From: Lars Ingebrigtsen <larsi@gnus.org>
Date: Sun, 19 Jul 2020 20:41:28 +0000 (+0200)
Subject: Make timeouts work more reliably in url-retrieve-synchronously
X-Git-Tag: emacs-28.0.90~6963
X-Git-Url: http://git.eshelyaron.com/gitweb/?a=commitdiff_plain;h=5d2bc1543925960009167988523d76fc6d4c805a;p=emacs.git

Make timeouts work more reliably in url-retrieve-synchronously

* lisp/url/url.el (url-retrieve-synchronously): Ensure that the
processes have been killed on timeouts before returning the buffer
(bug#34607).
---

diff --git a/lisp/url/url.el b/lisp/url/url.el
index 12a8a9c2e21..367af1b5a90 100644
--- a/lisp/url/url.el
+++ b/lisp/url/url.el
@@ -300,8 +300,15 @@ how long to wait for a response before giving up."
 	      (when quit-flag
 		(delete-process proc))
               (setq proc (and (not quit-flag)
-			      (get-buffer-process asynch-buffer)))))))
-      asynch-buffer)))
+			      (get-buffer-process asynch-buffer))))))
+        ;; On timeouts, make sure we kill any pending processes.
+        ;; There may be more than one if we had a redirect.
+        (when (process-live-p proc)
+          (delete-process proc))
+        (when-let ((aproc (get-buffer-process asynch-buffer)))
+          (when (process-live-p aproc)
+            (delete-process aproc)))))
+    asynch-buffer))
 
 ;; url-mm-callback called from url-mm, which requires mm-decode.
 (declare-function mm-dissect-buffer "mm-decode"