This is still a bit hacky and eventually we should rework the way we handle
optional dependencies (especially with regard to builtin/system libs), but
it's a simple first step.
Fixes#39219.
The insertion order for dictionaries is only a language feature for
Python 3.6/3.7+ implementations, and not prior to that.
This ensures that the engine won't be rebuilt if the order of detected
modules changes in any way, as the `OrderedDict` should guarantee
inerstion order.
The existence of `SCsub` is checked instead. This file is required for
all modules, and prevents the build system to leave modules without
`config.py` undetected, leading to silently ignoring the module during
compilation.
This patch adds ability to include external, user-defined C++ modules
to be compiled as part of Godot via `custom_modules` build option
which can be passed to `scons`.
```
scons platform=x11 tools=yes custom_modules="../project/modules"
```
Features:
- detects all available modules under `custom_modules` directory the
same way as it does for built-in modules (not recursive);
- works with both relative and absolute paths on the filesystem;
- multiple search paths can be specified as a comma-separated list.
Module custom documentation and editor icons collection and generation
process is adapted to work with absolute paths needed by such modules.
Also fixed doctool bug mixing absolute and relative paths respectively.
Implementation details:
- `env.module_list` is a dictionary now, which holds both module name as
key and either a relative or absolute path to a module as a value.
- `methods.detect_modules` is run twice: once for built-in modules, and
second for external modules, all combined later.
- `methods.detect_modules` was not doing what it says on the tin. It is
split into `detect_modules` which collects a list of available modules
and `write_modules` which generates `register_types` sources for each.
- whether a module is built-in or external is distinguished by relative
or absolute paths respectively. `custom_modules` scons converter
ensures that the path is absolute even if relative path is supplied,
including expanding user paths and symbolic links.
- treats the parent directory as if it was Godot's base directory, so
that there's no need to change include paths in cases where custom
modules are included as dependencies in other modules.
Some required changes are made:
- locally imported SCons-specific packages within the method;
- `global` variables converted to `nonlocal` (used in nested functions).
Configured for a max line length of 120 characters.
psf/black is very opinionated and purposely doesn't leave much room for
configuration. The output is mostly OK so that should be fine for us,
but some things worth noting:
- Manually wrapped strings will be reflowed, so by using a line length
of 120 for the sake of preserving readability for our long command
calls, it also means that some manually wrapped strings are back on
the same line and should be manually merged again.
- Code generators using string concatenation extensively look awful,
since black puts each operand on a single line. We need to refactor
these generators to use more pythonic string formatting, for which
many options are available (`%`, `format` or f-strings).
- CI checks and a pre-commit hook will be added to ensure that future
buildsystem changes are well-formatted.
Scons' `Environment.subst()` does that, and was already used in the
other place where we query an env variable (`env["LINK"]` in x11 code).
Fixes `3.2` iOS build after cherry-pick of #36559 (previously it only
ran for GCC code, not iOS's Clang), and the same issue would likely
affect `master` if iOS builds were enabled right now.
We already had `MODULE_*_ENABLED` defines but only in the modules
environment, and a few custom `*_ENABLED` defines in the main env
when we needed the information in core.
Now this is defined in a single header which can be included in the
files that need this information.
Unify pack file version and magic to avoid hardcoded literals.
`version.py` now always includes `patch` even for the first release in
a new stable branch (e.g. 3.2). The public name stays without the patch
number, but `Engine.get_version_info()` already included `patch == 0`,
and we can remove some extra handling of undefined `VERSION_PATCH` this
way.
Co-authored-by: Rémi Verschelde <rverschelde@gmail.com>
The new 'split_libmodules=yes' option is useful to work around linker
command line size limitations when linking a huge number of objects.
We're currently over 64k chars when linking libmodules.a on Windows
with MinGW, which triggers issues as seen in #30892.
Even on Linux, we can also reach linker command line size limitations
by adding more custom modules.
We force this option to True for MinGW on Windows, which fixes#30892.
Additional changes to lib splitting:
- Fix linking of the split module libs with interdependent symbols,
hacking our way into LINKCOM and SHLINKCOM to set the `--start-group`
and `--end-group` flags.
- Fix Python 3 compatibility in `methods.split_lib()`.
- Drop seemingly obsolete condition for 'msys' on 'posix'.
