Use filename as fallback for date

Why:

 * when migrating from systems like jekylla and awestruct to hugo
   you might have hundreds if not thousands of blog entries using the
   filename as a way to set the date and slug part of page.
   i.e. 2017-08-01-my-life-as-a-dog.md would have a date of 2017-08-01 and
   a url called my-life-as-a-dog.md.

This change addreses the need by:

 * introduce a setting called `useFilenameDateAsFallback` that defaults to true,
   which will derive date and slug from the filename.

docs/content/getting-started/configuration.md

@@ -245,10 +245,6 @@ disablePathToLower =          false
 enableEmoji =                 false
 # Show a placeholder instead of the default value or an empty string if a translation is missing
 enableMissingTranslationPlaceholders = false
-# Regex to use for dates in filenames.
-filenameDateFallbackPattern = "(?P<year>\\d{4})\\-(?P<month>\\d{2})\\-(?P<day>\\d{2})"
-# Time format for custom dates in filenames. Only used with `useFilenameDateAsFallback` and must match `filenameDateFallbackPattern`
-filenameDateFallbackFormat = "2006-01-02"
 footnoteAnchorPrefix =        ""
 footnoteReturnLinkContents =  ""
 # google analytics tracking id
@@ -302,8 +298,8 @@ theme =                       ""
 title =                       ""
 # if true, use /filename.html instead of /filename/
 uglyURLs =                    false
-#if true, use dates in filenames e.g. 2017-01-31-mypostname.md
-useFilenameDateAsFallback =  false
+# use dates in filenames e.g. 2017-01-31-mypostname.md
+useFilenameDateAsFallback =  true
 # verbose output
 verbose =                     false
 # verbose logging

docs/content/variables/page.md

@@ -37,7 +37,11 @@ See [`.Scratch`](/functions/scratch/) for page-scoped, writable variables.
 
 `.Date`
 : the date associated with the page; `.Date` pulls from the `date` field in a content's front matter.
-If you're migrating content to Hugo, you may have content with dates in the filename. For example `2017-01-31-myblog.md`.  You can optionally enable the `useFilenameDateAsFallback` configuration option. This will attempt to parse the datestamp in the filename and use it as a fallback to providing a date variable in the front matter.
+If no `date` field present Hugo will try and extract the date from the filename. For example `2017-01-31-myblog.md`. It will also set `.Slug` to contain the path without the
+date prefix.'
+
+You can disable the automatic date detection and slug modification by setting `useFilenameDateAsFallback` configuration option to false. 
+
 See also `.ExpiryDate`, `.PublishDate`, and `.Lastmod`.
 
 `.Description`

hugolib/config.go

@@ -127,8 +127,7 @@ func loadDefaultSettingsFor(v *viper.Viper) {
 	v.SetDefault("enableEmoji", false)
 	v.SetDefault("pygmentsCodeFencesGuessSyntax", false)
 	v.SetDefault("useModTimeAsFallback", false)
-	v.SetDefault("filenameDateFallbackPattern", nil)
-	v.SetDefault("filenameDateFallbackFormat", "2006-01-02")
+	v.SetDefault("useFilenameDateAsFallback", true)
 	v.SetDefault("defaultContentLanguage", "en")
 	v.SetDefault("defaultContentLanguageInSubdir", false)
 	v.SetDefault("enableMissingTranslationPlaceholders", false)

hugolib/page.go

@@ -44,6 +44,9 @@ import (
 var (
 	cjk = regexp.MustCompile(`\p{Han}|\p{Hangul}|\p{Hiragana}|\p{Katakana}`)
 
+	// date expression used to find iso date in filename
+	dateExp = regexp.MustCompile("(?P<year>\\d{4})\\-(?P<month>\\d{2})\\-(?P<day>\\d{2})-(?P<slug>.*)\\..*")
+
 	// This is all the kinds we can expect to find in .Site.Pages.
 	allKindsInPages = []string{KindPage, KindHome, KindSection, KindTaxonomy, KindTaxonomyTerm}
 
