I've added a simple edge-smoothing function to debug island's generator. It
doesn't handle all edge cases (pun intended), but the proof of concept is
sound. I just wish I could release this...
I've also added exception checks to region.cpp
I've also adjusted the TODO file, which really shouldn't be committed.
Still, it's there now, so it stays. I don't think the logger is going to
be possible any time soon, so I'll probably look into the disconnection
handler.
Project compiles and runs, multiplayer works, but the map is not visible
There are a number of major changes to the map system, beth directly from the
jam branch and from the splicing process. The changes that are listed here have
been noted as they were added to the index for committing.
* tile_sheet.*pp moved from 'common/graphics/' to 'common/map/'
* Minor method and member name changes to TileSheet
* TileSheet has a lua API, but it isn't used anywhere
NOTE: Nothing uses both lua and graphics, but a theoretical editor might
* Region API's glue functions have been changed from triggers to simple dummy
methods. These should simply be over written.
* RegionPagerBase::GetRegion(int x, int y) now snaps it's parameters
Presicely why is unknown, but I do remember there was a bug without it
* RegionPagerLua has been rewritten to use the Region API's methods, rather
than the triggers.
* RegionPagerLua no longer stores the map's save directory, or passes it to the
the Region API's methods.
* RegionPagerLua::luaState renamed to RegionPagerLua::lua
conforms to changes elsewhere
* Removed the directory glue functions from the RegionPager API
* region_pager_api.hpp preprocessor guard changed
* Adjusted makefiles to account for TileSheet's movement
The APIs have access to the solid data, and I fixed a bug: Basically, the
template parameter for std::bitset expects an integer representing the
number of bits to hold, but I initially misread it as the number of bytes.
This has been corrected.
I've also added a sandy beach to the generated island. I'm tempted to
start working on some simple generators soon.
region_api.cpp had a bug, where a glue function's name was used twice. It
was an easy catch, but there was an issue in the new script, where I was
counting from 0 instead of 1. As a result, I was chasing a segfault for 5
hours.
I've changed some naming concentions in the lua APIs. I've also made a few
other tweaks, like region_pager_api.cpp delegating to the passed
RegionPager object. This won't explicitly run, becuase there's still a few
more changes needed.
The changes were fairly easy, and I was actually able to find and fix an
out-of-sequence bug: The destructors for the AccountManager and the
CharacterManager were using SQL code after the call to SQLite3_close_v2(),
so I added and called UnloadAll() for both of them.
Kayne Ruse