On Fri, 14 Mar 2008, Joel Sherrill wrote: > If you all think this is the right approach, ack me and I will > get the powerpc cases done first and submit a patch for review.
Janis will need to confirm the right approach. The aim is: * If no -mcpu option in the multilib flags, run the test (with the -mcpu option from dg-options getting added). * If there is a -mcpu option in the multilib flags, and it's the same as that in dg-options, run the test. * If there is a -mcpu option in the multilib flags, and it's different from that in dg-options, skip the test because the multilib flags would override dg-options. -- Joseph S. Myers [EMAIL PROTECTED]
