3
0
mirror of https://github.com/ergochat/ergo.git synced 2024-11-22 20:09:41 +01:00

more lenient casefolding for skeletons

This commit is contained in:
Shivaram Lingamneni 2019-02-03 02:45:02 -05:00
parent e7399ba2b5
commit c34d9e0b72
2 changed files with 17 additions and 10 deletions

View File

@ -9,8 +9,10 @@ import (
"strings" "strings"
"github.com/oragono/confusables" "github.com/oragono/confusables"
"golang.org/x/text/cases"
"golang.org/x/text/language"
"golang.org/x/text/secure/precis" "golang.org/x/text/secure/precis"
"golang.org/x/text/unicode/norm" "golang.org/x/text/width"
) )
const ( const (
@ -126,14 +128,6 @@ func isBoring(name string) bool {
return true return true
} }
var skeletonCasefolder = precis.NewIdentifier(precis.FoldWidth, precis.LowerCase(), precis.Norm(norm.NFC))
// similar to Casefold, but exempts the bidi rule, because skeletons may
// mix scripts strangely
func casefoldSkeleton(str string) (string, error) {
return iterateFolding(skeletonCasefolder, str)
}
// Skeleton produces a canonicalized identifier that tries to catch // Skeleton produces a canonicalized identifier that tries to catch
// homoglyphic / confusable identifiers. It's a tweaked version of the TR39 // homoglyphic / confusable identifiers. It's a tweaked version of the TR39
// skeleton algorithm. We apply the skeleton algorithm first and only then casefold, // skeleton algorithm. We apply the skeleton algorithm first and only then casefold,
@ -146,5 +140,16 @@ func Skeleton(name string) (string, error) {
if !isBoring(name) { if !isBoring(name) {
name = confusables.Skeleton(name) name = confusables.Skeleton(name)
} }
return casefoldSkeleton(name)
// XXX the confusables table includes some, but not all, fullwidth->standard
// mappings for latin characters. do a pass of explicit width folding,
// same as PRECIS:
name = width.Fold.String(name)
// internationalized lowercasing for skeletons; this is much more lenient than
// Casefold. In particular, skeletons are expected to mix scripts (which may
// violate the bidi rule). We also don't care if they contain runes
// that are disallowed by PRECIS, because every identifier must independently
// pass PRECIS --- we are just further canonicalizing the skeleton.
return cases.Lower(language.Und).String(name), nil
} }

View File

@ -173,4 +173,6 @@ func TestSkeleton(t *testing.T) {
t.Errorf("we must protect against cyrillic homoglyph attacks") t.Errorf("we must protect against cyrillic homoglyph attacks")
} }
// should not raise an error:
skeleton("けらんぐ")
} }