#!/bin/sh

cache()
{
	name="/tmp/librune/rtype/$(basename "$1")"
	if test ! -f "$name"
	then
		mkdir -p /tmp/librune/rtype
		wget -q "$1" -O "$name"
	fi
}

cd "${0%/*}/.."

readonly URL1='https://www.unicode.org/Public/UCD/latest/ucd/PropList.txt'
readonly URL2='https://www.unicode.org/Public/UCD/latest/ucd/DerivedCoreProperties.txt'
readonly URL3='https://www.unicode.org/Public/UCD/latest/ucd/emoji/emoji-data.txt'
readonly URL4='https://www.unicode.org/Public/UCD/latest/ucd/DerivedNormalizationProps.txt'
readonly URL5='https://www.unicode.org/Public/UCD/latest/ucd/extracted/DerivedBinaryProperties.txt'

cache "$URL1" &
cache "$URL2" &
cache "$URL3" &
cache "$URL4" &
cache "$URL5" &
wait

props1='
bidi_c=Bidi_Control
dash=Dash
dep=Deprecated
dia=Diacritic
ext=Extender
hex=Hex_Digit
idbo=IDS_Binary_Operator
id_compat_math_continue=ID_Compat_Math_Continue
id_compat_math_start=ID_Compat_Math_Start
ideo=Ideographic
loe=Logical_Order_Exception
pat_syn=Pattern_Syntax
pcm=Prepended_Concatenation_Mark
qmark=Quotation_Mark
radical=Radical
sd=Soft_Dotted
sterm=Sentence_Terminal
term=Terminal_Punctuation
uideo=Unified_Ideograph
vs=Variation_Selector
wspace=White_Space
'

props2='
alpha=Alphabetic
cased=Cased
ci=Case_Ignorable
cwcf=Changes_When_Casefolded
cwcm=Changes_When_Casemapped
cwl=Changes_When_Lowercased
cwt=Changes_When_Titlecased
cwu=Changes_When_Uppercased
di=Default_Ignorable_Code_Point
gr_base=Grapheme_Base
gr_ext=Grapheme_Extend
idc=ID_Continue
ids=ID_Start
incb=Indic_Conjunct_Break
lower=Lowercase
math=Math
upper=Uppercase
xidc=XID_Continue
xids=XID_Start
'

props3='
ebase=Emoji_Modifier_Base
ecomp=Emoji_Component
emod=Emoji_Modifier
emoji=Emoji
epres=Emoji_Presentation
extpic=Extended_Pictographic
'

props4='
cwkcf=Changes_When_NFKC_Casefolded
'

props5='
bidi_m=Bidi_Mirrored
'

manual='
ahex=ASCII_Hex_Digit
idst=IDS_Trinary_Operator
idsu=IDS_Unary_Operator
join_c=Join_Control
nchar=Noncharacter_Code_Point
pat_ws=Pattern_White_Space
ri=Regional_Indicator
'

gen()
{
	local p=${1%%=*}
	gawk -M -v prop=${1#*=} -v word=$2 -v short=$p \
		-f gen/rtype-prop.awk /tmp/librune/rtype/$3 \
		>lib/rtype/rprop_${2}_${p}.c
	printf 'DONE rprop_%s_%s()\n' $2 $p >&2
}

for prop in $props1
do
	gen $prop is PropList.txt &
done

for prop in $props2
do
	gen $prop is DerivedCoreProperties.txt &
done

for prop in $props3
do
	gen $prop is emoji-data.txt &
done

for prop in $props4
do
	gen $prop is DerivedNormalizationProps.txt &
done

for prop in $props5
do
	gen $prop is DerivedBinaryProperties.txt &
done

printf '[[unsequenced]] bool rprop_is_%s(rune);\n' \
	$(printf '%s\n' $props1 $props2 $props3 $props4 $props5 | cut -d= -f1) \
| gawk '
	/PROP PREDICATES END/ { no = 0 }
	FILENAME != "-" && !no { print }
	FILENAME == "-" { funcs[++i] = $0 }

	/PROP PREDICATES START/ {
		no = 1
		asort(funcs)
		for (i = 1; i <= length(funcs); i++)
			print funcs[i]
	}
' - include/rtype.h | sponge include/rtype.h

wait
for prop in $manual
do
	shrt=${prop%%=*}
	printf 'Function rprop_is_%s() implemented manually\n' $shrt >&2
done