replace regex use with nom

This reduces the binary size noticeably on my macOS machine (#70):

                             unstripped  stripped
1  before switching to clap    11.1 MiB   6.7 MiB
2  after switching to clap     11.4 MiB   6.9 MiB
3  without regex               10.1 MiB   5.9 MiB
This commit is contained in:
Scott Lamb 2020-04-17 23:02:02 -07:00
parent e8eb764b90
commit af9e568344
10 changed files with 476 additions and 367 deletions

73
Cargo.lock generated
View File

@ -15,15 +15,6 @@ dependencies = [
"const-random",
]
[[package]]
name = "aho-corasick"
version = "0.7.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "58fb5e95d83b38284460a5fda7d6470aa0b8844d283a0b614b8535e880800d2d"
dependencies = [
"memchr",
]
[[package]]
name = "ansi_term"
version = "0.9.0"
@ -1031,6 +1022,19 @@ version = "1.4.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "e2abad23fbc42b3700f2f279844dc832adb2b2eb069b2df918f455c4e18cc646"
[[package]]
name = "lexical-core"
version = "0.6.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d7043aa5c05dd34fb73b47acb8c3708eac428de4545ea3682ed2f11293ebd890"
dependencies = [
"arrayvec 0.4.12",
"cfg-if",
"rustc_version",
"ryu",
"static_assertions",
]
[[package]]
name = "libc"
version = "0.2.66"
@ -1214,8 +1218,8 @@ dependencies = [
"lazy_static",
"libc",
"log",
"nom 5.1.1",
"parking_lot",
"regex",
"time 0.1.42",
]
@ -1243,7 +1247,6 @@ dependencies = [
"prettydiff",
"protobuf",
"protobuf-codegen-pure",
"regex",
"rusqlite",
"smallvec 1.1.0",
"tempdir",
@ -1287,11 +1290,11 @@ dependencies = [
"moonfire-ffmpeg",
"mylog",
"nix",
"nom 5.1.1",
"openssl",
"parking_lot",
"protobuf",
"reffers",
"regex",
"reqwest",
"ring",
"rusqlite",
@ -1386,6 +1389,17 @@ dependencies = [
"version_check 0.1.5",
]
[[package]]
name = "nom"
version = "5.1.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "0b471253da97532da4b61552249c521e01e736071f71c1a4f7ebbfbf0a06aad6"
dependencies = [
"lexical-core",
"memchr",
"version_check 0.9.1",
]
[[package]]
name = "num"
version = "0.2.1"
@ -1879,18 +1893,6 @@ dependencies = [
"stable_deref_trait",
]
[[package]]
name = "regex"
version = "1.3.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b5508c1941e4e7cb19965abef075d35a9a8b5cdf0846f30b4050e9b55dc55e87"
dependencies = [
"aho-corasick",
"memchr",
"regex-syntax",
"thread_local",
]
[[package]]
name = "regex-automata"
version = "0.1.8"
@ -1900,12 +1902,6 @@ dependencies = [
"byteorder",
]
[[package]]
name = "regex-syntax"
version = "0.6.13"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "e734e891f5b408a29efbf8309e656876276f49ab6a6ac208600b4419bd893d90"
[[package]]
name = "remove_dir_all"
version = "0.5.2"
@ -2298,6 +2294,12 @@ version = "1.1.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "dba1a27d3efae4351c8051072d619e3ade2820635c3958d826bfea39d59b54c8"
[[package]]
name = "static_assertions"
version = "0.3.4"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "7f3eb36b47e512f8f1c9e3d10c2c1965bc992bd9cdb024fa581e2194501c83d3"
[[package]]
name = "strsim"
version = "0.8.0"
@ -2463,15 +2465,6 @@ dependencies = [
"unicode-width",
]
[[package]]
name = "thread_local"
version = "1.0.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d40c6d1b69745a6ec6fb1ca717914848da4b44ae29d9b3080cbee91d72a69b14"
dependencies = [
"lazy_static",
]
[[package]]
name = "time"
version = "0.1.42"
@ -2869,7 +2862,7 @@ version = "0.10.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "3bb43f70885151e629e2a19ce9e50bd730fd436cfd4b666894c9ce4de9141164"
dependencies = [
"nom",
"nom 4.2.3",
]
[[package]]

View File

@ -43,11 +43,11 @@ memchr = "2.0.2"
memmap = "0.7"
mylog = { git = "https://github.com/scottlamb/mylog" }
nix = "0.16.1"
nom = "5.1.1"
openssl = "0.10"
parking_lot = { version = "0.9", features = [] }
protobuf = { git = "https://github.com/stepancheg/rust-protobuf" }
reffers = "0.6.0"
regex = "1.0"
ring = "0.14.6"
rusqlite = "0.21.0"
serde = { version = "1.0", features = ["derive"] }

View File

@ -17,5 +17,5 @@ lazy_static = "1.0"
libc = "0.2"
log = "0.4"
parking_lot = { version = "0.9", features = [] }
regex = "1.0"
nom = "5.1.1"
time = "0.1"

