0.2.8.0 final: Release process and bugs
-
- Dr Z Level
- Posts: 2246
- Joined: Sun Mar 20, 2005 4:03 pm
- Location: IM: luke@dashjr.org
About Windows, AA dedicated practically cannot be run at the moment...z-man wrote:Belenus: Could you post the full Windows error? You may be bitten here by our security fix.
About OpenBSD, I hope we can find a solution for the path errors soon, would prefer to have it running under OpenBSD rather than Windows =)
OpenBSD content moved: http://forums.armagetronad.net/viewtopi ... 4727#44727
Lets hope it isn't too far away...Luke-Jr wrote:Summary: We screwed up on handling directories. Let's fix this for 0.2.8.1.
You do not have the required permissions to view the files attached to this post.
Last edited by belenus on Sun Mar 19, 2006 11:25 am, edited 1 time in total.
- bel
No I found out who is guilty, and you'll like it Microsoft.Luke-Jr wrote:Summary: We screwed up on handling directories. Let's fix this for 0.2.8.1.
We're using _mkdir to create directories. The documentation says, unless it is clarified in one of the references, that on error, errno is set to EEXIST if the directory OR DEVICE already exists. And really, that's true if you try to create "E:" or "D:" on my Windows 2000 box and apparently also for "C:" on my test installations, but it is set to EACCES if I try to make "C:" on my real machine. My user profiles reside on on that, so all my installations worked fine.
Oh, and of course I do have write rights on C:. My system isn't even on C:. There is no logical reason to protect C: on my box.
There was absolutely no way we could have anticipated that. Calling the release 0.2.8.0 and not rc5 was the right decision, we need users to adopt the version fast.
So, what now? The fix will be a workaround for an obviously broken system, we shouldn't bump the release number. I'll to test builds (call them 0.2.8.0-1 for now) and push them here for testing.
The workaround will be to reinstall the strange backward searching in mkdir_recurse, that's silly, but it seemed to work.
belenus: the workaround is to run from the build directory as you already did, it seems we RELY on binreloc to work. For the other error, try replacing "file:///etc/xml/catalog" in gParser.cpp with "file:///var/db/xmlcatalog".
I'm currently installing freeBSD (didn't grok the OpenBSD partitioning), I hope I can see all those errors myself.
No error after the above change, but how do I load my configurations and map file that are in /etc/armagetronad-ded.....? And where are the statistics and logs stored? Can't find em... =)z-man wrote:belenus: the workaround is to run from the build directory as you already did, it seems we RELY on binreloc to work. For the other error, try replacing "file:///etc/xml/catalog" in gParser.cpp with "file:///var/db/xmlcatalog".
I'm currently installing freeBSD (didn't grok the OpenBSD partitioning), I hope I can see all those errors myself.
If you need help with OpenBSD partitioning (which is really easy once you read the FAQ about installation) you can also ask me. =)
- bel
Test fix builds.
Edit: withdrawn.
Edit: withdrawn.
Last edited by Z-Man on Sun Mar 19, 2006 10:51 am, edited 1 time in total.
Windows server now works... thanks, now the OpenBSD stuff?z-man wrote:Test fix builds.
I mean, it works when started from the compile dir, but that is no permanent solution for a production server....
Well, once thing OpenBSD is known for (aside from security and stability) is the quality of the manpages once installed... the FAQ isn't bad either. And, you can never stop learningz-man wrote:Thanks for the offer on OpenBSD The reason I tried FreeBSD first is that there, I didn't have to RTFM. It doesn't have release songs, though.
- bel
You could try to specify the right paths to the data files on the command line, like
You should be able to override everything that way. If that doesn't work, I have to point out that OpenBSD is not officially supported in this release
All four fixed Windows builds are now on my harddisk, plus a source "patch". The internal version is still 0.2.8.0, only the installer version reflects the change. I'm not really sure whether the "This is a workaround of an OS bug, hence no new version" is the right way to go, so I'm waiting for at least one team member to say "yes, that's absolutely right, up to aabeta and sourceforge with it!" and every team member has veto rights for the next hours.
Aww, what am I saying? The patch is attached, I'm going to upload the builds to aabeta anyway for everyone to get. Veto rights still apply, I'll unlist the builds if they are exercised. SF waits til tomorrow.
Edit: withdrawn patch.
Code: Select all
armagetronad-dedicated
--datadir /usr/local/share/games/armagetronad-dedicated
--configdir /etc/games/armagetronad-dedicated
--vardir <wherever-you-want-it>
All four fixed Windows builds are now on my harddisk, plus a source "patch". The internal version is still 0.2.8.0, only the installer version reflects the change. I'm not really sure whether the "This is a workaround of an OS bug, hence no new version" is the right way to go, so I'm waiting for at least one team member to say "yes, that's absolutely right, up to aabeta and sourceforge with it!" and every team member has veto rights for the next hours.
Aww, what am I saying? The patch is attached, I'm going to upload the builds to aabeta anyway for everyone to get. Veto rights still apply, I'll unlist the builds if they are exercised. SF waits til tomorrow.
Edit: withdrawn patch.
Last edited by Z-Man on Sun Mar 19, 2006 10:51 am, edited 1 time in total.
Well, yeah, but... its practically working... there are just some weird path errors... would it be really hard and much work to make it work? (For someone who knows what hes doing. )z-man wrote:You should be able to override everything that way. If that doesn't work, I have to point out that OpenBSD is not officially supported in this release
- bel
- Tank Program
- Forum & Project Admin, PhD
- Posts: 6711
- Joined: Thu Dec 18, 2003 7:03 pm
Well, I don't know what I'm doing with respect to BSD, and you probably don't know too much about the internals, and the next showstopper may be just around the corner. And overriding the paths is really my last idea, the path handling sources are too messy to even think about remotely patching them.
Can't that wait for 0.2.8.1? I'll have a chance to look at the problems first hand until then. Running a server from the build directory isn't that bad, my test servers do that all of the time.
Edit: the fixed builds are on aabeta. Those who downloaded the patch only to find that it's no use for them should go there instead
Can't that wait for 0.2.8.1? I'll have a chance to look at the problems first hand until then. Running a server from the build directory isn't that bad, my test servers do that all of the time.
Edit: the fixed builds are on aabeta. Those who downloaded the patch only to find that it's no use for them should go there instead
Last edited by Z-Man on Sat Mar 18, 2006 11:53 pm, edited 1 time in total.
How can I use the rcd script with the build directory, I don't want to run the server as root and want it automaticly restarted when it ends after idle time... (because of dynIP changes)z-man wrote: Running a server from the build directory isn't that bad, my test servers do that all of the time.
- bel
Well, first thing, don't build as root, use a specially created user. I assume regular users are allowed to compile?
Second, write a small wrapper script, something simple like
or the other way round if you want to continue after crashes or fatal exits.
Second, write a small wrapper script, something simple like
Code: Select all
#!/bin/sh
while ./armagetronad-dedicated; do true; done
Content moved: http://forums.armagetronad.net/viewtopi ... 4728#44728
Last edited by belenus on Sun Mar 19, 2006 11:27 am, edited 1 time in total.
- bel