xazax

Hmm I wonder if it worth, to give users the flexibility, to create custom settings for groups. For example "featureX: true", and script could query that. Of course script could also check for group name/id, but then what's with inherited groups also check for those?

I vote for static linking. It hasn't got a big size. Anyways I think a command instead of @adjgmlvl would be needed (well, not really needed, but it would be nice to not loose functionality), to add a player to a group. I hope this does not mean IO load on each atcommand. I feel a bit sad, we can not port other confs to use libconfig =(. Edit: jman and gepard was faster adjgmlvl concerns are answerd

612

603

592

585

Thanks for the clarifications, yeah, thinking about it, one value for atcommand only makes more sense. It would be a pain to manually disable charcommands for every command in players' groups. Anyways does porting the whole conf folder to libconf involved in this project? It is not in the progress list.

Is multiple inheritance supported? if so, [ ] should be used, to reflect it is a list actually (like aliases). atcommand: true, atcommand: false? I assume all atcommands are off by default. Does false means disable an inherited command? In that case we can warn users if there is no such command enabled in the base groups ( because it may be a typo ). I guess [False, true] is for atcommand, charcommand. If only one value present it is both at and charcommand I guess. I like the concept so far. Anyways server owners (don't need to be a core dev) out there. Someone who feels like working on default/recommended groups I'm bet Gepard would welcome if you'd share it

In short: priority. Right now we are in cleaning up phase. Want to triage and fix as many old eA bugs as possible before starting to add new content again.

If there is a new bonus noone force you to use it. So ballance is not an argument here IMO. However the smaller the code the easier to maintain. This is a very strong reason to keep everything that is unofficial and not widely used away from rA.

575

567

565

562

Considering trailing semicolons are optional, I vote yes for libconfig. It may also help to make new atcommand alias reading cleaner.

Well, I don't really know if semicolons in a config file is desirable. Users may think it is a language and somehow interpreted and use experssions. And it does not support multi threading. Of course rA is not multi threaded right now, but at least do not create new obstacles for someone who decide to work on this area in the future. (Although probably we will not want to access a config file from multiple threads, so my first argument is more important.) Anyways it would be awesome if the config lib we choose would natively support importing configs. Edit: Just found out, it supports @include

558

546

541

532

529

We are no longer support old versions of visual studio. We do not want to maintain that amount of project files. However you can download and install the express version of a newer visual studio. It is small, free and easy to find.

519

515

511