Re: Emacs port to gcc -fcheck-pointer-bounds

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

[-- Attachment #1: Type: text/plain, Size: 707 bytes --]

On 12/13/2017 08:20 AM, Eli Zaretskii wrote:
> using a (fake) pointer is only marginally safer than using an
> integer, isn't it?

The fake pointer catches (at compile-time) common faults like the one 
the attached patch fixes, where an int was passed where a Lisp_Object 
was expected. These are the most important faults that 
--enable-check-lisp-object-type catches.

We could say that the fake pointer is only marginally safer, in the 
sense that --enable-check-lisp-object-type is only marginally safer than 
--disable-check-lisp-object-type. However, this marginal safety is 
useful; and once you have the fake pointer, 
--enable-check-lisp-object-type doesn't buy much extra safety that is 
useful.


[-- Attachment #2: 0001-Fix-type-typo-on-Solaris.patch --]
[-- Type: text/x-patch, Size: 794 bytes --]

From c662e2d4fc3678d1ea6eda16541b82bc88f0890b Mon Sep 17 00:00:00 2001
From: Paul Eggert <eggert@cs.ucla.edu>
Date: Thu, 1 Dec 2016 22:45:07 -0800
Subject: [PATCH] Fix type typo on Solaris

* src/sysdep.c (system_process_attributes) [SOLARIS2 && HAVE_PROCFS]:
Fix type mismatch, caught by --enable-check-lisp-object-type.
---
 src/sysdep.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/sysdep.c b/src/sysdep.c
index 892e97626b..257634292b 100644
--- a/src/sysdep.c
+++ b/src/sysdep.c
@@ -3371,7 +3371,7 @@ system_process_attributes (Lisp_Object pid)
     nread = 0;
   else
     {
-      record_unwind_protect (close_file_unwind, fd);
+      record_unwind_protect_int (close_file_unwind, fd);
       nread = emacs_read (fd, &pinfo, sizeof pinfo);
     }
 
-- 
2.14.3