Skip to content

Commit

Permalink
Showing 2 changed files with 26 additions and 25 deletions.
5 changes: 3 additions & 2 deletions scripts/generate.rb
Original file line number Diff line number Diff line change
@@ -167,10 +167,11 @@ def generate_coderange_list
end

ranges = unicode_src[/CodeRanges\[\]\s+=\s+\{(.*?)\}\;/m, 1].scan(/CR_(\w+)/).flatten
out = ranges.map do |range|

out = ranges.take(14).map{|range|[range.tr('_', '').downcase, range]} +
ranges.drop(14).map do |range|
name = range =~ /Age_(\d)_(\d)/ ? "age=#{$1}.#{$2}" : range.tr('_', '').downcase
name = cr_map.delete(range) || name

([name] + aliases[name].to_a).map{|n|[n, range]}
end.flatten(1)

46 changes: 23 additions & 23 deletions src/org/jcodings/unicode/UnicodeProperties.java
Original file line number Diff line number Diff line change
@@ -23,31 +23,27 @@

public class UnicodeProperties {
static final CodeRangeEntry[]CodeRangeTable = new CodeRangeEntry[] {
new CodeRangeEntry("graphemeclusterbreaklf", "CR_NEWLINE"),
new CodeRangeEntry("alphabetic", "CR_Alpha"),
new CodeRangeEntry("newline", "CR_NEWLINE"),
new CodeRangeEntry("alpha", "CR_Alpha"),
new CodeRangeEntry("blank", "CR_Blank"),
new CodeRangeEntry("cc", "CR_Cntrl"),
new CodeRangeEntry("control", "CR_Cntrl"),
new CodeRangeEntry("nd", "CR_Digit"),
new CodeRangeEntry("decimalnumber", "CR_Digit"),
new CodeRangeEntry("cntrl", "CR_Cntrl"),
new CodeRangeEntry("digit", "CR_Digit"),
new CodeRangeEntry("graph", "CR_Graph"),
new CodeRangeEntry("lowercase", "CR_Lower"),
new CodeRangeEntry("lower", "CR_Lower"),
new CodeRangeEntry("print", "CR_Print"),
new CodeRangeEntry("p", "CR_Punct"),
new CodeRangeEntry("punctuation", "CR_Punct"),
new CodeRangeEntry("whitespace", "CR_Space"),
new CodeRangeEntry("wspace", "CR_Space"),
new CodeRangeEntry("uppercase", "CR_Upper"),
new CodeRangeEntry("asciihexdigit", "CR_XDigit"),
new CodeRangeEntry("ahex", "CR_XDigit"),
new CodeRangeEntry("punct", "CR_Punct"),
new CodeRangeEntry("space", "CR_Space"),
new CodeRangeEntry("upper", "CR_Upper"),
new CodeRangeEntry("xdigit", "CR_XDigit"),
new CodeRangeEntry("word", "CR_Word"),
new CodeRangeEntry("alnum", "CR_Alnum"),
new CodeRangeEntry("inbasiclatin", "CR_ASCII"),
new CodeRangeEntry("any", "CR_Any"),
new CodeRangeEntry("assigned", "CR_Assigned"),
new CodeRangeEntry("c", "CR_C"),
new CodeRangeEntry("other", "CR_C"),
new CodeRangeEntry("cntrl", "CR_Cntrl"),
new CodeRangeEntry("cc", "CR_Cntrl"),
new CodeRangeEntry("control", "CR_Cntrl"),
new CodeRangeEntry("cf", "CR_Cf"),
new CodeRangeEntry("format", "CR_Cf"),
new CodeRangeEntry("cn", "CR_Cn"),
@@ -81,12 +77,14 @@ public class UnicodeProperties {
new CodeRangeEntry("nonspacingmark", "CR_Mn"),
new CodeRangeEntry("n", "CR_N"),
