asgardius/virtualx-engine
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
virtualx-engine/modules/gdnative/gdnative_builders.py

266 lines
8.3 KiB
Python
Raw Normal View History

Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
"""Functions used to generate source files during build time
All such functions are invoked in a subprocess on Windows to prevent build flakiness.
"""
import json
from platform_methods import subprocess_main
def _spaced(e):
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
return e if e[-1] == "*" else e + " "
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
def _build_gdnative_api_struct_header(api):
out = [
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
"/* THIS FILE IS GENERATED DO NOT EDIT */",
"#ifndef GODOT_GDNATIVE_API_STRUCT_H",
"#define GODOT_GDNATIVE_API_STRUCT_H",
"",
"#include <gdnative/gdnative.h>",
"#include <android/godot_android.h>",
Renaming all ARVR nodes to XR
2020-04-08 16:47:36 +02:00
"#include <xr/godot_xr.h>",
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
"#include <nativescript/godot_nativescript.h>",
"#include <net/godot_net.h>",
"#include <pluginscript/godot_pluginscript.h>",
"#include <videodecoder/godot_videodecoder.h>",
"",
"#ifdef __cplusplus",
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
'extern "C" {',
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
"#endif",
"",
"enum GDNATIVE_API_TYPES {",
"\tGDNATIVE_" + api["core"]["type"] + ",",
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
]
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
for ext in api["extensions"]:
out += ["\tGDNATIVE_EXT_" + ext["type"] + ","]
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
out += ["};", ""]
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
def generate_extension_struct(name, ext, include_version=True):
ret_val = []
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
if ext["next"]:
ret_val += generate_extension_struct(name, ext["next"])
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
ret_val += [
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
"typedef struct godot_gdnative_ext_"
+ name
+ ("" if not include_version else ("_{0}_{1}".format(ext["version"]["major"], ext["version"]["minor"])))
+ "_api_struct {",
"\tunsigned int type;",
"\tgodot_gdnative_api_version version;",
"\tconst godot_gdnative_api_struct *next;",
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
]
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
for funcdef in ext["api"]:
args = ", ".join(["%s%s" % (_spaced(t), n) for t, n in funcdef["arguments"]])
ret_val.append("\t%s(*%s)(%s);" % (_spaced(funcdef["return_type"]), funcdef["name"], args))
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
ret_val += [
"} godot_gdnative_ext_"
+ name
+ ("" if not include_version else ("_{0}_{1}".format(ext["version"]["major"], ext["version"]["minor"])))
+ "_api_struct;",
"",
]
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
return ret_val
[GDNative] add initial core 1.1 extension
2018-08-30 19:18:55 +02:00
def generate_core_extension_struct(core):
ret_val = []
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
if core["next"]:
ret_val += generate_core_extension_struct(core["next"])
Remove trailing whitespace With `sed -i $(rg -l '[[:blank:]]*$' -g'!thirdparty') -e 's/[[:blank:]]*$//g'` (+ manual revert of some thirdparty code under `platform/android`).
2018-11-20 11:14:07 +01:00
[GDNative] add initial core 1.1 extension
2018-08-30 19:18:55 +02:00
ret_val += [
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
"typedef struct godot_gdnative_core_"
CI: Install master version of psf/black Until https://github.com/psf/black/pull/1328 makes it in a stable release, we have to use the latest from Git. Apply new style fixes done by latest black.
2020-07-26 19:38:10 +02:00
+ "{0}_{1}".format(core["version"]["major"], core["version"]["minor"])
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
+ "_api_struct {",
"\tunsigned int type;",
"\tgodot_gdnative_api_version version;",
"\tconst godot_gdnative_api_struct *next;",
[GDNative] add initial core 1.1 extension
2018-08-30 19:18:55 +02:00
]
Remove trailing whitespace With `sed -i $(rg -l '[[:blank:]]*$' -g'!thirdparty') -e 's/[[:blank:]]*$//g'` (+ manual revert of some thirdparty code under `platform/android`).
2018-11-20 11:14:07 +01:00
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
for funcdef in core["api"]:
args = ", ".join(["%s%s" % (_spaced(t), n) for t, n in funcdef["arguments"]])
ret_val.append("\t%s(*%s)(%s);" % (_spaced(funcdef["return_type"]), funcdef["name"], args))
Remove trailing whitespace With `sed -i $(rg -l '[[:blank:]]*$' -g'!thirdparty') -e 's/[[:blank:]]*$//g'` (+ manual revert of some thirdparty code under `platform/android`).
