My guess is: Walter is changing all the test builds to use system boost [now that boost.py can do that] - but some testboxes don't have system-boost installed yet.
[Per previous e-mail - he is still working on changes to the buildbot tests..] Satish On Fri, 16 Apr 2010, Barry Smith wrote: > > Well obviously something is wrong with that option and needs to be fixed. > What is it? > > Barry > > On Apr 16, 2010, at 2:35 PM, Satish Balay wrote: > > > On Fri, 16 Apr 2010, Barry Smith wrote: > > > > > > > > Satish, > > > > > > All the nightly builds like > > > http://www.geodynamics.org:8009/PETSc%20dev%20x86%20linux%20single%20nosched%20binbot/builds/879/step-shell_2/1 > > > failed last night due to boost issues. Could you please take a look at > > > them > > > and see if this is because of some change you made in the boost.py? > > > > I think walter at geodynamics - who sent in the patch for --with-boost=1 > > to autosearch system-install of boost is redoing these tests to use > > system-install boost. > > > > Previous builds [from last week] uses '--with-boost=1 > > --download-boost=1' but todays uses only '--with-boost=1' > > > > http://www.geodynamics.org:8009/PETSc%20dev%20x86%20linux%20single%20nosched%20binbot/builds/865/step-shell_2/1 > > > > satish > > >
