On Tue, Aug 5, 2008 at 5:26 PM, Aaron Mavrinac <[EMAIL PROTECTED]> wrote: > AI developers, could you please take a look at the following: > > http://thousandparsec.net/wiki/Single_Player#AI_Clients > > and let me know if (and how) this is doable with your client? At a > minimum, I need to be able to start and stop it non-interactively from > a shell script, specifying the port to connect to, the ruleset to use > (if necessary), and a username for the client to use (if possible). > Some way of probing whether it started and connected successfully > would be nice also.
For daneel-ai: * start from a shell script should be easy * stopping might be hard, unless you gather the process number and send it SIGINT or something. I'm not entirely sure how you want to see this supported otherwise? * port and username are collected together with password and server (localhost I suppose) in a connection URI that can be passed on the command line * ruleset is a bit harder. daneel-ai uses configuration files to specify behavior. There is no theoretical link between configuration files and rulesets, although usually a configuration file is only used for a single ruleset. However one ruleset might (and will) have different configuration files - different strategies, you could say. Would it be possible to keep a ruleset->configurations mapping and then, after making a ruleset selection, make a second selection from the configurations as parameter or something? The problem here is that the options for configuration files are depending on the selection of a ruleset, I don't think you support that? * I'm not sure about probing. However, if it doesn't start or connect, it should just raise an exception and quit. If you want true interprocess communication, I could make some modifications or add a wrapper script I guess. No idea how to start on that though. Iwanowitch _______________________________________________ tp-devel mailing list [email protected] http://www.thousandparsec.net/tp/mailman.php/listinfo/tp-devel
