X-Git-Url: https://cgit.sxemacs.org/?p=gnus;a=blobdiff_plain;f=lisp%2Fshr.el;h=6124ff7d8d2c9bf1f94a2cbd133d3cb70a5efa65;hp=e97ca1da47a85df57d7f1bee44618f6257be6fae;hb=64d972eb2840e696731a81b8a71afa1934715dc7;hpb=724b5c8f365a9f1c549bd639c29048aa8627a478 diff --git a/lisp/shr.el b/lisp/shr.el index e97ca1da4..6124ff7d8 100644 --- a/lisp/shr.el +++ b/lisp/shr.el @@ -52,12 +52,23 @@ fit these criteria." :group 'shr :type 'regexp) +(defvar shr-content-function nil + "If bound, this should be a function that will return the content. +This is used for cid: URLs, and the function is called with the +cid: URL as the argument.") + +(defvar shr-width 70 + "Frame width to use for rendering.") + +;;; Internal variables. + (defvar shr-folding-mode nil) (defvar shr-state nil) (defvar shr-start nil) (defvar shr-indentation 0) - -(defvar shr-width 70) +(defvar shr-inhibit-images nil) +(defvar shr-list-mode nil) +(defvar shr-content-cache nil) (defvar shr-map (let ((map (make-sparse-keymap))) @@ -69,6 +80,76 @@ fit these criteria." (define-key map "\r" 'shr-browse-url) map)) +;; Public functions and commands. + +;;;###autoload +(defun shr-insert-document (dom) + (setq shr-content-cache nil) + (let ((shr-state nil) + (shr-start nil)) + (shr-descend (shr-transform-dom dom)))) + +(defun shr-copy-url () + "Copy the URL under point to the kill ring. +If called twice, then try to fetch the URL and see whether it +redirects somewhere else." + (interactive) + (let ((url (get-text-property (point) 'shr-url))) + (cond + ((not url) + (message "No URL under point")) + ;; Resolve redirected URLs. + ((equal url (car kill-ring)) + (url-retrieve + url + (lambda (a) + (when (and (consp a) + (eq (car a) :redirect)) + (with-temp-buffer + (insert (cadr a)) + (goto-char (point-min)) + ;; Remove common tracking junk from the URL. + (when (re-search-forward ".utm_.*" nil t) + (replace-match "" t t)) + (message "Copied %s" (buffer-string)) + (copy-region-as-kill (point-min) (point-max))))))) + ;; Copy the URL to the kill ring. + (t + (with-temp-buffer + (insert url) + (copy-region-as-kill (point-min) (point-max)) + (message "Copied %s" url)))))) + +(defun shr-show-alt-text () + "Show the ALT text of the image under point." + (interactive) + (let ((text (get-text-property (point) 'shr-alt))) + (if (not text) + (message "No image under point") + (message "%s" text)))) + +(defun shr-browse-image () + "Browse the image under point." + (interactive) + (let ((url (get-text-property (point) 'shr-image))) + (if (not url) + (message "No image under point") + (message "Browsing %s..." url) + (browse-url url)))) + +(defun shr-insert-image () + "Insert the image under point into the buffer." + (interactive) + (let ((url (get-text-property (point) 'shr-image))) + (if (not url) + (message "No image under point") + (message "Inserting %s..." url) + (url-retrieve url 'shr-image-fetched + (list (current-buffer) (1- (point)) (point-marker)) + t)))) + +;;; Utility functions. + (defun shr-transform-dom (dom) (let ((result (list (pop dom)))) (dolist (arg (pop dom)) @@ -81,12 +162,6 @@ fit these criteria." (push (shr-transform-dom sub) result))) (nreverse result))) -;;;###autoload -(defun shr-insert-document (dom) - (let ((shr-state nil) - (shr-start nil)) - (shr-descend (shr-transform-dom dom)))) - (defun shr-descend (dom) (let ((function (intern (concat "shr-tag-" (symbol-name (car dom))) obarray))) (if (fboundp function) @@ -101,10 +176,43 @@ fit these criteria." ((listp (cdr sub)) (shr-descend sub))))) -(defun shr-tag-p (cont) - (shr-ensure-paragraph) - (shr-generic cont) - (shr-ensure-paragraph)) +(defun shr-insert (text) + (when (eq shr-state 'image) + (insert "\n") + (setq shr-state nil)) + (cond + ((eq shr-folding-mode 'none) + (insert text)) + (t + (let ((first t) + column) + (when (and (string-match "\\`[ \t\n]" text) + (not (bolp))) + (insert " ")) + (dolist (elem (split-string text)) + (when (and (bolp) + (> shr-indentation 0)) + (shr-indent)) + ;; The shr-start is a special variable that is used to pass + ;; upwards the first point in the buffer where the text really + ;; starts. + (unless shr-start + (setq shr-start (point))) + (insert elem) + (when (> (current-column) shr-width) + (if (not (search-backward " " (line-beginning-position) t)) + (insert "\n") + (delete-char 1) + (insert "\n") + (put-text-property (1- (point)) (point) 'shr-break t) + (when (> shr-indentation 0) + (shr-indent)) + (end-of-line))) + (insert " ")))))) + +(defun shr-ensure-newline () + (unless (zerop (current-column)) + (insert "\n"))) (defun shr-ensure-paragraph () (unless (bobp) @@ -119,17 +227,8 @@ fit these criteria." (insert "\n") (insert "\n\n"))))) -(defun shr-tag-b (cont) - (shr-fontize-cont cont 'bold)) - -(defun shr-tag-i (cont) - (shr-fontize-cont cont 'italic)) - -(defun shr-tag-u (cont) - (shr-fontize-cont cont 'underline)) - -(defun shr-tag-s (cont) - (shr-fontize-cont cont 'strike-through)) +(defun shr-indent () + (insert (make-string shr-indentation ? ))) (defun shr-fontize-cont (cont &rest types) (let (shr-start) @@ -141,17 +240,6 @@ fit these criteria." (let ((overlay (make-overlay start end))) (overlay-put overlay 'face type))) -(defun shr-tag-a (cont) - (let ((url (cdr (assq :href cont))) - (start (point)) - shr-start) - (shr-generic cont) - (widget-convert-button - 'link (or shr-start start) (point) - :help-echo url) - (put-text-property (or shr-start start) (point) 'keymap shr-map) - (put-text-property (or shr-start start) (point) 'shr-url url))) - (defun shr-browse-url () "Browse the URL under point." (interactive) @@ -160,99 +248,6 @@ fit these criteria." (message "No link under point") (browse-url url)))) -(defun shr-copy-url () - "Copy the URL under point to the kill ring. -If called twice, then try to fetch the URL and see whether it -redirects somewhere else." - (interactive) - (let ((url (get-text-property (point) 'shr-url))) - (cond - ((not url) - (message "No URL under point")) - ;; Resolve redirected URLs. - ((equal url (car kill-ring)) - (url-retrieve - url - (lambda (a) - (when (and (consp a) - (eq (car a) :redirect)) - (with-temp-buffer - (insert (cadr a)) - (goto-char (point-min)) - ;; Remove common tracking junk from the URL. - (when (re-search-forward ".utm_.*" nil t) - (replace-match "" t t)) - (message "Copied %s" (buffer-string)) - (copy-region-as-kill (point-min) (point-max))))))) - ;; Copy the URL to the kill ring. - (t - (with-temp-buffer - (insert url) - (copy-region-as-kill (point-min) (point-max)) - (message "Copied %s" url)))))) - -(eval-and-compile - (defalias 'shr-encode-url-chars - ;; Neither Emacs 22 nor XEmacs provides this function. - (if (fboundp 'browse-url-url-encode-chars) - 'browse-url-url-encode-chars - (lambda (text chars) - "URL-encode the chars in TEXT that match CHARS. -CHARS is a regexp-like character alternative (e.g., \"[)$]\")." - (let ((encoded-text (copy-sequence text)) - (s 0)) - (while (setq s (string-match chars encoded-text s)) - (setq encoded-text - (replace-match (format "%%%x" - (string-to-char - (match-string 0 encoded-text))) - t t encoded-text) - s (1+ s))) - encoded-text))))) - -(defun shr-tag-img (cont) - (when (and (> (current-column) 0) - (not (eq shr-state 'image))) - (insert "\n")) - (let ((start (point-marker))) - (let ((alt (cdr (assq :alt cont))) - (url (cdr (assq :src cont)))) - (when (zerop (length alt)) - (setq alt "[img]")) - (cond - ((and shr-blocked-images - (string-match shr-blocked-images url)) - (insert alt)) - ((url-is-cached (shr-encode-url-chars url "[&)$ ]")) - (shr-put-image (shr-get-image-data url) (point) alt)) - (t - (insert alt) - (url-retrieve url 'shr-image-fetched - (list (current-buffer) start (point-marker)) - t))) - (insert " ") - (put-text-property start (point) 'keymap shr-map) - (put-text-property start (point) 'shr-alt alt) - (put-text-property start (point) 'shr-image url) - (setq shr-state 'image)))) - -(defun shr-show-alt-text () - "Show the ALT text of the image under point." - (interactive) - (let ((text (get-text-property (point) 'shr-alt))) - (if (not text) - (message "No image under point") - (message "%s" text)))) - -(defun shr-browse-image () - "Browse the image under point." - (interactive) - (let ((url (get-text-property (point) 'shr-image))) - (if (not url) - (message "No image under point") - (message "Browsing %s..." url) - (browse-url url)))) - (defun shr-image-fetched (status buffer start end) (when (and (buffer-name buffer) (not (plist-get status :error))) @@ -302,64 +297,6 @@ CHARS is a regexp-like character alternative (e.g., \"[)$]\")." image))) image))) -(defun shr-tag-pre (cont) - (let ((shr-folding-mode 'none)) - (shr-ensure-newline) - (shr-generic cont) - (shr-ensure-newline))) - -(defun shr-tag-blockquote (cont) - (shr-ensure-paragraph) - (let ((shr-indentation (+ shr-indentation 4))) - (shr-generic cont)) - (shr-ensure-paragraph)) - -(defun shr-ensure-newline () - (unless (zerop (current-column)) - (insert "\n"))) - -(defun shr-insert (text) - (when (eq shr-state 'image) - (insert "\n") - (setq shr-state nil)) - (cond - ((eq shr-folding-mode 'none) - (insert text)) - (t - (let ((first t) - column) - (when (and (string-match "\\`[ \t\n]" text) - (not (bolp))) - (insert " ")) - (dolist (elem (split-string text)) - (setq column (current-column)) - (when (> column 0) - (cond - ((and (or (not first) - (eq shr-state 'space)) - (> (+ column (length elem) 1) shr-width)) - (insert "\n")) - ((not first) - (insert " ")))) - (setq first nil) - (when (and (bolp) - (> shr-indentation 0)) - (shr-indent)) - ;; The shr-start is a special variable that is used to pass - ;; upwards the first point in the buffer where the text really - ;; starts. - (unless shr-start - (setq shr-start (point))) - (insert elem)) - (setq shr-state nil) - (when (and (string-match "[ \t\n]\\'" text) - (not (bolp))) - (insert " ") - (setq shr-state 'space)))))) - -(defun shr-indent () - (insert (make-string shr-indentation ? ))) - (defun shr-get-image-data (url) "Get image data for URL. Return a string with image data." @@ -372,7 +309,95 @@ Return a string with image data." (search-forward "\r\n\r\n" nil t)) (buffer-substring (point) (point-max)))))) -(defvar shr-list-mode nil) +(defun shr-heading (cont &rest types) + (shr-ensure-paragraph) + (apply #'shr-fontize-cont cont types) + (shr-ensure-paragraph)) + +;;; Tag-specific rendering rules. + +(defun shr-tag-p (cont) + (shr-ensure-paragraph) + (shr-generic cont) + (shr-ensure-paragraph)) + +(defun shr-tag-b (cont) + (shr-fontize-cont cont 'bold)) + +(defun shr-tag-i (cont) + (shr-fontize-cont cont 'italic)) + +(defun shr-tag-em (cont) + (shr-fontize-cont cont 'bold)) + +(defun shr-tag-u (cont) + (shr-fontize-cont cont 'underline)) + +(defun shr-tag-s (cont) + (shr-fontize-cont cont 'strike-through)) + +(defun shr-tag-a (cont) + (let ((url (cdr (assq :href cont))) + (start (point)) + shr-start) + (shr-generic cont) + (widget-convert-button + 'link (or shr-start start) (point) + :help-echo url) + (put-text-property (or shr-start start) (point) 'keymap shr-map) + (put-text-property (or shr-start start) (point) 'shr-url url))) + +(defun shr-tag-img (cont) + (when (and (> (current-column) 0) + (not (eq shr-state 'image))) + (insert "\n")) + (let ((start (point-marker))) + (let ((alt (cdr (assq :alt cont))) + (url (cdr (assq :src cont)))) + (when (zerop (length alt)) + (setq alt "[img]")) + (cond + ((and (not shr-inhibit-images) + (string-match "\\`cid:" url)) + (let ((url (substring url (match-end 0))) + image) + (if (or (not shr-content-function) + (not (setq image (funcall shr-content-function url)))) + (insert alt) + (shr-put-image image (point) alt)))) + ((or shr-inhibit-images + (and shr-blocked-images + (string-match shr-blocked-images url))) + (setq shr-start (point)) + (let ((shr-state 'space)) + (if (> (length alt) 8) + (shr-insert (substring alt 0 8)) + (shr-insert alt)))) + ((url-is-cached (browse-url-url-encode-chars url "[&)$ ]")) + (shr-put-image (shr-get-image-data url) (point) alt)) + (t + (insert alt) + (ignore-errors + (url-retrieve url 'shr-image-fetched + (list (current-buffer) start (point-marker)) + t)))) + (insert " ") + (put-text-property start (point) 'keymap shr-map) + (put-text-property start (point) 'shr-alt alt) + (put-text-property start (point) 'shr-image url) + (setq shr-state 'image)))) + +(defun shr-tag-pre (cont) + (let ((shr-folding-mode 'none)) + (shr-ensure-newline) + (shr-generic cont) + (shr-ensure-newline))) + +(defun shr-tag-blockquote (cont) + (shr-ensure-paragraph) + (let ((shr-indentation (+ shr-indentation 4))) + (shr-generic cont)) + (shr-ensure-paragraph)) (defun shr-tag-ul (cont) (shr-ensure-paragraph) @@ -418,20 +443,46 @@ Return a string with image data." (defun shr-tag-h6 (cont) (shr-heading cont)) -(defun shr-heading (cont &rest types) - (shr-ensure-paragraph) - (apply #'shr-fontize-cont cont types) - (shr-ensure-paragraph)) +;;; Table rendering algorithm. +;; Table rendering is the only complicated thing here. We do this by +;; first counting how many TDs there are in each TR, and registering +;; how wide they think they should be ("width=45%", etc). Then we +;; render each TD separately (this is done in temporary buffers, so +;; that we can use all the rendering machinery as if we were in the +;; main buffer). Now we know how much space each TD really takes, so +;; we then render everything again with the new widths, and finally +;; insert all these boxes into the main buffer. (defun shr-tag-table (cont) (shr-ensure-paragraph) (setq cont (or (cdr (assq 'tbody cont)) cont)) - (let* ((columns (shr-column-specs cont)) + (let* ((shr-inhibit-images t) + ;; Find all suggested widths. + (columns (shr-column-specs cont)) + ;; Compute how many characters wide each TD should be. (suggested-widths (shr-pro-rate-columns columns)) + ;; Do a "test rendering" to see how big each TD is (this can + ;; be smaller (if there's little text) or bigger (if there's + ;; unbreakable text). (sketch (shr-make-table cont suggested-widths)) - (sketch-widths (shr-table-widths sketch (length suggested-widths)))) - (shr-insert-table (shr-make-table cont sketch-widths t) sketch-widths))) + (sketch-widths (shr-table-widths sketch suggested-widths))) + ;; Then render the table again with these new "hard" widths. + (shr-insert-table (shr-make-table cont sketch-widths t) sketch-widths)) + ;; Finally, insert all the images after the table. The Emacs buffer + ;; model isn't strong enough to allow us to put the images actually + ;; into the tables. + (dolist (elem (shr-find-elements cont 'img)) + (shr-tag-img (cdr elem)))) + +(defun shr-find-elements (cont type) + (let (result) + (dolist (elem cont) + (cond ((eq (car elem) type) + (push elem result)) + ((consp (cdr elem)) + (setq result (nconc (shr-find-elements (cdr elem) type) result))))) + (nreverse result))) (defun shr-insert-table (table widths) (shr-insert-table-ruler widths) @@ -446,11 +497,20 @@ Return a string with image data." (insert "|\n")) (dolist (column row) (goto-char start) - (let ((lines (split-string (nth 2 column) "\n"))) + (let ((lines (nth 2 column)) + (overlay-lines (nth 3 column)) + overlay overlay-line) (dolist (line lines) + (setq overlay-line (pop overlay-lines)) (when (> (length line) 0) (end-of-line) (insert line "|") + (dolist (overlay overlay-line) + (let ((o (make-overlay (- (point) (nth 0 overlay) 1) + (- (point) (nth 1 overlay) 1))) + (properties (nth 2 overlay))) + (while properties + (overlay-put o (pop properties) (pop properties))))) (forward-line 1))) ;; Add blank lines at padding at the bottom of the TD, ;; possibly. @@ -467,24 +527,47 @@ Return a string with image data." (insert (make-string (aref widths i) ?-) ?+)) (insert "\n")) -(defun shr-table-widths (table length) - (let ((widths (make-vector length 0))) +(defun shr-table-widths (table suggested-widths) + (let* ((length (length suggested-widths)) + (widths (make-vector length 0)) + (natural-widths (make-vector length 0))) (dolist (row table) (let ((i 0)) (dolist (column row) (aset widths i (max (aref widths i) (car column))) - (incf i)))) + (aset natural-widths i (max (aref natural-widths i) + (cadr column))) + (setq i (1+ i))))) + (let ((extra (- (reduce '+ suggested-widths) + (reduce '+ widths))) + (expanded-columns 0)) + (when (> extra 0) + (dotimes (i length) + ;; If the natural width is wider than the rendered width, we + ;; want to allow the column to expand. + (when (> (aref natural-widths i) (aref widths i)) + (setq expanded-columns (1+ expanded-columns)))) + (dotimes (i length) + (when (> (aref natural-widths i) (aref widths i)) + (aset widths i (min + (1+ (aref natural-widths i)) + (+ (/ extra expanded-columns) + (aref widths i)))))))) widths)) (defun shr-make-table (cont widths &optional fill) (let ((trs nil)) (dolist (row cont) (when (eq (car row) 'tr) - (let ((i 0) - (tds nil)) - (dolist (column (cdr row)) - (when (memq (car column) '(td th)) + (let ((tds nil) + (columns (cdr row)) + (i 0) + column) + (while (< i (length widths)) + (setq column (pop columns)) + (when (or (memq (car column) '(td th)) + (null column)) (push (shr-render-td (cdr column) (aref widths i) fill) tds) (setq i (1+ i)))) @@ -493,11 +576,18 @@ Return a string with image data." (defun shr-render-td (cont width fill) (with-temp-buffer - (let ((shr-width width) - (shr-indentation 0)) - (shr-generic cont)) - (while (re-search-backward "\n *$" nil t) - (delete-region (match-beginning 0) (match-end 0))) + (let ((cache (cdr (assoc (cons width cont) shr-content-cache)))) + (if cache + (insert cache) + (let ((shr-width width) + (shr-indentation 0)) + (shr-generic cont)) + (delete-region + (point) + (+ (point) + (skip-chars-backward " \t\n"))) + (push (cons (cons width cont) (buffer-string)) + shr-content-cache))) (goto-char (point-min)) (let ((max 0)) (while (not (eobp)) @@ -506,23 +596,70 @@ Return a string with image data." (forward-line 1)) (when fill (goto-char (point-min)) - (while (not (eobp)) - (end-of-line) - (when (> (- width (current-column)) 0) - (insert (make-string (- width (current-column)) ? ))) - (forward-line 1))) - (list max (count-lines (point-min) (point-max)) (buffer-string))))) + ;; If the buffer is totally empty, then put a single blank + ;; line here. + (if (zerop (buffer-size)) + (insert (make-string width ? )) + ;; Otherwise, fill the buffer. + (while (not (eobp)) + (end-of-line) + (when (> (- width (current-column)) 0) + (insert (make-string (- width (current-column)) ? ))) + (forward-line 1)))) + (if fill + (list max + (count-lines (point-min) (point-max)) + (split-string (buffer-string) "\n") + (shr-collect-overlays)) + (list max + (shr-natural-width)))))) + +(defun shr-natural-width () + (goto-char (point-min)) + (let ((current 0) + (max 0)) + (while (not (eobp)) + (end-of-line) + (setq current (+ current (current-column))) + (unless (get-text-property (point) 'shr-break) + (setq max (max max current) + current 0)) + (forward-line 1)) + max)) + +(defun shr-collect-overlays () + (save-excursion + (goto-char (point-min)) + (let ((overlays nil)) + (while (not (eobp)) + (push (shr-overlays-in-region (point) (line-end-position)) + overlays) + (forward-line 1)) + (nreverse overlays)))) + +(defun shr-overlays-in-region (start end) + (let (result) + (dolist (overlay (overlays-in start end)) + (push (list (if (> start (overlay-start overlay)) + (- end start) + (- end (overlay-start overlay))) + (if (< end (overlay-end overlay)) + 0 + (- end (overlay-end overlay))) + (overlay-properties overlay)) + result)) + (nreverse result))) (defun shr-pro-rate-columns (columns) (let ((total-percentage 0) (widths (make-vector (length columns) 0))) (dotimes (i (length columns)) - (incf total-percentage (aref columns i))) + (setq total-percentage (+ total-percentage (aref columns i)))) (setq total-percentage (/ 1.0 total-percentage)) (dotimes (i (length columns)) (aset widths i (max (truncate (* (aref columns i) total-percentage - shr-width)) + (- shr-width (1+ (length columns))))) 10))) widths)) @@ -539,8 +676,8 @@ Return a string with image data." (string-match "\\([0-9]+\\)%" width)) (aset columns i (/ (string-to-number (match-string 1 width)) - 100.0))))) - (setq i (1+ i)))))) + 100.0)))) + (setq i (1+ i))))))) columns)) (defun shr-count (cont elem) @@ -554,7 +691,8 @@ Return a string with image data." (let ((max 0)) (dolist (row cont) (when (eq (car row) 'tr) - (setq max (max max (shr-count (cdr row) 'td))))) + (setq max (max max (+ (shr-count (cdr row) 'td) + (shr-count (cdr row) 'th)))))) max)) (provide 'shr)