kts
|
9f32631fe5
|
Added rolls test program. This reads rolls.tsc in and attempts to do some basic attack/defend rolls for Tiles. Also added the basic structure for the VM function precompiler. Next modification will likely be converting the current VM precompiling code into using a buffer rather than reading from the file handle directly. This will allow for forward reading of characters in addition to backwards, thereby allowing for easier expression parsing (at least towards what I am conceiving of it thus far).
|
2014-05-17 16:28:24 -07:00 |
|
kts
|
42de0d491c
|
Fixed up potential bugs and warnings revealed by clang.
|
2014-04-29 13:08:08 -07:00 |
|
kts
|
9bf0e3c7fa
|
newTile no longer takes x, y, and z params. Default values are set to 0 and are changed/set upon addTileToMap call.
|
2014-04-07 10:08:04 -07:00 |
|
kts
|
b45862f351
|
Windows PTime now uses native (non-SDL) millisecond precision for acquiring and delaying time.
|
2014-04-03 19:54:29 -07:00 |
|
kts
|
902e21ebc6
|
After far too long of a delay, a git commit. The updates pertain to vm_compile and implements the basic parsing of virtual machine C source files into global, group, and local scopes. At the moment it does not convert to op codes - it only identifies variables declared outside of functions for the given scope.
|
2014-04-02 00:37:31 -07:00 |
|
kts
|
fa7a158ec9
|
The Mystery of the git Commit
|
2014-03-14 14:51:41 -07:00 |
|
kts
|
e3941fa552
|
Updated devcpp project. Engine runs fine on win64.
|
2014-03-13 16:29:13 -07:00 |
|
kts
|
c1397f62fc
|
Fixed addTablePair bug wherein a new TablePair would overwrite the old rather than add itself to the end of the TablePair linked list. Also moved timer.c/.h from engine/sdl/ to engine/. If compiling on Windows, getPTime and doNanoSleep still require SDL.
|
2014-03-13 16:13:35 -07:00 |
|
kts
|
0842a2252e
|
Updated xcode project to compile the engine. tile_edtior and old engine probably will not compile in xcode at the moment, but this will be remedied later. Need to make modules directory definable in some way.
|
2014-03-13 04:10:59 -07:00 |
|
kts
|
6b0ef6fd83
|
Lots of fun abounded with segfaulting. Fixed it eventually - was accessing a state after it had already been freed, thus wreaking havoc in the next state. Maps and Tiles implemented now. Animation is a thing as well, but still a work in progress. It will likely be managed entirely by the Tile's doThink VM function. With modules now (partially loading) and Maps/Tiles figured out, the limited virtual machine will soon take shape.
|
2014-03-13 03:25:17 -07:00 |
|
kts
|
b33b2a534c
|
Added some (pretty bad) module loading code. At the moment modules/module_name/CONF is loaded (if not, it bails) into g_module_conf and the list of sprites and default scale is acquired. From this, g_module_spritesheets is malloced for (spritesheet count) * sizeof(struct Spritesheet*) and Spritesheets are created and loaded into their respective positions. NOTE: Spritesheets should be set up by the renderer in part - probably via function pointer, similar to Elements. After these are loaded, ../modules/module_name/module_name.tsd is attempted to be loaded as well. Not much error checking done yet, so eggshells abound. Module CONF error checking will probably be handled by some function when we first generate the list of modules in modules/ dir. ALSO, Spritesheet loading does not check sprites/CONF for width, height and columns - this will be fixed on next commit.
|
2014-03-12 03:46:46 -07:00 |
|
kts
|
d42b12885f
|
Added a load module state that provides a button list of modules found in the ../modules directory. When clicking on one, it will attempt to load ../modules/MODULE_NAME/MODULE_NAME.tsd into struct Data* g_tile_data. If it succeeds, it switches to a test state where I will be tweaking stuff like tile movement and similar. I will also add spritesheet loading in the load module state.
|
2014-03-11 23:56:23 -07:00 |
|
kts
|
722ceb5747
|
Reorganized the project root, deleted unnecessary files, moved data.c/.h, fio.c/.h, llist.c/.h into common and old code into old/ directory. All modules will now reside on the modules/ directory.
|
2014-03-10 12:41:49 -07:00 |
|
kts
|
53ffd0fce3
|
Added vm.c/.h in test. Implemented a good portion of the bytecode execution system. Need to implement non int data type operations. The VM is coming together\!
|
2014-03-06 23:54:19 -08:00 |
Vex
|
e9f2e33cf8
|
Created armor and weapons sprite sheets for Xibalba. Created system with Tiers(materials) on Y-axis and Types on X-axis. (ex. body slots) Also, created spell sprite sheet with system of Elements on Y and types (cone, bolt, ball, lode and combinations) on X.
