ref: 74d91a0021de012908cfdb35fb61a1473a376130
dir: /lib/date/parse.myr/
use std
use "types"
use "names"
use "date"
use "zoneinfo"
pkg date =
type parsefail = union
`Doublefmt char
`Badsep (char, char)
`Badfmt char
`Badzone byte[:]
`Badname byte[:]
`Badchar
`Badampm
`Shortint
`Badint
;;
/* date i/o */
const parsefmt : (f : byte[:], s: byte[:] -> std.result(instant, parsefail))
const parsefmtl : (f : byte[:], s: byte[:] -> std.result(instant, parsefail))
const parsefmtz : (f : byte[:], s: byte[:], tz : byte[:] -> std.result(instant, parsefail))
;;
const UnixJulianDiff = 719468
const parsefmt = {f, s; -> strparse(f, s, "", false)}
const parsefmtl = {f, s; -> strparse(f, s, "local", true)}
const parsefmtz = {f, s, tz; -> strparse(f, s, tz, true)}
type parsedtz = union
`None
`Off duration
`Name byte[:]
;;
const __init__ = {
var fail : parsefail
std.fmtinstall(std.typeof(fail), failfmt)
}
const strparse = {f, s, tz, replace
var d, err
var seen
d = [.year = 0]
err = `std.None
seen = std.mkbs()
s = filldate(&d, f, s, seen, &err)
std.bsfree(seen)
d.actual -= (d.tzoff : std.time)
match err
| `std.Some e: -> `std.Err e
| `std.None: /* no error, we're ok */
;;
if replace
d = mkinstant(d.actual, tz)
;;
-> `std.Ok d
}
const filldate = {d, f, s, seen, err
var fc, sc, z, am
z = ""
am = `std.None
while f.len != 0
(fc, f) = std.charstep(f)
if fc == '%'
(fc, f) = std.charstep(f)
if std.bshas(seen, fc)
err# = `std.Some `Doublefmt fc
-> s
;;
std.bsput(seen, fc)
match fc
/* named things */
| 'a': s = indexof(&d.day, s, _names.abbrevday[:], err)
| 'A': s = indexof(&d.day, s, _names.fullday[:], err)
| 'b': s = indexof(&d.mon, s, _names.abbrevmon[:], err)
| 'B': s = indexof(&d.mon, s, _names.fullmon[:], err)
| 'c': s = filldate(d, "%Y-%m-%d", s, seen, err)
| 'C':
s = intval(&d.year, s, 2, 2, err)
d.year += 1900
| 'd': s = intval(&d.day, s, 2, 2, err)
| 'D': s = filldate(d, "%m/%d/%y", s, seen, err)
| 'e': s = intval(&d.day, s, 1, 2, err)
| 'F': s = filldate(d, "%y-%m-%d", s, seen, err)
| 'h': s = indexof(&d.day, s, _names.abbrevmon[:], err)
| 'H': s = intval(&d.h, s, 1, 2, err)
| 'I': s = intval(&d.h, s, 1, 2, err)
| 'k': s = intval(&d.h, s, 1, 2, err)
| 'l': s = intval(&d.h, s, 1, 2, err)
| 'm': s = intval(&d.mon, s, 1, 2, err)
| 'M': s = intval(&d.m, s, 1, 2, err)
| 'n': s = matchstr(s, "\n", err)
| 'p': s = matchampm(d, s, &am, err)
| 'P': s = matchampm(d, s, &am, err)
| 'r': s = filldate(d, "%H:%M:%S %P", s, seen, err)
| 'R': s = filldate(d, "%H:%M %P", s, seen, err)
| 's': s = intval(&d.actual, s, 1, 64, err)
| 'S': s = intval(&d.s, s, 1, 2, err)
| 't': s = eatspace(s)
| 'u': s = intval(&d.wday, s, 1, 1, err)
| 'x': s = filldate(d, Datefmt, s, seen, err)
| 'X': s = filldate(d, Timefmt, s, seen, err)
| 'y': s = intval(&d.year, s, 1, 2, err)
d.year += 1900
| 'Y':
s = intval(&d.year, s, 1, 4, err)
| 'z': s = tzoffset(&d.tzoff, s, err)
| 'Z': (s, z) = tzstring(d, s, err)