new CodeRangeEntry("number", "CR_N"),
new CodeRangeEntry("digit", "CR_Digit"),
new CodeRangeEntry("nd", "CR_Digit"),
new CodeRangeEntry("decimalnumber", "CR_Digit"),
new CodeRangeEntry("nl", "CR_Nl"),
new CodeRangeEntry("letternumber", "CR_Nl"),
new CodeRangeEntry("no", "CR_No"),
new CodeRangeEntry("othernumber", "CR_No"),
new CodeRangeEntry("punct", "CR_Punct"),
new CodeRangeEntry("p", "CR_Punct"),
new CodeRangeEntry("punctuation", "CR_Punct"),
new CodeRangeEntry("pc", "CR_Pc"),
new CodeRangeEntry("connectorpunctuation", "CR_Pc"),
new CodeRangeEntry("pd", "CR_Pd"),
@@ -120,9 +118,9 @@ public class UnicodeProperties {
new CodeRangeEntry("zs", "CR_Zs"),
new CodeRangeEntry("spaceseparator", "CR_Zs"),
new CodeRangeEntry("math", "CR_Math"),
new CodeRangeEntry("alpha", "CR_Alpha"),
new CodeRangeEntry("lower", "CR_Lower"),
new CodeRangeEntry("upper", "CR_Upper"),
new CodeRangeEntry("alphabetic", "CR_Alpha"),
new CodeRangeEntry("lowercase", "CR_Lower"),
new CodeRangeEntry("uppercase", "CR_Upper"),
new CodeRangeEntry("cased", "CR_Cased"),
new CodeRangeEntry("caseignorable", "CR_Case_Ignorable"),
new CodeRangeEntry("ci", "CR_Case_Ignorable"),
@@ -413,7 +411,8 @@ public class UnicodeProperties {
new CodeRangeEntry("osge", "CR_Osage"),
new CodeRangeEntry("tangut", "CR_Tangut"),
new CodeRangeEntry("tang", "CR_Tangut"),
new CodeRangeEntry("space", "CR_Space"),
new CodeRangeEntry("whitespace", "CR_Space"),
new CodeRangeEntry("wspace", "CR_Space"),
new CodeRangeEntry("bidicontrol", "CR_Bidi_Control"),
new CodeRangeEntry("bidic", "CR_Bidi_Control"),
new CodeRangeEntry("joincontrol", "CR_Join_Control"),
@@ -428,7 +427,8 @@ public class UnicodeProperties {
new CodeRangeEntry("omath", "CR_Other_Math"),
new CodeRangeEntry("hexdigit", "CR_Hex_Digit"),
new CodeRangeEntry("hex", "CR_Hex_Digit"),
new CodeRangeEntry("xdigit", "CR_XDigit"),
new CodeRangeEntry("asciihexdigit", "CR_XDigit"),
new CodeRangeEntry("ahex", "CR_XDigit"),
new CodeRangeEntry("otheralphabetic", "CR_Other_Alphabetic"),
new CodeRangeEntry("oalpha", "CR_Other_Alphabetic"),
new CodeRangeEntry("ideographic", "CR_Ideographic"),
@@ -496,7 +496,7 @@ public class UnicodeProperties {
new CodeRangeEntry("age=9.0", "CR_Age_9_0"),
new CodeRangeEntry("graphemeclusterbreakprepend", "CR_Grapheme_Cluster_Break_Prepend"),
new CodeRangeEntry("graphemeclusterbreakcr", "CR_Grapheme_Cluster_Break_CR"),
new CodeRangeEntry("newline", "CR_NEWLINE"),
new CodeRangeEntry("graphemeclusterbreaklf", "CR_NEWLINE"),
new CodeRangeEntry("graphemeclusterbreakcontrol", "CR_Grapheme_Cluster_Break_Control"),
new CodeRangeEntry("graphemeextend", "CR_Grapheme_Extend"),
new CodeRangeEntry("grext", "CR_Grapheme_Extend"),

0 comments on commit 9cb42f5

Please sign in to comment.