The tutorial is super useful, I'd be a lot farther behind without it, so thanks again for making it. I'm glad to see it get a couple tweaks and not get forgotten about as many things like that do. If you rewrite it, I would separate out the system stuff from the EQEmu stuff, maybe into separate tutorials. I skipped like half the steps because they didn't pertain to me (hyper-v, freebsd, no web editing--yet, mysql CLI, ...).
Sorry it wasn't clear, I didn't need to rebuild shared_memory (or, maybe I do need to, but I didn't), I just had to run it again. IIRC, the first time I ran it (step 17) was before the updates/maps/etc. step was done. After everything was set, I ran it again and it generated output (without rebuilding). It still exited without appearing to do anything, though. It doesn't seem like a daemon, does it generate a cache file or something? Still seems weird.
The shell script I'm working on is intended to do everything after the OS install. I plan to host it online so you can just wget-and-forget. One would probably only need to tweak one line to change it from pkg to yum/apt for dependencies, and add a getopts flag, otherwise it's already pretty hands-off. I figure I'll be doing that a lot as I learn/break stuff so I wanted an easy way to start over. I'll get it into Git soon, it's hobbled by eqemu_update issues right now.
Speaking of which, I also wrote a new update script to try and test some stuff eqemu_update wasn't doing right (as far as I could tell). I haven't touched Perl (on purpose) for 10+ years so it was a learning experience. The script does most of its stuff asynchronously--fetching, cloning, building, etc., and buttons it all up at the end. But it isn't as smart as eqemu_update currently, with db_version and whatnot, so I probably won't go forward with it. The "threading" was interesting to write, basically background system() calls with semaphores, but it's pretty fragile.
The character_bind issue seems like it's maybe just a bad join call or something, with a null array member or something like that, generating an extra "," at the end instead of closing/finishing the query. Just a guess.
|