On Fri, Jul 28, 2006 at 12:24:20PM -0700, Jonathan Davis wrote:
> Right before the "make CC="gcc ${BUILD32}" CXX="g++ ${BUILD32}"
> libbz2.a" command, the book says "Recompile the package using a
> non-shared library and test it." I took that to mean that I needed to
> run "make check" after that make command, even though the book never
> explicitly says to run the command "make check" in this section.
>
> I guess that I misunderstood the directions. There are no errors from
> any of the make commands that are explicitly listed. However, even if
> "make check" wasn't meant to be run, it seems odd that it would
> fail....
Now that we only make libbz2.a, I'm not surprised that 'make check'
needs to build something else to run the tests.
>
>
> After looking at the make commands again, I just tried a modified "make
> check":
>
> make CC="gcc ${BUILD32}" CXX="g++ ${BUILD32} check
>
> and this version of "make check" works just fine. My guess is that
> the make file is being stupid without the extra parameters about
> 32-bit stuff in spite of the fact that the 32-bit version was just
> compiled. I don't know if you want to add this check to the book, but
> it appears to work while a simple "make check" does not.
>
At the moment, I'm not sure how much the current book is actually
testing in multilib bzip2. Jim's results for Sparc show 6 runs of
cmp for each size of bzip2. My earlier multilib results from ppc64
are not accessible at the moment (I'm not going to mount the
filesystem they are on until my patch to e2fsprogs for the temp
system has survived the full build, and because I only uploaded my
'check' logs to my LFS space I can't look there) but I had the
impression that no tests were run. Maybe when I look at the current
build I will see that everything is fine, and I might even remember
that these logs also need to be uploaded.
Like me, Joe didn't uploaded the regular logs so I'm still unsure
if the tests do run for 'make libbz2.a'. Certainly, if 'make check'
is needed as a separate command it will need the BUILD32 stuff, but
then we might be better reverting to 'make' with BUILD32. I'll take
a look at the Makefile if I need to after reviewing results from my
current build.
Bzip2 is like many packages that don't use autoconf, and does
things its own sweet way. Fortunately, the more troublesome ones
are in blfs territory so we don't need to speak about them very
often.
Ken
--
das eine Mal als Tragödie, das andere Mal als Farce
_______________________________________________
Clfs-support mailing list
[email protected]
http://lists.cross-lfs.org/cgi-bin/mailman/listinfo/clfs-support
