Make string sorting (e.g. ByTitle, ByLinkTitle and ByParam) language …

…aware

Fixes #2180

hugolib/site.go

@@ -739,7 +739,12 @@ func (s *SiteInfo) Sites() page.Sites {
 }
 
 // Current returns the currently rendered Site.
+// If that isn't set yet, which is the situation before we start rendering,
+// if will return the Site itself.
 func (s *SiteInfo) Current() page.Site {
+	if s.s.h.currentSite == nil {
+		return s
+	}
 	return s.s.h.currentSite.Info
 }
 

hugolib/taxonomy.go

@@ -18,6 +18,7 @@ import (
 	"sort"
 
 	"github.com/gohugoio/hugo/compare"
+	"github.com/gohugoio/hugo/langs"
 
 	"github.com/gohugoio/hugo/resources/page"
 )
@@ -40,6 +41,15 @@ type Taxonomy map[string]page.WeightedPages
 // Important because you can't order a map.
 type OrderedTaxonomy []OrderedTaxonomyEntry
 
+// getOneOPage returns one page in the taxonomy,
+// nil if there is none.
+func (t OrderedTaxonomy) getOneOPage() page.Page {
+	if len(t) == 0 {
+		return nil
+	}
+	return t[0].Pages()[0]
+}
+
 // OrderedTaxonomyEntry is similar to an element of a Taxonomy, but with the key embedded (as name)
 // e.g:  {Name: Technology, page.WeightedPages: TaxonomyPages}
 type OrderedTaxonomyEntry struct {
@@ -72,11 +82,18 @@ func (i Taxonomy) TaxonomyArray() OrderedTaxonomy {
 
 // Alphabetical returns an ordered taxonomy sorted by key name.
 func (i Taxonomy) Alphabetical() OrderedTaxonomy {
+	ia := i.TaxonomyArray()
+	p := ia.getOneOPage()
+	if p == nil {
+		return ia
+	}
+	currentSite := p.Site().Current()
+	coll := langs.GetCollator(currentSite.Language())
+	coll.Lock()
+	defer coll.Unlock()
 	name := func(i1, i2 *OrderedTaxonomyEntry) bool {
-		return compare.LessStrings(i1.Name, i2.Name)
+		return coll.CompareStrings(i1.Name, i2.Name) < 0
 	}
-
-	ia := i.TaxonomyArray()
 	oiBy(name).Sort(ia)
 	return ia
 }

langs/language.go

@@ -19,6 +19,9 @@ import (
 	"sync"
 	"time"
 
+	"golang.org/x/text/collate"
+	"golang.org/x/text/language"
+
 	"github.com/pkg/errors"
 
 	"github.com/gohugoio/hugo/common/htime"
@@ -80,8 +83,9 @@ type Language struct {
 	// TODO(bep) do the same for some of the others.
 	translator    locales.Translator
 	timeFormatter htime.TimeFormatter
-
-	location *time.Location
+	tag           language.Tag
+	collator      *Collator
+	location      *time.Location
 
 	// Error during initialization. Will fail the buld.
 	initErr error
@@ -111,6 +115,18 @@ func NewLanguage(lang string, cfg config.Provider) *Language {
 		}
 	}
 
+	var coll *Collator
+	tag, err := language.Parse(lang)
+	if err == nil {
+		coll = &Collator{
+			c: collate.New(tag),
+		}
+	} else {
+		coll = &Collator{
+			c: collate.New(language.English),
+		}
+	}
+
 	l := &Language{
 		Lang:       lang,
 		ContentDir: cfg.GetString("contentDir"),
@@ -119,6 +135,8 @@ func NewLanguage(lang string, cfg config.Provider) *Language {
 		params:        params,
 		translator:    translator,
 		timeFormatter: htime.NewTimeFormatter(translator),
+		tag:           tag,
+		collator:      coll,
 	}
 
 	if err := l.loadLocation(cfg.GetString("timeZone")); err != nil {
@@ -275,6 +293,10 @@ func GetLocation(l *Language) *time.Location {
 	return l.location
 }
 
