Jump to content

[Zero] Playerbot (Classic)


Guest blueboy_

Recommended Posts

It is good to see that you are still updating portalzero. smile

I've sent you a PM regarding portalzero and last updates to mangos zero.

Thanks for the heads up on the switch. I took a short break to recharge my batteries, but now I'm back and eager to support the new backports ;) I've replied to your PM and I have already made some progress. If you haven't received my reply, I'll re-send it

Cheers

Link to comment
Share on other sites

Good News Guys,

I've got the zeronew-ai port to the new cmangos-classic core fully working. This flavour of playerbot will work only with the code being developed on the cmangos-classic 'proper' branch.

@cala and anyone else interested in testing the code, feedback and contributions are always appreciated.

There is a great deal of activity at present with the 'proper' branch, so I will rely on you guys keeping me informed on any issues you encounter concerning playerbot. We will then do our best to resolve these as soon as possible.

I hope to make the playerbot code available shortly as a separate repo on the 'blueboy' github site and it will be called 'portalclassic'.

Hope this helps

Link to comment
Share on other sites

Interesting. When does this occurs? Is all the time or only sometime?

Yes it does on mine. There didn't appear to be anything obvious on the server console to explain the lag. I will try it tomorrow with vanilla core code to see if it still occurs. I did have my PS3 on and maybe that was hogging the network. I'll get back to you on this.

@EDIT: When the lag occurs, the server console appears to be spammed with network statistic info.

DETAIL_LOG("Send all time packets count: " UI64FMTD " bytes: " UI64FMTD " avr.count/sec: %f avr.bytes/sec: %f time: %u", sendPacketCount, sendPacketBytes, float(sendPacketCount) / fullTime, float(sendPacketBytes) / fullTime, uint32(fullTime));

DETAIL_LOG("Send last min packets count: " UI64FMTD " bytes: " UI64FMTD " avr.count/sec: %f avr.bytes/sec: %f", sendLastPacketCount, sendLastPacketBytes, float(sendLastPacketCount) / minTime, float(sendLastPacketBytes) / minTime);

I traced this to void WorldSession::SendPacket(WorldPacket const* packet) in WorldSession.cpp. This is conditional code that is only executed if MANGOS_DEBUG is active. So normally this would be ignored on a production server. As a test I commented out this code and the lag stopped. However, this code is present in all versions of MaNGOS core, so it isn't the root cause for this lag. I will continue to test the code and search for a reason.

Cheers

Link to comment
Share on other sites

Good job. I have personnal matters do deal with this WE, so it will be difficult for me to test anything before monday. However, I'll look into recents commits if some affects functions or methods calling void WorldSession::SendPacket(WorldPacket const* packet)

Thanks ! :)

Link to comment
Share on other sites

Hi cala,

I've spent the weekend trying to track down the cause for this 'lag' issue. There is no apparent lagging with core code up to commit (z1898). Several commits after this involve the inclusion of the movemaps code. I built a server using code for commit (z1913) and everything ground to a halt after a few minutes. To prove that the problem is with the movemaps code, I disabled this feature in mangosd.conf and the lag went away. I know that playerbot code works fine with the movemap implementation for 'mangos-wotlk', so there is no reason why they shouldn't work together here too.

The path taken by bots returning to the player is very elongated, and this has already been commented on by xfurry here, with reference to npcs;

@EDIT The information on this post has been removed as it referred to code not supported by this forum. Sorry for the inconvenience: Please use https://github.com/blueboy/portalzero (master or zeronew-ai branches)

I have posted here to ask you and anyone else interested in testing the playerbot code to confirm (or not) this issue. I'm sure if this issue effects the playerbot code, it will also cause problems elsewhere.

I will work on a port for the 'mangos-tbc' code to check that this issue does not occur here too.

Regards

blueboy

Link to comment
Share on other sites

There is indeed a problem with mmaps, big enough to crash the client. Probably to overload server too as you stated, but I did not see it. Extracted mmaps seem to be wrong (very wrong). Until they are fixed, I will test playerbot without mmaps and tell you if everything is ok with playerbot. :)

Link to comment
Share on other sites

Hi cala,

I agree with you completely. It's never nice to suggest disabling code, but until this is resolved, it seems to be the best solution. I would still be interested to know how the playerbot code performs, so I can improve the code and start porting across some of the later features ;)

Link to comment
Share on other sites

  • 1 month later...

Hi Guys,

You might notice some changes with the playerbot repo https://github.com/blueboy/portalzero

After recent changes with the core, the playerbot code would no longer properly merge. The conflicts were related to code format changes in core and the resulting merge (after conflict resolution) did not reflect the current status of the mangoszero core code. As such I was forced to remove the portalzero code and make a fresh port into the current mangoszero core code. I have removed the 'zeronew-ai' branch and the new portalzero 'master' branch represents a merge between the old 'master' & 'zeronew-ai' branches.

Hope this helps

Link to comment
Share on other sites

  • madmax unpinned this topic

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. Privacy Policy Terms of Use