* refactor(Core/ChatHandler): Cleanup retriving session and checking for valid session
* redundant
* make it clear what a "valid" session is
* HasValidSession -> HasSession
* Differentiate from "valid" which in this case would mean the player is in world which isnt checked here
* refactor(Core/World): Move SendGMText to WorldSession and allow `fmt`
- Move SendGMText from World to WorldSession
- Make SendGMText use fmt
- Make SendGMText parse acore_string entries
* Update cs_message.cpp
* tokenize the string only once
* Move to chathandler
* Update WorldSession.cpp
* make sure we have a session
* Config - Deprecated for 3 years
* ChatCommand - Deprecated for 2 years
* getLevel() - Deprecated for 2 years
* SendGlobalText - Deprecated for 8 years +
* feat(Core/Chat): Provide a fully-formed protocol for addons to interact with GM commands
* Send success/fail state, allow interleaving, and indicate end of output. Add framework for supporting non-human-readable output in commands.
* cherry-pick commit (508c9d2fc1)
This PR implements a well-formed protocol for addons to communicate with the server, outlined below:
- All communication happens over the addon channel (`LANG_ADDON` in terms of the core, `CHAT_MSG_ADDON`/`SendAddonMessage` for the client). The prefix used for all messages is `AzerothCore` (in client terms - in core terms, every message starts with `AzerothCore\t`).
- In each message, the first character is the opcode. The following four characters are a unique identifier for the invocation in question, and will be echoed back by the server in every message related to that invocation. Following is the message body, if any.
- The following opcodes are supported:
- Client to server:
- `p` - Ping request. The core will always respond by ACKing with the passed identifier. No body.
- `i` or `h` - Command invocation. The message body is the command text without prefix. `i` requests machine-readable output, `h` requests human-readable.
- Server to client:
- `a` - ACK. The first message sent in response to any invocation (before any output). No body.
- `m` - Message. Sent once per line of output the server generates. Body = output line.
- `o` - OK. Indicates that the command finished processing with no errors. No body.
- `f` - Failed. Indicates that command processing is done, but there was an error. No body.
Expected overhead is minimal, and this integrates seamlessly with existing command scripts (no changes necessary).
PS: There's also a client-side addon library that exposes this protocol in a developer-friendly way over at https://github.com/azerothcore/LibAzerothCore-1.0
---------
Co-authored-by: Treeston <14020072+Treeston@users.noreply.github.com>
* Fix CanExecuteCommand only being executed on command failure.
* Updated for style guide.
* Removed pointless hook call.
* Modified the name of the hook to better reflect function and moved the call to trigger for non-existent commands.
* Updated to OnTryExecuteCommand because Foe has a special place in my heart.
* Forgot to change these 2 areas.
