Fixes exported property modified values lost when creating a placeholder script instance with a failed script compilation
- Object set/get will call PlaceHolderScriptInstance's new fallback set/get methods as a last resort. This way, placeholder script instances can keep the values for storage or until the script is compiled successfuly.
- Script::can_instance() will only return true if a real script instance can be created. Otherwise, in the case of placeholder script instances, it will return false.
- Object::set_script(script) is now in charge of requesting the creation of placeholder script instances. It's no longer Script::instance_create(owner)'s duty.
- PlaceHolderScriptInstance has a new method set_build_failed(bool) to determine whether it should call into its script methods or not.
- Fixed a few problems during reloading of C# scripts.
- Refactored all builder (make_*) functions into separate Python modules along to the build tree
- Introduced utility function to wrap all invocations on Windows, but does not change it elsewhere
- Introduced stub to use the builders module as a stand alone script and invoke a selected function
There is a problem with file handles related to writing generated content (*.gen.h and *.gen.cpp)
on Windows, which randomly causes a SHARING VIOLATION error to the compiler resulting in flaky
builds. Running all such content generators in a new subprocess instead of directly inside the
build script works around the issue.
Yes, I tried the multiprocessing module. It did not work due to conflict with SCons on cPickle.
Suggested workaround did not fully work either.
Using the run_in_subprocess wrapper on osx and x11 platforms as well for consistency. In case of
running a cross-compilation on Windows they would still be used, but likely it will not happen
in practice. What counts is that the build itself is running on which platform, not the target
platform.
Some generated files are written directly in an SConstruct or SCsub file, before the parallel build starts. They don't need to be written in a subprocess, apparently, so I left them as is.
This commit makes operator[] on Vector const and adds a write proxy to it. From
now on writes to Vectors need to happen through the .write proxy. So for
instance:
Vector<int> vec;
vec.push_back(10);
std::cout << vec[0] << std::endl;
vec.write[0] = 20;
Failing to use the .write proxy will cause a compilation error.
In addition COWable datatypes can now embed a CowData pointer to their data.
This means that String, CharString, and VMap no longer use or derive from
Vector.
_ALWAYS_INLINE_ and _FORCE_INLINE_ are now equivalent for debug and non-debug
builds. This is a lot faster for Vector in the editor and while running tests.
The reason why this difference used to exist is because force-inlined methods
used to give a bad debugging experience. After extensive testing with modern
compilers this is no longer the case.
* Prototype faster function call ux.
* Work on general search ux.
* Able to create nodes from search.
* Show class for variables but not methods.
* Get actions search working.
* Descriptions now show for both methods and properties.
* Enable zooming on mouse wheel up and down.
* Make the drag trigger on right mouse button.
* Search now shows for action visual script nodes.
* Able to search visual node names.
* Search works better.
* Change zooming scale to hide artifacts better.
* Remove zoom changes
* Select from base should check properties too like the other functions.
* Seq_connect flag is needed to set sequence lines correctly.
* Remove comment
* Code cleanup with function names and arguments.
* Use brief description for search descriptions.
* Clean and fix bug with input nodes connecting with sequence lines.
* Add a warning and fix some edge conditions with sequence into data lines and vice versa.
* Don't search functions when pulling from a sequence node.
* Don't show actions when pulling from a data line.
* Set set and get properties.
* Convert visual script operators to the correct type
* Create a function preset finds only functions.
* Singletons can now find functions.
* Add shift-a for generic search.
* Add brief descriptions for Visual Script nodes.
* Search boxes can now filter names.
* Add bigger hit zones to node connect.
* For the drop zones, make all the rect2 areas the same size.
* Function names in visual script node should be lower case so that search works better.
* Use the convention of capitalize() for set, set, visual script nodes and methods.
* Make search more general. Ignore "_" and make case-insensitive. Also made the search window smaller and remove extra info from search
* Make type_cast use the connecting node's type and remove use of found variable.
* Fix case where you call an instance's call function where it becomes an invalid call.
* Make get_visual_node_names use a set of filters, move action creations and fix bug with sequence node connections.
* Make the window bigger.
* Make connect_data and connect_seq more robust.
* Add icons to search items.
* Add vs constructors in shift-a menu.
* Operators, builtins and constructors show type name. Fix several problems with port connections.
* In shift-a mode search everything.
* Code cleanup
* Work on autocompleting the type.
* Use type guess in action creation.
* Check if type hint string exists in object variables when generating the visual script search.
* Add the hint to SceneTree.
* Add original type detection.
* Make type casting great again. This puts the type casted base type as the data output type string hint.
* Pass the type in a VisualScriptFunctionCall too.
* Set the base type correctly in VisualScriptFunctionGet and VisualScriptFunctionSet using hint string.
* Make sure the instance is passed in VisualScriptPropertySet.
* Restore search on the node's type.
* Remove dependencies from graph_edit.
* Remove dependencies from property_selector and name the class visual_script_property_selector.
* Extract hot zones into a function.
* Move hot_zones constants into default theme.
* Bigger capture zones.
* Clean messy port_grab_distance variables.
* Remove RMB functionality.
* Remove memory leak on showing visual script descriptions.
* Read the port_grab_distance constants on enter tree and theme changed.
I've removed the section about being unable to export games using C# - as you are now able to do this, as long as the export templates are installed. Also, I've made a few minor grammar tweaks.
- Use data type struct from the parser.
- Avail from type hints when type can't be guessed.
- Consider inner classes and other scripts when looking for candidates.
- Allow type hints to be completed.
- Use type information to infer completion candidates.
- Show typed function signature in tooltip.
- Add type hints when completing declaration from virtual functions
(optional).
- Typed assignment (built-in, native, and script).
- Cast (built-in conversion; native and script checks).
- Check type of functions arguments on call.
- Check type of members on set.
- Resolve types for all identifiers.
- Error when identifier is not found.
- Match return type and error when not returning a value when it should.
- Check unreachable code (code after sure return).
- Match argument count and types for function calls.
- Determine if return type of function call matches the assignment.
- Do static type check with match statement when possible.
- Use type hints to determine export type.
- Check compatibility between type hint and explicit export type.
On macOS, it is common to install packages like Mono through the third-party
package-manager Homebrew. This commit simply adds an additional path to
where Homebrew installs the Mono framework.
Theora and WebM video streams were mistakenly imported with a ResourceImporter,
but those imported ogvstr and webmstr were simply links to the local resource.
While that works fine in the editor, it no longer works when exporting a game
as the "source" ogv and webm files are ommitted and only the ogvstr and webmstr
references were exported.
As discussed with @reduz, it doesn't make sense to import videos, as we only
intend to play them back and not modify them/access their raw data. As such we
use a ResourceFormatLoader instead of an importer, to load the file on the fly.
ogv and webm files linked to this loader are now considered as resources, and
thus exported.
Note: The Theora and WebM loaders lack any kind of validity check beyond the
existence of the target file, but it was already the case with the importer.
Better checks and error reports could be added, but those loaders will eventually
be obsoleted by GDNative plugins anyway.
Fixes#14954.
- Add (Csc/Vbc/Fsc)ToolExe environment variables when running Mono's MSBuild.
- Fix directory for the 'mono_assemblies_output_dir' argument being created with the '#' top level directory token as part of its name.
- Allow to build with 'mono_static=yes' on Unix without specifying a mono prefix. The build script will try to find the mono prefix using the output from pkg-config.
