* Versioned to 1.92 due to the bugfix
* Warning, if compiled in debug mode I have crap in there that makes stuff run slow
git-svn-id: svn://rtsoft.com/rtsvn/projects/RTDink@1632 353e56fe-9613-0410-8469-b96ad8e6f29c
* (html5) -s PRECISE_F32=2 was added to the compile, fixes "drunk midi". However, it hurts performance. This will be removed after FMOD fixes their stuff (see http://www.fmod.org/questions/question/midi-sounds-drunk-in-html5/ )
* (html5) Enabled Proton's persistent support for dink/dmod installs and saves
git-svn-id: svn://rtsoft.com/rtsvn/projects/RTDink@1621 353e56fe-9613-0410-8469-b96ad8e6f29c
* Added option to disable screen scroll transitions, one of my Android test devices has the slowest glread in the world, it's like 1 full second to walk to the next screen because of it. This might make dmods that
use timing (Myster island camera sequence for example) easier though, not sure.
* Fixed issue where behavior of two load_sequence commands in a row on the same sequence differed from 1.08 (in 1.08, max frame is set to the second one as long as the first one wasn't actually loaded yet. key now shows up in TGKA)
* Fixed regression where offsets in Mayhem were incorrect on the hand mouse pointer
* Upped max frames per seq to 100 instead of 50, fixes Echoes of the Ancient. I should probably make it dynamic but changing one const number is much easier!
* Now gives a clear warning in the log.txt if a seq tries to go beyond what it can handle and will truncate it (better than breaking the whole sequence as it did before)
* Added support for a weird bmp header type, fixes incorrect palette issue in the dmod The Orb of Darkness
* Fixed life/exp/health to correctly interpolate between values like the original dink, huge thanks to Dan's detailed bug report on this, he even made a mod which made testing the fix a breeze!
git-svn-id: svn://rtsoft.com/rtsvn/projects/RTDink@1520 353e56fe-9613-0410-8469-b96ad8e6f29c
* DMOD installer progress now switches to showing MB instead of K if the size is big
* BUGFIX: Fixed map loading bug that could crash the game (I added this one recently trying to clean up code to use more consts rather than magic #s, but there is always the risk I stupidely break something!)
* load_tile no longer instantly takes effect but requires a draw_screen or moving screens, this matches how 1.08 worked. Loading a save state or resizing the window will cause it to happen early, but hey, close enough
* BUGFIX: Fixed extra nasty bug where logic on certain things like charging your magic would pause the amount of time you used TAB to skip time
git-svn-id: svn://rtsoft.com/rtsvn/projects/RTDink@1517 353e56fe-9613-0410-8469-b96ad8e6f29c
* Dialog box rendering was off a few pixels due to me forgetting to remove some tests I did, fixed
* Black at index 255 and white at index 0 is now forced during bmp loading. Windows does it, photoshop doesn't, but older versions of Dink (directx) seemed to do
it so going with that
* Invalid sprite sent to freeze/unfreeze will no longer crash the game (bug was also in original 1.08)
* Added hardening to check valid input on many script functions, will stop "sometimes" crashes based on bad scripting as well as log them
* Crash logging should show correct function names. I mean it this time!
- Some mobile related files added now that I'm starting to work on that side too
git-svn-id: svn://rtsoft.com/rtsvn/projects/RTDink@1512 353e56fe-9613-0410-8469-b96ad8e6f29c
