[lfs-support] error in vim inn chapter 6.61

Ken Moffat zarniwhoop at ntlworld.com
Sat Apr 21 19:23:26 PDT 2012


On Sun, Apr 22, 2012 at 06:27:48AM +0530, ankit vishwakarma wrote:
> the following error occur while executing make test. please provide help.
> 
> 
> Press ENTER or type command to continue
> 
 No idea if that is odd - I route the output from the tests to a
file.  Seems unusual, maybe someone else who actually runs the tests
without dumping the results to a file would like to comment ?

 But either way, not something to worry about.

> # Sleep a moment to avoid that the xterm title is messed up
> ../vim -u unix.vim -U NONE --noplugin -s dotest.in test51.in
> 3,6c3
> < NewGroup       xxx term=bold
> <                    cterm=italic
> <                    ctermfg=2
> <                    ctermbg=3
> ---
> > NewGroup       xxx term=bold cterm=italic ctermfg=2 ctermbg=3
> 10,11c7
> < Group3         xxx term=underline
> <                    cterm=bold
> ---
> > Group3         xxx term=underline cterm=bold
> 
 This is probably the failure - the results have different
formatting from the expected results.

[...]
> 
> Test results:
> test51 FAILED
> TEST FAILURE

 For years, I saw one or more failures in the vim tests, when I
could be bothered to run them, but oddly enough for LFS-7.0 and
LFS-7.1 it has always managed to pass the tests on those builds
where I run them (x86_64, using the previous release of LFS).

 But does it matter ?  Perhaps (in an infinite universe, *anything*
*might* make a difference), but unlikely.  The book got very worked
up about running tests in LFS-6.0 or whenever 'pure LFS' was
introduced (understandably, after the different / unrepeatable
results some people saw in earlier versions), but the reality is
that test suites are rarely decisive and mostly concerned with
testing 'corner cases'.

 My own experience is that from time to time I see more failures in
toolchain tests, then at other times the failures no longer happen.
The only real way to make sense of that is to build LFS-svn
frequently, run the tests each time, and then try to work out what
changed (e.g. the host kernel).  Mostly, that is NOT worth the effort
[ those who differ are welcome to spend their time doing this, I
prefer to find out if the resulting LFS system can build a system
which I find useful ]

 If you get one failure in a non-toolchain package, it's very
unlikely to make a difference.

 If you get more than 10 failures in a *toolchain* package, it
*might* indicate a problem.

 If you get 100 or more failures in the same package, you have a
problem.  For intermediate numbers of failures, your guess is as
good as mine ;-)  So, asking about failures is fine, but expecting
100%  lack of test failures is unrealistic.

 Yes, I know Bruce's posted results often show perfection - I just
wish I had his machines or his environment.  I often see a few test
failures - some are obviously because of what I do differently (no
static libs), others seem to occur for no obvious reason.

 Summary: don't panic, and don't expect perfection (this is
software, no software is bug-free and that includes the testsuites) -
particularly if you are running in a virtual machine, or using an
older kernel, or a kernel with a lot of distro patches.

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



More information about the lfs-support mailing list