Hopefully fixes issue #127.
Wew, was some work! Conflicts: src/modes/battle/battle_finish.cpp
Closes #123 Conflicts: src/modes/battle/battle_finish.cpp
Made the GL calls be done after SDL_SetVideoMode() as it seemed to be needed on mac. Also tested it on Windows and Linux.
…cave. Now Kalya will clearly show where the exit is if the monsters haven't been all beaten. I also reduced the time the stone sign image is shown.
…tive ones. The correct object copy is already ensured by luabind anyway. And using plain member (autovars) makes the lua memory being freed more easily.
Threshold, joystick input enabled and the joy quit button are now configurable through the gui. I also prettified the settings.lua file written in the user config space.
- The save mode now clean the save_game1 table before reopening a save game script - It also disable slots where no data or invalid data is present at load time, as requested by qubodup. - When saving, it will warn whether a slot has got no data, or invalid data.
'clean' the lua global tables they're going to use before opening their script. This will permit to get rid of a lot of potential flaws linked to the lua stack data handling...
- Made the thread pointers be removed at file closure, to let lua handle them properly. - Fixed a potential crash in a table data type lookup. - Properly test whether a file is open before writing in it. - Properly check the number types as it couldn't do it before, and this could be useful for floats, for instance. - Added the foundation permitting to get rid of the content of a global table, thus making it safer to place new data in it. I'll use that in the next commits.
…cave1-2 stone sign dialogue. This shouldn't have led to crash but will look better anyway when playing using Kalya.
It is faster and will avoid names collisions that might have led to crashes.
Linked to #112: As shirish reported having 'virtual function called' based segfault, and as the bug happens within the function that reload the sprite, I've suspected the map sprite reference badly taken has a map object one, or something similar. If it's not the case, at least this will be a cause less to see that bug, or it will throw the error from within a c++ function, which is more talkative with gdb. A more general global variable hunt is also to be made also to possibly get rid of that bug as well as the others.
… the inventory list.
For me, the quest log is now actually very complete!!
as per #97, a couple remaining quest menu changes Features: + Added location and sub location image / description support + Added Completion Descriptions. This is a summary description that shows up when you complete the quest (in aqua! oooooOOooo....). note that in order to support this the description text box has been changed from 400px height to 200px height. Bug Fix: - Quest List had some problems with the pointer and "completed" quests. This should now be fixed - An edge case where a completed quest that had not been read yet was causing graphical inconsistencies. This is now fixed
…encounter. Also move slightly a tree that was annoying.
It's so great to see those in use. :)
- Changed a bit the world map cursor y offset. - Made the menu main options only loaded once, thus not resetting the cursor position on Reset() calls, and made the entries without content be disabled. - Add better Quest log entries, now the \n in text boxes is working fine. (Thanks Ikarus!!)