1 ;;; gnus-html.el --- Render HTML in a buffer.
3 ;; Copyright (C) 2010 Free Software Foundation, Inc.
5 ;; Author: Lars Magne Ingebrigtsen <larsi@gnus.org>
8 ;; This file is part of GNU Emacs.
10 ;; GNU Emacs is free software: you can redistribute it and/or modify
11 ;; it under the terms of the GNU General Public License as published by
12 ;; the Free Software Foundation, either version 3 of the License, or
13 ;; (at your option) any later version.
15 ;; GNU Emacs is distributed in the hope that it will be useful,
16 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
17 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 ;; GNU General Public License for more details.
20 ;; You should have received a copy of the GNU General Public License
21 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
25 ;; The idea is to provide a simple, fast and pretty minimal way to
26 ;; render HTML (including links and images) in a buffer, based on an
27 ;; external HTML renderer (i.e., w3m).
31 (eval-when-compile (require 'cl))
32 (eval-when-compile (require 'mm-decode))
35 (defcustom gnus-html-cache-directory (nnheader-concat gnus-directory "html-cache/")
36 "Where Gnus will cache images it downloads from the web."
41 (defcustom gnus-html-cache-size 500000000
42 "The size of the Gnus image cache."
47 (defcustom gnus-html-frame-width 70
48 "What width to use when rendering HTML."
53 (defcustom gnus-blocked-images "."
54 "Images that have URLs matching this regexp will be blocked."
59 (defcustom gnus-max-image-proportion 0.7
60 "How big pictures displayed are in relation to the window they're in.
61 A value of 0.7 means that they are allowed to take up 70% of the
62 width and height of the window. If they are larger than this,
63 and Emacs supports it, then the images will be rescaled down to
70 (defun gnus-article-html (handle)
71 (let ((article-buffer (current-buffer)))
73 (narrow-to-region (point) (point))
76 (let* ((coding-system-for-read 'utf-8)
77 (coding-system-for-write 'utf-8)
78 (default-process-coding-system
79 (cons coding-system-for-read coding-system-for-write))
80 (charset (mail-content-type-get (mm-handle-type handle)
83 (setq charset (mm-charset-to-coding-system charset))
84 (not (eq charset 'ascii)))
86 (mm-decode-coding-string (buffer-string) charset)
88 (mm-enable-multibyte))))
89 (call-process-region (point-min) (point-max)
91 nil article-buffer nil
98 "-t" (format "%s" tab-width)
99 "-cols" (format "%s" gnus-html-frame-width)
100 "-o" "display_image=on"
102 (gnus-html-wash-tags))))
104 (defvar gnus-article-mouse-face)
106 (defun gnus-html-wash-tags ()
107 (let (tag parameters string start end images url)
108 (mm-url-decode-entities)
109 (goto-char (point-min))
110 (while (re-search-forward "<pre_int> *</pre_int>\n" nil t)
111 (replace-match "" t t))
112 (goto-char (point-min))
113 (while (re-search-forward "<\\([^ />]+\\)\\([^>]*\\)>" nil t)
114 (setq tag (match-string 1)
115 parameters (match-string 2)
116 start (match-beginning 0))
117 (when (plusp (length parameters))
118 (set-text-properties 0 (1- (length parameters)) nil parameters))
119 (delete-region start (point))
120 (when (search-forward (concat "</" tag ">") nil t)
121 (delete-region (match-beginning 0) (match-end 0)))
124 ;; Fetch and insert a picture.
125 ((equal tag "img_alt")
126 (when (string-match "src=\"\\([^\"]+\\)" parameters)
127 (setq url (match-string 1 parameters))
128 (gnus-message 8 "gnus-html-wash-tags: fetching image URL %s" url)
129 (if (string-match "^cid:\\(.*\\)" url)
130 ;; URLs with cid: have their content stashed in other
131 ;; parts of the MIME structure, so just insert them
133 (let ((handle (mm-get-content-id
134 (setq url (match-string 1 url))))
138 (setq image (gnus-create-image (buffer-string)
141 (let ((string (buffer-substring start end)))
142 (delete-region start end)
143 (gnus-put-image image (gnus-string-or string "*")))))
144 ;; Normal, external URL.
145 (unless (gnus-html-image-url-blocked-p
147 (if (buffer-live-p gnus-summary-buffer)
148 (with-current-buffer gnus-summary-buffer
150 gnus-blocked-images))
151 (let ((file (gnus-html-image-id url))
153 (when (string-match "height=\"?\\([0-9]+\\)" parameters)
154 (setq height (string-to-number (match-string 1 parameters))))
155 (when (string-match "width=\"?\\([0-9]+\\)" parameters)
156 (setq width (string-to-number (match-string 1 parameters))))
157 ;; Don't fetch images that are really small. They're
158 ;; probably tracking pictures.
159 (when (and (or (null height)
163 (if (file-exists-p file)
164 ;; It's already cached, so just insert it.
165 (let ((string (buffer-substring start end)))
166 ;; Delete the ALT text.
167 (delete-region start end)
168 (gnus-html-put-image file (point) string))
169 ;; We don't have it, so schedule it for fetching
172 (set-marker (make-marker) start)
178 (when (string-match "href=\"\\([^\"]+\\)" parameters)
179 (setq url (match-string 1 parameters))
180 (gnus-message 8 "gnus-html-wash-tags: fetching link URL %s" url)
181 (gnus-article-add-button start end
184 (let ((overlay (gnus-make-overlay start end)))
185 (gnus-overlay-put overlay 'evaporate t)
186 (gnus-overlay-put overlay 'gnus-button-url url)
187 (gnus-put-text-property start end 'gnus-string url)
188 (when gnus-article-mouse-face
189 (gnus-overlay-put overlay 'mouse-face gnus-article-mouse-face)))))
190 ;; The upper-case IMG_ALT is apparently just an artifact that
191 ;; should be deleted.
