Also:
recreate the window when enabling resizable window, same workaround I applied to my mod for jacob1/The-Powder-Toy#24
recreate the window when toggling fullscreen, to work around some bugs where the window is sized incorrectly or the mouse position doesn't match where the cursor is
Don't capture mouse cursor when this is a debug build
Update save format to optionally store type as two bits
PIPE now stores element in ctype
Disallow uploading saves using two bytes in type or other fields
update save format to store pmapbits and automatically convert data
When making a local save, then stamping and loading the stamp (without doing anything else in between), issues with the code would cause the unsigned int in the json to turn into a normal int. It would then think the authors data from the stamp was from something else and append it. For users that do an excessive amount of stamping, this causes duplication in the authors links that shouldn't have ever happened.
The logging is saved inside the bson data in all online saves, local saves, stamps, and clipboard pieces. It is loaded back when reloading each of those.
See #474 for the format of the data. It is the same format for the bson data. Note that "links" is an array of objects. It can be recursive.
There is some effort to not duplicate information, we don't care if you loaded a stamp 10 times or if you are using the clipboard in your own save. Extra information is mostly not saved for your own stuff, only when you take material from other saves.
Press ctrl+a in debug builds to show what info it is currently saving in Client. Also enabled in snapshots for now.
There is one unrelated change in here, which fixes a crash pointed out by QuanTech. It was also save related and it was too close to the other changes to separate it into another commit. It fixes a crash when saving signs with invalid unicode. the BSON library doesn't like this, it was returning an error but we ignored it, which caused a crash. It now notices those errors. I also had to update several Serialize calls to check if it actually returned save data, or else it then would have started crashing there instead.
Also some debug prints were removed
It is undefined behavior what happens when you try to call pthread_join on a NULL thread (which will happen if you close the game before opening the save browser). Seems to be harmless, but make sure we don't do that just in case --jacob1
Fixes cross compilers, TODO: I don't think scons can check if a flag exists but might support adding a test for that, even if through hacky means, like CheckFramework
The crash when opening saves on mac is definitely something to do with the threading, and i'd rather just test this first to see if it fixes it
Copied from my mod almost entirely as-is, with the changes to HTTP.cpp included, also added locks into Download.cpp even though it is probably overkill
something is up with the pthreads library (dynamic compiles still broken), probably windows libraries all need to be recompiled and distributed separately like I did for visual studio
cajun is a hard to use library with many necessary features missing. It also has been causing the windows version, and probably the mac version, to randomly crash.
Other json stuff still uses cajun (TODO: remove)
you can get current OS / build information, exe name, restart tpt, open a link in the web browser, and use clipboard copy/paste functions
Also remove some older Platform.h file which wasn't really needed or used
by rounding temperatures to nearest instead of always flooring.
Effect on DLAY delay:
Room temperature offset - loaded as whole number of degrees C, so rounding to nearest deg C gives the same number as flooring.
Full - loaded as whole number of K = original whole deg C - 0.15. Rounding to nearest deg C means the whole deg C delay from before saving is used as the number of frames to delay, instead of deg C - 1 if flooring was used.
There are *no* more spacing issues. I used tabs exclusively.
Also, I made changes off of the develop branch.
Here's the Ruby regex that I used:
text.gsub!(/(( |\t)*)if ?\((.*)\) ?{?\n?( |\t)*(free ?\(|delete |delete\[\] )(.*)\)?;/, "\\1\\5\\6;")
I've left some of the less important items, like SearchView.cpp "'then' statement is equivalent to the 'else' statement", and RequestBroker::Request::~Request, because I don't feel like spending a few days entirely rewriting those files at the moment (which is what I'd end up doing if I started fixing minor problems and refactoring...)
GameSave::readOPS - not changed. At some point we may have to move to a larger type for element IDs (probably two or four bytes), but PT_NUM isn't likely to be raised to the maximum value of that type immediately, so this check will be needed then. There should be an elements[partsData[i]].Enabled check in there too, but it might be a bit difficult - I'm not sure how to access a Simulation object from GameSave::readOPS...
Notes on changes:
Graphics::textsize, Element_FRZW::Element_FRZW - typos
Button::Draw - the extra case was originally used to invert the icon (draw it in black instead of in white) when the button was clicked. However, the icon colour is now automatically set depending on the background colour. (Note similar conditions "if(Enabled) { if(isButtonDown || (isTogglable && toggle)) " near the start of the function - same logic but in a different place, setting icon colour indirectly).
Simulation::transform_save - unused redundant function, everything uses GameSave::Transform which does much the same thing.
PreviewView::NotifySaveChanged - should be height==YRES/2, it's checking whether the preview image is the correct size, and resizing it if it isn't.
Element_FWRK::update - no idea why that line was there, even though it was my commit that originally added it...
