Add new function 'dom-inner-text'

This replaces 'dom-text' and 'dom-texts', and is both more correct and
more efficient than them.

* lisp/dom.el (dom-text, dom-texts): Make obsolete in favor of...
(dom-inner-text--1, dom-inner-text): ... these new functions.  Update
callers.

* doc/lispref/text.texi (Document Object Model): Update documentation to
refer to 'dom-inner-text'.

* etc/NEWS: Announce this change.

(cherry picked from commit 2e53c7d08ba0de2468ec25cf9a9d7079604f9409)

diff --git a/doc/lispref/text.texi b/doc/lispref/text.texi
index 0d293d14f93b2c5725ad700e243ce5640a5cc233..b8437d954f7c61f4997cd4a72367c6283ef065aa 100644 (file)
--- a/doc/lispref/text.texi
+++ b/doc/lispref/text.texi
@@ -5729,14 +5729,10 @@ Return all the non-string children of the node.
 @item dom-attributes @var{node}
 Return the key/value pair list of attributes of the node.
 
-@item dom-text @var{node}
-Return all the textual elements of the node as a concatenated string.
-
-@item dom-texts @var{node}
+@item dom-inner-text @var{node}
 Return all the textual elements of the node, as well as the textual
 elements of all the children of the node, recursively, as a
-concatenated string.  This function also takes an optional separator
-to be inserted between the textual elements.
+concatenated string.
 
 @item dom-parent @var{dom} @var{node}
 Return the parent of @var{node} in @var{dom}.

diff --git a/lisp/dom.el b/lisp/dom.el
index c65dfd66dcc90ba734b57524c95a9776cd9ed616..89f450415ea2b45f6c8248a036c327ac99d1175c 100644 (file)
--- a/lisp/dom.el
+++ b/lisp/dom.el
@@ -75,10 +75,12 @@ A typical attribute is `href'."
 
 (defun dom-text (node)
   "Return all the text bits in the current node concatenated."
+  (declare (obsolete 'dom-inner-text "31.1"))
   (mapconcat #'identity (cl-remove-if-not #'stringp (dom-children node)) " "))
 
 (defun dom-texts (node &optional separator)
   "Return all textual data under NODE concatenated with SEPARATOR in-between."
+  (declare (obsolete 'dom-inner-text "31.1"))
   (if (eq (dom-tag node) 'script)
       ""
     (mapconcat
@@ -93,6 +95,25 @@ A typical attribute is `href'."
      (dom-children node)
      (or separator " "))))
 
+(defun dom-inner-text--1 (node)
+  (dolist (child (dom-children node))
+    (cond
+     ((stringp child) (insert child))
+     ((memq (dom-tag child) '(script comment)))
+     (t (dom-inner-text--1 child)))))
+
+(defun dom-inner-text (node)
+  "Return all textual data under NODE as a single string."
+  (let ((children (dom-children node)))
+    (if (and (length= children 1)
+             (stringp (car children)))
+        ;; Copy the string content when returning to be consistent with
+        ;; the other branch of this `if' expression.
+        (copy-sequence (car children))
+    (with-work-buffer
+      (dom-inner-text--1 node)
+      (buffer-string)))))
+
 (defun dom-child-by-tag (dom tag)
   "Return the first child of DOM that is of type TAG."
   (assoc tag (dom-children dom)))

diff --git a/lisp/gnus/nnatom.el b/lisp/gnus/nnatom.el
index b4a37b4588588e0a0ca34f6c947d383029465728..6e34751df1769dc104fa0db49f454f3ead51575f 100644 (file)
--- a/lisp/gnus/nnatom.el
+++ b/lisp/gnus/nnatom.el
@@ -110,7 +110,8 @@
 
 (defun nnatom--dom-line (node)
   "Return NODE's text as a single, whitespace-trimmed line."
-  (string-trim (replace-regexp-in-string "[\r\n]+" " " (dom-text node) t)))
+  (string-trim (replace-regexp-in-string
+                "[\r\n]+" " " (dom-inner-text node) t)))
 
