Re: [Nel] invisible users

[Vianney Lecroart]

Hi Michael,

> Sometimes, (Not Always!) the first user to log on after a fresh full
services restart cannot see the second.  It seems the
> problem is with new players not recieving the data for the old ones- it
happens with not so fresh servers more often and
> sometimes with greater numbers of users, though the 2nd and 3rd can
usually see eachother.  I'm pretty familiar with the
> frontend and position service code having mostly rewritten them for our
own ends and added several services and iI
> cant find any bugs in it.  This problem happens with unmodified nel and
snowballs2 code.  Has anyone else
> experienced this?

Hum, quite strange. I know a bug in the Net lib that happens very rarely.
Sometimes, the lib seems to lost some messages (in TCP, yes... we know...).
It's very very strange we can't reproduce the bug. We add a incremental
number in each packet to see when we lost something, and when it happens, we
log a nlwarning and should displays some thing like : "ERR: stop in the
buffer %u and should be %u" or "WRN: LNETL1: !!!LOST A MESSAGE!!! I received
the message number %u but I'm waiting the message number %u (cnx %s)"
It's perhaps a lost in a stack during thread communication but anyway, it s
happen very rarerly and we ll have to fix that quickly when we'll found a
way to reproduce the problem.
If you have this log message, your problem is this one, otherwise, I don't
really know what it could be. I'll create a little stress test that
connect/disconnect some clients, some services automatically with a front
end and a login system to see the robustness of the system.

> WRN 1272 unified_network.cpp 830 : HNETL5: can't find service WS to send
message CC     <-- could this be it?

WS is welcome service, it works with the Login System to identify players
account. The message says that you didn't launch the WS so the front end
can't send info to it.
(I don't think that problem comes from that)

Vianney Lecroart
---
lead network programmer / nevrax.com
icq#: 6870415
homepage: http://ace.planet-d.net
www.geekcode.com: GCS/E d- s+++: a-- C+++$ UL++ P- L+++>+$ E+>- W++ N+ o? K-
w++$ O- M- V- PS- PE? Y PGP t 5? X+ R- tv++ b- DI D+ G e++ h+ r-- y?