internal/number: fix bug in info lookup
fallback needs to be computed for both modes.
Latn data is elided from the tables, so check
for latin script at the root.
Change-Id: I6db954b9b76761380fec3b90fde962449536fa23
Reviewed-on: https://go-review.googlesource.com/81755
Run-TryBot: Marcel van Lohuizen <mpvl@golang.org>
Reviewed-by: Nigel Tao <nigeltao@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
diff --git a/internal/number/number.go b/internal/number/number.go
index ea341c7..336c0d8 100644
--- a/internal/number/number.go
+++ b/internal/number/number.go
@@ -31,9 +31,9 @@
if !ok {
// Take the value for the default numbering system. This is by far the
// most common case as an alternative numbering system is hardly used.
- if p&0x80 == 0 {
+ if p&0x80 == 0 { // Latn digits.
pSymIndex = p
- } else {
+ } else { // Non-Latn or multiple numbering systems.
// Take the first entry from the alternatives list.
data := langToAlt[p&^0x80]
pSymIndex = data.symIndex
@@ -43,8 +43,22 @@
langIndex := compactIndex
ns := system
outerLoop:
- for {
- if p&0x80 == 0 {
+ for ; ; p = langToDefaults[langIndex] {
+ if langIndex == 0 {
+ // The CLDR root defines full symbol information for all
+ // numbering systems (even though mostly by means of
+ // aliases).
+ // Fall back to the default entry for Latn if there is
+ // no data for the numbering system of this language.
+ if ns == 0 {
+ break
+ }
+ // Fall back to Latin and start from the original
+ // language. See
+ // http://unicode.org/reports/tr35/#Locale_Inheritance.
+ ns = numLatn
+ langIndex = compactIndex
+ } else if p&0x80 == 0 {
if ns == 0 {
// The index directly points to the symbol data.
pSymIndex = p
@@ -52,30 +66,16 @@
}
// Move to the parent and retry.
langIndex = int(internal.Parent[langIndex])
- }
- // The index points to a list of symbol data indexes.
- for _, e := range langToAlt[p&^0x80:] {
- if int(e.compactTag) != langIndex {
- if langIndex == 0 {
- // The CLDR root defines full symbol information for all
- // numbering systems (even though mostly by means of
- // aliases). This means that we will never fall back to
- // the default of the language. Also, the loop is
- // guaranteed to terminate as a consequence.
- ns = numLatn
- // Fall back to Latin and start from the original
- // language. See
- // http://unicode.org/reports/tr35/#Locale_Inheritance.
- langIndex = compactIndex
- } else {
+ } else {
+ // The index points to a list of symbol data indexes.
+ for _, e := range langToAlt[p&^0x80:] {
+ if int(e.compactTag) != langIndex {
// Fall back to parent.
langIndex = int(internal.Parent[langIndex])
+ } else if e.system == ns {
+ pSymIndex = e.symIndex
+ break outerLoop
}
- break
- }
- if e.system == ns {
- pSymIndex = e.symIndex
- break outerLoop
}
}
}
diff --git a/internal/number/number_test.go b/internal/number/number_test.go
index 8717ead..cbc28ab 100644
--- a/internal/number/number_test.go
+++ b/internal/number/number_test.go
@@ -31,6 +31,10 @@
{"de-CH-oxendict", SymGroup, "’", '9'}, // inherits from de-CH (no compact index)
{"de-CH-u-nu-deva", SymGroup, "’", '\u096f'}, // miss -> latn -> de-CH
+ {"bn-u-nu-beng", SymGroup, ",", '\u09ef'},
+ {"bn-u-nu-deva", SymGroup, ",", '\u096f'},
+ {"bn-u-nu-latn", SymGroup, ",", '9'},
+
{"pa", SymExponential, "E", '9'},
// "×۱۰^" -> U+00d7 U+06f1 U+06f0^"
