aboutsummaryrefslogtreecommitdiff
path: root/lib/std/time.zig
diff options
context:
space:
mode:
authorAndrew Kelley <andrew@ziglang.org>2019-09-26 01:54:45 -0400
committerGitHub <noreply@github.com>2019-09-26 01:54:45 -0400
commit68bb3945708c43109c48bda3664176307d45b62c (patch)
treeafb9731e10cef9d192560b52cd9ae2cf179775c4 /lib/std/time.zig
parent6128bc728d1e1024a178c16c2149f5b1a167a013 (diff)
parent4637e8f9699af9c3c6cf4df50ef5bb67c7a318a4 (diff)
downloadzig-68bb3945708c43109c48bda3664176307d45b62c.tar.gz
zig-68bb3945708c43109c48bda3664176307d45b62c.zip
Merge pull request #3315 from ziglang/mv-std-lib
Move std/ to lib/std/
Diffstat (limited to 'lib/std/time.zig')
-rw-r--r--lib/std/time.zig221
1 files changed, 221 insertions, 0 deletions
diff --git a/lib/std/time.zig b/lib/std/time.zig
new file mode 100644
index 0000000000..caf6c31b31
--- /dev/null
+++ b/lib/std/time.zig
@@ -0,0 +1,221 @@
+const builtin = @import("builtin");
+const std = @import("std.zig");
+const assert = std.debug.assert;
+const testing = std.testing;
+const os = std.os;
+const math = std.math;
+
+pub const epoch = @import("time/epoch.zig");
+
+/// Spurious wakeups are possible and no precision of timing is guaranteed.
+pub fn sleep(nanoseconds: u64) void {
+ if (os.windows.is_the_target) {
+ const ns_per_ms = ns_per_s / ms_per_s;
+ const big_ms_from_ns = nanoseconds / ns_per_ms;
+ const ms = math.cast(os.windows.DWORD, big_ms_from_ns) catch math.maxInt(os.windows.DWORD);
+ os.windows.kernel32.Sleep(ms);
+ return;
+ }
+ const s = nanoseconds / ns_per_s;
+ const ns = nanoseconds % ns_per_s;
+ std.os.nanosleep(s, ns);
+}
+
+/// Get the posix timestamp, UTC, in seconds
+/// TODO audit this function. is it possible to return an error?
+pub fn timestamp() u64 {
+ return @divFloor(milliTimestamp(), ms_per_s);
+}
+
+/// Get the posix timestamp, UTC, in milliseconds
+/// TODO audit this function. is it possible to return an error?
+pub fn milliTimestamp() u64 {
+ if (os.windows.is_the_target) {
+ //FileTime has a granularity of 100 nanoseconds
+ // and uses the NTFS/Windows epoch
+ var ft: os.windows.FILETIME = undefined;
+ os.windows.kernel32.GetSystemTimeAsFileTime(&ft);
+ const hns_per_ms = (ns_per_s / 100) / ms_per_s;
+ const epoch_adj = epoch.windows * ms_per_s;
+
+ const ft64 = (u64(ft.dwHighDateTime) << 32) | ft.dwLowDateTime;
+ return @divFloor(ft64, hns_per_ms) - -epoch_adj;
+ }
+ if (os.wasi.is_the_target and !builtin.link_libc) {
+ var ns: os.wasi.timestamp_t = undefined;
+
+ // TODO: Verify that precision is ignored
+ const err = os.wasi.clock_time_get(os.wasi.CLOCK_REALTIME, 1, &ns);
+ assert(err == os.wasi.ESUCCESS);
+
+ const ns_per_ms = 1000;
+ return @divFloor(ns, ns_per_ms);
+ }
+ if (os.darwin.is_the_target) {
+ var tv: os.darwin.timeval = undefined;
+ var err = os.darwin.gettimeofday(&tv, null);
+ assert(err == 0);
+ const sec_ms = tv.tv_sec * ms_per_s;
+ const usec_ms = @divFloor(tv.tv_usec, us_per_s / ms_per_s);
+ return @intCast(u64, sec_ms + usec_ms);
+ }
+ var ts: os.timespec = undefined;
+ //From what I can tell there's no reason clock_gettime
+ // should ever fail for us with CLOCK_REALTIME,
+ // seccomp aside.
+ os.clock_gettime(os.CLOCK_REALTIME, &ts) catch unreachable;
+ const sec_ms = @intCast(u64, ts.tv_sec) * ms_per_s;
+ const nsec_ms = @divFloor(@intCast(u64, ts.tv_nsec), ns_per_s / ms_per_s);
+ return sec_ms + nsec_ms;
+}
+
+/// Multiples of a base unit (nanoseconds)
+pub const nanosecond = 1;
+pub const microsecond = 1000 * nanosecond;
+pub const millisecond = 1000 * microsecond;
+pub const second = 1000 * millisecond;
+pub const minute = 60 * second;
+pub const hour = 60 * minute;
+
+/// Divisions of a second
+pub const ns_per_s = 1000000000;
+pub const us_per_s = 1000000;
+pub const ms_per_s = 1000;
+pub const cs_per_s = 100;
+
+/// Common time divisions
+pub const s_per_min = 60;
+pub const s_per_hour = s_per_min * 60;
+pub const s_per_day = s_per_hour * 24;
+pub const s_per_week = s_per_day * 7;
+
+/// A monotonic high-performance timer.
