GNU bug report logs - #79219
[PATCH] Insert * on TAB with completion-pcm-leading-wildcard=t

Previous Next

To reply to this bug, email your comments to 79219 AT debbugs.gnu.org.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Spencer Baugh <sbaugh <at> janestreet.com>
To: bug-gnu-emacs <at> gnu.org
Subject: [PATCH] Insert * on TAB with completion-pcm-leading-wildcard=t
Date: Mon, 11 Aug 2025 12:44:12 -0400

[Message part 1 (text/plain, inline)]

Tags: patch


Improve completion-pcm-leading-wildcard by actually inserting
a * when try-completion runs.  completion-pcm--merge-try will
automatically remove the wildcard anywhere that it has no
matches, so the * is inserted only where it actually had an
effect.

This makes the behavior more explicit and consistent, and
improves performance in the case where completion-styles is
configured such that we first do partial-completion with
completion-pcm-leading-wildcard=nil, and then fall back to
completion-pcm-leading-wildcard=t: explicitly inserting the *
avoids the need to fallback.

My site has been running with this patch for 4 months now and it's
worked fine.  In my experience it's very useful.

In GNU Emacs 30.1.90 (build 14, x86_64-pc-linux-gnu, X toolkit, cairo
 version 1.15.12, Xaw scroll bars) of 2025-08-11 built on
 igm-qws-u22796a
Repository revision: 2828c98aa29c66798fcb3fa67ccc3ed706024340
Repository branch: emacs-30
Windowing system distributor 'The X.Org Foundation', version 11.0.12011000
System Description: Rocky Linux 8.10 (Green Obsidian)

Configured using:
 'configure --with-x-toolkit=lucid --without-gpm --without-gconf
 --without-gsettings --without-selinux --without-imagemagick
 --with-modules --with-gif=no --with-cairo --with-rsvg
 --without-compress-install --with-tree-sitter
 --with-native-compilation=aot
 PKG_CONFIG_PATH=/usr/local/home/garnish/libtree-sitter/0.22.6-1/lib/pkgconfig/'

[0001-Insert-on-TAB-with-completion-pcm-leading-wildcard-t.patch (text/patch, attachment)]

Message #8 received at 79219 <at> debbugs.gnu.org (full text, mbox):

From: Dmitry Gutov <dmitry <at> gutov.dev>
To: Spencer Baugh <sbaugh <at> janestreet.com>, 79219 <at> debbugs.gnu.org
Cc: Stefan Monnier <monnier <at> iro.umontreal.ca>
Subject: Re: bug#79219: [PATCH] Insert * on TAB with
 completion-pcm-leading-wildcard=t
Date: Fri, 15 Aug 2025 04:34:25 +0300

Spencer, hi,

On 11/08/2025 19:44, Spencer Baugh wrote:
> Improve completion-pcm-leading-wildcard by actually inserting
> a * when try-completion runs.  completion-pcm--merge-try will
> automatically remove the wildcard anywhere that it has no
> matches, so the * is inserted only where it actually had an
> effect.

Would it be difficult to add a test for this?

I can see that you authored some of most recent ones in 
minibuffer-tests.el, but completion-pcm-leading-wildcard does not seem 
to be exercised in our suite yet.

Message #11 received at 79219 <at> debbugs.gnu.org (full text, mbox):

From: Spencer Baugh <sbaugh <at> janestreet.com>
To: Dmitry Gutov <dmitry <at> gutov.dev>
Cc: Stefan Monnier <monnier <at> iro.umontreal.ca>, 79219 <at> debbugs.gnu.org
Subject: Re: bug#79219: [PATCH] Insert * on TAB with
 completion-pcm-leading-wildcard=t
Date: Mon, 18 Aug 2025 15:01:32 -0400

[Message part 1 (text/plain, inline)]

Dmitry Gutov <dmitry <at> gutov.dev> writes:

> Spencer, hi,
>
> On 11/08/2025 19:44, Spencer Baugh wrote:
>> Improve completion-pcm-leading-wildcard by actually inserting
>> a * when try-completion runs.  completion-pcm--merge-try will
>> automatically remove the wildcard anywhere that it has no
>> matches, so the * is inserted only where it actually had an
>> effect.
>
> Would it be difficult to add a test for this?
>
> I can see that you authored some of most recent ones in
> minibuffer-tests.el, but completion-pcm-leading-wildcard does not seem
> to be exercised in our suite yet.

Certainly, easy enough to do a basic test, attached.

Though, after working on this some more, I think there's a better thing
to be doing here, which should always apply to pcm rather than be
conditioned on completion-pcm-leading-wildcard: insert a * when a non-*
wildcard is matching against an empty string.

For example, when doing (completion-pcm-try-completion "a/b" '("a/bbb"
"a-test/bbb") nil 3) return "a*/bbb" instead of the current "a/bbb".

This matches up with this comment in completion-pcm--merge-completions:

  ;; FIXME: in some cases, it may be necessary to turn an
  ;; `any' into a `star' because the surrounding context has
  ;; changed such that string->pattern wouldn't add an `any'
  ;; here any more.

I tried doing that change instead, and it works pretty nicely, but
before landing it I found I had to fix bug#79265 first.  So let's settle
that one first then come back to this bug.

[0001-add-some-leading-wildcard-tests.patch (text/x-patch, inline)]

From e3e8e2f90f33f03127e091388305251d26e5c233 Mon Sep 17 00:00:00 2001
From: Spencer Baugh <sbaugh <at> janestreet.com>
Date: Mon, 18 Aug 2025 14:51:33 -0400
Subject: [PATCH] add some leading-wildcard tests

---
 test/lisp/minibuffer-tests.el | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

diff --git a/test/lisp/minibuffer-tests.el b/test/lisp/minibuffer-tests.el
index 59b72899e22..2b22ee20b3f 100644
--- a/test/lisp/minibuffer-tests.el
+++ b/test/lisp/minibuffer-tests.el
@@ -332,6 +332,22 @@ completion-pcm-test-8
                   "" '("fooxbar" "fooybar") nil 0)
                  '("foobar" . 3))))
 
+(ert-deftest completion-pcm-test-9 ()
+  ;; By default completions must match the begining of the string.
+  (should (equal (completion-pcm-try-completion
+                  "oo" '("foo") nil 2)
+                 nil))
+  ;; But not with `completion-pcm-leading-wildcard'.
+  (should (equal (let ((completion-pcm-leading-wildcard t))
+                   (completion-pcm-try-completion
+                    "oo" '("foo") nil 2))
+                 '("foo" . 3)))
+  ;; The wildcard is actually inserted into the text.
+  (should (equal (let ((completion-pcm-leading-wildcard t))
+                   (completion-pcm-try-completion
+                    "oo" '("foo" "goo") nil 2))
+                 '("*oo" . 1))))
+
 (ert-deftest completion-substring-test-1 ()
   ;; One third of a match!
   (should (equal
-- 
2.43.7

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.