我在Rust中使用Piston开发了一个迷宫游戏,以自学图形/ UI编程。该游戏通常可以正常运行,但是当我使用大型迷宫(例如120 x 72矩形)运行游戏时,游戏在几分钟后崩溃,并出现内存分配错误。
一个简化的示例如下:
extern crate graphics;
extern crate opengl_graphics;
extern crate piston;
extern crate piston_window;
use opengl_graphics::OpenGL;
use piston::event_loop::*;
use piston::input::*;
use piston_window::{PistonWindow, WindowSettings};
const BLACK: [f32; 4] = [0.0, 0.0, 0.0, 1.0];
const WHITE: [f32; 4] = [1.0, 1.0, 1.0, 1.0];
fn main() {
let opengl = OpenGL::V3_2;
let mut window: PistonWindow = WindowSettings::new("maze", [800, 600])
.opengl(opengl)
.exit_on_esc(true)
.build()
.unwrap();
let mut events = Events::new(EventSettings::new());
while let Some(event) = events.next(&mut window) {
if let Some(_args) = event.render_args() {
window.draw_2d(&event, |c, gl| {
graphics::clear(BLACK, gl);
for _row in 0..72 {
for _col in 0..120 {
let color = WHITE;
let box_rect =
graphics::rectangle::rectangle_by_corners(0.0, 0.0, 10.0, 10.0);
graphics::rectangle(color, box_rect, c.transform, gl);
}
}
});
}
}
}
运行此命令时,出现内存分配错误,然后进程中止:
$ time cargo run --release
Compiling maze v0.1.0 (/home/isaac/prog/rust/test-maze)
Finished release [optimized] target(s) in 5.28s
Running `target/release/maze`
memory allocation of 7927496704 bytes failedAborted (core dumped)
real 2m24.317s
user 1m29.515s
sys 0m3.644s
在调试器中运行它,我得到以下回溯:
(gdb) backtrace
#0 __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:51
#1 0x00007ffff71e1801 in __GI_abort () at abort.c:79
#2 0x0000555555708fc7 in std::sys::unix::abort_internal () at src/libstd/sys/unix/mod.rs:157
#3 0x000055555570419d in rust_oom () at src/libstd/alloc.rs:211
#4 0x0000555555717f37 in alloc::alloc::handle_alloc_error () at src/liballoc/alloc.rs:224
#5 0x00005555556eaf7d in <alloc::vec::Vec<T> as alloc::vec::SpecExtend<T, I>>::spec_extend ()
#6 0x00005555556e64aa in <gfx_core::handle::Manager<R>>::extend ()
#7 0x00005555556da204 in <gfx_device_gl::Device as gfx_core::Device>::pin_submitted_resources ()
#8 0x000055555559d654 in <gfx::encoder::Encoder<R, C>>::flush ()
#9 0x000055555558f8af in maze::main ()
#10 0x000055555558bb73 in std::rt::lang_start::{{closure}} ()
#11 0x0000555555704913 in std::rt::lang_start_internal::{{closure}} () at src/libstd/rt.rs:49
#12 std::panicking::try::do_call () at src/libstd/panicking.rs:297
#13 0x000055555570927a in __rust_maybe_catch_panic () at src/libpanic_unwind/lib.rs:92
#14 0x0000555555705466 in std::panicking::try () at src/libstd/panicking.rs:276
#15 std::panic::catch_unwind () at src/libstd/panic.rs:388
#16 std::rt::lang_start_internal () at src/libstd/rt.rs:48
#17 0x000055555558fae2 in main ()
我正在使用Ubuntu Linux 18.04。
是否应以不同的方式编写我的程序以防止出现此问题?活塞有问题吗?
答案 0 :(得分:5)
问题位于此处:
let mut events = Events::new(EventSettings::new());
while let Some(event) = events.next(&mut window) {
相反,应该使用此代码:
while let Some(event) = window.next() {
事实证明,window.next()
运行了events.next(&mut window)
省略的一些清理步骤,这些清理步骤释放了相关的内存。据我所知,文档中没有对此进行提及,并且示例使用了两种模式,而没有指出差异。
答案 1 :(得分:2)
这大约是8 GB的分配空间;太大而无法理解。 8 GB可能与您的系统RAM有关。等于其一半或两倍,具体取决于您的系统设置。
问题必须在循环中并且必须很小。循环不是很大,是吗?关于小例子,这是一件好事。看起来好像是事件代码。我不知道为什么。
这是一个似乎没有内存泄漏的版本:
extern crate graphics;
extern crate opengl_graphics;
extern crate piston;
extern crate piston_window;
use opengl_graphics::OpenGL;
use piston::event_loop::*;
use piston::input::*;
use piston_window::{PistonWindow, WindowSettings};
const BLACK: [f32; 4] = [0.0, 0.0, 0.0, 1.0];
const WHITE: [f32; 4] = [1.0, 1.0, 1.0, 1.0];
fn main() {
let opengl = OpenGL::V3_2;
let mut window: PistonWindow = WindowSettings::new("maze", [800, 600])
.opengl(opengl)
.exit_on_esc(true)
.build()
.unwrap();
while let Some(event) = window.next() {
window.draw_2d(&event, |c, gl| {
graphics::clear(BLACK, gl);
for _row in 0..72 {
for _col in 0..120 {
let color = WHITE;
let box_rect =
graphics::rectangle::rectangle_by_corners(0.0, 0.0, 10.0, 10.0);
graphics::rectangle(color, box_rect, c.transform, gl);
}
}
});
}
}