View File

@ -29,6 +29,7 @@
// along with this program. If not, see <http://www.gnu.org/licenses/>.
pub mod clock;
pub mod time;
mod error;
pub mod strutil;

View File

@ -28,10 +28,13 @@
// You should have received a copy of the GNU General Public License
// along with this program. If not, see <http://www.gnu.org/licenses/>.
use lazy_static::lazy_static;
use regex::Regex;
use nom::IResult;
use nom::branch::alt;
use nom::bytes::complete::{tag, take_while1};
use nom::character::complete::space0;
use nom::combinator::{map, map_res, opt};
use nom::sequence::{delimited, tuple};
use std::fmt::Write as _;
use std::str::FromStr as _;
static MULTIPLIERS: [(char, u64); 4] = [
// (suffix character, power of 2)
@ -58,32 +61,33 @@ pub fn encode_size(mut raw: i64) -> String {
encoded
}
fn decode_sizepart(input: &str) -> IResult<&str, i64> {
map(
tuple((
map_res(take_while1(|c: char| c.is_ascii_digit()),
|input: &str| i64::from_str_radix(input, 10)),
opt(alt((
nom::combinator::value(1<<40, tag("T")),
nom::combinator::value(1<<30, tag("G")),
nom::combinator::value(1<<20, tag("M")),
nom::combinator::value(1<<10, tag("K"))
)))
)),
|(n, opt_unit)| n * opt_unit.unwrap_or(1)
)(input)
}
fn decode_size_internal(input: &str) -> IResult<&str, i64> {
nom::multi::fold_many1(
delimited(space0, decode_sizepart, space0),
0,
|sum, i| sum + i)(input)
}
/// Decodes a human-readable size as output by encode_size.
pub fn decode_size(encoded: &str) -> Result<i64, ()> {
let mut decoded = 0i64;
lazy_static! {
static ref RE: Regex = Regex::new(r"\s*([0-9]+)([TGMK])?,?\s*").unwrap();
}
let mut last_pos = 0;
for cap in RE.captures_iter(encoded) {
let whole_cap = cap.get(0).unwrap();
if whole_cap.start() > last_pos {
return Err(());
}
last_pos = whole_cap.end();
let mut piece = i64::from_str(&cap[1]).map_err(|_| ())?;
if let Some(m) = cap.get(2) {
let m = m.as_str().as_bytes()[0] as char;
for &(some_m, n) in &MULTIPLIERS {
if some_m == m {
piece *= 1i64<<n;
break;
}
}
}
decoded += piece;
}
if last_pos < encoded.len() {
let (remaining, decoded) = decode_size_internal(encoded).map_err(|_e| ())?;
if !remaining.is_empty() {
return Err(());
}
Ok(decoded)
@ -130,6 +134,7 @@ mod tests {
#[test]
fn test_decode() {
assert_eq!(super::decode_size("100M").unwrap(), 100i64 << 20);
assert_eq!(super::decode_size("100M 42").unwrap(), (100i64 << 20) + 42);
}
#[test]

