On Mon, Nov 16, 2009 at 03:19:41PM -0800, Brock Pytlik wrote: > [email protected] wrote: > >On Mon, Nov 16, 2009 at 02:03:19PM -0800, Brock Pytlik wrote: > >>Since this didn't grab anyone last week, I thought I'd try again. > >>http://cr.opensolaris.org/~bpytlik/ips-11805-v4/ > >> > >>J - If you could take a look at the signature bit in the manifest > >>class, that's the only big change. > > > >Offline, I went over these changes with Brock. The code generally > >looked fine, but I had a performance question about the code in > >manifest.set_content(). I'm ok with this change, as long as we can > >verify that there's no serious regression with set_content(). I believe > >Shawn also asked to see a benchmark back in October > >([email protected]). > > > >-j > I remember doing this before and sending the results to the IRC > channel, but here they are again. Using the old method, doing 1000 > set contents on a package the size of SUNWipkg took between 14.3 and > 14.8 seconds, using this code, it took between 14.7 and 15.2 > seconds. I believe this is good enough to not be a serious > regression and will put back momentarily.
That is correct. For this project, only silly regressions are permitted. -j _______________________________________________ pkg-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
