bug#60146: file-exists-in-trash-p needs better name or semantics

bug#60146: file-exists-in-trash-p needs better name or semantics

[Paul Eggert]

The recently-added function (file-exists-in-trash-p FILE) is poorly 
named since it's not really related to trash - it's simply checking for 
the existence of a directory entry named FILE.

How about extending file-exists-p instead? (file-exists-p FILE t) would 
be like (file-exists-p FILE) except it would not follow symlinks. This 
extension can be implemented via a single system call on POSIX systems, 
and this would be more efficient and would avoid a race in the current 
implementation of file-exists-in-trash-p. (Though of course pretty much 
any use of this new function makes one vulnerable to races....)

If extending file-exists-p is too much, at least please rename 
file-exists-in-trash-p to something like files--exists-nofollow-p, to 
indicate that it's private to files.el and to say better what it means.

bug#60146: file-exists-in-trash-p needs better name or semantics

[Eli Zaretskii]

> Date: Fri, 16 Dec 2022 21:17:13 -0800
> From: Paul Eggert <eggert@cs.ucla.edu>
> 
> The recently-added function (file-exists-in-trash-p FILE) is poorly 
> named since it's not really related to trash - it's simply checking for 
> the existence of a directory entry named FILE.
> 
> How about extending file-exists-p instead? (file-exists-p FILE t) would 
> be like (file-exists-p FILE) except it would not follow symlinks. This 
> extension can be implemented via a single system call on POSIX systems, 
> and this would be more efficient and would avoid a race in the current 
> implementation of file-exists-in-trash-p. (Though of course pretty much 
> any use of this new function makes one vulnerable to races....)
> 
> If extending file-exists-p is too much, at least please rename 
> file-exists-in-trash-p to something like files--exists-nofollow-p, to 
> indicate that it's private to files.el and to say better what it means.

I don't really mind renaming that function, but the reason I called it
like I did was that apparently no one needed such a functionality in
Emacs until now, except in this obscure use case of moving to trash.
That seems to tell me that extending file-exists-p would be a solution
waiting for the problem, something that we don't like doing.

In any case, if we do decide to extend file-exists-p, it would need to
be done on master, as that is not a trivial change, which has to be
done in C.

Lars, Stefan, WDYT?

bug#60146: file-exists-in-trash-p needs better name or semantics

[Michael Albinus]

Paul Eggert <eggert@cs.ucla.edu> writes:

Hi Paul,

> How about extending file-exists-p instead? (file-exists-p FILE t)
> would be like (file-exists-p FILE) except it would not follow
> symlinks. This extension can be implemented via a single system call
> on POSIX systems, and this would be more efficient and would avoid a
> race in the current implementation of file-exists-in-trash-p. (Though
> of course pretty much any use of this new function makes one
> vulnerable to races....)

Pls don't do it in the emacs-29 branch. file-exists-p checks for file
name handlers, with all consequences for a changed signature.

Best regards, Michael.

bug#60146: file-exists-in-trash-p needs better name or semantics

[Stefan Monnier via Bug reports for GNU Emacs, the Swiss army knife of text editors]

> I don't really mind renaming that function, but the reason I called it
> like I did was that apparently no one needed such a functionality in
> Emacs until now, except in this obscure use case of moving to trash.
> That seems to tell me that extending file-exists-p would be a solution
> waiting for the problem, something that we don't like doing.

Hmm... arguing about names, eh?
Let me throw that hat into the ring: IIUC this is a variant of
`file-exists-p` (currently) used only when dealing the trash, so I'd
call it `trash--file-exists-p`.

> In any case, if we do decide to extend file-exists-p, it would need to
> be done on master, as that is not a trivial change, which has to be
> done in C.

Indeed, I see no need for that on emacs-29.


        Stefan

bug#60146: file-exists-in-trash-p needs better name or semantics

[Eli Zaretskii]

> From: Stefan Monnier <monnier@iro.umontreal.ca>
> Cc: Paul Eggert <eggert@cs.ucla.edu>,  Lars Ingebrigtsen <larsi@gnus.org>,
>   60146@debbugs.gnu.org
> Date: Sat, 17 Dec 2022 10:51:38 -0500
> 
> Let me throw that hat into the ring: IIUC this is a variant of
> `file-exists-p` (currently) used only when dealing the trash, so I'd
> call it `trash--file-exists-p`.

