 98 and 2000, but for some unknown reason not on XP. All I get on XP is a cryptic "Installation problem, try to reinstall" error message with no hint whatsoever what the problem is. That's why the build is not on sourceforge. I'm hoping to get more information from XP-savy people. The checksum is a good idea, I'll compare it when I'm home.
User avatar
Lucifer
Project Developer
Posts: 8640
Joined: Sun Aug 15, 2004 3:32 pm
Location: Republic of Texas
Contact:

Post by Lucifer »

Got a manifest file? I don't see how it would make a difference, but here's a link:

http://www.codeproject.com/cpp/AddXpStyle.asp

NSIS can generate them, or does by default, or something like that.
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 »

Yes, I'm using XP.
User avatar
Z-Man
God & Project Admin
Posts: 11585
Joined: Sun Jan 23, 2005 6:01 pm
Location: Cologne
Contact:

Post by Z-Man »

From Lucifer's link it seems a manifest file is only for modifying application appearance. Is it required for all programs that open a window?
Post Reply