GNU bug report logs - #11927
shred.c i686-specific warning from gcc-4.7 on fedora 17

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 11927 in the body.
You can then email your comments to 11927 AT debbugs.gnu.org in the normal way.

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

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

From: Jim Meyering <jim <at> meyering.net>
To: bug-coreutils <at> gnu.org
Subject: shred.c i686-specific warning from gcc-4.7 on fedora 17
Date: Thu, 12 Jul 2012 23:27:15 +0200

[mostly for the record, and so I don't forget.
 I'm not ready to push this just yet. ]

Building shred with fedora 17 and its stock gcc 4.7 on i686,
I see this warning/error:

    CC     shred.o
  shred.c: In function 'dopass':
  shred.c:501:14: error: assuming signed overflow does not occur when simplifying conditional to constant [-Werror=strict-overflow]
  shred.c:504:18: error: assuming signed overflow does not occur when simplifying conditional to constant [-Werror=strict-overflow]
  cc1: all warnings being treated as errors
  make[3]: *** [shred.o] Error 1
  make[3]: Leaving directory `/f12/home/meyering/w/co/cu/src'
  make[2]: *** [all] Error 2
  make[2]: Leaving directory `/f12/home/meyering/w/co/cu/src'
  make[1]: *** [all-recursive] Error 1
  make[1]: Leaving directory `/f12/home/meyering/w/co/cu'
  make: *** [all] Error 2

I didn't really understand the cause but ended up
writing the following patch to avoid the problem
while making the code slightly more readable:


diff --git a/src/shred.c b/src/shred.c
index 7a28260..322bcf1 100644
--- a/src/shred.c
+++ b/src/shred.c
@@ -485,10 +485,11 @@ dopass (int fd, char const *qname, off_t *sizep, int type,

       offset += soff;

+      bool done = offset == size;
+
       /* Time to print progress? */
-      if (n
-          && ((offset == size && *previous_human_offset)
-              || thresh <= (now = time (NULL))))
+      if (n && ((done && *previous_human_offset)
+                || thresh <= (now = time (NULL))))
         {
           char offset_buf[LONGEST_HUMAN_READABLE + 1];
           char size_buf[LONGEST_HUMAN_READABLE + 1];
@@ -498,8 +499,7 @@ dopass (int fd, char const *qname, off_t *sizep, int type,
             = human_readable (offset, offset_buf,
                               human_floor | human_progress_opts, 1, 1);

-          if (offset == size
-              || !STREQ (previous_human_offset, human_offset))
+          if (done || !STREQ (previous_human_offset, human_offset))
             {
               if (size < 0)
                 error (0, 0, _("%s: pass %lu/%lu (%s)...%s"),
@@ -516,7 +516,7 @@ dopass (int fd, char const *qname, off_t *sizep, int type,
                     = human_readable (size, size_buf,
                                       human_ceiling | human_progress_opts,
                                       1, 1);
-                  if (offset == size)
+                  if (done)
                     human_offset = human_size;
                   error (0, 0, _("%s: pass %lu/%lu (%s)...%s/%s %d%%"),
                          qname, k, n, pass_string, human_offset, human_size,

Message #8 received at 11927 <at> debbugs.gnu.org (full text, mbox):

From: Paul Eggert <eggert <at> cs.ucla.edu>
To: Jim Meyering <jim <at> meyering.net>
Cc: 11927 <at> debbugs.gnu.org
Subject: Re: bug#11927: shred.c i686-specific warning from gcc-4.7 on fedora 17
Date: Thu, 12 Jul 2012 16:29:02 -0700

On 07/12/2012 02:27 PM, Jim Meyering wrote:
> I didn't really understand the cause

That one is one of my least favorite warnings
because sometimes it warns about real bugs but
often, as in your case, the warning is either
bogus or so hard to figure out that it might as
well be bogus.  I see that you and I and others have
been filing GCC bug reports in this area:

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51309
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48267
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52904
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34515
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52560
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49234

In this particular case the shred.c rewrite seems harmless.

Message #13 received at 11927-done <at> debbugs.gnu.org (full text, mbox):

From: Jim Meyering <jim <at> meyering.net>
To: Paul Eggert <eggert <at> cs.ucla.edu>
Cc: 11927-done <at> debbugs.gnu.org
Subject: Re: bug#11927: shred.c i686-specific warning from gcc-4.7 on fedora 17
Date: Sun, 15 Jul 2012 11:36:16 +0200

Paul Eggert wrote:
> On 07/12/2012 02:27 PM, Jim Meyering wrote:
>> I didn't really understand the cause
>
> That one is one of my least favorite warnings
> because sometimes it warns about real bugs but
> often, as in your case, the warning is either
> bogus or so hard to figure out that it might as
> well be bogus.  I see that you and I and others have
> been filing GCC bug reports in this area:
>
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51309
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48267
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52904
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34515
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52560
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49234
>
> In this particular case the shred.c rewrite seems harmless.

Thanks for the review.
I've just hit it again, so have pushed this:

From b255b9a68a0dcec113a193daa602bdcbda2658db Mon Sep 17 00:00:00 2001
From: Jim Meyering <meyering <at> redhat.com>
Date: Sun, 15 Jul 2012 11:31:18 +0200
Subject: [PATCH] build: shred.c: avoid i686-specific gcc -Wstrict-overflow
 warning

* src/shred.c: Avoid gcc -Wstrict-overflow warning.
Addresses http://bugs.gnu.org/11927
---
 src/shred.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/src/shred.c b/src/shred.c
index 7a28260..322bcf1 100644
--- a/src/shred.c
+++ b/src/shred.c
@@ -485,10 +485,11 @@ dopass (int fd, char const *qname, off_t *sizep, int type,

       offset += soff;

+      bool done = offset == size;
+
       /* Time to print progress? */
-      if (n
-          && ((offset == size && *previous_human_offset)
-              || thresh <= (now = time (NULL))))
+      if (n && ((done && *previous_human_offset)
+                || thresh <= (now = time (NULL))))
         {
           char offset_buf[LONGEST_HUMAN_READABLE + 1];
           char size_buf[LONGEST_HUMAN_READABLE + 1];
@@ -498,8 +499,7 @@ dopass (int fd, char const *qname, off_t *sizep, int type,
             = human_readable (offset, offset_buf,
                               human_floor | human_progress_opts, 1, 1);

-          if (offset == size
-              || !STREQ (previous_human_offset, human_offset))
+          if (done || !STREQ (previous_human_offset, human_offset))
             {
               if (size < 0)
                 error (0, 0, _("%s: pass %lu/%lu (%s)...%s"),
@@ -516,7 +516,7 @@ dopass (int fd, char const *qname, off_t *sizep, int type,
                     = human_readable (size, size_buf,
                                       human_ceiling | human_progress_opts,
                                       1, 1);
-                  if (offset == size)
+                  if (done)
                     human_offset = human_size;
                   error (0, 0, _("%s: pass %lu/%lu (%s)...%s/%s %d%%"),
                          qname, k, n, pass_string, human_offset, human_size,
--
1.7.11.2.194.g7bdb748

Previous Next

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