329
base/time.rs Normal file
View File

@ -0,0 +1,329 @@
// This file is part of Moonfire NVR, a security camera network video recorder.
// Copyright (C) 2016-2020 The Moonfire NVR Authors
//
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// In addition, as a special exception, the copyright holders give
// permission to link the code of portions of this program with the
// OpenSSL library under certain conditions as described in each
// individual source file, and distribute linked combinations including
// the two.
//
// You must obey the GNU General Public License in all respects for all
// of the code used other than OpenSSL. If you modify file(s) with this
// exception, you may extend this exception to your version of the
// file(s), but you are not obligated to do so. If you do not wish to do
// so, delete this exception statement from your version. If you delete
// this exception statement from all source files in the program, then
// also delete it here.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU General Public License for more details.
//
// You should have received a copy of the GNU General Public License
// along with this program. If not, see <http://www.gnu.org/licenses/>.
//! Time and durations for Moonfire NVR's internal format.
use failure::{Error, bail, format_err};
use nom::branch::alt;
use nom::bytes::complete::{tag, take_while_m_n};
use nom::combinator::{map, map_res, opt};
use nom::sequence::{preceded, tuple};
use std::ops;
use std::fmt;
use std::str::FromStr;
use time;
type IResult<'a, I, O> = nom::IResult<I, O, nom::error::VerboseError<&'a str>>;
pub const TIME_UNITS_PER_SEC: i64 = 90_000;
/// A time specified as 90,000ths of a second since 1970-01-01 00:00:00 UTC.
#[derive(Clone, Copy, Default, Eq, Ord, PartialEq, PartialOrd)]
pub struct Time(pub i64);
/// Returns a parser for a `len`-digit non-negative number which fits into an i32.
fn fixed_len_num<'a>(len: usize) -> impl Fn(&'a str) -> IResult<&'a str, i32> {
map_res(take_while_m_n(len, len, |c: char| c.is_ascii_digit()),
|input: &str| i32::from_str_radix(input, 10))
}
/// Parses `YYYY-mm-dd` into pieces.
fn parse_datepart(input: &str) -> IResult<&str, (i32, i32, i32)> {
tuple((
fixed_len_num(4),
preceded(tag("-"), fixed_len_num(2)),
preceded(tag("-"), fixed_len_num(2))
))(input)
}
/// Parses `HH:MM[:SS[:FFFFF]]` into pieces.
fn parse_timepart(input: &str) -> IResult<&str, (i32, i32, i32, i32)> {
let (input, (hr, _, min)) = tuple((fixed_len_num(2), tag(":"), fixed_len_num(2)))(input)?;
let (input, stuff) = opt(tuple((
preceded(tag(":"), fixed_len_num(2)),
opt(preceded(tag(":"), fixed_len_num(5)))
)))(input)?;
let (sec, opt_subsec) = stuff.unwrap_or((0, None));
Ok((input, (hr, min, sec, opt_subsec.unwrap_or(0))))
}
/// Parses `Z` (UTC) or `{+,-,}HH:MM` into a time zone offset in seconds.
fn parse_zone(input: &str) -> IResult<&str, i32> {
alt((
nom::combinator::value(0, tag("Z")),
map(
tuple((
opt(nom::character::complete::one_of(&b"+-"[..])),
fixed_len_num(2),
tag(":"),
fixed_len_num(2)
)),
|(sign, hr, _, min)| {
let off = hr * 3600 + min * 60;
if sign == Some('-') { off } else { -off }
})
))(input)
}
impl Time {
pub fn new(tm: time::Timespec) -> Self {
Time(tm.sec * TIME_UNITS_PER_SEC + tm.nsec as i64 * TIME_UNITS_PER_SEC / 1_000_000_000)
}
pub const fn min_value() -> Self { Time(i64::min_value()) }
pub const fn max_value() -> Self { Time(i64::max_value()) }
/// Parses a time as either 90,000ths of a second since epoch or a RFC 3339-like string.
///
/// The former is 90,000ths of a second since 1970-01-01T00:00:00 UTC, excluding leap seconds.
///
/// The latter is a date such as `2006-01-02T15:04:05`, followed by an optional 90,000ths of
/// a second such as `:00001`, followed by an optional time zone offset such as `Z` or
/// `-07:00`. A missing fraction is assumed to be 0. A missing time zone offset implies the
/// local time zone.
pub fn parse(input: &str) -> Result<Self, Error> {
// First try parsing as 90,000ths of a second since epoch.
match i64::from_str(input) {
Ok(i) => return Ok(Time(i)),
Err(_) => {},
}
// If that failed, parse as a time string or bust.
let (remaining, ((tm_year, tm_mon, tm_mday), opt_time, opt_zone)) =
tuple((parse_datepart,
opt(preceded(tag("T"), parse_timepart)),
opt(parse_zone)))(input)
.map_err(|e| match e {
nom::Err::Incomplete(_) => format_err!("incomplete"),
nom::Err::Error(e) | nom::Err::Failure(e) => {
format_err!("{}", nom::error::convert_error(input, e))
}
})?;
if remaining != "" {
bail!("unexpected suffix {:?} following time string", remaining);
}
let (tm_hour, tm_min, tm_sec, subsec) = opt_time.unwrap_or((0, 0, 0, 0));
let mut tm = time::Tm {
tm_sec,
tm_min,
tm_hour,
tm_mday,
tm_mon,
tm_year,
tm_wday: 0,
tm_yday: 0,
tm_isdst: -1,
tm_utcoff: 0,
tm_nsec: 0,
};
if tm.tm_mon == 0 {
bail!("time {:?} has month 0", input);
}
tm.tm_mon -= 1;
if tm.tm_year < 1900 {
bail!("time {:?} has year before 1900", input);
}
tm.tm_year -= 1900;
// The time crate doesn't use tm_utcoff properly; it just calls timegm() if tm_utcoff == 0,
