From: Eli Zaretskii <eliz@gnu.org>
To: Stefan Monnier <monnier@iro.umontreal.ca>
Cc: enometh@meer.net, 74208@debbugs.gnu.org
Subject: bug#74208: 31.0.50; minibuffer read-file-name-default mutates global value of default-directory incorrectly
Date: Sat, 23 Nov 2024 14:25:42 +0200 [thread overview]
Message-ID: <864j3ym915.fsf@gnu.org> (raw)
In-Reply-To: <jwvr07iv413.fsf-monnier+emacs@gnu.org> (message from Stefan Monnier on Sun, 10 Nov 2024 22:40:27 -0500)
Ping! How should we make progress with this issue?
> From: Stefan Monnier <monnier@iro.umontreal.ca>
> Cc: eliz@gnu.org, 74208@debbugs.gnu.org
> Date: Sun, 10 Nov 2024 22:40:27 -0500
>
> >>> The problem is that the parameter DIR == "http://example.com" was
> >>> getting bound to default-directory with undesirable results.
> >
> >> It can also have desirable results sometimes (depends on the URL at point
> >> and what kind of support for URLs you have in your Emacs, admittedly).
> >> Do you happen to have a more realistic scenario than your "echo foo"
> >> that lets us better judge the severity of the problem, and maybe other
> >> ways to solve the actual problem?
> >
> > Are you saying it is ok for emacs to make all calls to `call-prcoess'
> > to unconditionally fail --- when emacs is waiting for input at the
> > minibuffer? the calls to call-process can come from anywhere, from
> > timers, or from outside via emacsclient, etc.
>
> `default-directory` can point to random places, including non-existing
> directories, HTTP places (e.g. if you enable `url-handler-mode`), so
> code like timers need to take this possibility into account, yes.
> The ffap minibuffer discussed here is just one of many other ways to get
> into this situation.
>
> > I already posted that any hypothetical completion facilites are
> > singularly useless for the task of ffap (url) -- to get the url at
> > point into the minibuffer and maybe manually edit it. If these
> > facilities exist and you are using them, I'd like to see an example.
>
> diff --git a/lisp/url/url-handlers.el b/lisp/url/url-handlers.el
> index 9edc7865a74..17013345c25 100644
> --- a/lisp/url/url-handlers.el
> +++ b/lisp/url/url-handlers.el
> @@ -375,18 +375,48 @@ url-insert-file-contents-literally
> (kill-buffer buffer)
> nil))
>
> -(defun url-file-name-completion (url _directory &optional _predicate)
> - ;; Even if it's not implemented, it's not an error to ask for completion,
> - ;; in case it's available (bug#14806).
> - ;; (error "Unimplemented")
> - url)
> +(defun url-file-name-completion (url directory &optional predicate)
> + (let ((all (url-file-name-all-completions url directory)))
> + (if (null all)
> + ;; If `url' is the empty string, don't return nil, so as to prevent
> + ;; partial-completion from recursing into the parent directory.
> + (if (equal url "") url)
> + (try-completion url all predicate))))
> (put 'file-name-completion 'url-file-handlers #'url-file-name-completion)
>
> -(defun url-file-name-all-completions (_file _directory)
> - ;; Even if it's not implemented, it's not an error to ask for completion,
> - ;; in case it's available (bug#14806).
> - ;; (error "Unimplemented")
> - nil)
> +(defvar url-handler-temp-buf)
> +
> +(defun url-file-name-all-completions (file directory)
> + ;; FIXME: Cache the "directory" buffers between completion requests.
> + (let ((buf (get-file-buffer directory)))
> + (unless buf
> + (setq buf (ignore-errors (find-file-noselect directory)))
> + (when buf
> + (with-current-buffer buf
> + (set (make-local-variable 'url-handler-temp-buf) t))))
> + (when buf
> + (unwind-protect
> + (with-current-buffer buf
> + (save-excursion
> + (let ((all ())
> + (case-fold-search t)
> + ;; FIXME: Handle URL-quoting.
> + (regexp (format "<a href=\"\\(%s[^\"]+\\)\"" file)))
> + (goto-char (point-min))
> + (while (re-search-forward regexp nil t)
> + (let ((url (match-string 1)))
> + (unless (string-match
> + "\\`\\(?:\\.\\.\\|[#?/]\\|[-a-z]+:/\\)\\|" url)
> + ;; It's a relative URL.
> + (when (string-match "[#?]\\|/\\(.\\)" url)
> + (setq url (substring url (or (match-beginning 1)
> + (match-beginning 0)))))
> + ;; FIXME: Handle URL-unquoting.
> + (push url all))))
> + all)))
> + (and (buffer-live-p buf)
> + (buffer-local-value 'url-handler-temp-buf buf)
> + (kill-buffer buf))))))
> (put 'file-name-all-completions
> 'url-file-handlers #'url-file-name-all-completions)
>
>
>
next prev parent reply other threads:[~2024-11-23 12:25 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-11-05 2:09 bug#74208: 31.0.50; minibuffer read-file-name-default mutates global value of default-directory incorrectly Madhu
2024-11-09 11:11 ` Eli Zaretskii
2024-11-09 16:11 ` Stefan Monnier via Bug reports for GNU Emacs, the Swiss army knife of text editors
2024-11-10 0:57 ` Madhu
2024-11-10 6:00 ` Eli Zaretskii
2024-11-10 7:00 ` Madhu
2024-11-10 10:00 ` Eli Zaretskii
2024-11-10 10:41 ` Madhu
2024-11-10 10:45 ` Eli Zaretskii
2024-11-10 11:17 ` Madhu
2024-11-10 11:26 ` Eli Zaretskii
2024-11-10 16:54 ` Stefan Monnier via Bug reports for GNU Emacs, the Swiss army knife of text editors
2024-11-11 2:17 ` Madhu
2024-11-11 3:40 ` Stefan Monnier via Bug reports for GNU Emacs, the Swiss army knife of text editors
2024-11-23 12:25 ` Eli Zaretskii [this message]
2024-11-23 14:56 ` Madhu
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: https://www.gnu.org/software/emacs/
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=864j3ym915.fsf@gnu.org \
--to=eliz@gnu.org \
--cc=74208@debbugs.gnu.org \
--cc=enometh@meer.net \
--cc=monnier@iro.umontreal.ca \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://git.savannah.gnu.org/cgit/emacs.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).