+func GetCollator(l *Language) *Collator {
+	return l.collator
+}
+
 func (l *Language) loadLocation(tzStr string) error {
 	location, err := time.LoadLocation(tzStr)
 	if err != nil {
@@ -284,3 +306,16 @@ func (l *Language) loadLocation(tzStr string) error {
 
 	return nil
 }
+
+type Collator struct {
+	sync.Mutex
+	c *collate.Collator
+}
+
+// CompareStrings compares a and b.
+// It returns -1 if a < b, 1 if a > b and 0 if a == b.
+// Note that the Collator is not thread safe, so you may want
+// to aquire a lock on it before calling this method.
+func (c *Collator) CompareStrings(a, b string) int {
+	return c.c.CompareString(a, b)
+}

langs/language_test.go

@@ -14,10 +14,13 @@
 package langs
 
 import (
+	"sync"
 	"testing"
 
 	qt "github.com/frankban/quicktest"
 	"github.com/gohugoio/hugo/config"
+	"golang.org/x/text/collate"
+	"golang.org/x/text/language"
 )
 
 func TestGetGlobalOnlySetting(t *testing.T) {
@@ -47,3 +50,59 @@ func TestLanguageParams(t *testing.T) {
 	c.Assert(lang.Params()["p1"], qt.Equals, "p1p")
 	c.Assert(lang.Get("p1"), qt.Equals, "p1cfg")
 }
+
+func TestCollator(t *testing.T) {
+
+	c := qt.New(t)
+
+	var wg sync.WaitGroup
+
+	coll := &Collator{c: collate.New(language.English, collate.Loose)}
+
+	for i := 0; i < 10; i++ {
+		wg.Add(1)
+		go func() {
+			coll.Lock()
+			defer coll.Unlock()
+			defer wg.Done()
+			for j := 0; j < 10; j++ {
+				k := coll.CompareStrings("abc", "def")
+				c.Assert(k, qt.Equals, -1)
+			}
+		}()
+	}
+	wg.Wait()
+
+}
+
+func BenchmarkCollator(b *testing.B) {
+	s := []string{"foo", "bar", "éntre", "baz", "qux", "quux", "corge", "grault", "garply", "waldo", "fred", "plugh", "xyzzy", "thud"}
+
+	doWork := func(coll *Collator) {
+		for i := 0; i < len(s); i++ {
+			for j := i + 1; j < len(s); j++ {
+				_ = coll.CompareStrings(s[i], s[j])
+			}
+		}
+	}
+
+	b.Run("Single", func(b *testing.B) {
+		coll := &Collator{c: collate.New(language.English, collate.Loose)}
+		for i := 0; i < b.N; i++ {
+			doWork(coll)
+		}
+	})
+
+	b.Run("Para", func(b *testing.B) {
+		b.RunParallel(func(pb *testing.PB) {
+			coll := &Collator{c: collate.New(language.English, collate.Loose)}
+
+			for pb.Next() {
+				coll.Lock()
+				doWork(coll)
+				coll.Unlock()
+			}
+		})
+	})
+
+}

