build problem LFS 6.5

Ken Moffat zarniwhoop73 at googlemail.com
Sat Nov 28 10:06:26 PST 2009


2009/11/28 stosss <stosss at gmail.com>:
> I am using the lfs 6.3 last version LiveCD. The computer does not have
> anything installed on it.
> I started over. I deleted the partitions on the hard drive and created
> new ones and started at the beginning of the book. I have been running
> the commands manually by copy and paste at the cli. Everything
> appeared to be fine until now.
>
> lfs 6.5 chapter 6.15 GCC-4.4.1
>
> here is the result of make -k check
> ../gcc-4.4.1/contrib/test_summary | grep -A7 Summ
>
>                === g++ Summary ===
>
> # of expected passes            384
> # of unexpected failures        16815
> # of unexpected successes       53
> # of expected failures          100
> # of unresolved testcases       372
> # of unsupported tests          366
> --
>                === gcc Summary ===
>
> # of expected passes            938
> # of unexpected failures        39876
> # of unexpected successes       30
> # of expected failures          147
> # of unresolved testcases       8338
> # of unsupported tests          1721
>
> I would think this is a big problem. Will someone please confirm this
> is a problem. If this is a problem how do I get passed it? This is
> probably where the trouble has been all along.
> --

 Definitely a problem (specifically, the vast numbers of unexpected
failures, and probably also the large number of unresolved testcases.

  I normally log everything, then when tests fail I can look at what it
says in the log to tell me which tests were marked as 'FAIL'..  I don't
understand enough about how gcc uses dejagnu to run its testsuite
to tell you which files to look in for error messages, but I guess that
somewhere in the testsuite directories within the build directory (so,
gcc-build/gcc/testsuite is the first place to look) are files containing
messages from when the tests were run.

 Find the messages, then look at them to see what problems are
reported many times (might only be one error causing this, or it
might be several).

ĸen
-- 
After tragedy, and farce, "OMG poneys!"



More information about the lfs-support mailing list