@@ -1079,14 +1082,31 @@ func (p *Page) update(f interface{}) error {
 		}
 	}
 
-	if p.Date.IsZero() && p.s.Cfg.GetString("filenameDateFallbackPattern") != "" {
-		dateExp := regexp.MustCompile(p.s.Cfg.GetString("filenameDateFallbackPattern"))
-		dateString := dateExp.FindString(p.File.Path())
-		filenameDate, err := time.Parse(p.s.Cfg.GetString("filenameDateFallbackFormat"), dateString)
-		if err == nil {
-			p.s.Log.DEBUG.Printf("using filename date as fallback for page %s", p.File.Path())
-			p.Date = filenameDate
-			p.Params["date"] = p.Date
+	if p.Date.IsZero() && p.s.Cfg.GetBool("useFilenameDateAsFallback") {
+		match := dateExp.FindStringSubmatch(p.File.Path())
+
+		if match != nil {
+			year := match[1]
+			month := match[2]
+			day := match[3]
+			slug := match[4]
+
+			filenameDate, err := time.Parse("2006-01-02", year+"-"+month+"-"+day)
+
+			if err == nil {
+				p.Date = filenameDate
+				p.Params["date"] = p.Date
+				if p.Slug == "" {
+					p.Slug = slug
+					p.s.Log.DEBUG.Printf("Using filename date (%s) and slug (%s) as fallback for page %s", p.Date, p.Slug, p.File.Path())
+
+				} else {
+					p.s.Log.DEBUG.Printf("Using filename date (%s) as fallback for page %s", p.Date, p.File.Path())
+
+				}
+			} else {
+				p.s.Log.WARN.Printf("File has what looks like a date, but the date is invalid for page %s.", p.File.Path())
+			}
 		}
 	}
 

hugolib/page_test.go

@@ -516,6 +516,7 @@ func checkPageDate(t *testing.T, page *Page, time time.Time) {
 	if page.Date != time {
 		t.Fatalf("Page date is: %s.  Expected: %s", page.Date, time)
 	}
+
 }
 
 func checkTruncation(t *testing.T, page *Page, shouldBe bool, msg string) {
@@ -855,22 +856,80 @@ func TestPageWithDate(t *testing.T) {
 	d, _ := time.Parse(time.RFC3339, "2013-05-17T16:59:30Z")
 
 	checkPageDate(t, p, d)
+
 }
 
-func TestPageWithDateInFilename(t *testing.T) {
+func TestDateAsFallbackWithDateInFilenameDisabled(t *testing.T) {
 	t.Parallel()
 	cfg, fs := newTestCfg()
 
+	cfg.Set("useFilenameDateAsFallback", false)
+
 	writeSource(t, fs, filepath.Join("content", "2017-01-31-simple.md"), simplePageNoDate)
 
 	s := buildSingleSite(t, deps.DepsCfg{Fs: fs, Cfg: cfg}, BuildCfg{SkipRender: true})
 
 	require.Len(t, s.RegularPages, 1)
 
 	p := s.RegularPages[0]
-	d, _ := time.Parse(time.RFC3339, "2017-01-31")
+
+	assert.True(t, p.Date.IsZero(), "page date should be empty as no date in file nor filename.")
+
+	assert.Equal(t, "", p.Slug, "slug for page should not be set as filename as date is disabled.")
+}
+
+func TestDateAsFallbackWithDateInFilename(t *testing.T) {
+	t.Parallel()
+	cfg, fs := newTestCfg()
+
+	writeSource(t, fs, filepath.Join("content", "2017-01-31-simple.md"), simplePageNoDate)
+
+	s := buildSingleSite(t, deps.DepsCfg{Fs: fs, Cfg: cfg}, BuildCfg{SkipRender: true})
+
+	require.Len(t, s.RegularPages, 1)
+
+	p := s.RegularPages[0]
+	d, err := time.Parse("2006-01-02", "2017-01-31")
+
+	assert.Equal(t, err, nil)
 
 	checkPageDate(t, p, d)
+
+	assert.Equal(t, "simple", p.Slug, "slug for page should be set as there is both date and title in filename")
+}
+
+func TestDateAsFallbackWithOutDateInFilename(t *testing.T) {
+	t.Parallel()
+	cfg, fs := newTestCfg()
+
+	writeSource(t, fs, filepath.Join("content", "simple.md"), simplePageNoDate)
+
+	s := buildSingleSite(t, deps.DepsCfg{Fs: fs, Cfg: cfg}, BuildCfg{SkipRender: true})
+
+	require.Len(t, s.RegularPages, 1)
+
+	p := s.RegularPages[0]
+
+	assert.True(t, p.Date.IsZero(), "page date should be empty as no date in filename")
+	assert.Equal(t, "", p.Slug, "page slug should not be set as no date in filename")
+
+}
+
+func TestDateAsFallbackWithInvalidDateInFilename(t *testing.T) {
+	t.Parallel()
+	cfg, fs := newTestCfg()
+
+	writeSource(t, fs, filepath.Join("content", "2017-31-31-simple.md"), simplePageNoDate)
+
+	s := buildSingleSite(t, deps.DepsCfg{Fs: fs, Cfg: cfg}, BuildCfg{SkipRender: true})
+
+	require.Len(t, s.RegularPages, 1)
+
+	p := s.RegularPages[0]
+
+	assert.True(t, p.Date.IsZero(), "page date should be empty as no valid date in filename")
+	assert.Equal(t, "", p.Slug, "page slug should not be set as no valid date in filename")
+
 }
 
 func TestWordCountWithAllCJKRunesWithoutHasCJKLanguage(t *testing.T) {