On Tue, 2010-04-06 at 07:04 -0600, Gary Thomas wrote: > On 04/03/2010 07:44 PM, Gary Thomas wrote: > > On 04/03/2010 05:11 PM, Tom Rini wrote: > >> On Sat, 2010-04-03 at 11:22 -0600, Gary Thomas wrote: > >>> On 04/03/2010 10:40 AM, Tom Rini wrote: > >>>> On Sat, 2010-04-03 at 09:28 -0600, Gary Thomas wrote: > >>>> > >>>>> That said, this seems to be a [Make] parallelism problem, not clear > >>>>> why/how. > >>>>> If I just rerun 'bitbake firefox' which restarts the compile step, > >>>>> it will > >>>>> run to completion. I do normally use -j4, but disabling this did > >>>>> not seem > >>>>> to make any difference. > >>>>> > >>>>> Note: the previous version I built (3.5.2) did not exhibit this > >>>>> behaviour. > >>>> > >>>> Well that's depressing since we just added the best we can tell correct > >>>> magic to build firefox with make -jN. If you put PARALLEL_MAKE = "" in > >>>> the .bb, does it complete this time? > >>>> > >>> > >>> No change, I'm afraid. > >> > >> Interesting. From log.do_compile, was make -jN taken out? I kinda fear > >> not... > >> > > > > I see only -j1 in the logs > > > > I'm not getting anywhere with this. True, it builds, but only > after a second attempt (the first always fails as reported). > > Perhaps you'd like to try? Just build firefox (3.6.2 is what > I'm using) for a PowerPC target (I use MACHINE=efika, > DISTRO="angstrom-2008.1") > You'll need the attached patches to work around some of the va_list bugs.
So, funny thing. Locally on an older snapshot of OE, 3.6.0 built, for e500v2 machines, unpatched. I'll try more combinations (doing MACHINE=calamari on current now) as time and CPU cycles permit. > > Thanks > -- Tom Rini <[email protected]> Mentor Graphics Corporation _______________________________________________ Openembedded-devel mailing list [email protected] http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