2018-11-20 11:14:07 +01:00
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
ret_val += [
"} godot_gdnative_core_"
+ "{0}_{1}".format(core["version"]["major"], core["version"]["minor"])
+ "_api_struct;",
"",
]
Remove trailing whitespace With `sed -i $(rg -l '[[:blank:]]*$' -g'!thirdparty') -e 's/[[:blank:]]*$//g'` (+ manual revert of some thirdparty code under `platform/android`).
2018-11-20 11:14:07 +01:00
[GDNative] add initial core 1.1 extension
2018-08-30 19:18:55 +02:00
return ret_val
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
for ext in api["extensions"]:
name = ext["name"]
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
out += generate_extension_struct(name, ext, False)
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
if api["core"]["next"]:
out += generate_core_extension_struct(api["core"]["next"])
[GDNative] add initial core 1.1 extension
2018-08-30 19:18:55 +02:00
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
out += [
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
"typedef struct godot_gdnative_core_api_struct {",
"\tunsigned int type;",
"\tgodot_gdnative_api_version version;",
"\tconst godot_gdnative_api_struct *next;",
"\tunsigned int num_extensions;",
"\tconst godot_gdnative_api_struct **extensions;",
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
]
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
for funcdef in api["core"]["api"]:
args = ", ".join(["%s%s" % (_spaced(t), n) for t, n in funcdef["arguments"]])
out.append("\t%s(*%s)(%s);" % (_spaced(funcdef["return_type"]), funcdef["name"], args))
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
out += [
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
"} godot_gdnative_core_api_struct;",
"",
"#ifdef __cplusplus",
"}",
"#endif",
"",
"#endif // GODOT_GDNATIVE_API_STRUCT_H",
"",
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
]
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
return "\n".join(out)
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
def _build_gdnative_api_struct_source(api):
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
out = ["/* THIS FILE IS GENERATED DO NOT EDIT */", "", "#include <gdnative_api_struct.gen.h>", ""]
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
def get_extension_struct_name(name, ext, include_version=True):
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
return (
"godot_gdnative_ext_"
+ name
+ ("" if not include_version else ("_{0}_{1}".format(ext["version"]["major"], ext["version"]["minor"])))
+ "_api_struct"
)
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
def get_extension_struct_instance_name(name, ext, include_version=True):
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
return (
"api_extension_"
+ name
+ ("" if not include_version else ("_{0}_{1}".format(ext["version"]["major"], ext["version"]["minor"])))
+ "_struct"
)
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
def get_extension_struct_definition(name, ext, include_version=True):
ret_val = []
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
if ext["next"]:
ret_val += get_extension_struct_definition(name, ext["next"])
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
ret_val += [
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
"extern const "
+ get_extension_struct_name(name, ext, include_version)
+ " "
+ get_extension_struct_instance_name(name, ext, include_version)
+ " = {",
"\tGDNATIVE_EXT_" + ext["type"] + ",",
"\t{" + str(ext["version"]["major"]) + ", " + str(ext["version"]["minor"]) + "},",
"\t"
+ (
Replace more occurrences of NULL with nullptr
2020-04-02 14:41:58 +02:00
"nullptr"
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
if not ext["next"]
else ("(const godot_gdnative_api_struct *)&" + get_extension_struct_instance_name(name, ext["next"]))
)
+ ",",
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
]
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
for funcdef in ext["api"]:
ret_val.append("\t%s," % funcdef["name"])
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
ret_val += ["};\n"]
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
return ret_val
Remove trailing whitespace With `sed -i $(rg -l '[[:blank:]]*$' -g'!thirdparty') -e 's/[[:blank:]]*$//g'` (+ manual revert of some thirdparty code under `platform/android`).
2018-11-20 11:14:07 +01:00
[GDNative] add initial core 1.1 extension
2018-08-30 19:18:55 +02:00
def get_core_struct_definition(core):
ret_val = []
Remove trailing whitespace With `sed -i $(rg -l '[[:blank:]]*$' -g'!thirdparty') -e 's/[[:blank:]]*$//g'` (+ manual revert of some thirdparty code under `platform/android`).
2018-11-20 11:14:07 +01:00
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
if core["next"]:
ret_val += get_core_struct_definition(core["next"])
Remove trailing whitespace With `sed -i $(rg -l '[[:blank:]]*$' -g'!thirdparty') -e 's/[[:blank:]]*$//g'` (+ manual revert of some thirdparty code under `platform/android`).