// mktime() otherwise. If a zone is specified, use the timegm path and a manual offset.
// If no zone is specified, use the tm_utcoff path. This is pretty lame, but follow the
// chrono crate's lead and just use 0 or 1 to choose between these functions.
let sec = if let Some(off) = opt_zone {
tm.to_timespec().sec + i64::from(off)
} else {
tm.tm_utcoff = 1;
tm.to_timespec().sec
};
Ok(Time(sec * TIME_UNITS_PER_SEC + i64::from(subsec)))
}
/// Convert to unix seconds by floor method (rounding down).
pub fn unix_seconds(&self) -> i64 { self.0 / TIME_UNITS_PER_SEC }
}
impl std::str::FromStr for Time {
type Err = Error;
fn from_str(s: &str) -> Result<Self, Self::Err> { Self::parse(s) }
}
impl ops::Sub for Time {
type Output = Duration;
fn sub(self, rhs: Time) -> Duration { Duration(self.0 - rhs.0) }
}
impl ops::AddAssign<Duration> for Time {
fn add_assign(&mut self, rhs: Duration) { self.0 += rhs.0 }
}
impl ops::Add<Duration> for Time {
type Output = Time;
fn add(self, rhs: Duration) -> Time { Time(self.0 + rhs.0) }
}
impl ops::Sub<Duration> for Time {
type Output = Time;
fn sub(self, rhs: Duration) -> Time { Time(self.0 - rhs.0) }
}
impl fmt::Debug for Time {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
// Write both the raw and display forms.
write!(f, "{} /* {} */", self.0, self)
}
}
impl fmt::Display for Time {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
let tm = time::at(time::Timespec{sec: self.0 / TIME_UNITS_PER_SEC, nsec: 0});
let zone_minutes = tm.tm_utcoff.abs() / 60;
write!(f, "{}:{:05}{}{:02}:{:02}", tm.strftime("%FT%T").or_else(|_| Err(fmt::Error))?,
self.0 % TIME_UNITS_PER_SEC,
if tm.tm_utcoff > 0 { '+' } else { '-' }, zone_minutes / 60, zone_minutes % 60)
}
}
/// A duration specified in 1/90,000ths of a second.
/// Durations are typically non-negative, but a `moonfire_db::db::CameraDayValue::duration` may be
/// negative.
#[derive(Clone, Copy, Debug, Default, Eq, Ord, PartialEq, PartialOrd)]
pub struct Duration(pub i64);
impl Duration {
pub fn to_tm_duration(&self) -> time::Duration {
time::Duration::nanoseconds(self.0 * 100000 / 9)
}
}
impl fmt::Display for Duration {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
let mut seconds = self.0 / TIME_UNITS_PER_SEC;
const MINUTE_IN_SECONDS: i64 = 60;
const HOUR_IN_SECONDS: i64 = 60 * MINUTE_IN_SECONDS;
const DAY_IN_SECONDS: i64 = 24 * HOUR_IN_SECONDS;
let days = seconds / DAY_IN_SECONDS;
seconds %= DAY_IN_SECONDS;
let hours = seconds / HOUR_IN_SECONDS;
seconds %= HOUR_IN_SECONDS;
let minutes = seconds / MINUTE_IN_SECONDS;
seconds %= MINUTE_IN_SECONDS;
let mut have_written = if days > 0 {
write!(f, "{} day{}", days, if days == 1 { "" } else { "s" })?;
true
} else {
false
};
if hours > 0 {
write!(f, "{}{} hour{}", if have_written { " " } else { "" },
hours, if hours == 1 { "" } else { "s" })?;
have_written = true;
}
if minutes > 0 {
write!(f, "{}{} minute{}", if have_written { " " } else { "" },
minutes, if minutes == 1 { "" } else { "s" })?;
have_written = true;
}
if seconds > 0 || !have_written {
write!(f, "{}{} second{}", if have_written { " " } else { "" },
seconds, if seconds == 1 { "" } else { "s" })?;
}
Ok(())
}
}
impl ops::Add for Duration {
type Output = Duration;
fn add(self, rhs: Duration) -> Duration { Duration(self.0 + rhs.0) }
}
impl ops::AddAssign for Duration {
fn add_assign(&mut self, rhs: Duration) { self.0 += rhs.0 }
}
impl ops::SubAssign for Duration {
fn sub_assign(&mut self, rhs: Duration) { self.0 -= rhs.0 }
}
#[cfg(test)]
mod tests {
use super::{Duration, Time, TIME_UNITS_PER_SEC};
#[test]
fn test_parse_time() {
std::env::set_var("TZ", "America/Los_Angeles");
time::tzset();
let tests = &[
("2006-01-02T15:04:05-07:00", 102261550050000),
("2006-01-02T15:04:05:00001-07:00", 102261550050001),
("2006-01-02T15:04:05-08:00", 102261874050000),
("2006-01-02T15:04:05", 102261874050000), // implied -08:00
("2006-01-02T15:04", 102261873600000), // implied -08:00
("2006-01-02T15:04:05:00001", 102261874050001), // implied -08:00
("2006-01-02T15:04:05-00:00", 102259282050000),
("2006-01-02T15:04:05Z", 102259282050000),
("2006-01-02-08:00", 102256992000000), // implied -08:00
("2006-01-02", 102256992000000), // implied -08:00
("2006-01-02Z", 102254400000000),
("102261550050000", 102261550050000),
];
for test in tests {
assert_eq!(test.1, Time::parse(test.0).unwrap().0, "parsing {}", test.0);
}
}
#[test]
fn test_format_time() {
std::env::set_var("TZ", "America/Los_Angeles");
time::tzset();
assert_eq!("2006-01-02T15:04:05:00000-08:00", format!("{}", Time(102261874050000)));
}
#[test]
fn test_display_duration() {
let tests = &[
// (output, seconds)
("0 seconds", 0),
("1 second", 1),
("1 minute", 60),
("1 minute 1 second", 61),
("2 minutes", 120),
("1 hour", 3600),
("1 hour 1 minute", 3660),
("2 hours", 7200),
("1 day", 86400),
("1 day 1 hour", 86400 + 3600),
("2 days", 2 * 86400),
];
for test in tests {
assert_eq!(test.0, format!("{}", Duration(test.1 * TIME_UNITS_PER_SEC)));
}
}
}

