artofnoise wrote:

TIP: For the Visual C++ crashes, I keep a shortcut to the EQ\Logs directory. Every time I crash, I open it and copy-paste the files to a new folder with a few keywords so I remember what was happening prior to the crash.

Yes, this is tedious, but until the EQ launcher/client has an auto-detect routine at launch, it appears to be the only way to provide data to the dev team.

JChan, do you think it would help to add a routine to the client to automate this? It seems like it might benefit troubleshooting on your end since most people do not submit these files if manual is the only option. Something like: on Launch check Logs (dump files), if date/time stamp > than previous launch/patch = send files to EQ Dev team. This would not necessarily provide you with what they were doing at the time, but at least you would see more dump files. More dump files might provide a greater chance of seeing patterns.

I can think of a few ways to rank these dump submissions automatically so your team doesn't get overwhelmed:

- looking at the volume (total #) of crashes on that client in the last "x" days

- lower importance to files with date/time stamp more than 30 minutes ago (ie. if someone crashes and relaunches quickly it is less likely they just exited the game in some odd fashion or had a power failure and gave up for the day)

Just a thought.
An auto-detect is definitely something that we'll start the ball rolling on. As for the automatic ranking, we'll have to think about this a bit more. Ideally we'd be able to put a bit more data around the crashes and be able to target systems that we think are susceptible and be able to change these rankings on the fly.



More...