Release process for 0.2.8_beta2

Luke-Jr · Post by **Luke-Jr** » Mon Sep 12, 2005 2:16 am

TiTnAsS wrote:i guess ctrl + escape minamizes now?

Ctrl-Escape has always been the same as the "Windows" key.

TiTnAsS · Post by **TiTnAsS** » Mon Sep 12, 2005 5:00 am

Error in  in
E:\Development\AADevel\0_2_8_beta2\armagetronad-0.2.8_beta2-win
\src\render\rFont.cpp:65 :   
Font file textures/font.png could not be loaded.

Then right after:

Code: Select all

Internal error in  in
E:\Development\AADevel\0_2_8_beta2\armagetronad-0.2.8_beta2-win
\src\tools\tHeap.cpp:145 :
         Element is not in this heap!
Please send a bug report!

Then right after:

Code: Select all

Microsoft Visual C++ Runtime Library:

Runtime error!

Program: D:\Program Files\Armagetron Advanced\armagetron.exe


This applixcation has requested the Runtime to be terminated in an unusual way.
Please contact the applications support team for more information.

This all happened when i tryed to make it windowed screen, thats the second time its happened.[/code]

TiTnAsS · Post by **TiTnAsS** » Mon Sep 12, 2005 6:59 am

z-man wrote: I uploaded my builds to aabeta. TnA, that's here for Windows.

wtf "operation timed out"...was trying to link bob to it...

EDIT: ah lukes net is down.

EDIT EDIT: its back now

Luke-Jr · Post by **Luke-Jr** » Mon Sep 12, 2005 8:52 am

TiTnAsS wrote:
z-man wrote: I uploaded my builds to aabeta. TnA, that's here for Windows.
wtf "operation timed out"...was trying to link bob to it...

EDIT: ah lukes net is down.

EDIT EDIT: its back now

Actually, the absense of my primary connection should leave the secondary one working (except when ISP is down, which wasn't the case)... so I suspect it could just be you put in a request when the system was too busy to respond in under whatever your time out limit is. On that topic, anyone know how to get RAM usage down under 1 GB? I can't seem to get far below 2 GB of usage... O.o;;

Post by **dlh** » Mon Sep 12, 2005 7:11 pm

Code: Select all

$ ping -c 5 beta.armagetronad.net
PING dashjr.org (24.145.163.67): 56 data bytes
64 bytes from 24.145.163.67: icmp_seq=1 ttl=46 time=2026.934 ms
64 bytes from 24.145.163.67: icmp_seq=2 ttl=46 time=1897.942 ms
64 bytes from 24.145.163.67: icmp_seq=3 ttl=46 time=1859.257 ms

--- dashjr.org ping statistics ---
5 packets transmitted, 3 packets received, 40% packet loss
round-trip min/avg/max/stddev = 1859.257/1928.044/2026.934/71.687 ms

Ouch.

Uploaded to sf, downloaded to beta.armagetronad.net. Took 15 minutes to type the command "mv armagetronad-0.2.8_beta2.macosx.dmg /var/www/aabeta/download/0.2.8/0.2.8_beta2".

Mirrors have already picked it up.

Post by **dlh** » Mon Sep 12, 2005 8:07 pm

Luke, stop renaming my files.

Luke-Jr · Post by **Luke-Jr** » Mon Sep 12, 2005 8:28 pm

nemostultae wrote:Luke, stop renaming my files.

Then don't name them incorrectly in the first place.

Website · Post by **Z-Man** » Mon Sep 12, 2005 11:00 pm

TnA: The second error is only a failure to recover from the first. The first error may be caused by someone using the font when rendering is actually disabled; I'll enter it into the database. A recording may be useful here; it should also catch the relevant system events. Use the new menu items, then rename the new .aarec file on your desktop to .zip and attach it here, please.

Something else I noticed in Windows: When I start AA in windowed mode, switch over to fullscreen mode, then alt-tab away from it, then back, and then enter windowed mode again, the window is minimized and can't be restored by normal means. The program is still running; sound and input are working. Rendering is activated, so the app must think it is currently active. The only thing that's missing is the window. It's possible to go back to fullscreen mode; Also, the "send to" feature of the NVidia window context menu also restores the window. The only place that actively could be responsible for this is the SDL_WM_IconifyWindow call in sr_Activate; but disabling it has no effect on the behavior. Strange, strange. Can anyone please check whether this is just a quirk of my PC?

Jonathan · Post by **Jonathan** » Mon Sep 12, 2005 11:19 pm

Which version, so I can test it? I'll try the current version on it (0.2.7.0 or 0.2.7.1) anyway.

Website · Post by **Z-Man** » Mon Sep 12, 2005 11:32 pm

Since this is the beta2 thread, the version defaults to 0.2.8_beta2

Edit: and Windows is Windows 2000, last SP and all patches from more than a month ago installed (I don't keep track of what Windowsupdate is doing).

Edit2: The reason the server autopackage did not work for me was that I had a QUIT command in ~/.armagetronad-dedicated/var/settings_custom.cfg. I left it there from the last testing session

I reenabled the download on aabeta. I'm putting both autopackages on SF now.

Jonathan · Post by **Jonathan** » Tue Sep 13, 2005 12:20 am

OK, but it may also apply to other versions, right? The PC I can test on runs XP SP2 and (coincidentally) has a Radeon 9600.

TiTnAsS · Post by **TiTnAsS** » Tue Sep 13, 2005 1:12 am

z-man wrote:TnA: The second error is only a failure to recover from the first. The first error may be caused by someone using the font when rendering is actually disabled; I'll enter it into the database. A recording may be useful here; it should also catch the relevant system events. Use the new menu items, then rename the new .aarec file on your desktop to .zip and attach it here, please.

Can't reproduce it, i'll try with the other one later though...

Website · Post by **Z-Man** » Tue Sep 13, 2005 8:09 am

Jonathan: Yes, It would be valuable information if your system is affected (and if older versions of AA do the same); since it's the NVidia software that can bring the window back, it may very well be a bug there that caused it to disappear in the first place. If the bug does not happen for you, we can be quite sure that's the case.
Edit: My win98 is affected, too. But older versions of AA (0.2.7.0 and one from May) don't show it, so it is related to something we do. Binary search time!
Edit^2: Verdict: It's SDL. Copying over all the DLLs from 0.2.7.1 into the current version fixes the problem. I'll check our options.

TnA: No hurry. I'll grep for possible causes.
Edit: found one! If an error message is to be printed ("You've been kicked", "Map load failure", "Master server does not answer", these things) and the app is not active, I get the font load failure. Easy to fix.

Website · Post by **Z-Man** » Tue Sep 13, 2005 2:11 pm

Jonathan: Say, that timer code of yours you were talking about is not for Windows by coincidence? If so, would you share it?

Jonathan · Post by **Jonathan** » Tue Sep 13, 2005 3:20 pm

No, I said I used gettimeofday. But gettimeofday appears to be a good Google keyword to find ugly and non-perfect workarounds for (switching to) Linux or such.