On 07/12/2015 12:34 AM, Ken Moffat wrote:
On Sat, Jul 11, 2015 at 03:39:01PM +0200, Frans de Boer wrote:
On 07/11/2015 01:49 PM, Frans de Boer wrote:
On 07/11/2015 10:23 AM, Frans de Boer wrote:
On 07/11/2015 12:20 AM, Bruce Dubbs wrote:
Frans de Boer wrote:
Ok, I did exactly the same under OpenSuSE 13.2 - modified where
required
- and now all tests work like a charm.
What were you using before?
-- Bruce
Sorry, I was not exactly precise in my wording above.
I was using the same machine and made the build of coreutils without
using the LFS environment. Made it as a replacement of the distribution
coreutils-8.23. So, no LFS environment was used.
Frans
Btw, executing '/tools/bin/seq 1 inf | /tools/bin/head -n2' - both
within and outside the LFS environment - also does not terminate seq.
The problem already exist within the Tool Chain (TC).
I will check again every command needed to build the TC. Maybe there is
something wrong there. That - of course - might take some time.
Frans.
Maybe it is just my system.
running 'seq 1 inf | head -n2' yields different results depending upon the
environment. It's expected output is
1
2
and returns to prompt.
Within X, KConsole as normal user: Ok
Within X, Kconsole as root : fail
System console as normal user : fail
System console as root : fail
The above was with coreutils-8.23!!
I've been testing 8.23 on a couple of my recent systems (I don't
have any with 8.24) and your test works as expected. On one system,
I used urxvt and the tty, on the other I only used konsole.
This is not an asy problem to google for, but I did find one 10-year
old post re cygwin which offers one possible reason.
https://www.cygwin.com/ml/cygwin/2005-02/msg00248.html
In that case, the suggestion is that somewhere within your
environment you might be ignoring SIGPIPE. Examples in that post,
and the quickest thing to try would be the first line of the
examples:
trap - PIPE # restore default handling to SIGPIPE
From that, looking to see if anything happened to SIGPIPE _recently_
in the kernel, I came upon
http://www.rpmfind.net/linux/RPM/opensuse/13.2/x86_64/systemd-logger-210-25.5.4.x86_64.html
The relevant line in the ChangeLog is NOT recent (Feb 2012, Systemd
41) but anyway:
+ SIGPIPE is ignored by default. This can be disabled with
IgnoreSIGPIPE=no in unit files.
So, are any / all of these systems you tested using systemd ?
ĸen
I am no fan of systemd - brings to many problems without being able to
debug it properly - but it's what is provided by default by my
distribution. Still, I do have some questions:
- If systemd would be the problem, all my machines use systemd (OpenSuSE
13.2) without exception, why is there no issue on the 32-bit machines
(my servers)?
- Also, if 64-bit systemd is the culprit, in which unit file should the
IgnoreSIGPIPE=no being inserted?
- And - last but not least - why does it work correctly in an
X-environment under regular user control?
The good news is that the problem does not seem to appear - as user root
using the system console - in my virtual and finished (64-bit) LFS
environment. The virtual machine has no systemd!
Maybe incompatibility problem number n+1 with systemd?
Option: I inserted "sed -i '/tests\/misc\/seq-precision.sh/ d'
Makefile.in" in the build script to be able to continue the build
process without error.
Regards,
Frans.
--
http://lists.linuxfromscratch.org/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
Do not top post on this list.
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?
http://en.wikipedia.org/wiki/Posting_style
