[Small issue] If you pause bots, they get disconnected due to inactivity.

[The-Gig]

Description:
This is just a small issue because it's only relevant to a few development/testing scenarios, however I write it down anyway, as it actually happened to me while doing a few tests.
If you "pause" bots, they will be "disconnected" after three minutes, due to inactivity. In case you are doing some testing which requires to keep them quiet, stationary and unaggressive for a long time, this can be a problem.

How to reproduce:

• load a map with devmap [mapname]
• add one or more bots
• bot_pause 1
• wait three minutes (you can speed up things by tweaking timescale value)
• see bots disconnect with "timed out" in console.
  UPDATE: It looks like the above happens with non-team-based modes (I tested gametype 0): in team-based modes (I tested 3, 5 and 8) the bot disconnects earlier (like one minute and half?) and with "server command overflow" in console instead of "timed out".

Suggested behavior:
I guess the inactivity check routine should verify whether the client is a bot, and whether bot_pause<>0: if both are true, the client should not be disconnected, assuming that there is a developer/tester which is purposedly testing something.
Or maybe just ignore bots at all?

[NeonKnightOA]

Does the bug still appear in b53?

[NeonKnightOA]

Oh, wait, does it happen in any gamemode? Because I remember Sago adding something about killing the bots if all the players are killed in Elimination/eCTF/LMS gamemodes.

[The-Gig]

I think somewhere in commit history there were Start killing bots in Elimination and CTF Elimination if all humans have been eliminated. and If all humans are killed start killing the bots in Last Man Standing.. I'd like to have more infos about how it works from @sago007 ... (it was in mentioned changelog.md, but I can't find it anymore in that file).

However that one seems about killing bots in Elimination in order for humans to wait less before respawning (by forcing a random winner by randomly killing bots? :-/ Or by following some criteria like the health left at that moment? I don't know...)" ... while this issue instead looks like related to some mechanism meant to kick inactive human players (which have their connection interrupted?), but being mistakenly also applied to bots.

I'll try to make a few tests later.

[NeonKnightOA]

The way that works is: in Elimination/eCTF/LMS, when all humans are killed in the round, bots slowly die (at a rate of 5HP/s).
This is something that should either be optional or, at least, restricted to multiplayer, where the implementation actually makes sense.

Restricting things to multiplayer is one of the things I should learn to do soon.

[The-Gig]

About the topic of this issue (bots kicked), I just tested in B53, following the example of the opening post, and the bot still got kicked after three minutes.

About the side topic of the "killing bots in eliminitation mode when no human players are left", the current implementation seems good enough to me, as it looks like if the teams were composed only of bots since the start, then the feature doesn't kick in. I should do a few more test to check whether it considers "since the start" of the whole match or of the single round.
While it may be nice for people a way to disable this feature (a dmflags, maybe?) if they wish, I don't think in this case multiplayer or singleplayer should make a difference: people may not like to wait for long for bots "to finish without him" even in singleplayer mode, isn't it?

[NeonKnightOA]

I would still make that feature optional.

[The-Gig]

I would still make that feature optional.

That's fine for me. Maybe even the 5hp/sec is a bit too much, as maybe 3hp/sec might alrady be enough... But my tests of today were too few to tell for sure.
How much I would like some other users to come here and tell their thoughts about stuff...