+/// Timer.start() must be called to initialize the struct, which captures
+/// the counter frequency on windows and darwin, records the resolution,
+/// and gives the user an opportunity to check for the existnece of
+/// monotonic clocks without forcing them to check for error on each read.
+/// .resolution is in nanoseconds on all platforms but .start_time's meaning
+/// depends on the OS. On Windows and Darwin it is a hardware counter
+/// value that requires calculation to convert to a meaninful unit.
+pub const Timer = struct {
+ ///if we used resolution's value when performing the
+ /// performance counter calc on windows/darwin, it would
+ /// be less precise
+ frequency: switch (builtin.os) {
+ .windows => u64,
+ .macosx, .ios, .tvos, .watchos => os.darwin.mach_timebase_info_data,
+ else => void,
+ },
+ resolution: u64,
+ start_time: u64,
+
+ const Error = error{TimerUnsupported};
+
+ ///At some point we may change our minds on RAW, but for now we're
+ /// sticking with posix standard MONOTONIC. For more information, see:
+ /// https://github.com/ziglang/zig/pull/933
+ const monotonic_clock_id = os.CLOCK_MONOTONIC;
+ /// Initialize the timer structure.
+ //This gives us an opportunity to grab the counter frequency in windows.
+ //On Windows: QueryPerformanceCounter will succeed on anything >= XP/2000.
+ //On Posix: CLOCK_MONOTONIC will only fail if the monotonic counter is not
+ // supported, or if the timespec pointer is out of bounds, which should be
+ // impossible here barring cosmic rays or other such occurrences of
+ // incredibly bad luck.
+ //On Darwin: This cannot fail, as far as I am able to tell.
+ pub fn start() Error!Timer {
+ var self: Timer = undefined;
+
+ if (os.windows.is_the_target) {
+ self.frequency = os.windows.QueryPerformanceFrequency();
+ self.resolution = @divFloor(ns_per_s, self.frequency);
+ self.start_time = os.windows.QueryPerformanceCounter();
+ } else if (os.darwin.is_the_target) {
+ os.darwin.mach_timebase_info(&self.frequency);
+ self.resolution = @divFloor(self.frequency.numer, self.frequency.denom);
+ self.start_time = os.darwin.mach_absolute_time();
+ } else {
+ //On Linux, seccomp can do arbitrary things to our ability to call
+ // syscalls, including return any errno value it wants and
+ // inconsistently throwing errors. Since we can't account for
+ // abuses of seccomp in a reasonable way, we'll assume that if
+ // seccomp is going to block us it will at least do so consistently
+ var ts: os.timespec = undefined;
+ os.clock_getres(monotonic_clock_id, &ts) catch return error.TimerUnsupported;
+ self.resolution = @intCast(u64, ts.tv_sec) * u64(ns_per_s) + @intCast(u64, ts.tv_nsec);
+
+ os.clock_gettime(monotonic_clock_id, &ts) catch return error.TimerUnsupported;
+ self.start_time = @intCast(u64, ts.tv_sec) * u64(ns_per_s) + @intCast(u64, ts.tv_nsec);
+ }
+
+ return self;
+ }
+
+ /// Reads the timer value since start or the last reset in nanoseconds
+ pub fn read(self: *Timer) u64 {
+ var clock = clockNative() - self.start_time;
+ if (os.windows.is_the_target) {
+ return @divFloor(clock * ns_per_s, self.frequency);
+ }
+ if (os.darwin.is_the_target) {
+ return @divFloor(clock * self.frequency.numer, self.frequency.denom);
+ }
+ return clock;
+ }
+
+ /// Resets the timer value to 0/now.
+ pub fn reset(self: *Timer) void {
+ self.start_time = clockNative();
+ }
+
+ /// Returns the current value of the timer in nanoseconds, then resets it
+ pub fn lap(self: *Timer) u64 {
+ var now = clockNative();
+ var lap_time = self.read();
+ self.start_time = now;
+ return lap_time;
+ }
+
+ fn clockNative() u64 {
+ if (os.windows.is_the_target) {
+ return os.windows.QueryPerformanceCounter();
+ }
+ if (os.darwin.is_the_target) {
+ return os.darwin.mach_absolute_time();
+ }
+ var ts: os.timespec = undefined;
+ os.clock_gettime(monotonic_clock_id, &ts) catch unreachable;
+ return @intCast(u64, ts.tv_sec) * u64(ns_per_s) + @intCast(u64, ts.tv_nsec);
+ }
+};
+
+test "sleep" {
+ sleep(1);
+}
+
+test "timestamp" {
+ const ns_per_ms = (ns_per_s / ms_per_s);
+ const margin = 50;
+
+ const time_0 = milliTimestamp();
+ sleep(ns_per_ms);
+ const time_1 = milliTimestamp();
+ const interval = time_1 - time_0;
+ testing.expect(interval > 0 and interval < margin);
+}
+
+test "Timer" {
+ const ns_per_ms = (ns_per_s / ms_per_s);
+ const margin = ns_per_ms * 150;
+
+ var timer = try Timer.start();
+ sleep(10 * ns_per_ms);
+ const time_0 = timer.read();
+ testing.expect(time_0 > 0 and time_0 < margin);
+
+ const time_1 = timer.lap();
+ testing.expect(time_1 >= time_0);
+
+ timer.reset();
+ testing.expect(timer.read() < time_1);
+}