- Drop the unnecessary 'split_drivers' as the drivers lib is no longer
too big since we moved all thirdparty builds to modules.
Co-authored-by: Hein-Pieter van Braam-Stewart <hp@tmm.cx>
When checking mingw-w64 version, at least on debian, the regex being used returned 86 because the name of the binary in debian starts with x86_64-w64 so we use the dumpversion option that gcc has. This fixes not compiling because gcc versions < 7 don't have some checks like shadow-local
Also added support for SCons project-absolute paths (starting with #) and
warning about duplicates in add_source_files(), and fixed
default_controller_mappings.gen.cpp being included twice after first build
due to *.cpp globbing.
Part of #30270.
It's the recommended way to set those, and is more portable
(automatically prepends -D for GCC/Clang and /D for MSVC).
We still use CPPFLAGS for some pre-processor flags which are not
defines.
Also include website URL and make it configurable via version.py
together with the rest of the engine branding.
Add mention to MIT license in --help output.
Many contributors (me included) did not fully understand what CCFLAGS,
CXXFLAGS and CPPFLAGS refer to exactly, and were thus not using them
in the way they are intended to be.
As per the SCons manual: https://www.scons.org/doc/HTML/scons-user/apa.html
- CCFLAGS: General options that are passed to the C and C++ compilers.
- CFLAGS: General options that are passed to the C compiler (C only;
not C++).
- CXXFLAGS: General options that are passed to the C++ compiler. By
default, this includes the value of $CCFLAGS, so that setting
$CCFLAGS affects both C and C++ compilation.
- CPPFLAGS: User-specified C preprocessor options. These will be
included in any command that uses the C preprocessor, including not
just compilation of C and C++ source files [...], but also [...]
Fortran [...] and [...] assembly language source file[s].
TL;DR: Compiler options go to CCFLAGS, unless they must be restricted
to either C (CFLAGS) or C++ (CXXFLAGS). Preprocessor defines go to
CPPFLAGS.
If it needs to be hardcoded (for the sake of reproducible builds),
it should be together with the other hardcoded version info.
And yeah, two months in, let's move to 2019.
so that godot package builds reproducibly
in spite of indeterministic filesystem readdir order
and http://bugs.python.org/issue30461
See https://reproducible-builds.org/ for why this is good.
Sort font input file list, so that builtin_fonts.gen.h
is created in a reproducible way
Sort list of platforms, so that editor/register_exporters.gen.cpp
is created in a reproducible way
Sort list of source files, so that .a files and resulting godot binaries
are created in a reproducible way
This value becomes part of get_version_info output,
but if it is changing every year without any other change,
it cannot be a useful indicator of anything.
Using a constant value, makes the package build reproducible.
See https://reproducible-builds.org/ for why this is good.
Previously the compiler would use system headers located at
/System/Library/Frameworks, which could result in compilation failures
due to the headers not always being up-to-date in regards to the
latest installed macOS SDK headers that come with Xcode.
Fix the issue by passing the SDK path via the -isysroot option to the
compiler and linker invocations.
If no custom SDK path is given, the build system queries the SDK path
via xcrun --show-sdk-path, which returns something similar to
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/
/Developer/SDKs/MacOSX.sdk/
Querying via xcrun is now also done for iphone (and simulator)
platforms as well.
Here is an example of a compilation failure message due to outdated
headers:
platform/osx/os_osx.mm:1421:41: error: use of undeclared identifier 'NSAppKitVersionNumber10_12'; did you mean 'NSAppKitVersionNumber'?
if (floor(NSAppKitVersionNumber) >= NSAppKitVersionNumber10_12) {
^~~~~~~~~~~~~~~~~~~~~~~~~~
NSAppKitVersionNumber
/System/Library/Frameworks/AppKit.framework/Headers/NSApplication.h:26:28: note: 'NSAppKitVersionNumber' declared here
- 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.
Rémi Verschelde
Andrii Doroshenko (Xrayez)
unknown
Nickolai Korshunov
Juan Linietsky
Joost Heitbrink
santouits
hbina085
Juan Linietsky
Hendrikto
marxin
George Marques
Naoto Kondo
lupoDharkael
K. S. Ernest (iFire) Lee
Bernhard M. Wiedemann
Alexandru Croitor
dragmz
Viktor Ferenczi
karroffel