diff options
Diffstat (limited to 'admin')
| -rw-r--r-- | admin/FOR-RELEASE | 39 |
1 files changed, 3 insertions, 36 deletions
diff --git a/admin/FOR-RELEASE b/admin/FOR-RELEASE index 40fb0168793..84c236fe426 100644 --- a/admin/FOR-RELEASE +++ b/admin/FOR-RELEASE | |||
| @@ -32,6 +32,9 @@ invalid pointer from string_free_list. | |||
| 32 | 32 | ||
| 33 | ** Fix up url-ldap.el. | 33 | ** Fix up url-ldap.el. |
| 34 | 34 | ||
| 35 | ** url/*.el has lots of `(declare (special ...))' which | ||
| 36 | are meaningless. What's that trying to do? | ||
| 37 | |||
| 35 | * BUGS | 38 | * BUGS |
| 36 | 39 | ||
| 37 | ** Incomplete overlay mouse-face highlight bug (Ralf Angeli, Oct 18) | 40 | ** Incomplete overlay mouse-face highlight bug (Ralf Angeli, Oct 18) |
| @@ -62,42 +65,6 @@ further. | |||
| 62 | I think in the near future we will see more of this problem, so it might be | 65 | I think in the near future we will see more of this problem, so it might be |
| 63 | time to make anfe-ftp more intelligent. | 66 | time to make anfe-ftp more intelligent. |
| 64 | 67 | ||
| 65 | ** Bug in url-http-parse-headers, reported in | ||
| 66 | From: Vivek Dasmohapatra <vivek@zeus.com> | ||
| 67 | Date: Tue, 28 Sep 2004 16:13:13 +0100 | ||
| 68 | |||
| 69 | Fetching a url with url-retrieve can reult in an anrbitrary buffer | ||
| 70 | being killed if a 401 (or possibly a 407) result is encountered: | ||
| 71 | |||
| 72 | url-http-parse-headers calls url-http-handle-authentication, | ||
| 73 | which can call url-retrieve. | ||
| 74 | |||
| 75 | This results in the current buffer being killed, and a new http buffer | ||
| 76 | being generated. However, when the old http buffer is killed, emacs | ||
| 77 | picks the top buffer from the list as the new current buffer, so by the | ||
| 78 | time we get to the end of url-http-parse-headers, _that_ buffer is marked | ||
| 79 | as dead even though it is not necessarily a url buffer, so next time the | ||
| 80 | url libraries reap their dead buffers, an innocent bystander buffer is | ||
| 81 | killed instead (and an obsolete http buffer may be left lying around too). | ||
| 82 | |||
| 83 | A possible fix (which I am currently using) is to call set-buffer | ||
| 84 | on the return value of url-http-parse-headers: | ||
| 85 | |||
| 86 | (case url-http-response-status | ||
| 87 | (401 | ||
| 88 | ;; The request requires user authentication. The response | ||
| 89 | ;; MUST include a WWW-Authenticate header field containing a | ||
| 90 | ;; challenge applicable to the requested resource. The | ||
| 91 | ;; client MAY repeat the request with a suitable | ||
| 92 | ;; Authorization header field. | ||
| 93 | (url-mark-buffer-as-dead (current-buffer)) | ||
| 94 | (set-buffer (url-http-handle-authentication nil))) | ||
| 95 | etc .... | ||
| 96 | |||
| 97 | which makes sure that it is the right http buffer that is current when | ||
| 98 | we come to mark the http buffers as dead. | ||
| 99 | |||
| 100 | |||
| 101 | 68 | ||
| 102 | * GTK RELATED BUGS | 69 | * GTK RELATED BUGS |
| 103 | 70 | ||