GCC - chapter 6.14 help with the logs

Ken Moffat ken at linuxfromscratch.org
Mon Feb 9 07:10:12 PST 2009


On Mon, Feb 09, 2009 at 09:41:23AM +0100, Luca Di Stefano wrote:
> 
> Hello!
> 
> I'm going again through the LFS book, and I wanted to know if this output, when checking the GCC can be acceptable.
> 
[...]
>         === gcc Summary ===
> 
> 
> 
> # of expected passes        49398
> 
> # of unexpected failures    4
> 
> # of unexpected successes    2
> 
[..]
>         === libmudflap tests ===
> 
> 
> 
> 
> 
> Running target unix
> 
> FAIL: libmudflap.c++/pass41-frag.cxx execution test
> 
> --
> 
>         === libmudflap Summary ===
> 
> 
> 
> # of expected passes        1844
> 
> # of unexpected failures    5
> 
[...]
>         === libstdc++ Summary ===
> 
> 
> 
> # of expected passes        4942
> 
> # of unexpected failures    1
> 


> What else do you need to say if I can continue or not? The results seem
> close to the ones described in the logs, except for some unexpected
> failures...
> 
 "except for some unexpected failures" - the unexpected failures are
the reason for running the tests!

 Is this the 6.4 book ?  If it is, I wouldn't worry about
libmudflap, I always get several failures with gcc-4.3 versions.

 The important thing is the paragraph which reads
| A few unexpected failures cannot always be avoided. The GCC
| developers are usually aware of these issues, but have not resolved
| them yet. In particular, the libmudflap tests are known be
| particularly problematic as a result of a bug in GCC
| (http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20003). Unless the test
| results are vastly different from those at the above URL, it is safe
| to continue. 

 If you have the FAIL messages, it might be helpful to record them
here so that later builders can say "oh, yes".  Noting your host
kernel version and distro, and processor type, might also be
relevant (although it's usually only glibc which gives different
results on different results on different cpus).

 In my case, the only failures I saw on my most recent build of 6.4
were in libmudflap (5 in pass41-frag.cxx), on a single processor
athlon64, using a 2.6.28 kernel for the build.

 You have an extra 4 failures in gcc, and 1 in libstdc++.  My
rule-of-thumb is that unexpected failures in single figures are
nothing to worry about.  From experience, thousands of failures
indicate something seriously wrong.

 FWIW, I think you have nothing to worry about.

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



More information about the lfs-support mailing list