emacs-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: C-g while exiting the minibuffer


From: Stefan Monnier
Subject: Re: C-g while exiting the minibuffer
Date: Sun, 01 Dec 2013 17:13:12 -0500
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/24.3.50 (gnu/linux)

>> But the problem remains of what to do with a C-g interrupting a unwind
>> form: in the case of an Fset_window_configuration in an unwind form, the
>> intention is to make that "no matter what happens, we end up recovering
>> the original state", but a C-g at the wrong time will break this promise.
> I'm afraid that C-g at the wrong time might produce an inconsistent
> state which IMO seems more fatal than a non-original state.

An inconsistent state would also be a non-original state, so yes, an
inconsistent would also break the promise.

>> - checking cycles here gives us no guarantee since the caller can do
>> (set-window-prev-buffers w bufs)
>> (setcdr bufs bufs)
>> and you again end up with a cycle in your window-prev-buffers.
> Right.  So we'd need a safe implementation of delq and I wouldn't know
> how to do that reasonably via DOLIST.

Indeed, the FOREACH macro was designed for "read-only" traversals, just
like the DOLIST macro.  I don't think the hare/tortoise trick is
inapplicable, tho: it just requires more care (since the deletions of
Fdelq might give the tortoise shortcuts that risk making the tortoise
catch up with the hare).  See patch below.

> I'd rather use a simple routine to find out whether the original
> argument list of delq is infinite (if we think this could be a real
> problem).

I think the hare/tortoise is a "simple routine to find out if the list is
infinite".


        Stefan


=== modified file 'src/fns.c'
--- src/fns.c   2013-11-29 19:47:58 +0000
+++ src/fns.c   2013-12-01 22:02:46 +0000
@@ -1537,15 +1537,12 @@
 the value of a list `foo'.  */)
   (register Lisp_Object elt, Lisp_Object list)
 {
-  register Lisp_Object tail, prev;
-  register Lisp_Object tem;
+  Lisp_Object tail, tortoise, prev = Qnil;
+  bool skip;
 
-  tail = list;
-  prev = Qnil;
-  while (CONSP (tail))
+  FOR_EACH_TAIL (tail, list, tortoise, skip)
     {
-      CHECK_LIST_CONS (tail, list);
-      tem = XCAR (tail);
+      Lisp_Object tem = XCAR (tail);
       if (EQ (elt, tem))
        {
          if (NILP (prev))
@@ -1555,8 +1552,6 @@
        }
       else
        prev = tail;
-      tail = XCDR (tail);
-      QUIT;
     }
   return list;
 }

=== modified file 'src/lisp.h'
--- src/lisp.h  2013-11-30 09:25:31 +0000
+++ src/lisp.h  2013-12-01 22:05:02 +0000
@@ -4443,6 +4443,20 @@
       memory_full (SIZE_MAX);                                 \
   } while (0)
 
+/* Loop over all tails of a list, checking for cycles.
+   FIXME: Make tortoise and n internal declarations.
+   FIXME: Unroll the loop body so we don't need `n'.  */
+#define FOR_EACH_TAIL(hare, list, tortoise, n) \
+  for (tortoise = hare = (list), n = true;                     \
+       CONSP (hare);                                           \
+       (hare = XCDR (hare), n = !n,                            \
+       (n                                                      \
+        ? ((EQ (hare, tortoise)                                \
+            && (xsignal1 (Qcircular_list, (list)), 0)))        \
+        /* Move tortoise before the next iteration, in case */ \
+        /* the next iteration does an Fsetcdr.  */             \
+        : (tortoise = XCDR (tortoise), 0))))
+
 /* Do a `for' loop over alist values.  */
 
 #define FOR_EACH_ALIST_VALUE(head_var, list_var, value_var)            \




reply via email to

[Prev in Thread] Current Thread [Next in Thread]