The bug was being caused by a lack of a new player object being sent to
some clients. Fixed this by adding in the sending code, and preventing the
new entries being "registered" until the end of the HandleJoinRequest()
method.
Hopefully I can simply abstract away some of this code soon.
The bug was being caused by not assigning default values to the player
objects' internals. I've added a quick fix to these structs, but I think
it's best to expand them into proper classes now.
I've divided the server's source into several different files. This should hopefully make it a bit easier to manage.
Please note that I do know there's an exception being thrown by the client
in this build.
I've created a separate branch for this because it's gonna be a bitch to
get compiled, and then I'll still have to ensure that the client & server
are work together correctly.
This build does not compile.
I've added the handle and avatar fields back into the PlayerInfo section
in the network code, because I need to be able to load a specific file
when a new player is created. This wasn't forseen, but it's fine. i'm
leaving the fields in ClientInfo as well, because LobbyMenu is using them
to login to the server.
PlayerIndex is now a shared parameter.
I've shifted some code around in InWorld, however the overall logic is the
same.
This build (as well as the last) does not compile.
The serialization code is now using macros to simplify each line, and to
prevent errors. It should be noted that, apart from the region content,
the serialization and deserialization code is essentially identical.
The list in todo.txt are a list of tasks that will be needed to complete
the login/user accounts overhaul, and will hopefully make the server a
persistent system.
Once the server is persistent (via the database) I can start on the combat
system.
This branch was originally intended to rewrite the login system, but I've discovered some serious problems that could not wait. See the previous commit for more information. Hopefully I can get back to this soon.
I believe this version deserves a new release tag.
Coordinates are supposed to be stored by their x & y, but the sizes of the
tile sheets were getting mixed in as well. When trying to store a region
at (0, 20), it was being stored at (0, 260).
Another visual bug inside the tile sheet's rendering function masked this
issue until now.
Another thing to note is that I've removed an incredibly complex system
for updating the client's map. The new system may seem complex at first,
but it is straight forward compared to what it replaced.
I've also fixed a few logarithmic lagging points in the code.
After refactoring the map system, this new build uses the preprocessor macros to define the sizes of each region object. These macros, which are defined in region.hpp, were being used anyway; these modifications simply speed up the process by cutting out a lot of the fat.
It seems that the serialization code has a bug in it. I was expecting
something like this. When the server tries to send the region content, it
exits. I'll try and find the cause of the error, but I'm committing my
changes anyway.
The region's width, height and depth are all defined by preprocessor
macros. The rest of the map system has been updated to match. The
programs proper need to be updated as well. It would be a good idea to
include the macros' values as part of the initial communication protocols,
so that the clients don't connect to a server that is using the wrong
sized regions.
This algorithm is designed to find the distance of a certain region,
however it may have been better to do a quick check, rather than worrying
how many regions were loaded at once.
I've also added in some debug code to the map classes, because I was
hunting down a std::bad_alloc beingthrown. Turns out I forgot to set the
map sizes in the client's InWorld constructor. I'm committing the fix, and
the debug code.
I've added lua hooks for both pager functor classes. Hopefully, I haven't
missed any corner cases, because it took me a while to hunt everything
down. One issue is that the map's save directory needs to be set in the
Format class, but it'll do for now. I'll review this again when I've got
more than one map running at one time.
There should be enough here for a lua-driven map generator to be
implemented, even if it's a bit rough. I think I'll test this out in the
editor eventually, but getting the base branch's network map code going
comes first.
The current process is extremely convulted, so I need to document
everything that I've done so far, including C++ and lua functions.
Kayne Ruse