Prashant Srinivasan wrote: > Chris Zhu wrote: >> >> I reordered the patch sequence of ruby-sparc-perf.patch and >> ruby-dtrace-1.8.6.patch, as ruby-dtrace-1.8.6.patch is only for x86, >> so the source code will be different in x86 and sparc after patching >> ruby-dtrace-1.8.6.patch, although gpatch can find the right place to >> patch the following patches, I prefer to putting it as the last >> patch which will make the code clear in the new ruby patch version. >> > > Chris, > Come to think of it, the order of application doesn't matter since the > perf patch is exclusively for SPARC , and the DTrace patch is > exclusively for x64. We should be fine as long as the architecture > specific patches are applied after the other patches so that the other > patches can be arch-neutral. Actually the perf patch is patched on both x86 and sparc, although it just affect the performance of sparc. There's no problem of the sequence of patches of ruby-sparc-perf.patch and ruby-dtrace-1.8.6.patch, although they patched eval.c both, gpatch will find the right place by source code matching, but I think put ruby-dtrace-1.8.6.patch in the end will make it clearer.
> > On the testing front - it would be nice to have rorbench tested on > p287. We caught a SEGV in Ruby the last time I tested this app on Ruby. > What rorbench ? > thanks > -ps > > > > >