2018-11-20 11:14:07 +01:00
[GDNative] add initial core 1.1 extension
2018-08-30 19:18:55 +02:00
ret_val += [
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
"extern const godot_gdnative_core_"
CI: Install master version of psf/black Until https://github.com/psf/black/pull/1328 makes it in a stable release, we have to use the latest from Git. Apply new style fixes done by latest black.
2020-07-26 19:38:10 +02:00
+ "{0}_{1}_api_struct api_{0}_{1}".format(core["version"]["major"], core["version"]["minor"])
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
+ " = {",
"\tGDNATIVE_" + core["type"] + ",",
"\t{" + str(core["version"]["major"]) + ", " + str(core["version"]["minor"]) + "},",
"\t"
+ (
Replace more occurrences of NULL with nullptr
2020-04-02 14:41:58 +02:00
"nullptr"
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
if not core["next"]
else (
"(const godot_gdnative_api_struct *)& api_{0}_{1}".format(
core["next"]["version"]["major"], core["next"]["version"]["minor"]
)
)
)
+ ",",
[GDNative] add initial core 1.1 extension
2018-08-30 19:18:55 +02:00
]
Remove trailing whitespace With `sed -i $(rg -l '[[:blank:]]*$' -g'!thirdparty') -e 's/[[:blank:]]*$//g'` (+ manual revert of some thirdparty code under `platform/android`).
2018-11-20 11:14:07 +01:00
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
for funcdef in core["api"]:
ret_val.append("\t%s," % funcdef["name"])
Remove trailing whitespace With `sed -i $(rg -l '[[:blank:]]*$' -g'!thirdparty') -e 's/[[:blank:]]*$//g'` (+ manual revert of some thirdparty code under `platform/android`).
2018-11-20 11:14:07 +01:00
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
ret_val += ["};\n"]
Remove trailing whitespace With `sed -i $(rg -l '[[:blank:]]*$' -g'!thirdparty') -e 's/[[:blank:]]*$//g'` (+ manual revert of some thirdparty code under `platform/android`).
2018-11-20 11:14:07 +01:00
[GDNative] add initial core 1.1 extension
2018-08-30 19:18:55 +02:00
return ret_val
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
for ext in api["extensions"]:
name = ext["name"]
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
out += get_extension_struct_definition(name, ext, False)
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
out += ["", "const godot_gdnative_api_struct *gdnative_extensions_pointers[] = {"]
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
for ext in api["extensions"]:
name = ext["name"]
out += ["\t(godot_gdnative_api_struct *)&api_extension_" + name + "_struct,"]
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
out += ["};\n"]
Remove trailing whitespace With `sed -i $(rg -l '[[:blank:]]*$' -g'!thirdparty') -e 's/[[:blank:]]*$//g'` (+ manual revert of some thirdparty code under `platform/android`).
2018-11-20 11:14:07 +01:00
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
if api["core"]["next"]:
out += get_core_struct_definition(api["core"]["next"])
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
out += [
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
"extern const godot_gdnative_core_api_struct api_struct = {",
"\tGDNATIVE_" + api["core"]["type"] + ",",
"\t{" + str(api["core"]["version"]["major"]) + ", " + str(api["core"]["version"]["minor"]) + "},",
GDNative: merge API structs, bump version of merged structs.
2020-05-31 12:52:21 +02:00
"\t"
+ (
"nullptr, "
if not api["core"]["next"]
else (
"(const godot_gdnative_api_struct *)& api_{0}_{1},".format(
api["core"]["next"]["version"]["major"], api["core"]["next"]["version"]["minor"]
)
)
),
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
"\t" + str(len(api["extensions"])) + ",",
"\tgdnative_extensions_pointers,",
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
]
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
for funcdef in api["core"]["api"]:
out.append("\t%s," % funcdef["name"])
out.append("};\n")
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
return "\n".join(out)
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
def build_gdnative_api_struct(target, source, env):
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
with open(source[0], "r") as fd:
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
api = json.load(fd)
header, source = target
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
with open(header, "w") as fd:
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
fd.write(_build_gdnative_api_struct_header(api))
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
with open(source, "w") as fd:
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
fd.write(_build_gdnative_api_struct_source(api))
SCons: Format buildsystem files with psf/black 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.
2020-03-30 08:28:32 +02:00
if __name__ == "__main__":
Running builder (content generator) functions in subprocesses on Windows - 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.
2018-03-17 23:23:55 +01:00
subprocess_main(globals())
Reference in a new issue Copy permalink