Bug #19007: Unicode tables differences from Unicode.org 14.0 data

• Status: Closed
• Assignee: duerst (Martin Dürst)
• Target version: -
• ruby -v: 3.2.0 6898984f1cd
• Backport: 2.7: DONTNEED, 3.0: DONTNEED, 3.1: DONTNEED

[nobu (Nobuyoshi Nakada)]

I found the header in Unicode Emoji 14.0 data files had changed slightly (and again at 15.0), but enc/unicode/case-folding.rb didn't follow it.
Then I fixed it and rebuilt the headers under enc/unicode/14.0.0, name2ctype.h had diffences from the master, as bellow.

CR_Lower, CR_Cased and CR_Other_Lowercase just seem misses in the previous operation, and no problems.

But U+11720..U+11721 in CR_Grapheme_Cluster_Break_SpacingMark is absent in the original data of the Unicode.org.
According to @naruse's investigation, it was removed at the commit Update to Unicode 14.0.0, while U+11720 is still SpacingMark in the latest https://www.unicode.org/reports/tr29/.

diff --git a/enc/unicode/14.0.0/name2ctype.h b/enc/unicode/14.0.0/name2ctype.h
index 99a3eeca190..f49e5cd7273 100644
--- a/enc/unicode/14.0.0/name2ctype.h
+++ b/enc/unicode/14.0.0/name2ctype.h
@@ -1565,7 +1565,7 @@ static const OnigCodePoint CR_Graph[] = {
 
 /* 'Lower': [[:Lower:]] */
 static const OnigCodePoint CR_Lower[] = {
-	664,
+	668,
 	0x0061, 0x007a,
 	0x00aa, 0x00aa,
 	0x00b5, 0x00b5,
@@ -2196,6 +2196,10 @@ static const OnigCodePoint CR_Lower[] = {
 	0x105a3, 0x105b1,
 	0x105b3, 0x105b9,
 	0x105bb, 0x105bc,
+	0x10780, 0x10780,
+	0x10783, 0x10785,
+	0x10787, 0x107b0,
+	0x107b2, 0x107ba,
 	0x10cc0, 0x10cf2,
 	0x118c0, 0x118df,
 	0x16e60, 0x16e7f,
@@ -12651,7 +12655,7 @@ static const OnigCodePoint CR_Math[] = {
 
 /* 'Cased': Derived Property */
 static const OnigCodePoint CR_Cased[] = {
-	151,
+	155,
 	0x0041, 0x005a,
 	0x0061, 0x007a,
 	0x00aa, 0x00aa,
@@ -12763,6 +12767,10 @@ static const OnigCodePoint CR_Cased[] = {
 	0x105a3, 0x105b1,
 	0x105b3, 0x105b9,
 	0x105bb, 0x105bc,
+	0x10780, 0x10780,
+	0x10783, 0x10785,
+	0x10787, 0x107b0,
+	0x107b2, 0x107ba,
 	0x10c80, 0x10cb2,
 	0x10cc0, 0x10cf2,
 	0x118a0, 0x118df,
@@ -22615,7 +22623,7 @@ static const OnigCodePoint CR_Extender[] = {
 
 /* 'Other_Lowercase': Binary Property */
 static const OnigCodePoint CR_Other_Lowercase[] = {
-	20,
+	24,
 	0x00aa, 0x00aa,
 	0x00ba, 0x00ba,
 	0x02b0, 0x02b8,
@@ -22636,6 +22644,10 @@ static const OnigCodePoint CR_Other_Lowercase[] = {
 	0xa770, 0xa770,
 	0xa7f8, 0xa7f9,
 	0xab5c, 0xab5f,
+	0x10780, 0x10780,
+	0x10783, 0x10785,
+	0x10787, 0x107b0,
+	0x107b2, 0x107ba,
 }; /* CR_Other_Lowercase */
 
 /* 'Other_Uppercase': Binary Property */
@@ -37049,7 +37061,7 @@ static const OnigCodePoint CR_Grapheme_Cluster_Break_Extend[] = {
 
 /* 'Grapheme_Cluster_Break_SpacingMark': Grapheme_Cluster_Break=SpacingMark */
 static const OnigCodePoint CR_Grapheme_Cluster_Break_SpacingMark[] = {
-	161,
+	160,
 	0x0903, 0x0903,
 	0x093b, 0x093b,
 	0x093e, 0x0940,
@@ -37183,7 +37195,6 @@ static const OnigCodePoint CR_Grapheme_Cluster_Break_SpacingMark[] = {
 	0x116ac, 0x116ac,
 	0x116ae, 0x116af,
 	0x116b6, 0x116b6,
-	0x11720, 0x11721,
 	0x11726, 0x11726,
 	0x1182c, 0x1182e,
 	0x11838, 0x11838,

[nobu (Nobuyoshi Nakada)]

https://github.com/nobu/ruby/tree/emoji

[nobu (Nobuyoshi Nakada)]

• Related to Feature #18639: Update Unicode data to Unicode Version 15.0.0 added

[nobu (Nobuyoshi Nakada)]

nobu (Nobuyoshi Nakada) wrote:

But U+11720..U+11721 in CR_Grapheme_Cluster_Break_SpacingMark is absent in the original data of the Unicode.org.
According to @naruse's investigation, it was removed at the commit [Update to Unicode 14.0.0], while U+11720 is still SpacingMark in the latest https://www.unicode.org/reports/tr29/.

Read https://www.unicode.org/reports/tr29/#SpacingMark again, and found U+11720 and U+11721 are listed as exceptions.

Exceptions: The following (which have General_Category = Spacing_Mark and would otherwise be included) are specifically excluded:

That means these are removed intentionally.

[nobu (Nobuyoshi Nakada)]

• Subject changed from Unicode tables differences from Unicode.org 14.0 data and removed properties since 13.0 to Unicode tables differences from Unicode.org 14.0 data

[duerst (Martin Dürst)]

The wrong properties were fixed for version 14.0.0 with commit e31d645. This issue should stay open until we are sure what caused the wrong properties in the first place.

[hsbt (Hiroshi SHIBATA)]

@duerst (Martin Dürst) Is there any action for Ruby 3.2 related this? If there is nothing to do for Ruby 3.2, I'll remove this from Ruby 3.2 milestone.

[duerst (Martin Dürst)]

• Target version deleted (3.2)

[duerst (Martin Dürst)]

• Related to Feature #18037: Update Unicode data to Unicode Version 14.0.0 added

[duerst (Martin Dürst)]

hsbt (Hiroshi SHIBATA) wrote in #note-6:

@duerst (Martin Dürst) Is there any action for Ruby 3.2 related this? If there is nothing to do for Ruby 3.2, I'll remove this from Ruby 3.2 milestone.

No, nothing should be needed for Ruby 3.2. I removed it myself.

BTW, can we add a "3.3" target version?

[hsbt (Hiroshi SHIBATA)]

Thanks. I did create the 3.3 milestone

[jeremyevans0 (Jeremy Evans)]

@duerst (Martin Dürst) Did this issue reoccur in the update to Unicode 15? If not, do you think this can be closed?

[nobu (Nobuyoshi Nakada)]

• Status changed from Open to Closed