bug#59506: 29.0.50; (time-since nil) return format change

[Paul Eggert <eggert@cs.ucla.edu>]

[-- Attachment #1: Type: text/plain, Size: 534 bytes --]

On 2022-11-23 05:09, Eli Zaretskii wrote:
> Paul, any comments?

lsp-mode was buggy even with traditional Emacs timestamps like (0 0 0 0) 
because it mishandled millisecond counts. With new Emacs timestamps like 
(0 . 1000000000) lsp-mode signaled an error, exposing the bug. The 
problem has been fixed in bleeding-edge lsp-mode here:

https://github.com/emacs-lsp/lsp-mode/commit/e5a6274fac076f0d74a6ca892467be66a90a9649

and to forestall similar problems elsewhere I installed the attached 
patch into GNU Emacs master on Savannah.

[-- Attachment #2: 0001-Work-around-lsp-mode-compatibility-bug.patch --]
[-- Type: text/x-patch, Size: 903 bytes --]

From 55f6f1c82a35f0589d3dbdd1f32fae7ea9a758d8 Mon Sep 17 00:00:00 2001
From: Paul Eggert <eggert@cs.ucla.edu>
Date: Thu, 24 Nov 2022 12:56:22 -0800
Subject: [PATCH] Work around lsp-mode compatibility bug

* src/timefns.c (Ftime_subtract): Respect current-time-list when
args are eq (Bug#59506).
---
 src/timefns.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/timefns.c b/src/timefns.c
index eed2edf1cc..9beec1ce38 100644
--- a/src/timefns.c
+++ b/src/timefns.c
@@ -1194,7 +1194,7 @@ DEFUN ("time-subtract", Ftime_subtract, Stime_subtract, 2, 2, 0,
      quicker while we're at it.  This means (time-subtract X X) does
      not signal an error if X is not a valid time value, but that's OK.  */
   if (BASE_EQ (a, b))
-    return timespec_to_lisp ((struct timespec) {0});
+    return make_lisp_time ((struct timespec) {0});
 
   return time_arith (a, b, true);
 }
-- 
2.37.2