resources/page/integration_test.go

@@ -70,3 +70,69 @@ date: "2020-02-01"
 	b.AssertFileContent("public/en/index.html", "0|February, 2020|Pages(1)1|January, 2020|Pages(1)")
 	b.AssertFileContent("public/fr/index.html", "0|février, 2020|Pages(1)1|janvier, 2020|Pages(1)")
 }
+
+func TestPagesSortCollation(t *testing.T) {
+
+	files := `
+-- config.toml --
+defaultContentLanguage = 'en'
+defaultContentLanguageInSubdir = true
+[languages]
+[languages.en]
+title = 'My blog'
+weight = 1
+[languages.fr]
+title = 'Mon blogue'
+weight = 2
+[languages.nn]
+title = 'Bloggen min'
+weight = 3
+-- content/p1.md --
+---
+title: "zulu"
+date: "2020-01-01"
+param1: "xylophone"
+tags: ["xylophone", "éclair", "zulu", "emma"]
+---
+-- content/p2.md --
+---
+title: "émotion"
+date: "2020-01-01"
+param1: "violin"
+---
+-- content/p3.md --
+---
+title: "alpha"
+date: "2020-01-01"
+param1: "éclair"
+---
+-- layouts/index.html --
+ByTitle: {{ range site.RegularPages.ByTitle }}{{ .Title }}|{{ end }}
+ByLinkTitle: {{ range site.RegularPages.ByLinkTitle }}{{ .Title }}|{{ end }}
+ByParam: {{ range site.RegularPages.ByParam "param1" }}{{ .Params.param1 }}|{{ end }}
+Tags Alphabetical: {{  range site.Taxonomies.tags.Alphabetical }}{{ .Term }}|{{ end }}
+GroupBy: {{ range site.RegularPages.GroupBy "Title" }}{{ .Key }}|{{ end }}
+{{ with (site.GetPage "p1").Params.tags }}
+Sort: {{  sort . }}
+ByWeight: {{ range site.RegularPages.ByWeight }}{{ .Title }}|{{ end }}
+{{ end }}
+
+	`
+
+	b := hugolib.NewIntegrationTestBuilder(
+		hugolib.IntegrationTestConfig{
+			T:           t,
+			TxtarString: files,
+			NeedsOsFS:   true,
+		}).Build()
+
+	b.AssertFileContent("public/en/index.html", `
+ByTitle: alpha|émotion|zulu|
+ByLinkTitle: alpha|émotion|zulu|
+ByParam: éclair|violin|xylophone
+Tags Alphabetical: éclair|emma|xylophone|zulu|
+GroupBy: alpha|émotion|zulu|
+Sort: [éclair emma xylophone zulu]
+ByWeight: alpha|émotion|zulu|
+`)
+}

resources/page/pagegroup.go

@@ -53,13 +53,16 @@ type mapKeyByInt struct{ mapKeyValues }
 
 func (s mapKeyByInt) Less(i, j int) bool { return s.mapKeyValues[i].Int() < s.mapKeyValues[j].Int() }
 
-type mapKeyByStr struct{ mapKeyValues }
+type mapKeyByStr struct {
+	less func(a, b string) bool
+	mapKeyValues
+}
 
 func (s mapKeyByStr) Less(i, j int) bool {
-	return compare.LessStrings(s.mapKeyValues[i].String(), s.mapKeyValues[j].String())
+	return s.less(s.mapKeyValues[i].String(), s.mapKeyValues[j].String())
 }
 
-func sortKeys(v []reflect.Value, order string) []reflect.Value {
+func sortKeys(examplePage Page, v []reflect.Value, order string) []reflect.Value {
 	if len(v) <= 1 {
 		return v
 	}
@@ -72,10 +75,12 @@ func sortKeys(v []reflect.Value, order string) []reflect.Value {
 			sort.Sort(mapKeyByInt{v})
 		}
 	case reflect.String:
+		stringLess, close := collatorStringLess(examplePage)
+		defer close()
 		if order == "desc" {
-			sort.Sort(sort.Reverse(mapKeyByStr{v}))
+			sort.Sort(sort.Reverse(mapKeyByStr{stringLess, v}))
 		} else {
-			sort.Sort(mapKeyByStr{v})
+			sort.Sort(mapKeyByStr{stringLess, v})
 		}
 	}
 	return v
@@ -161,7 +166,7 @@ func (p Pages) GroupBy(key string, order ...string) (PagesGroup, error) {
 		tmp.SetMapIndex(fv, reflect.Append(tmp.MapIndex(fv), ppv))
 	}
 
-	sortedKeys := sortKeys(tmp.MapKeys(), direction)
+	sortedKeys := sortKeys(p[0], tmp.MapKeys(), direction)
 	r := make([]PageGroup, len(sortedKeys))
 	for i, k := range sortedKeys {
 		r[i] = PageGroup{Key: k.Interface(), Pages: tmp.MapIndex(k).Interface().(Pages)}
@@ -213,7 +218,7 @@ func (p Pages) GroupByParam(key string, order ...string) (PagesGroup, error) {
 	}
 
 	var r []PageGroup
-	for _, k := range sortKeys(tmp.MapKeys(), direction) {
+	for _, k := range sortKeys(p[0], tmp.MapKeys(), direction) {
 		r = append(r, PageGroup{Key: k.Interface(), Pages: tmp.MapIndex(k).Interface().(Pages)})
 	}