Avoid calling strings.Fields multiple times with same content

This should be a relief for big sites.

diff --git a/helpers/content.go b/helpers/content.go
index 491e61c0d30cba575f09834f6f4930377a699fa1..3b33b6b174fcc62e952030325699de2a7fdc5603 100644 (file)
--- a/helpers/content.go
+++ b/helpers/content.go
@@ -275,8 +275,7 @@ func TruncateWords(s string, max int) string {
 
 // TruncateWordsToWholeSentence takes content and an int
 // and returns entire sentences from content, delimited by the int.
-func TruncateWordsToWholeSentence(s string, max int) string {
-       words := strings.Fields(s)
+func TruncateWordsToWholeSentence(words []string, max int) string {
        if max > len(words) {
                return strings.Join(words, " ")
        }

diff --git a/hugolib/page.go b/hugolib/page.go
index 543b40a78afbec3fe15686c14a304a3e9cf39031..ddb1bfb2c58acc1af6a8d26c3115fcebbd92d98f 100644 (file)
--- a/hugolib/page.go
+++ b/hugolib/page.go
@@ -65,6 +65,8 @@ type Page struct {
        rawContent          []byte
        contentShortCodes   map[string]string
        plain               string // TODO should be []byte
+       plainWords          []string
+       plainInit           sync.Once
        renderingConfig     *helpers.Blackfriday
        renderingConfigInit sync.Once
        PageMeta
@@ -97,12 +99,22 @@ type Position struct {
 type Pages []*Page
 
 func (p *Page) Plain() string {
-       if len(p.plain) == 0 {
-               p.plain = helpers.StripHTML(string(p.Content))
-       }
+       p.initPlain()
        return p.plain
 }
 
+func (p *Page) PlainWords() []string {
+       p.initPlain()
+       return p.plainWords
+}
+
+func (p *Page) initPlain() {
+       p.plainInit.Do(func() {
+               p.plain = helpers.StripHTML(string(p.Content))
+               p.plainWords = strings.Fields(p.plain)
+       })
+}
+
 func (p *Page) IsNode() bool {
        return false
 }
@@ -178,9 +190,11 @@ func (p *Page) setSummary() {
        } else {
                // If hugo defines split:
                // render, strip html, then split
-               plain := strings.Join(strings.Fields(p.Plain()), " ")
-               p.Summary = helpers.BytesToHTML([]byte(helpers.TruncateWordsToWholeSentence(plain, helpers.SummaryLength)))
-               p.Truncated = len(p.Summary) != len(plain)
+               p.Summary = helpers.BytesToHTML([]byte(helpers.TruncateWordsToWholeSentence(p.PlainWords(), helpers.SummaryLength)))
+
+               // todo bep - check if the Plain() can be trimmed earlier
+               p.Truncated = len(p.Summary) != len(strings.Trim(p.Plain(), "\n\r "))
+
        }
 }
 
@@ -322,7 +336,7 @@ func (p *Page) ReadFrom(buf io.Reader) (err error) {
 }
 
 func (p *Page) analyzePage() {
-       p.WordCount = helpers.TotalWords(p.Plain())
+       p.WordCount = len(p.PlainWords())
        p.FuzzyWordCount = int((p.WordCount+100)/100) * 100
        p.ReadingTime = int((p.WordCount + 212) / 213)
 }