Fix typo in comment, remove unnecessary escaping, and simplify toms (#…

…45885)

stdlib/Dates/src/periods.jl

@@ -160,7 +160,7 @@ coarserperiod(::Type{Month}) = (Year, 12)
     CompoundPeriod
 
 A `CompoundPeriod` is useful for expressing time periods that are not a fixed multiple of
-smaller periods. For example, \"a year and a  day\" is not a fixed number of days, but can
+smaller periods. For example, "a year and a  day" is not a fixed number of days, but can
 be expressed using a `CompoundPeriod`. In fact, a `CompoundPeriod` is automatically
 generated by addition of different period types, e.g. `Year(1) + Day(1)` produces a
 `CompoundPeriod` result.
@@ -482,7 +482,7 @@ const zero_or_fixedperiod_seed = UInt === UInt64 ? 0x5b7fc751bba97516 : 0xeae0fd
 const nonzero_otherperiod_seed = UInt === UInt64 ? 0xe1837356ff2d2ac9 : 0x170d1b00
 otherperiod_seed(x::OtherPeriod) = iszero(value(x)) ? zero_or_fixedperiod_seed : nonzero_otherperiod_seed
 # tons() will overflow for periods longer than ~300,000 years, implying a hash collision
-# which is relatively harmless given how infrequent such periods should appear
+# which is relatively harmless given how infrequently such periods should appear
 Base.hash(x::FixedPeriod, h::UInt) = hash(tons(x), h + zero_or_fixedperiod_seed)
 # Overflow can also happen here for really long periods (~8e17 years)
 Base.hash(x::Year, h::UInt) = hash(12 * value(x), h + otherperiod_seed(x))
@@ -511,11 +511,7 @@ toms(c::Millisecond) = value(c)
 toms(c::Second)      = 1000 * value(c)
 toms(c::Minute)      = 60000 * value(c)
 toms(c::Hour)        = 3600000 * value(c)
-toms(c::Day)         = 86400000 * value(c)
-toms(c::Week)        = 604800000 * value(c)
-toms(c::Month)       = 86400000.0 * 30.436875 * value(c)
-toms(c::Quarter)     = 86400000.0 * 91.310625 * value(c)
-toms(c::Year)        = 86400000.0 * 365.2425 * value(c)
+toms(c::Period)      = 86400000 * days(c)
 toms(c::CompoundPeriod) = isempty(c.periods) ? 0.0 : Float64(sum(toms, c.periods))
 tons(x)              = toms(x) * 1000000
 tons(x::Microsecond) = value(x) * 1000