View File

@ -30,7 +30,6 @@ openssl = "0.10"
parking_lot = { version = "0.9", features = [] }
prettydiff = "0.3.1"
protobuf = { git = "https://github.com/stepancheg/rust-protobuf" }
regex = "1.0"
rusqlite = "0.21.0"
smallvec = "1.0"
tempdir = "0.3"

View File

@ -30,199 +30,17 @@
use crate::coding::{append_varint32, decode_varint32, unzigzag32, zigzag32};
use crate::db;
use failure::{Error, bail, format_err};
use lazy_static::lazy_static;
use failure::{Error, bail};
use log::trace;
use regex::Regex;
use std::ops;
use std::fmt;
use std::ops::Range;
use std::str::FromStr;
use time;
pub const TIME_UNITS_PER_SEC: i64 = 90000;
pub use base::time::TIME_UNITS_PER_SEC;
pub const DESIRED_RECORDING_DURATION: i64 = 60 * TIME_UNITS_PER_SEC;
pub const MAX_RECORDING_DURATION: i64 = 5 * 60 * TIME_UNITS_PER_SEC;
/// A time specified as 90,000ths of a second since 1970-01-01 00:00:00 UTC.
#[derive(Clone, Copy, Default, Eq, Ord, PartialEq, PartialOrd)]
pub struct Time(pub i64);
impl Time {
pub fn new(tm: time::Timespec) -> Self {
Time(tm.sec * TIME_UNITS_PER_SEC + tm.nsec as i64 * TIME_UNITS_PER_SEC / 1_000_000_000)
}
pub const fn min_value() -> Self { Time(i64::min_value()) }
pub const fn max_value() -> Self { Time(i64::max_value()) }
/// Parses a time as either 90,000ths of a second since epoch or a RFC 3339-like string.
///
/// The former is 90,000ths of a second since 1970-01-01T00:00:00 UTC, excluding leap seconds.
///
/// The latter is a string such as `2006-01-02T15:04:05`, followed by an optional 90,000ths of
/// a second such as `:00001`, followed by an optional time zone offset such as `Z` or
/// `-07:00`. A missing fraction is assumed to be 0. A missing time zone offset implies the
/// local time zone.
pub fn parse(s: &str) -> Result<Self, Error> {
lazy_static! {
static ref RE: Regex = Regex::new(r#"(?x)
^
([0-9]{4})-([0-9]{2})-([0-9]{2})T([0-9]{2}):([0-9]{2}):([0-9]{2})
(?::([0-9]{5}))?
(Z|[+-]([0-9]{2}):([0-9]{2}))?
$"#).unwrap();
}
// First try parsing as 90,000ths of a second since epoch.
match i64::from_str(s) {
Ok(i) => return Ok(Time(i)),
Err(_) => {},
}
// If that failed, parse as a time string or bust.
let c = RE.captures(s).ok_or_else(|| format_err!("unparseable time {:?}", s))?;
let mut tm = time::Tm{
tm_sec: i32::from_str(c.get(6).unwrap().as_str()).unwrap(),
tm_min: i32::from_str(c.get(5).unwrap().as_str()).unwrap(),
tm_hour: i32::from_str(c.get(4).unwrap().as_str()).unwrap(),
tm_mday: i32::from_str(c.get(3).unwrap().as_str()).unwrap(),
tm_mon: i32::from_str(c.get(2).unwrap().as_str()).unwrap(),
tm_year: i32::from_str(c.get(1).unwrap().as_str()).unwrap(),
tm_wday: 0,
tm_yday: 0,
tm_isdst: -1,
tm_utcoff: 0,
tm_nsec: 0,
};
if tm.tm_mon == 0 {
bail!("time {:?} has month 0", s);
}
tm.tm_mon -= 1;
if tm.tm_year < 1900 {
bail!("time {:?} has year before 1900", s);
}
tm.tm_year -= 1900;
// The time crate doesn't use tm_utcoff properly; it just calls timegm() if tm_utcoff == 0,
// mktime() otherwise. If a zone is specified, use the timegm path and a manual offset.
// If no zone is specified, use the tm_utcoff path. This is pretty lame, but follow the
// chrono crate's lead and just use 0 or 1 to choose between these functions.
let sec = if let Some(zone) = c.get(8) {
tm.to_timespec().sec + if zone.as_str() == "Z" {
0
} else {
let off = i64::from_str(c.get(9).unwrap().as_str()).unwrap() * 3600 +
i64::from_str(c.get(10).unwrap().as_str()).unwrap() * 60;
if zone.as_str().as_bytes()[0] == b'-' { off } else { -off }
}
} else {
tm.tm_utcoff = 1;
tm.to_timespec().sec
};
let fraction = if let Some(f) = c.get(7) { i64::from_str(f.as_str()).unwrap() } else { 0 };
Ok(Time(sec * TIME_UNITS_PER_SEC + fraction))
}
/// Convert to unix seconds by floor method (rounding down).
pub fn unix_seconds(&self) -> i64 { self.0 / TIME_UNITS_PER_SEC }
}
impl ops::Sub for Time {
type Output = Duration;
fn sub(self, rhs: Time) -> Duration { Duration(self.0 - rhs.0) }
}
impl ops::AddAssign<Duration> for Time {
fn add_assign(&mut self, rhs: Duration) { self.0 += rhs.0 }
