protocol looks good to me. I would just suggest using static const std::string = "COMMANDNAME". #defines are not typesafe and have no benefits in this case (IMHO). Otherwise the protocol looks good to me. I am also pro ignoring translation of motd and therelike. The other messages should imho be templates that are translated inside widelands.
I agree with hjd that a testsuite would be desirable. The only way to do this properly is to run the "server" locally and propagate it with "fake" user and game data, then let the client run some automated tests against this. This would be easiest if the internet protocol where a stand alone library that is only linked against widelands. Obviously that would be desirable for all parts of widelands but in this case we can really achieve this because it is a small seperated part inside the code - especially after the refactoring. Could you try this out, peter? -- You received this bug notification because you are a member of Widelands Developers, which is a bug assignee. https://bugs.launchpad.net/bugs/916545 Title: About depending on GGZ packages Status in Widelands: Confirmed Status in “widelands” package in Debian: New Bug description: I noticed a bug report in Debian which basically suggests to drop the GGZ packages because it is unmaintained upstreams. Since Widelands depends on ggz, this of course affects us there (not to mention Debian derivaties like Ubuntu and Linux Mint). I know the ggz dependency has been discussed a bit before, but we should probably figure out what the long-term plan is. To manage notifications about this bug go to: https://bugs.launchpad.net/widelands/+bug/916545/+subscriptions _______________________________________________ Mailing list: https://launchpad.net/~widelands-dev Post to : [email protected] Unsubscribe : https://launchpad.net/~widelands-dev More help : https://help.launchpad.net/ListHelp