192 ((equal tag "IMG_ALT")
193 (delete-region start end))
194 ;; Whatever. Just ignore the tag.
198 (goto-char (point-min))
199 ;; The output from -halfdump isn't totally regular, so strip
200 ;; off any </pre_int>s that were left over.
201 (while (re-search-forward "</pre_int>" nil t)
202 (replace-match "" t t))
204 (gnus-html-schedule-image-fetching (current-buffer) (nreverse images)))))
206 (defun gnus-html-schedule-image-fetching (buffer images)
207 (gnus-message 8 "gnus-html-schedule-image-fetching: buffer %s, images %s"
209 (let* ((url (caar images))
210 (process (start-process
215 "-o" (gnus-html-image-id url)
217 (process-kill-without-query process)
218 (set-process-sentinel process 'gnus-html-curl-sentinel)
219 (gnus-set-process-plist process (list 'images images
222 (defun gnus-html-image-id (url)
223 (expand-file-name (sha1 url) gnus-html-cache-directory))
225 (defun gnus-html-curl-sentinel (process event)
226 (when (string-match "finished" event)
227 (let* ((images (gnus-process-get process 'images))
228 (buffer (gnus-process-get process 'buffer))
230 (file (gnus-html-image-id (car spec))))
231 (when (and (buffer-live-p buffer)
232 ;; If the position of the marker is 1, then that
233 ;; means that the text it was in has been deleted;
234 ;; i.e., that the user has selected a different
235 ;; article before the image arrived.
236 (not (= (marker-position (cadr spec)) (point-min))))
237 (with-current-buffer buffer
238 (let ((inhibit-read-only t)
239 (string (buffer-substring (cadr spec) (caddr spec))))
240 (delete-region (cadr spec) (caddr spec))
241 (gnus-html-put-image file (cadr spec) string))))
243 (gnus-html-schedule-image-fetching buffer images)))))
245 (defun gnus-html-put-image (file point string)
246 (when (gnus-graphic-display-p)
247 (let* ((image (ignore-errors
248 (gnus-create-image file)))
250 (if (featurep 'xemacs)
251 (cons (glyph-width image) (glyph-height image))
252 (image-size image t)))))
256 ;; Kludge to avoid displaying 30x30 gif images, which
257 ;; seems to be a signal of a broken image.
258 (not (and (if (featurep 'xemacs)
261 (eq (if (featurep 'xemacs)
262 (let ((data (cdadar (specifier-spec-list
263 (glyph-image image)))))
266 (plist-get (cdr image) :type))
271 (gnus-put-image (gnus-html-rescale-image image file size)
272 (gnus-string-or string "*"))
275 (when (fboundp 'find-image)
276 (gnus-put-image (find-image
277 '((:type xpm :file "lock-broken.xpm")))
278 (gnus-string-or string "*")))
281 (defun gnus-html-rescale-image (image file size)
282 (if (or (not (fboundp 'imagemagick-types))
283 (not (get-buffer-window (current-buffer))))
285 (let* ((width (car size))
287 (edges (window-pixel-edges (get-buffer-window (current-buffer))))
288 (window-width (truncate (* gnus-max-image-proportion
289 (- (nth 2 edges) (nth 0 edges)))))
290 (window-height (truncate (* gnus-max-image-proportion
291 (- (nth 3 edges) (nth 1 edges)))))
293 (when (> height window-height)
294 (setq image (or (create-image file 'imagemagick nil
295 :height window-height)
297 (setq size (image-size image t)))
298 (when (> (car size) window-width)
300 (create-image file 'imagemagick nil
305 (defun gnus-html-prune-cache ()
308 (dolist (file (directory-files gnus-html-cache-directory t nil t))
309 (let ((attributes (file-attributes file)))
310 (unless (nth 0 attributes)
311 (incf total-size (nth 7 attributes))
312 (push (list (time-to-seconds (nth 5 attributes))
313 (nth 7 attributes) file)
315 (when (> total-size gnus-html-cache-size)
316 (setq files (sort files (lambda (f1 f2)
317 (< (car f1) (car f2)))))
319 (when (> total-size gnus-html-cache-size)
320 (decf total-size (cadr file))
321 (delete-file (nth 2 file)))))))
323 (defun gnus-html-image-url-blocked-p (url blocked-images)
324 "Find out if URL is blocked by BLOCKED-IMAGES."
325 (let ((ret (and blocked-images
326 (string-match blocked-images url))))
328 (gnus-message 8 "gnus-html-image-url-blocked-p: %s blocked by regex %s"
330 (gnus-message 9 "gnus-html-image-url-blocked-p: %s passes regex %s"
335 (defun gnus-html-prefetch-images (summary)
336 (let (blocked-images urls)
337 (when (buffer-live-p summary)
338 (with-current-buffer summary
339 (setq blocked-images gnus-blocked-images))
341 (while (re-search-forward "<img.*src=[\"']\\([^\"']+\\)" nil t)
342 (let ((url (match-string 1)))
343 (unless (gnus-html-image-url-blocked-p url blocked-images)
344 (unless (file-exists-p (gnus-html-image-id url))
346 (push (gnus-html-image-id url) urls)
349 (apply 'start-process
355 (process-kill-without-query process))))))
359 ;;; gnus-html.el ends here