}
impl ops::Add<Duration> for Time {
type Output = Time;
fn add(self, rhs: Duration) -> Time { Time(self.0 + rhs.0) }
}
impl ops::Sub<Duration> for Time {
type Output = Time;
fn sub(self, rhs: Duration) -> Time { Time(self.0 - rhs.0) }
}
impl fmt::Debug for Time {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
// Write both the raw and display forms.
write!(f, "{} /* {} */", self.0, self)
}
}
impl fmt::Display for Time {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
let tm = time::at(time::Timespec{sec: self.0 / TIME_UNITS_PER_SEC, nsec: 0});
let zone_minutes = tm.tm_utcoff.abs() / 60;
write!(f, "{}:{:05}{}{:02}:{:02}", tm.strftime("%FT%T").or_else(|_| Err(fmt::Error))?,
self.0 % TIME_UNITS_PER_SEC,
if tm.tm_utcoff > 0 { '+' } else { '-' }, zone_minutes / 60, zone_minutes % 60)
}
}
/// A duration specified in 1/90,000ths of a second.
/// Durations are typically non-negative, but a `db::CameraDayValue::duration` may be negative.
#[derive(Clone, Copy, Debug, Default, Eq, Ord, PartialEq, PartialOrd)]
pub struct Duration(pub i64);
impl Duration {
pub fn to_tm_duration(&self) -> time::Duration {
time::Duration::nanoseconds(self.0 * 100000 / 9)
}
}
impl fmt::Display for Duration {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
let mut seconds = self.0 / TIME_UNITS_PER_SEC;
const MINUTE_IN_SECONDS: i64 = 60;
const HOUR_IN_SECONDS: i64 = 60 * MINUTE_IN_SECONDS;
const DAY_IN_SECONDS: i64 = 24 * HOUR_IN_SECONDS;
let days = seconds / DAY_IN_SECONDS;
seconds %= DAY_IN_SECONDS;
let hours = seconds / HOUR_IN_SECONDS;
seconds %= HOUR_IN_SECONDS;
let minutes = seconds / MINUTE_IN_SECONDS;
seconds %= MINUTE_IN_SECONDS;
let mut have_written = if days > 0 {
write!(f, "{} day{}", days, if days == 1 { "" } else { "s" })?;
true
} else {
false
};
if hours > 0 {
write!(f, "{}{} hour{}", if have_written { " " } else { "" },
hours, if hours == 1 { "" } else { "s" })?;
have_written = true;
}
if minutes > 0 {
write!(f, "{}{} minute{}", if have_written { " " } else { "" },
minutes, if minutes == 1 { "" } else { "s" })?;
have_written = true;
}
if seconds > 0 || !have_written {
write!(f, "{}{} second{}", if have_written { " " } else { "" },
seconds, if seconds == 1 { "" } else { "s" })?;
}
Ok(())
}
}
impl ops::Add for Duration {
type Output = Duration;
fn add(self, rhs: Duration) -> Duration { Duration(self.0 + rhs.0) }
}
impl ops::AddAssign for Duration {
fn add_assign(&mut self, rhs: Duration) { self.0 += rhs.0 }
}
impl ops::SubAssign for Duration {
fn sub_assign(&mut self, rhs: Duration) { self.0 -= rhs.0 }
}
pub use base::time::Time;
pub use base::time::Duration;
/// An iterator through a sample index.
/// Initially invalid; call `next()` before each read.
@ -533,52 +351,6 @@ mod tests {
use super::*;
use crate::testutil::{self, TestDb};
#[test]
fn test_parse_time() {
testutil::init();
let tests = &[
("2006-01-02T15:04:05-07:00", 102261550050000),
("2006-01-02T15:04:05:00001-07:00", 102261550050001),
("2006-01-02T15:04:05-08:00", 102261874050000),
("2006-01-02T15:04:05", 102261874050000), // implied -08:00
("2006-01-02T15:04:05:00001", 102261874050001), // implied -08:00
("2006-01-02T15:04:05-00:00", 102259282050000),
("2006-01-02T15:04:05Z", 102259282050000),
("102261550050000", 102261550050000),
];
for test in tests {
assert_eq!(test.1, Time::parse(test.0).unwrap().0, "parsing {}", test.0);
}
}
#[test]
fn test_format_time() {
testutil::init();
assert_eq!("2006-01-02T15:04:05:00000-08:00", format!("{}", Time(102261874050000)));
}
#[test]
fn test_display_duration() {
testutil::init();
let tests = &[
// (output, seconds)
("0 seconds", 0),
("1 second", 1),
("1 minute", 60),
("1 minute 1 second", 61),
("2 minutes", 120),
("1 hour", 3600),
("1 hour 1 minute", 3660),
("2 hours", 7200),
("1 day", 86400),
("1 day 1 hour", 86400 + 3600),
("2 days", 2 * 86400),
];
for test in tests {
assert_eq!(test.0, format!("{}", Duration(test.1 * TIME_UNITS_PER_SEC)));
}
}
/// Tests encoding the example from design/schema.md.
#[test]
fn test_encode_example() {

