Debian Bug report logs - #533356
tar: broken behaviour with input/output pipe and compression

Toggle useless messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: Petr Salinger <Petr.Salinger@seznam.cz>

To: submit@bugs.debian.org

Subject: tar: broken behaviour with input/output pipe and compression

Date: Tue, 16 Jun 2009 21:58:16 +0200 (CEST)

Package: tar
Severity: important
Version: 1.22-1
Tags: patch
User: glibc-bsd-devel@lists.alioth.debian.org
Usertags: kfreebsd

Hi,

the current version of tar breaks build of gcj-4.4 on GNU/kFreeBSD.

https://buildd.debian.org/fetch.cgi?&pkg=gcj-4.4&ver=4.4.0-6j1&arch=kfreebsd-amd64&stamp=1245000137&file=log:

	uudecode -o - java-classes.tgz.uue | tar -C src -xvz
	libjava/classpath/tools/classes/gnu/classpath/tools/gjdoc/Main.class
	sbuild received SIGPIPE -- shutting down

It looks like related to #525437, #532570, #525818.

The problem appears when a grandchild tar is used.
In the "child_pid", the child_pid is zero, the kill() signals
every process in the process group of the calling process
instead of themselves.

Please apply patch bellow, it would also be nice
if you can inform upstream about this issue.

Thanks in advance

                        Petr

--- src/system.c~       2009-03-29 05:23:17.000000000 +0200
+++ src/system.c        2009-06-16 21:11:20.000000000 +0200
@@ -435,7 +435,7 @@

   if (WIFSIGNALED (wait_status))
     {
-      kill (child_pid, WTERMSIG (wait_status));
+      kill (getpid(), WTERMSIG (wait_status));
       exit_status = TAREXIT_FAILURE;
     }
   else if (WEXITSTATUS (wait_status) != 0)
@@ -573,7 +573,7 @@

   if (WIFSIGNALED (wait_status))
     {
-      kill (child_pid, WTERMSIG (wait_status));
+      kill (getpid(), WTERMSIG (wait_status));
       exit_status = TAREXIT_FAILURE;
     }
   else if (WEXITSTATUS (wait_status) != 0)

Message #10 received at 533356-forwarded@bugs.debian.org (full text, mbox, reply):

From: Bdale Garbee <bdale@gag.com>

To: bug-tar@gnu.org

Cc: 533356-forwarded@bugs.debian.org

Subject: [Fwd: Bug#533356: tar: broken behaviour with input/output pipe and compression]

Date: Sat, 20 Jun 2009 09:49:28 -0600

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

Please see the attached proposed patch from Petr Salinger.  Looks
reasonable to me... your thoughts?

Bdale

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

From: Petr Salinger <Petr.Salinger@seznam.cz>

To: submit@bugs.debian.org

Subject: Bug#533356: tar: broken behaviour with input/output pipe and compression

Date: Tue, 16 Jun 2009 21:58:16 +0200 (CEST)

Package: tar
Severity: important
Version: 1.22-1
Tags: patch
User: glibc-bsd-devel@lists.alioth.debian.org
Usertags: kfreebsd

Hi,

the current version of tar breaks build of gcj-4.4 on GNU/kFreeBSD.

https://buildd.debian.org/fetch.cgi?&pkg=gcj-4.4&ver=4.4.0-6j1&arch=kfreebsd-amd64&stamp=1245000137&file=log:

	uudecode -o - java-classes.tgz.uue | tar -C src -xvz
	libjava/classpath/tools/classes/gnu/classpath/tools/gjdoc/Main.class
	sbuild received SIGPIPE -- shutting down

It looks like related to #525437, #532570, #525818.

The problem appears when a grandchild tar is used.
In the "child_pid", the child_pid is zero, the kill() signals
every process in the process group of the calling process
instead of themselves.

Please apply patch bellow, it would also be nice
if you can inform upstream about this issue.

Thanks in advance

                        Petr

--- src/system.c~       2009-03-29 05:23:17.000000000 +0200
+++ src/system.c        2009-06-16 21:11:20.000000000 +0200
@@ -435,7 +435,7 @@

   if (WIFSIGNALED (wait_status))
     {
-      kill (child_pid, WTERMSIG (wait_status));
+      kill (getpid(), WTERMSIG (wait_status));
       exit_status = TAREXIT_FAILURE;
     }
   else if (WEXITSTATUS (wait_status) != 0)
@@ -573,7 +573,7 @@

   if (WIFSIGNALED (wait_status))
     {
-      kill (child_pid, WTERMSIG (wait_status));
+      kill (getpid(), WTERMSIG (wait_status));
       exit_status = TAREXIT_FAILURE;
     }
   else if (WEXITSTATUS (wait_status) != 0)

Message #11 received at 533356-forwarded@bugs.debian.org (full text, mbox, reply):

From: Sergey Poznyakoff <gray@gnu.org.ua>

To: "Bdale Garbee" <bdale@gag.com>

Cc: <bug-tar@gnu.org>, <533356-forwarded@bugs.debian.org>

Subject: Re: [Bug-tar] [Fwd: Bug#533356: tar: broken behaviour with input/output pipe and compression]

Date: Sat, 20 Jun 2009 20:42:49 +0300

Bdale Garbee <bdale@gag.com> ha escrit:

> Please see the attached proposed patch from Petr Salinger.  Looks
> reasonable to me... your thoughts?

Please, take a look at the repository, where it is fixed already.

Regards,
Sergey

Message #16 received at 533356@bugs.debian.org (full text, mbox, reply):

From: Petr Salinger <Petr.Salinger@seznam.cz>

To: 533356@bugs.debian.org

Subject: [ping] Bug#533356 fixed upstream - tar: broken behaviour with input/output pipe and compression

Date: Wed, 5 Aug 2009 10:01:01 +0200 (CEST)

Hi,

please could you integrate either proposed patch or the enhanced patch applied upstream

http://git.savannah.gnu.org/cgit/tar.git/commit/?id=af30244849448c775b3e447a84ba8b0249b4b697

It is really not GNU/kFreeBSD specific problem, moreover appearance is timing dependent.

Petr

Message #21 received at 533356-done@bugs.debian.org (full text, mbox, reply):

From: Petr Salinger <Petr.Salinger@seznam.cz>

To: 533356-done@bugs.debian.org

Subject: Bug#533356 fixed upstream - tar: broken behaviour with input/output pipe and compression

Date: Mon, 8 Nov 2010 08:43:25 +0100 (CET)

Version: 1.23-3

Have been for some time fixed in upstream,
may be even 1.23-1 have been fine.

Petr

Send a report that this bug log contains spam.