Re: master 7b50ed5: Ask if dir and subdir dired buffers be killed when deleting dir

[Tassilo Horn]

"Basil L. Contovounesios" <contovob@tcd.ie> writes:

Hi Basil,

>> branch: master
>> commit 7b50ed553faa6de6d51bf07d12d106ca61ab3ac4
>> Author: Tassilo Horn <tsdh@gnu.org>
>> Commit: Tassilo Horn <tsdh@gnu.org>
> [...]
>>     (dired-in-this-tree-p): Make obsolete in favor of file-in-directory-p
>>     which actually does what the name suggest whereas dired-in-this-tree-p
>>     is just string-matching on filenames which will fail with symlinks
>>     filenames including ./ or ../.
>
> There are still a few uses of dired-in-this-tree-p in dired-aux.el,
> which give rise to byte-compilation warnings.

Oh, I somehow missed those.

> Is the attached patch TRT, or are there any subtle differences between
> the functions that still need to be addressed?

Yes, I think they are equivalent in all cases dired cares about and
file-in-directory-p stays correct also in cases where
dired-in-this-tree-p will give a wrong answer.

But sorry, I've read your mail on my phone without noticing the patch.
I've done basically the same right now before actually reading your mail
again in a proper MUA.

I haven't done this:

--8<---------------cut here---------------start------------->8---
diff --git a/lisp/dired.el b/lisp/dired.el
index bb428e2198..8c7bff31cb 100644
--- a/lisp/dired.el
+++ b/lisp/dired.el
@@ -2912,10 +2912,10 @@ dired-unadvertise
 ;;; utility functions
 
 (defun dired-in-this-tree-p (file dir)
-  ;;"Is FILE part of the directory tree starting at DIR?"
+  "Return non-nil if FILE is part of the directory tree starting at DIR."
+  (declare (obsolete file-in-directory-p "28.1"))
   (let (case-fold-search)
     (string-match-p (concat "^" (regexp-quote dir)) file)))
-(make-obsolete 'dired-in-this-tree-p 'file-in-directory-p "28.1")
 (define-obsolete-function-alias 'dired-in-this-tree
   'dired-in-this-tree-p "27.1")
--8<---------------cut here---------------end--------------->8---

Can you tell me the difference between this and `make-obsolete'?  And
will that also trigger a warning during byte-compilation of code using
the function or only when it is actually called?  Feel free to do that
change if you think it is better.

Bye,
Tassilo