Two modules fail building with this option, so we force users to disable them.
Small cleanup based on 2.1 branch and cherry-pick 01e65c4555
to fix support for NDK r20 (fixes#30688).
In x11, windows and osx crash handlers, check project settings exists
before looking up the crash handler message setting.
Avoids crashing the crash handler when handling a crash outside project
settings lifetime. Instead omitting the configurable message and
continuing with trace dump.
(cherry picked from commit 63068e2ccd)
Godot supports many different compilers and for production releases we
have to support 3 currently: GCC8, Clang6, and MSVC2017. These compilers
all do slightly different things with -ffast-math and it is causing
issues now. See #24841, #24540, #10758, #10070. And probably other
complaints about physics differences between release and release_debug
builds.
I've done some performance comparisons on Linux x86_64. All tests are
ran 20 times.
Bunnymark: (higher is better)
(bunnies) min max stdev average
fast-math 7332 7597 71 7432
this pr 7379 7779 108 7621 (102%)
FPBench (gdscript port http://fpbench.org/) (lower is better)
(ms)
fast-math 15441 16127 192 15764
this pr 15671 16855 326 16001 (99%)
Float_add (adding floats in a tight loop) (lower is better)
(sec)
fast-math 5.49 5.78 0.07 5.65
this pr 5.65 5.90 0.06 5.76 (98%)
Float_div (dividing floats in a tight loop) (lower is better)
(sec)
fast-math 11.70 12.36 0.18 11.99
this pr 11.92 12.32 0.12 12.12 (99%)
Float_mul (multiplying floats in a tight loop) (lower is better)
(sec)
fast-math 11.72 12.17 0.12 11.93
this pr 12.01 12.62 0.17 12.26 (97%)
I have also looked at FPS numbers for tps-demo, 3d platformer, 2d
platformer, and sponza and could not find any measurable difference.
I believe that given the issues and oft-reported (physics) glitches on
release builds I believe that the couple of percent of tight-loop
floating point performance regression is well worth it.
This fixes#24540 and fixes#24841
(cherry picked from commit e5b335d367)
The two POSIX style crash handlers (OSX and X11) now remove their signal
handlers when they are destroyed.
Additonally if they are called while no OS singleton is set, they will
simply abort(). This should not happen now that they remove themselves,
but if a future change seperates OS object and crash handler lifetimes,
this may be easier to report/debug than hanging on SIGSEGV.
(cherry picked from commit 653b832422)
On X11 when we send an XResizeWindow request to the X server it is happy
to say it is done when the request has been handed over to the window
manager. The window manager itself may however take some time to
actually do the resize. Godot expects that a resize request is
immediate. To work around this issue we could implement the whole
_NET_WM_SYNC_REQUEST protocol. However this protocol does not fit very
well with the way we currently process X events and would when
implemented in the current framework still cause a 1 frame delay between
a resize request and the actual resize happening.
This fixes#21720
(cherry picked from commit 9a1deedb84)
This fixes the editor on X11 not getting put on the foreground when a
debugged project hits an error or breakpoint.
(cherry picked from commit 827cadafc8)
Fixes thread and process handles leak when running and killing project
from editor (caused by a missing CloseHandle call) plus a potential leak
when calling OS_Windows::execute with p_blocking and !r_pipe.
The leak could be easily observed with a Handles counter in Task Manager
(or Performance Monitor) for the Godot editor process.
(cherry picked from commit b1e0da455b)
Initialised relevant variables before stating thread,
to prevent a branch on uninitialised data.
Fixed race condition in polling that could miss a device change.
(cherry picked from commit fe4265ad46)
The rationale for keeping those shared by default is that they're typical
dependencies found on any Linux system, and it saves compilation time and
binary size to link their dynamically.
But since official builds default to all-builtin, and Debian/Ubuntu still
don't have libpng16 (which we now require) readily available on all their
supported releases, it's simpler to bundle all the things.
This does not change the fact that those dependencies *can* be unbundled
on Linux, it's only the default option changing.
(cherry picked from commit 1769cbc0e2)
For 3.0, also building by default against bundled openssl.
From August 1, 2019, Google Play requires that all new apps and app updates
include 64-bit versions, so we enable ARM64 by default.
IINM support for x86 and x86_64 is still be optional, so not enabling them
out of the box.
Part of #25030.
(cherry picked from commit 9e820cdf20)
Like arm64v8, this is only supported by API 21 and later,
so we enforce 21 as min API for x86_64.
Part of #25030.
(cherry picked from commit 7f4ee36469)
This fixes exporting the NvOptimusEnablement export when building with
MingW. This also adds the equivalent for AMD.
This fixes#23400
(cherry picked from commit 19d91f788d)
The code in pre.js and engine.js is a bit confusing to see in isolation,
since the files aren't valid JS files by themselves. This just adds some
explanatory text to both files.
Fixes#22937.
(cherry picked from commit 61d5513525)
Initialized the PID to -2, which will be the value returns in blocking-
mode where the PID is not available. (-1 was already taken to signify an
execution failure).
OS::execute will now properly return a non-OK error code when it fails
to execute the target file.
The documentation was rewritten to be very clear about the differences
between blocking and non-blocking mode.
Fixes#19056.
(cherry picked from commit f392650be2)
Max
bruvzg
mellondill
Rémi Verschelde
Leon Krause
Ibrahn Sahir
Bastiaan Olij
Hein-Pieter van Braam
Ruslan Mustakov
Marcelo Fernandez
Colin Kinloch
Xavier Sellier
Karen Schwane
volzhs
Pedro J. Estébanez
Marcin Zawiejski
Duy-Nguyen Ta
Bernhard Liebl
Andrea Beconcini
Juan Linietsky
Zaven Muradyan
M. Huri
Wojciech Milkowski