Re: new-flex-completion-style

[João Távora]

Stefan Monnier <address@hidden> writes:

> How 'bout we change the buffer name completion table so it specifies its
> own cycle-sort-function which uses buffer-list, then.  It seems like it
> would be a clear win (the buffer-list order is a much better heuristic
> than the length).

Hi again,

Here are two patches: the first does exactly what you propose above, the
second goes a little bit further, but I still think it's the right
thing...

Lightly tested stuff...

João

>From f170e728c1c1ac426bb73c63cc54ba2ae30afd1a Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jo=C3=A3o=20T=C3=A1vora?= <address@hidden>
Date: Mon, 18 Feb 2019 20:32:38 +0000
Subject: [PATCH 1/2] switch-to-buffer's completion table uses its own sorting

* src/minibuf.c (Finternal_complete_buffer): Add
Qcycle_sort_function to completion table's metadata.
(syms_of_minibuf): New symbol Qcycle_sort_function.
---
 src/minibuf.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/src/minibuf.c b/src/minibuf.c
index 321fda1ba8..b23e24c4bd 100644
--- a/src/minibuf.c
+++ b/src/minibuf.c
@@ -1801,7 +1801,9 @@ If FLAG is nil, invoke `try-completion'; if it is t, 
invoke
   else if (EQ (flag, Qlambda))
     return Ftest_completion (string, Vbuffer_alist, predicate);
   else if (EQ (flag, Qmetadata))
-    return list2 (Qmetadata, Fcons (Qcategory, Qbuffer));
+    return list3 (Qmetadata,
+                  Fcons (Qcategory, Qbuffer),
+                  Fcons (Qcycle_sort_function, Qidentity));
   else
     return Qnil;
 }
@@ -1922,6 +1924,8 @@ syms_of_minibuf (void)
   DEFSYM (Qactivate_input_method, "activate-input-method");
   DEFSYM (Qcase_fold_search, "case-fold-search");
   DEFSYM (Qmetadata, "metadata");
+  DEFSYM (Qcycle_sort_function, "cycle-sort-function");
+
   /* A frame parameter.  */
   DEFSYM (Qminibuffer_exit, "minibuffer-exit");
 
-- 
2.20.0

>From f8ce65482d86a136f1320dac89e71662256c2168 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jo=C3=A3o=20T=C3=A1vora?= <address@hidden>
Date: Mon, 18 Feb 2019 20:41:09 +0000
Subject: [PATCH 2/2] cycle-sort-function prevails in
 completion-all-sorted-completions

* lisp/minibuffer.el (completion-all-sorted-completions): Don't
re-sort if completion table has cycle-sort-function.
---
 lisp/minibuffer.el | 22 +++++++++++++---------
 1 file changed, 13 insertions(+), 9 deletions(-)

diff --git a/lisp/minibuffer.el b/lisp/minibuffer.el
index 7413be42eb..cc87ffaced 100644
--- a/lisp/minibuffer.el
+++ b/lisp/minibuffer.el
@@ -1246,20 +1246,23 @@ completion-all-sorted-completions
           (setq all (delete-dups all))
           (setq last (last all))
 
-          (setq all (if sort-fun (funcall sort-fun all)
-                      ;; Prefer shorter completions, by default.
-                      (sort all (lambda (c1 c2) (< (length c1) (length c2))))))
-          ;; Prefer recently used completions and put the default, if
-          ;; it exists, on top.
-          (when (minibufferp)
-            (let ((hist (symbol-value minibuffer-history-variable)))
-              (setq all (sort all
+          (cond
+           (sort-fun
+            (setq all (funcall sort-fun all)))
+           (t
+            ;; Prefer shorter completions, by default.
+            (setq all (sort all (lambda (c1 c2) (< (length c1) (length c2)))))
+            (when (minibufferp)
+              ;; Prefer recently used completions and put the default, if
+              ;; it exists, on top.
+              (let ((hist (symbol-value minibuffer-history-variable)))
+                (setq all
+                      (sort all
                             (lambda (c1 c2)
                               (cond ((equal c1 minibuffer-default) t)
                                     ((equal c2 minibuffer-default) nil)
                                     (t (> (length (member c1 hist))
-                                          (length (member c2 hist))))))))))
+                                          (length (member c2 hist))))))))))))
           ;; Cache the result.  This is not just for speed, but also so that
           ;; repeated calls to minibuffer-force-complete can cycle through
           ;; all possibilities.
-- 
2.20.0