Ok, it works. (Was Re: powerpc NPTL port)

Rob Landley rob at landley.net
Wed May 12 06:14:52 UTC 2010


On Tuesday 11 May 2010 13:20:38 Khem Raj wrote:
> On (11/05/10 13:07), Rob Landley wrote:
> > On Tuesday 11 May 2010 10:30:08 Khem Raj wrote:
> > > On (11/05/10 01:29), Rob Landley wrote:
> > > > Had to switch on several more things in your config to make the
> > > > native toolchain and busybox happy, but it built and ran.  And the
> > > > native toolchain in the system image built and ran a threaded test
> > > > program (/usr/src/thread- hello2.c), which seemed to work ok.
> > >
> > > cool. now you have the native build environment. Can you give a shot at
> > > running uclibc tests too. make -k check UCLIBC_ONLY=1
> >
> > I can, but you can too.  It's just an emulator. :)
>
> yes I will. However since you have the setup its less head wind for you
> thats why I asked.

Sure, just pointing out I couldn't get to it until now. :)

Ok, I tarred up the patched source tarball (out of build/packages/alt-uClibc), 
wget it into the virtual powerpc system, extracted it, cd into the thing, and 
ran "make -k check UCLIBC_ONLY=1".

The attached "out.txt" shows precisely how unhappy it got.  Line 4 shows it 
dying with a fatal error (there's no .config), but it went on for another 1000 
lines of random flailing before Fibber McGee's closet was finally empty.  (No, I 
didn't supply -j anything to the build, this is single-threaded build 
weirdness.)

More fundamentally, this test apparently doesn't test an installed uClibc, it 
wants to build one and then test the one it built.  Is there a "test the one 
installed on the host"...  An, there's a README in the test directory...  
TEST_INSTALLED_UCLIBC=1 and...

Page after page after page of errors, mostly because it can't find standard 
headers.  (But I can build software from the command line no problem.  With 
uClibc-0.9.31 this setup built dropbear, strace, zlib, ncurses, and python 
natively.)

The error message log is already too big to attach to a message, and still 
going...

Ok, under what circumstances does this test suite _not_ go insane?  Can I test 
something known working on i686-0.9.31 maybe?

Rob
-- 
Latency is more important than throughput. It's that simple. - Linus Torvalds
-------------- next part --------------
A non-text attachment was scrubbed...
Name: out.txt.zip
Type: application/zip
Size: 4023 bytes
Desc: not available
URL: <http://lists.busybox.net/pipermail/uclibc/attachments/20100512/b3c3ec62/attachment-0001.zip>


More information about the uClibc mailing list