[website] lobby server update / preparation of moderation changes
[website] lobby server update / preparation of moderation changes
a few minutes ago the lobby server was updated to the current version. This was tried already a several times before but was always reverted because fatal bugs were found. The main reason for the update is that we want to improve moderation at the lobbyserver including autohosts.
the update looks atm like it "works" but this needs some time and feedback.
If you notice issues or problems because of the update, please report to the bugtracker!
For details of the changes, see the git commit log.
thanks!
the update looks atm like it "works" but this needs some time and feedback.
If you notice issues or problems because of the update, please report to the bugtracker!
For details of the changes, see the git commit log.
thanks!
- MasterBel2
- Posts: 347
- Joined: 11 Apr 2016, 12:03
Re: [website] lobby server update
It seems no battles can be successfully started. On all hosts Spring crashes 6-7 secs in it seems.
Can the update be reverted asap?
Can the update be reverted asap?
- MasterBel2
- Posts: 347
- Joined: 11 Apr 2016, 12:03
Re: [website] lobby server update
It's working now, thanks abma for looking at it! :)
Re: [website] lobby server update
yeah, i noticed that to. couldn't find the reason why this happens, help is needed:
https://github.com/Yaribz/SPADS/issues/29
https://github.com/Yaribz/SPADS/issues/29
Re: [website] lobby server update
the updated version is running again / the bug which made spring crash is fixed.
please report any issues you notice!
please report any issues you notice!
Re: [website] lobby server update
If the policy is nowadays to just delegate the changelogs to git logs, then may I kindly ask to improve a bit on the git log messages themselves, for example, messages like these are not very specific:abma wrote: For details of the changes, see the git commit log.
@abma abma notify client about invalid data
8026cbc
@abma abma oops, safety
83d58dd
@abma abma more useful error message
6d81219
@abma abma unify log messages: session_id in braces
bc3e230
@abma abma mark OPENBATTLEEX deprecated / inform client
92eeef4
@abma abma remove/disconnect all clients
e9b7b90
@abma abma force utf8 charset in mysql tables
7cb8e54
@abma abma keep old messages only for 2 weeks
d013e2e
@abma abma delete users which didn't play after 30 days
d4326a0
Commits on May 02, 2015
@abma abma reduce calls to time.time() / cleanup
aceae37
@abma abma more cleanup
906f535
@abma abma cleanup
- Silentwings
- Posts: 3720
- Joined: 25 Oct 2008, 00:23
Re: [website] lobby server update
Ofc abma can decide for himself whether to write longer commit messages - but whenever there are changes to the lobbyserver that require a non-trivial reaction from users, they will be announced explicitly on these forums.
(Also, afaics, all of those commit messages, with the possible exception of "opps safety", are pretty clear.)
(Also, afaics, all of those commit messages, with the possible exception of "opps safety", are pretty clear.)
Re: [website] lobby server update
@jools:
most of the changes where code-refactoring / adding tests / porting to python3. nothing of interest for "common" people in spring community. the others should look through the commit log when interested.
a change which affects players / autohost owners is announced here: viewtopic.php?f=1&t=36402
most of the changes where code-refactoring / adding tests / porting to python3. nothing of interest for "common" people in spring community. the others should look through the commit log when interested.
a change which affects players / autohost owners is announced here: viewtopic.php?f=1&t=36402
Re: [website] lobby server update
I'm just saying that documentation is important, not only for the current user-base but for the future as well. It's quite hard to predict the future so it would be better to be transparent as much as you can.
- PepeAmpere
- Posts: 589
- Joined: 03 Jun 2010, 01:28
Re: [website] lobby server update / preparation of moderation changes
No longer able to create user-hosted game. I have lobby with without a slider for setting number of people for a battle. Is default set for 10 if i send no max value?
- Silentwings
- Posts: 3720
- Joined: 25 Oct 2008, 00:23
Re: [website] lobby server update / preparation of moderation changes
Bug reports to bug tracker, link in OP. Not being able to open a battle is possibly a bug, although if you use a non-standard lobby you might want to check that your lobby isn't incorrectly assuming the battle its asking to opening will have the same number of players that its requested to have. I can host a battle via SL as normal.
On botflags/maxplayers - nothing has changed lobby-side; all that happens is if a user without a botflag asks to open a battle with >10 player count, the max players is limited to 10 (by a single extra line of code, in the obvious way) before the battle is opened, and a message is sent by the server informing the user.
I think sending no max value would violate the protocol, so the result would be undefined behaviour.
On botflags/maxplayers - nothing has changed lobby-side; all that happens is if a user without a botflag asks to open a battle with >10 player count, the max players is limited to 10 (by a single extra line of code, in the obvious way) before the battle is opened, and a message is sent by the server informing the user.
I think sending no max value would violate the protocol, so the result would be undefined behaviour.
Re: [website] lobby server update / preparation of moderation changes
FYI: (hopefully) a bad bug was fixed which caused several autohosts to connect and disconnect repeatedly.
sorry for the restart at a busy time: it was a good test to confirm it is fixed.
sorry for the restart at a busy time: it was a good test to confirm it is fixed.
- FabriceFABS
- Posts: 354
- Joined: 28 Jul 2010, 16:20
Re: [website] lobby server update / preparation of moderation changes
No one can believe you did that for fun only. Thx for your work.abma wrote:FYI: (hopefully) a bad bug was fixed which caused several autohosts to connect and disconnect repeatedly.
sorry for the restart at a busy time: it was a good test to confirm it is fixed.