Rewrite org-hugo--sanitize-title; make it more robust

The org-hugo--sanitize-title function is re-written, now with added
test cases for even apparently pathological cases.

Now it is possible to have *, _ and ` characters in the post title.

ox-hugo.el

@@ -1769,31 +1769,36 @@ to ((name . \"foo\") (weight . 80))."
 (defun org-hugo--sanitize-title (info)
   "Return sanitized version of the title string parsed from INFO.
 
-- Remove bold, italics, monospace Markdown markup characters.
-- Do not escape underscore characters in the title.
+The title string is returned in a markup-free form.
 
 If exporting title is disabled by setting `org-export-with-title'
 to nil or using the OPTIONS keyword e.g. \"title:nil\", return
 nil.
 
 INFO is a plist used as a communication channel."
   (when (plist-get info :with-title)
-    (let* ((title (org-export-data (plist-get info :title) info))
-           ;; Sanitize title.. cannot do bold, italics, monospace in title
-           (title (replace-regexp-in-string "\\\\?`" "" title))
-           (title (replace-regexp-in-string "\\`__?\\|\\`\\*\\*?\\|__?\\'\\|\\*\\*?\\'" "" title))
-           (title (replace-regexp-in-string " __?\\|__? \\| \\*\\*?\\|\\*\\*? " " " title))
-           ;; Do not escape underscores in title; \_ -> _
-           (title (replace-regexp-in-string "\\\\_" "_" title))
+    (let* ((title (org-export-data-with-backend (plist-get info :title) 'html info))
+           ;; The `title' variable contains the htmlized version of
+           ;; the Org title string.  By using the HTML version, it
+           ;; will be easy to later strip off just the HTML tags.
+
            ;; Hugo does not render Markdown in the titles and so the
            ;; Blackfriday smartDashes conversion does not work there.
            ;; So do that here instead.
            ;; Convert "---" and "--" to EM DASH (—) and EN DASH (–) respectively.
            ;; Convert "..." to HORIZONTAL ELLIPSIS (…).
-           ;; The order of below two replacements is important!
+           ;; Below two replacements are order sensitive!
            (title (replace-regexp-in-string "---\\([^-]\\)" "—" title)) ;EM DASH
            (title (replace-regexp-in-string "--\\([^-]\\)" "–" title)) ;EN DASH
            (title (replace-regexp-in-string "\\.\\.\\." "…" title))) ;HORIZONTAL ELLIPSIS
+
+      ;; Remove "<FOO>..</FOO>" HTML tags, but retain the string
+      ;; wrapped in those tags.
+      ;; (message "ox-hugo sanitize title [dbg] title 1: %s" title)
+      (while (string-match "<\\(?1:[a-z]+\\)[^>]*>\\(?2:[^<]+\\)</\\1>" title)
+        (setq title (replace-match "\\2" nil nil title))
+        ;; (message "ox-hugo sanitize title [dbg] title 2: %s" title)
+        )
       title)))
 
 (defun org-hugo--transform-org-tags (tag-list info &optional no-prefer-hyphen)

test/site/content-org/all-posts.org

@@ -324,6 +324,39 @@ it is explicitly set to /empty/ using =:EXPORT_TITLE:=.
 This tests an =ox-hugo= feature that gets around an upstream
 limitation, where the Blackfriday /smartDashes/ rendering does not
 happen in post titles ({{{hugoissue(4175)}}}).
+** *Title with **lots** and *lots* of * aste*risks**
+:PROPERTIES:
+:EXPORT_FILE_NAME: title_with_asterisks
+:END:
+- All asterisk characters meant for Org =*bold*= markup are removed.
+- The rest of the asterisk characters are retained.
+** Title with /lots/ and lots / / / of forward/slashes
+:PROPERTIES:
+:EXPORT_FILE_NAME: title_with_forward_slashes
+:END:
+- All forward slash characters meant for Org =/italic/= markup are
+  removed.
+- The rest of the forward slash characters are retained.
+** _Title_ with __lots and _lots_ _ _ of under_scores__
+:PROPERTIES:
+:EXPORT_FILE_NAME: title_with_under_scores
+:END:
+- All underscore characters meant for Org =_underline_= markup are
+  removed.
+- The rest of the underscore characters are retained.
+** =Title with lots= of ` ` ` backticks` `foo =bar= ~ ~zoomba~
+:PROPERTIES:
+:EXPORT_FILE_NAME: title_with_backticks_equals_tildes
+:END:
+- All equal/tilde characters meant for Org =~code~= or ~=verbatim=~
+  markup are removed.
+- The rest of the equal/tilde characters are retained.
+** =Title with lots= of ` ` ` backticks` * asterisks *bold*, /italics/ =retain*,_,` in monospace= underscores _ _ _*foo*_ =bar= **
+:PROPERTIES:
+:EXPORT_FILE_NAME: title_with_asterisks_underscores_backticks
+:END:
+- All the Org markup characters are removed.
+- The "markup characters" not doing actual markup are retained.
 * Description meta-data with "quoted text"
 :PROPERTIES:
 :EXPORT_FILE_NAME: post-description-quotes

test/site/content/posts/ndash-and-mdash.md

@@ -1,5 +1,5 @@
 +++
-title = "ndash and mdash"
+title = "ndash `and` mdash"
 date = 2017-07-22
 tags = ["body"]
 draft = false

test/site/content/posts/title_with_asterisks.md

@@ -0,0 +1,8 @@
++++
+title = "Title with *lots and lots of * aste*risks**"
+tags = ["title"]
+draft = false
++++
+
+-   All asterisk characters meant for Org `*bold*` markup are removed.
+-   The rest of the asterisk characters are retained.

test/site/content/posts/title_with_asterisks_underscores_backticks.md

@@ -0,0 +1,8 @@
++++
+title = "Title with lots of ` ` ` backticks` * asterisks bold, italics retain*,_,` in monospace underscores _ _ foo bar **"
+tags = ["title"]
+draft = false
++++
+
+-   All the Org markup characters are removed.
+-   The "markup characters" not doing actual markup are retained.

test/site/content/posts/title_with_backticks_equals_tildes.md

@@ -0,0 +1,9 @@
++++
+title = "Title with lots of ` ` ` backticks` `foo bar ~ zoomba"
+tags = ["title"]
+draft = false
++++
+
+-   All equal/tilde characters meant for Org `~code~` or `=verbatim=`
+    markup are removed.
+-   The rest of the equal/tilde characters are retained.

test/site/content/posts/title_with_forward_slashes.md

@@ -0,0 +1,9 @@
++++
+title = "Title with lots and lots / / / of forward/slashes"
+tags = ["title"]
+draft = false
++++
+
+-   All forward slash characters meant for Org `/italic/` markup are
+    removed.
+-   The rest of the forward slash characters are retained.

test/site/content/posts/title_with_under_scores.md

@@ -0,0 +1,9 @@
++++
+title = "Title with _lots and _lots _ _ of under_scores__"
+tags = ["title"]
+draft = false
++++
+
+-   All underscore characters meant for Org `_underline_` markup are
+    removed.
+-   The rest of the underscore characters are retained.