X-Git-Url: https://plomlompom.com/repos/index.html?a=blobdiff_plain;ds=inline;f=TODO;h=02283ba3462cd27066aafb195733a4f9056d71fd;hb=294583d0345a39cbca8c6c64a673f023826ee3e3;hp=04a1f0ec233a82c736854f8ae7ae7ad2578a9762;hpb=2ed0513d8eb44cb658107d0eb3c770b36dd7071b;p=plomrogue
diff --git a/TODO b/TODO
index 04a1f0e..02283ba 100644
--- a/TODO
+++ b/TODO
@@ -2,33 +2,30 @@ Next planned steps in plomrogue development:
BOTH SERVER/CLIENT:
-- implement better server/client architecture, meeting these criteria:
- - keep interaction of server and client in the filesystem (as currently
- happens by use of plain text files and fifos) to allow for easy scripting
- - allow for specific requests of the client for specific server data (such as:
- "what path do you propose for the player character to get from A to B?")
- - reduce hard disk i/o for server/client interaction (use fifos only?)
- - further avoid exits that keep the server running and the lock file existing
- when it shouldn't
+- expand use of hardcoded_strings module(s)
-- check for return values of *printf()
+- make server and client communicate by specific world state info requests
+ in server/out, replacing server/worldstate
-SERVER:
+- quote escaping in token_from_line()
+
+- do something about common crashes caused by server/client miscommunication
-- implement field of view / line of sight and obstacles for those on the map
+- think about refactoring occurrences of "free(...); ... = NULL;"
+
+SERVER:
-- get rid of the hard-coding of the MAGIC MEAT item type and its use mode
+- grow amount of possible things from UINT8_MAX to UINT16_MAX
-- is it actually useful to define map object action ids in the config file?
+- refactor occurrences of "if(...) { init_empty_map(...); }"
CLIENT:
-- enable toggling of window borders
+- re-work unnecessary complex command / keybinding / server message mapping
-- enable toggling of showing "\n" in the window content either as newline breaks
- (as currently) or as " / " (without newline break) for more efficent screen
- estate use
+- only send "ITEM_HERE" query to server when the dependent window is open
-- does get_n_of_keybs() really need to be its own function? (only used once)
+BUILD PROCESS:
-- transform Keybindings struct series from a linked list into an array
+- erlehmann's redo scripts are not OSX-compliant, see:
+