Well, the function is defined in files.el, so starting the name with
"trash" would be frowned upon...

bug#60146: file-exists-in-trash-p needs better name or semantics

[Paul Eggert]

[-- Attachment #1: Type: text/plain, Size: 817 bytes --]

On 12/17/22 00:47, Eli Zaretskii wrote:
> I don't really mind renaming that function, but the reason I called it
> like I did was that apparently no one needed such a functionality in
> Emacs until now,

I have a bit of a different take. First, there's adequate functionality 
in Emacs now if we don't mind inefficiency, and it's already used; just 
call file-attributes. (Gnus does this, in nnmaildir-unlink at least, and 
I expect there are other uses.) Second, I wouldn't be surprised if other 
uses of file-exists-p have problems similar to the one you found in 
move-file-to-trash. Not that any of us have time to go scout for them 
right now.

It's a minor point. Still, the name file-exists-in-trash-p really needs 
to go somehow as it's not a name that should be user-visible. How about 
the attached patch?


[-- Attachment #2: 0001-Remove-file-exists-in-trash-p.patch --]
[-- Type: text/x-patch, Size: 2525 bytes --]

From adcf3112e72431e4c4a2610941afcc238e7dd243 Mon Sep 17 00:00:00 2001
From: Paul Eggert <eggert@cs.ucla.edu>
Date: Sat, 17 Dec 2022 11:54:46 -0800
Subject: [PATCH] Remove file-exists-in-trash-p

* lisp/files.el (file-exists-in-trash-p): Remove, as this name is
not suitable for users.  All uses replaced by file-attributes,
which is good enough here.
---
 lisp/files.el | 14 +++-----------
 1 file changed, 3 insertions(+), 11 deletions(-)

diff --git a/lisp/files.el b/lisp/files.el
index c74e7e808e4..149480a6f10 100644
--- a/lisp/files.el
+++ b/lisp/files.el
@@ -8467,14 +8467,6 @@ trash--hexify-table
 
 (declare-function system-move-file-to-trash "w32fns.c" (filename))
 
-(defun file-exists-in-trash-p (filename)
-  "Return non-nil if FILENAME exists in the trash.
-
-This is like `file-exists-p', but it also returns non-nil
-if FILENAME is a dangling symlink, to allow trashing such files."
-  (or (file-exists-p filename)
-      (file-symlink-p filename)))
-
 (defun move-file-to-trash (filename)
   "Move the file (or directory) named FILENAME to the trash.
 When `delete-by-moving-to-trash' is non-nil, this function is
@@ -8505,7 +8497,7 @@ move-file-to-trash
 	   (unless (file-directory-p trash-dir)
 	     (make-directory trash-dir t))
 	   ;; Ensure that the trashed file-name is unique.
-	   (if (file-exists-in-trash-p new-fn)
+	   (if (file-attributes new-fn)
 	       (let ((version-control t)
 		     (backup-directory-alist nil))
 		 (setq new-fn (car (find-backup-file-name new-fn)))))
@@ -8582,7 +8574,7 @@ move-file-to-trash
                  ;; We're checking further down whether the info file
                  ;; exists, but the file name may exist in the trash
                  ;; directory even if there is no info file for it.
-                 (when (file-exists-in-trash-p
+                 (when (file-attributes
                         (file-name-concat trash-files-dir files-base))
                    (setq overwrite t
                          files-base (file-name-nondirectory
@@ -8620,7 +8612,7 @@ move-file-to-trash
 		 (let ((delete-by-moving-to-trash nil)
 		       (new-fn (file-name-concat trash-files-dir files-base)))
                    (if (or (not is-directory)
-                           (not (file-exists-in-trash-p new-fn)))
+                           (not (file-attributes new-fn)))
                        (rename-file fn new-fn overwrite)
                      (copy-directory fn
                                      (file-name-as-directory new-fn)
-- 
2.38.1

bug#60146: file-exists-in-trash-p needs better name or semantics

[Eli Zaretskii]

> Date: Sat, 17 Dec 2022 11:57:39 -0800
> Cc: 60146@debbugs.gnu.org
> From: Paul Eggert <eggert@cs.ucla.edu>
> 
> I have a bit of a different take. First, there's adequate functionality 
> in Emacs now if we don't mind inefficiency, and it's already used; just 
> call file-attributes. (Gnus does this, in nnmaildir-unlink at least, and 
> I expect there are other uses.) Second, I wouldn't be surprised if other 
> uses of file-exists-p have problems similar to the one you found in 
> move-file-to-trash. Not that any of us have time to go scout for them 
> right now.
> 
> It's a minor point. Still, the name file-exists-in-trash-p really needs 
> to go somehow as it's not a name that should be user-visible. How about 
> the attached patch?

If you can verify that it works in the scenario of bug#59986, I'm okay
with this change.

bug#60146: file-exists-in-trash-p needs better name or semantics

[Paul Eggert]

On 12/17/22 12:09, Eli Zaretskii wrote:
> If you can verify that it works in the scenario of bug#59986, I'm okay
> with this change.

Thanks, I checked that the patch works in that scenario (dangling 
symlink in the trash directory), and installed it. Closing the bug report.