| '%': s = matchstr(s, "%", err)
| _: std.fatal("unknown format character %c\n", fc)
;;
else
(sc, s) = std.charstep(s)
if std.isspace(fc) && std.isspace(sc)
s = eatspace(s)
elif sc != fc
err# = `std.Some `Badsep (fc, sc)
-> s
;;
;;
match err#
| `std.Some _: -> s
| `std.None:
;;
;;
d.actual = recalc(d)
if z.len > 0
match _zoneinfo.findtzoff(z, d.actual)
| `std.Some o: d.tzoff = o
| `std.None: err# = `std.Some `Badzone z
;;
;;
match am
| `std.None:
| `std.Some true:
d.h %= 12
| `std.Some false:
d.h %= 12
d.h += 12
;;
-> s
}
const eatspace = {s
var c
while std.isspace(std.decode(s))
(c, s) = std.charstep(s)
;;
-> s
}
const indexof = {dst, s, set, err
for var i = 0; i < set.len; i++
if s.len >= set[i].len && std.eq(s[:set[i].len], set[i])
dst# = i
-> s[set[i].len:]
;;
;;
err# = `std.Some `Badname s
dst# = 0
-> s
}
const tzoffset = {dst, s, err
var sgn
var tzoff
if s.len < 1
err# = `std.Some `Badzone s
-> ""
;;
if std.eq(s[:1], "-")
sgn = -1
elif std.eq(s[:1], "+")
sgn = 1
else
err# = `std.Some `Badzone s
-> s
;;
s = intval(&tzoff, s[1:], 2, 4, err)
dst# = sgn * (tzoff / 100) * 3600 * 1_000_000 + (tzoff % 100) * 60 * 1_000_000
-> s
}
const tzstring = {d, s, err
var c, n
n = 0
while true
c = std.decode(s[n:])
if c != '/' && !std.isalnum(c)
break
;;
n += std.charlen(c)
;;
if n < d._tzbuf.len
std.slcp(d._tzbuf[:n], s[:n])
else
err# = `std.Some `Badzone s[:n]
;;
-> (s[n:], s[:n])
}
const matchstr = {s, str, err
if s.len <= str.len || !std.eq(s[:str.len], str)
err# = `std.Some `Badchar
-> s
;;
-> s[str.len:]
}
const matchampm = {d, s, am, err
if s.len < 2
err# = `std.Some `Badampm
-> s
;;
if std.eq(s[:2], "am") || std.eq(s[:2], "AM")
am# = `std.Some true
-> s[2:]
elif std.eq(s[:2], "pm") || std.eq(s[:2], "PM")
am# = `std.Some false
-> s[2:]
else
err# = `std.Some `Badampm
-> s
;;
}
generic intval = {dst : @a#, s : byte[:], min : @a, max : @a, err : std.option(parsefail)# -> byte[:] :: numeric,integral @a
var i, c, num
num = s
for i = 0; i < min; i++
(c, s) = std.charstep(s)
if !std.isdigit(c)
err# = `std.Some `Shortint
-> s
;;
;;
for i = min ; i < max; i++
c = std.decode(s)
if !std.isdigit(c)
break
;;
s = s[std.charlen(c):]
;;
num = num[:i]
match std.intparse(num)
| `std.Some v:
dst# = v
-> s
| `std.None:
err# = `std.Some `Badint
-> s
;;
}
const failfmt = {sb, ap, opt
match std.vanext(ap)
| `Doublefmt chr: std.sbfmt(sb, "saw duplicate format char '{}'", chr)
| `Badsep (e, f): std.sbfmt(sb, "expected separator '{}', found '{}'", e, f)
| `Badfmt chr: std.sbfmt(sb, "invalid format character '{}'", chr)
| `Badzone zone: std.sbfmt(sb, "unknown time zone '{}'", zone)
| `Badname name: std.sbfmt(sb, "could not find name '{}'", name)
| `Badchar: std.sbfmt(sb, "unexpected character in parsed string")
| `Badampm: std.sbfmt(sb, "invalid am/pm specifier")
| `Shortint: std.sbfmt(sb, "integer had too few characters")
| `Badint: std.sbfmt(sb, "integer could not be parsed")
;;
}