GNU bug report logs - #10038
[PATCH] tests: avoid rare fp failure in du/move-dir-while-traversing

Previous Next

Full log

View this message in rfc822 format

From: help-debbugs <at> gnu.org (GNU bug Tracking System)
To: Nix <nix <at> esperi.org.uk>
Subject: bug#10038: closed (Re: bug#10038: [PATCH] tests: avoid rare fp
 failure in du/move-dir-while-traversing)
Date: Sun, 13 Nov 2011 19:07:01 +0000

[Message part 1 (text/plain, inline)]

Your bug report

#10038: [PATCH] tests: avoid rare fp failure in du/move-dir-while-traversing

which was filed against the coreutils package, has been closed.

The explanation is attached below, along with your original report.
If you require more details, please reply to 10038 <at> debbugs.gnu.org.

-- 
10038: http://debbugs.gnu.org/cgi/bugreport.cgi?bug=10038
GNU Bug Tracking System
Contact help-debbugs <at> gnu.org with problems

[Message part 2 (message/rfc822, inline)]

From: Jim Meyering <jim <at> meyering.net>
To: Nix <nix <at> esperi.org.uk>
Cc: 10038-done <at> debbugs.gnu.org
Subject: Re: bug#10038: [PATCH] tests: avoid rare fp failure in
	du/move-dir-while-traversing
Date: Sun, 13 Nov 2011 20:06:06 +0100

Nix wrote:
> du/move-dir-while-traversing: boost the mkdir iteration count yet again.
> ---
>  tests/du/move-dir-while-traversing |    4 ++--
>  1 files changed, 2 insertions(+), 2 deletions(-)
>
> I just had an FP in du/move-dir-while-traversing, due to the traversal
> terminating too soon. My autobuilder is running on a 2.2GHz Nehalem,
> with the source tree located on a loopback ext4 filesystem on a tmpfs,
> so particularly rapid test execution is understandable.
>
> Patch obvious (I ran 200 trials of this just to avoid having to change
> the comment!):
>
> diff --git a/tests/du/move-dir-while-traversing b/tests/du/move-dir-while-traversing
> index 322a9a5..d0969fe 100755
> --- a/tests/du/move-dir-while-traversing
> +++ b/tests/du/move-dir-while-traversing
> @@ -66,8 +66,8 @@ long=d/e/f/g/h/i/j/k/l/m/n/o/p/q/r/s/t/u/v/w/x/y/z
>  # Five iterations was not enough in 2 of 7 "make -j20 check" runs on a
>  # 6/12-core system.  However, using "10", I saw no failure in 20 trials.
>  # Using 10 iterations was not enough, either.
> -# Using 30, I saw no failure in 200 trials.
> -for i in $(seq 30); do
> +# Using 50, I saw no failure in 200 trials.
> +for i in $(seq 50); do
>    mkdir -p $t/3/a/b/c/$i/$long || framework_failure_
>  done

Thanks.
I've applied that and pushed it along with a tiny admin patch.
(also tweaked your log message to conform)

From 47052ddbfa04b16d649ae0ecbaacfd3ce1f03c53 Mon Sep 17 00:00:00 2001
From: Jim Meyering <meyering <at> redhat.com>
Date: Sun, 13 Nov 2011 20:00:08 +0100
Subject: [PATCH 1/2] maint: adjust git hook to allow "copy: ..." on line 1 of
 commit message

* scripts/git-hooks/commit-msg: Add "copy" to the list of valid
command-name-like summary line prefixes.
---
 scripts/git-hooks/commit-msg |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/scripts/git-hooks/commit-msg b/scripts/git-hooks/commit-msg
index 96d5947..76b829d 100755
--- a/scripts/git-hooks/commit-msg
+++ b/scripts/git-hooks/commit-msg
@@ -42,7 +42,7 @@ check_msg()
       test|timeout|touch|tr|true|truncate|tsort|tty|uname|unexpand) ;;
       uniq|unlink|uptime|users|vdir|wc|who|whoami|yes) ;;
       # other tags
-      gnulib|tests|maint|doc|build|scripts) ;;
+      copy|gnulib|tests|maint|doc|build|scripts) ;;
       *) echo "invalid first word of summary line: $w"; fail=1;;
     esac
   done
--
1.7.8.rc2


From b8b6dd27a61dce58d6537dfbee65c7428f4ae3cd Mon Sep 17 00:00:00 2001
From: Nix <nix <at> esperi.org.uk>
Date: Sun, 13 Nov 2011 18:15:10 +0000
Subject: [PATCH 2/2] tests: avoid rare fp failure in
 du/move-dir-while-traversing

* tests/du/move-dir-while-traversing: Boost the mkdir iteration count
yet again.
---
 tests/du/move-dir-while-traversing |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/tests/du/move-dir-while-traversing b/tests/du/move-dir-while-traversing
index 322a9a5..d0969fe 100755
--- a/tests/du/move-dir-while-traversing
+++ b/tests/du/move-dir-while-traversing
@@ -66,8 +66,8 @@ long=d/e/f/g/h/i/j/k/l/m/n/o/p/q/r/s/t/u/v/w/x/y/z
 # Five iterations was not enough in 2 of 7 "make -j20 check" runs on a
 # 6/12-core system.  However, using "10", I saw no failure in 20 trials.
 # Using 10 iterations was not enough, either.
-# Using 30, I saw no failure in 200 trials.
-for i in $(seq 30); do
+# Using 50, I saw no failure in 200 trials.
+for i in $(seq 50); do
   mkdir -p $t/3/a/b/c/$i/$long || framework_failure_
 done

--
1.7.8.rc2

[Message part 3 (message/rfc822, inline)]

From: Nix <nix <at> esperi.org.uk>
To: bug-coreutils <at> gnu.org
Subject: [PATCH] tests: avoid rare fp failure in du/move-dir-while-traversing
Date: Sun, 13 Nov 2011 18:15:10 +0000

du/move-dir-while-traversing: boost the mkdir iteration count yet again.
---
 tests/du/move-dir-while-traversing |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

I just had an FP in du/move-dir-while-traversing, due to the traversal
terminating too soon. My autobuilder is running on a 2.2GHz Nehalem,
with the source tree located on a loopback ext4 filesystem on a tmpfs,
so particularly rapid test execution is understandable.

Patch obvious (I ran 200 trials of this just to avoid having to change
the comment!):

diff --git a/tests/du/move-dir-while-traversing b/tests/du/move-dir-while-traversing
index 322a9a5..d0969fe 100755
--- a/tests/du/move-dir-while-traversing
+++ b/tests/du/move-dir-while-traversing
@@ -66,8 +66,8 @@ long=d/e/f/g/h/i/j/k/l/m/n/o/p/q/r/s/t/u/v/w/x/y/z
 # Five iterations was not enough in 2 of 7 "make -j20 check" runs on a
 # 6/12-core system.  However, using "10", I saw no failure in 20 trials.
 # Using 10 iterations was not enough, either.
-# Using 30, I saw no failure in 200 trials.
-for i in $(seq 30); do
+# Using 50, I saw no failure in 200 trials.
+for i in $(seq 50); do
   mkdir -p $t/3/a/b/c/$i/$long || framework_failure_
 done
 
-- 
1.7.7.2.144.g3624f

Previous Next

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