Help !! Having some problem during the compilation of Coreutils and Perl on LFS-6.4

Ken Moffat ken at linuxfromscratch.org
Sun Mar 22 15:35:14 PDT 2009


On Mon, Mar 23, 2009 at 01:42:50AM +0800, Jason Munkee wrote:
> 
> I'm useing the LFS-6.3 LiveCD as host and following LFS-6.4.
> 
> I did find some discussion about the Coreutils patch problem
> but still don't know how to fix it!
> 
 Do you have a url for this ?  I don't know what you are talking
about, and your error is new to me.  I've built 6.4 (or most of it)
three times now and the only problem I've seen is tests hanging.

[snip the coreutils messages because I have no opinion on the error,
and anyway you seem to have continued beyond this to get:]

> ......and another error from the test of Perl

> ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
> ext/Sys/Syslog/t/syslog.......................................
> #   Failed test 'setlogsock() should return true: '''
> #   at ../ext/Sys/Syslog/t/syslog.t line 204.
> # Looks like you failed 1 test of 199.
> FAILED at test 176
> 
> 
> ext/IO/t/io_utf8..............................................ok
> ext/IO/t/io_xs................................................ok
> ext/IPC/SysV/t/ipcsysv........................................# cannot proceed: semget() error:
> # Looks like you planned 17 tests but ran 16.
> FAILED--expected 17 tests, saw 16
> ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
> 
> I'm not sure if these problems are critical or not?
> so I didn't dare to continue the build!
 So, the perl failures are enough to stop you, but not the coreutils
failures.  I've seen _many_ testsuite failures over the year,
particularly building on other architectures.  My general view is
that a few test failures are nothing to worry about.

 The last message is not one I've seen, but then perl is so old that
its build system predates most other software.  Maybe google knows
about semget() errors - I'm not interested enough to look!

 What I *will* say is that building on an *old* host (specifically,
an old kernel - the 6.3 CD) is likely to add more test failures.
At times in the past we've had people reporting all sorts of odd
failures on specific distros and kernels, particularly in glibc.

 The general advice for LFS is "stop at the first error".  My
personal opinion is that all testsuites are somewhat unreliable -
certainly, if you get a _lot_ of errors in the tests it usually means
something is wrong (either a missing package or missing symlinks).

 ISTR perl has _many_ tests, probably more than 1500.  A couple of
failures within them is annoying but it doesn't seem like a big deal.

 If it really worries you, take a look at whatever output was
created from the first perl failure (with luck it will have written
something in a file), and try google for the semget() error.

ĸen
-- 
das eine Mal als Tragödie, das andere Mal als Farce



More information about the lfs-support mailing list