 (defun nnatom--read-title (group)
   "Return the title of GROUP, or nil."
@@ -245,7 +246,7 @@ return the subject.  Otherwise, return nil."
                      (dom-print (dom-child-by-tag part 'div) nil t)
                      (buffer-substring-no-properties
                       (point-min) (point-max)))
-                 (dom-text part)))
+                 (dom-inner-text part)))
          (type (if (member type atypes) (concat "text/" type) type))
          (type (or (cdr (assoc type mtypes)) type)))
     (unless (string-blank-p part)

diff --git a/lisp/net/eww.el b/lisp/net/eww.el
index aa70163c63cdd67db2f4f20735ebc9fe84c67597..14bd66d5064964a6b60705d92a8126d3f7fa421b 100644 (file)
--- a/lisp/net/eww.el
+++ b/lisp/net/eww.el
@@ -1007,7 +1007,7 @@ This replaces the region with the preprocessed HTML."
   (plist-put eww-data :title
             (replace-regexp-in-string
              "^ \\| $" ""
-             (replace-regexp-in-string "[ \t\r\n]+" " " (dom-text dom))))
+             (replace-regexp-in-string "[ \t\r\n]+" " " (dom-inner-text dom))))
   (eww--after-page-change))
 
 (defun eww-display-raw (buffer &optional encode)
@@ -1178,7 +1178,7 @@ non-nil, don't actually compute a score; just call the callback."
         (setq score 2
               noscore t))
        ((eq (dom-tag node) 'a)
-        (setq score (- (length (split-string (dom-text node))))
+        (setq score (- (length (split-string (dom-inner-text node))))
               noscore t))
        (t
         (setq score -1))))
@@ -1203,7 +1203,7 @@ If EWW can't create a readable version, return nil instead."
          (when (and score (> score best-score)
                     ;; We set a lower bound to how long we accept that
                     ;; the readable portion of the page is going to be.
-                    (> (length (split-string (dom-texts node))) 100))
+                    (> (length (split-string (dom-inner-text node))) 100))
            (setq best-score score
                  best-node node))
          ;; Keep track of any <title> and <link> tags we find to include
@@ -1218,7 +1218,7 @@ If EWW can't create a readable version, return nil instead."
            ;; directly in our list in addition to as a child of some
            ;; other node in the list.  This is ok for <title> and <link>
            ;; tags, but might need changed if supporting other tags.
-           (let* ((inner-text (dom-texts node ""))
+           (let* ((inner-text (dom-inner-text node))
                   (new-node `(,(dom-tag node)
                               ,(dom-attributes node)
                               ,@(when (length> inner-text 0)
@@ -1250,7 +1250,7 @@ If EWW can't create a readable version, return nil instead."
                   most-negative-fixnum))
         ;; We set a lower bound to how long we accept that the
         ;; readable portion of the page is going to be.
-        (when (> (length (split-string (dom-texts highest))) 100)
+        (when (> (length (split-string (dom-inner-text highest))) 100)
          (setq result highest))))
     result))
 
@@ -1875,7 +1875,7 @@ See URL `https://developer.mozilla.org/en-US/docs/Web/HTML/Element/Input'.")
                'display (make-string (length value) ?*)))))))))
 
 (defun eww-tag-textarea (dom)
-  (let ((value (or (dom-text dom) ""))
+  (let ((value (or (dom-inner-text dom) ""))
        (lines (string-to-number (or (dom-attr dom 'rows) "10")))
        (width (string-to-number (or (dom-attr dom 'cols) "10")))
        start end form)
@@ -1951,7 +1951,7 @@ See URL `https://developer.mozilla.org/en-US/docs/Web/HTML/Element/Input'.")
     (dolist (elem (dom-by-tag dom 'option))
       (when (dom-attr elem 'selected)
        (nconc menu (list :value (dom-attr elem 'value))))
-      (let ((display (dom-text elem)))
+      (let ((display (dom-inner-text elem)))
        (setq max (max max (length display)))
        (push (list 'item
                    :value (dom-attr elem 'value)