Previously logging logic was scattered over OS class implementations
with plenty of duplication. Major changes in this commit:
- Extracted logging logic into a separate Logger hierarchy. It allows
easy configuration of logging mechanism depending on compile-time or
run-time configuration.
- Implemented RotatedFileLogger which is usually used with StdLogger,
providing persistency of logs. It is often important to be able to
obtain logs of the game even in production to be able to understand
what happened prior to some problem. On mobile there previously was
no way to obtain the logs aside from having the device connected to
your machine.
- flush() is not performed in release mode for every logged line. It
is only performed for errors.
Reduz optimized field indexing in 3c85703 but the changes didn't apply
to dictionary so this code remained untouched. However, the logic for
validity checking was changed but not updated for the dictionary case.
This implement branch prediction macros likely() and unlikely() like in
Linux. When using these macros please ensure that when you use them the
condition in the branch really is very, very likely or unlikely. Think
90+% of the time. Primarily useful for error checking. (And I implement
these macros for all our error checking macros now)
See this article for more information:
https://kernelnewbies.org/FAQ/LikelyUnlikely
There are more places where these macros may make sense in renderer and
physics engine. Placing them will come in another commit down the line.
Uninitialzed values in GDScript are of type NIL so not allowing null
comparisons did end up breaking some code.
This commit reenables NULL equality checks for all types. We're going to
have to figure out how to make this fast for the compiler later.
Rename user facing methods and variables as well as the corresponding
C++ methods according to the folloming changes:
* pos -> position
* rot -> rotation
* loc -> location
C++ variables are left as is.
We now allow booleanization of all types. This means that empty versions
of all types now evaluate to false. So a Vector2(0,0), Dictionary(),
etc.
This allows you to write GDScript like:
if not Dictionary():
print("Empty dict")
Booleanization can now also no longer fail. There is no more valid flag,
this changes Variant and GDNative API.
After a short discussion with @reduz and @karroffel we decided to make
all non number/number comparisons return type errors on comparisons.
Now bool == bool is allowed but Vector2 == Vector3 is a type error and
no longer 'not equal'. The same has been done for the != operators.
In addition I forgot to add some failures to some Object operators
meaning that there was a potential for a crasher.
operator= does not need to call reference() if the new value is of the
same type as the old. This saves us zeroing the Variant, This speeds
up reuse of a Variant in a loop by roughly 50%.
In an effort to make GDScript a little faster replace the double
switch() with a computed goto on compilers that set __GNUC__. For
compilers that don't support computed goto it will fall back to regular
switch/case statements.
In addition disable using boolean values in a mathematical context. Now
boolean values can only be compared with other booleans. Booleans will
also no longer be coerced to integers.
This PR replaces #11308 and fixes#11291
Previously godot_variant_new_object constructed Variant without
accounting for the fact that the Object can be a Reference, so refcount
was not increased and References were destructed prematurely.
Also, Reference::init_ref did not propagate refcount increment to the
script instance, which led to desync of refcount info on the script
side and Godot side.
Color::html now expands 3 and 4 digit hex values into 6 and 8 digit
values by repeating each digit. This is to bring it in line with how
html handles these values
fixes#10997
After discussing with @reduz and @akien-mga it was decided that we do
not allow assignments or declarations in if statements. This PR removes
the instances of this I could find by automated means.
This fixes a source of many compiler warnings regarding comparing the
enum VARIANT_TYPE to Variant::Type. This changes the local value to a
static const Variant::Type value rather than an unrelated enum, this
also saves us a cast.
When doing large string concatenations the default push_back on the
String class can slow down things quite a bit. This is because it
has to constantly reallocate the memory and copy the contents. This
StringBuilder class delays the concatenation until the size of the
resulting string is known.
c812c17633 introduces some extra gdscript bindings for signal discovery
and adds a binding for has_method() to Script objects. This method is
already provided by the ancestor Object.
This fixes the startup message:
ERROR: bind_methodfi: Class Script already has a method has_method
At: core/class_db.cpp:1178.
Those functions were added in #8277 but they did more harm than good. They're subtle, don't do what people think and requires users to understand the non-uniqueness of polar decomposition. They ended up confusing people.
Until we store additional information enough to make a unique polar decomposition, these functions shouldn't be a part of Basis.
- Makes vararg methods automatically use PROPERTY_USAGE_NIL_IS_VARIANT on return types
- Completely removes the ":type" suffix for method names. Virtual methods must use the MethodInfo constructors that takes Variant::Type or PropertyHint as the first parameter for the return type (with CLASS_INFO as a helper to get the PropertyInfo). Parameters must use PROPERTY_HINT_RESOURCE_TYPE and hint string.
- PROPERTY_USAGE_NIL_IS_VARIANT is no longer needed for parameters, because parameters cannot be void.
- Adds missing PROPERTY_USAGE_NIL_IS_VARIANT to virtual and built-in methods that return Variant.
- The Windows, UWP, Android (on Windows) and Linux builds are
tested with Scons 3.0 alpha using Python 3.
- OSX and iOS should hopefully work but are not tested since
I don't have a Mac.
- Builds using SCons 2.5 and Python 2 should not be impacted.
Currently we rely on some undefined behavior when Object->cast_to() gets
called with a Null pointer. This used to work fine with GCC < 6 but
newer versions of GCC remove all codepaths in which the this pointer is
Null. However, the non-static cast_to() was supposed to be null safe.
This patch makes cast_to() Null safe and removes the now redundant Null
checks where they existed.
It is explained in this article: https://www.viva64.com/en/b/0226/
- Fixes some single-dash leftovers that were missed in the previous commit
- Reorder the help output for clarity, and document missing options
- Drop obsolete options: --noop, --pack, --editor-scene, --level, --import, --import-script, --no-quit
- Improve error message on malformed arguments and do not display help on error
- Always use long form of arguments when starting a new Godot process from C++, for clarity and easy grepping
- Cleanup obsolete code here and there
Ruslan Mustakov
Hein-Pieter van Braam
Marcelo Fernandez
Rémi Verschelde
Elliott Sales de Andrade
Juan Linietsky
m4nu3lf
Ross Hadden
Cradmon
letheed
ABU MD. MARUF SARKER
SaracenOne
Karroffel
Indah Sylvia
Konstantin Zaitsev
Thomas Herzog
Bastiaan Olij
Poommetee Ketson
bncastle
Ignacio Etcheverry
Wilson E. Alvarez
Maxim Sheronov
Geequlim
Dylan Enloe
mrawlingst
Scayze
Evgeny Zuev
Ignacio Etcheverry
toger5
Ferenc Arn
Matthias Hoelzl
Henrik Andersson
Andrii Doroshenko (Xrayez)
volzhs
Pedro J. Estébanez