View File

@ -42,8 +42,6 @@
use byteorder::{BigEndian, WriteBytesExt};
use failure::{Error, bail};
use lazy_static::lazy_static;
use regex::bytes::Regex;
// See ISO/IEC 14496-10 table 7-1 - NAL unit type codes, syntax element categories, and NAL unit
// type classes.
@ -60,15 +58,28 @@ const NAL_UNIT_TYPE_MASK: u8 = 0x1F; // bottom 5 bits of first byte of unit.
///
/// TODO: detect invalid byte streams. For example, several 0x00s not followed by a 0x01, a stream
/// stream not starting with 0x00 0x00 0x00 0x01, or an empty NAL unit.
fn decode_h264_annex_b<'a, F>(data: &'a [u8], mut f: F) -> Result<(), Error>
fn decode_h264_annex_b<'a, F>(mut data: &'a [u8], mut f: F) -> Result<(), Error>
where F: FnMut(&'a [u8]) -> Result<(), Error> {
lazy_static! {
static ref START_CODE: Regex = Regex::new(r"(\x00{2,}\x01)").unwrap();
}
for unit in START_CODE.split(data) {
if !unit.is_empty() {
f(unit)?;
let start_code = &b"\x00\x00\x01"[..];
use nom::FindSubstring;
'outer: while let Some(pos) = data.find_substring(start_code) {
let mut unit = &data[0..pos];
data = &data[pos + start_code.len() ..];
// Have zero or more bytes that end in a start code. Strip out any trailing 0x00s and
// process the unit if there's anything left.
loop {
match unit.last() {
None => continue 'outer,
Some(b) if *b == 0 => { unit = &unit[..unit.len()-1]; },
Some(_) => break,
}
}
f(unit)?;
}
// No remaining start codes; likely a unit left.
if !data.is_empty() {
f(data)?;
}
Ok(())
}