|
2014-03-05 00:18:34 -08:00 |
|
kts
|
f9f65cd376
|
Added the linked list struct, LList, along with two functions for initializing and freeing a single LList. Also added the fio.c/.h files that will contain all functions pertaining to more complicated file or directory access. At the moment it provides two functions: dirToLList, which populates a linked list with the contents of the directory (as explained in the source), and fileExists, a simple function that checks if the file exists.
|
2014-03-03 22:48:00 -08:00 |
|
kts
|
1ebb045fef
|
Fixed the freezing issue on Windows - apparently it does not handle type conversions as fluidly as *nix. Changed SDL_GetTicks in getPTime to be assigned to int64_t rather than int. I suspect bugs still exist, but this will likely become a non-issue once a high precision timer(or as close as we can get) is implemented for Windows. Also it is a mystery why fps is capped to 60 - perhaps some Windows-specific OpenGL voodoo.
|
2014-03-03 13:14:23 -08:00 |
|
kts
|
452bdf4a1d
|
Cleaned up code in data.c and organized it. Current and future code will be formatted in adherence to the current draft of KCS. Also added datatype-specific writing and reading functions for TablePair values. This reduces the chance of unexpected behaviors when one datatype is interpreted as another.
|
2014-03-03 12:11:19 -08:00 |
|
kts
|
bbcd6cd0aa
|
More stuff added - how nondescript. In particular, data.c/.h now includes a function for loading in a .tsc, or timesynk configuration file, into a Table. This adheres to the information regarding the timesynk Configuration Syntax on the wiki. This feature has been implemented in the engine, as video width, height, and other similar options are loaded from the settings.tsc file relative to the engine. Also added a modules directory with a basic skeleton for the Xibalba module.
|
2014-03-02 22:10:49 -08:00 |
|
kts
|
1569e89715
|
Fixed elements rendering twice in state_menu. Added a moving background to state_menu for movement testing.
|
2014-02-26 18:43:05 -08:00 |
|
kts
|
2f596f4d50
|
Fixed PTime on OS X. OpenGL performance still terrible on OS X.
|
2014-02-26 16:36:12 -08:00 |
|
kts
|
ffa74bd574
|
Fixed the high precision timer and renamed it to PTime. It seems a bit silly to use nanoseconds for game tickrate, as many systems only have microsecond precision at best - but why not. Still have to add support for higher-than-millisecond precision on Windows - will likely run with microsecond, but more research is needed.
|
2014-02-26 14:22:02 -08:00 |
|
kts
|
c6efd2245b
|
Made some changes for the timer to work on OS X. It is likely unoptimized, but more research will be done. Also, for some unknown reason, the OpenGL renderer now runs slower than the terrible software renderer (in OS X). No clue why as of yet.
|
2014-02-25 03:31:27 -08:00 |
|
kts
|
9b5a1ed4b3
|
Added rendering code for sprite and fonts, made a pretty bad high precision timer - while being higher precision than SDL_GetTicks, it has a few problems that I must sort out. Also need to make it crossplatform.
|
2014-02-25 02:45:10 -08:00 |
|
kts
|
3634e0cfa3
|
Added engine as a target for the xcode project. SDL_Surface to OpenGL Texture function now packs pixel data according to the endianness of the machine, as acquired from a run-time test in the function. This should probably be checked for on start and have an appropriate system flag set.
|
2014-02-22 01:09:37 -08:00 |
|
kts
|
b3f50776e8
|
Lots of stuff added since last commit. ASCII-based Font system and Element system(similar to tile_editor)s now partially in place. State management now more flushed out, but still likely to be tweaked. Now sports two renderers - software-mode SDL and hardware-accelerated OpenGL mode. At the moment code is not in place to switch between the two renderers in-program (managed by a call to interfaceSetRenderer), but it will be added later with the appropriate freeing of surfaces and textures. A more generalized binary to .c/.h program was added to the new tools subdirectory - it takes a binary file as an argument and creates the appropriate .c/.h files. It also can handle the append flag, -a.
|
2014-02-22 00:20:44 -08:00 |
|
kts
|
478090d7c4
|
Began work on complete client rewrite within the engine/ subdirectory. This new client is written with game states in mind (i.e., main menu, net lobby, game) and a more expandable structure for multiple interfaces. Old code in the main directory will be removed entirely or reintegrated with the new engine. Majority of the code related to Tile(s) will be removed, as their properties are now dynamically loaded from TSD files (as generated by the tile_editor). Tiles will also have a local x and y coordinate for positioning within their currently occupied tile (or something very similar) - this will allow for smooth movement and more precise collision detection (according to quarters of tiles). Much more to come\!
|
2014-02-14 06:46:11 -08:00 |