View File

@ -34,7 +34,6 @@ use bytes::Bytes;
use crate::body::{Body, BoxedError};
use crate::json;
use crate::mp4;
use base64;
use bytes::{BufMut, BytesMut};
use core::borrow::Borrow;
use core::str::FromStr;
@ -46,12 +45,12 @@ use futures::sink::SinkExt;
use futures::future::{self, Future, TryFutureExt};
use futures::stream::StreamExt;
use http::{Request, Response, status::StatusCode};
use http_serve;
use http::header::{self, HeaderValue};
use lazy_static::lazy_static;
use log::{debug, info, warn};
use regex::Regex;
use serde_json;
use nom::IResult;
use nom::bytes::complete::{take_while1, tag};
use nom::combinator::{all_consuming, map, map_res, opt};
use nom::sequence::{preceded, tuple};
use std::collections::HashMap;
use std::cmp;
use std::fs;
@ -64,14 +63,6 @@ use tokio_tungstenite::tungstenite;
use url::form_urlencoded;
use uuid::Uuid;
lazy_static! {
/// Regex used to parse the `s` query parameter to `view.mp4`.
/// As described in `design/api.md`, this is of the form
/// `START_ID[-END_ID][@OPEN_ID][.[REL_START_TIME]-[REL_END_TIME]]`.
static ref SEGMENTS_RE: Regex =
Regex::new(r"^(\d+)(-\d+)?(@\d+)?(?:\.(\d+)?-(\d+)?)?$").unwrap();
}
type BoxedFuture = Box<dyn Future<Output = Result<Response<Body>, BoxedError>> +
Sync + Send + 'static>;
@ -202,41 +193,48 @@ struct Segments {
end_time: Option<i64>,
}
fn num<'a, T: FromStr>() -> impl Fn(&'a str) -> IResult<&'a str, T> {
map_res(take_while1(|c: char| c.is_ascii_digit()), FromStr::from_str)
}
impl Segments {
pub fn parse(input: &str) -> Result<Segments, ()> {
let caps = SEGMENTS_RE.captures(input).ok_or(())?;
let ids_start = i32::from_str(caps.get(1).unwrap().as_str()).map_err(|_| ())?;
let ids_end = match caps.get(2) {
Some(m) => i32::from_str(&m.as_str()[1..]).map_err(|_| ())?,
None => ids_start,
} + 1;
let open_id = match caps.get(3) {
Some(m) => Some(u32::from_str(&m.as_str()[1..]).map_err(|_| ())?),
None => None,
};
if ids_start < 0 || ids_end <= ids_start {
/// Parses the `s` query parameter to `view.mp4` as described in `design/api.md`.
/// Doesn't do any validation.
fn parse(i: &str) -> IResult<&str, Segments> {
// Parse START_ID[-END_ID] into Range<i32>.
// Note that END_ID is inclusive, but Ranges are half-open.
let (i, ids) = map(tuple((num::<i32>(), opt(preceded(tag("-"), num::<i32>())))),
|(start, end)| start .. end.unwrap_or(start) + 1)(i)?;
// Parse [@OPEN_ID] into Option<u32>.
let (i, open_id) = opt(preceded(tag("@"), num::<u32>()))(i)?;
// Parse [.[REL_START_TIME]-[REL_END_TIME]] into (i64, Option<i64>).
let (i, (start_time, end_time)) = map(
opt(preceded(tag("."), tuple((opt(num::<i64>()), tag("-"), opt(num::<i64>()))))),
|t| {
t.map(|(s, _, e)| (s.unwrap_or(0), e))
.unwrap_or((0, None))
})(i)?;
Ok((i, Segments { ids, open_id, start_time, end_time, }))
}
}
impl FromStr for Segments {
type Err = ();
fn from_str(s: &str) -> Result<Self, Self::Err> {
let (_, s) = all_consuming(Segments::parse)(s).map_err(|_| ())?;
if s.ids.end <= s.ids.start {
return Err(());
}
let start_time = caps.get(4).map_or(Ok(0), |m| i64::from_str(m.as_str())).map_err(|_| ())?;
if start_time < 0 {
return Err(());
if let Some(e) = s.end_time {
if e < s.start_time {
return Err(());
}
}
let end_time = match caps.get(5) {
Some(v) => {
let e = i64::from_str(v.as_str()).map_err(|_| ())?;
if e <= start_time {
return Err(());
}
Some(e)
},
None => None
};
Ok(Segments {
ids: ids_start .. ids_end,
open_id,
start_time,
end_time,
})
Ok(s)
}
}
@ -422,7 +420,7 @@ impl ServiceInner {
let (key, value) = (key.borrow(), value.borrow());
match key {
"s" => {
let s = Segments::parse(value).map_err(
let s = Segments::from_str(value).map_err(
|()| plain_response(StatusCode::BAD_REQUEST,
format!("invalid s parameter: {}", value)))?;
debug!("stream_view_mp4: appending s={:?}", s);
@ -1078,6 +1076,7 @@ mod tests {
use futures::future::FutureExt;
use log::info;
use std::collections::HashMap;
use std::str::FromStr;
use super::Segments;
struct Server {
@ -1233,25 +1232,25 @@ mod tests {
fn test_segments() {
testutil::init();
assert_eq!(Segments{ids: 1..2, open_id: None, start_time: 0, end_time: None},
Segments::parse("1").unwrap());
Segments::from_str("1").unwrap());
assert_eq!(Segments{ids: 1..2, open_id: Some(42), start_time: 0, end_time: None},
Segments::parse("1@42").unwrap());
Segments::from_str("1@42").unwrap());
assert_eq!(Segments{ids: 1..2, open_id: None, start_time: 26, end_time: None},
Segments::parse("1.26-").unwrap());
Segments::from_str("1.26-").unwrap());
assert_eq!(Segments{ids: 1..2, open_id: Some(42), start_time: 26, end_time: None},
Segments::parse("1@42.26-").unwrap());
Segments::from_str("1@42.26-").unwrap());
assert_eq!(Segments{ids: 1..2, open_id: None, start_time: 0, end_time: Some(42)},
Segments::parse("1.-42").unwrap());
Segments::from_str("1.-42").unwrap());
assert_eq!(Segments{ids: 1..2, open_id: None, start_time: 26, end_time: Some(42)},
Segments::parse("1.26-42").unwrap());
Segments::from_str("1.26-42").unwrap());
assert_eq!(Segments{ids: 1..6, open_id: None, start_time: 0, end_time: None},
Segments::parse("1-5").unwrap());
Segments::from_str("1-5").unwrap());
assert_eq!(Segments{ids: 1..6, open_id: None, start_time: 26, end_time: None},
Segments::parse("1-5.26-").unwrap());
Segments::from_str("1-5.26-").unwrap());
assert_eq!(Segments{ids: 1..6, open_id: None, start_time: 0, end_time: Some(42)},
Segments::parse("1-5.-42").unwrap());
Segments::from_str("1-5.-42").unwrap());
assert_eq!(Segments{ids: 1..6, open_id: None, start_time: 26, end_time: Some(42)},
Segments::parse("1-5.26-42").unwrap());
Segments::from_str("1-5.26-42").unwrap());
}
#[tokio::test]