On Tue, 4 May 2004, csieh wrote: > Jan, > > On Tue, 4 May 2004, Jan Iven wrote: > > > >>>>> "Connie" == Connie Sieh <[log in to unmask]> writes: > > > > Connie> Scientific Linux (SL) Release Candidate 2 May 3, 2004 > > Connie> This is only a summary of the changes made since the release of April 26, > > Connie> 2004. > > > > Connie> Release is expected next week. > > > > Connie, thanks for all your work and the nice documentation. I feel I > > owe you some comments, even if I haven't managed to convert our own > > distribution to SL (yet), lack of time but not of interest. > > > > Thanks. > > > * you are shipping a number of zz_* RPMS in the "base" SL distribution > > - I understood that this prefix was actually used for Fermi-specific > > customizations in the past, and I wonder whether these should appear > > in the general release: > > > > All of the zz's that we included are of a "general" nature. Most fix bugs > that any audience will need. Also most are NOT installed by default. > Letting the end user decide. > > The zz prefix was used for all the "fixes" that we did. The original > reason for zz was that the installer on the "old" releases did not have > enough of a "requires,prereq" function and in alot of cases we needed > these to be installed after "something". The installer installed things > that is did not have any special ordering in alphabetical order. And > "zz" is very far down the alphabet. And if they all have the same naming > covention it is easy to find them. Some RPMS we fix by fixing the > "original" rpm. I do not like this way as much as they "fixit" rpms > mostly because "the original was changed" and many seem to have a problem > with that. You also have to keep up with the change when errata come out. > With the "fixit" rpms the change is very contained. > > > zz_desktop_tweeks-1.1-1 > > zz_inittab_change-1.0-4 > > zz_kernel_clean-1.0-2 > > zz_libg2c.a_change-3.2.3-1 > > zz_no_colorls-1.0-1 > > zz_sendmail_accept-1.0-2 > > > > * You also seem to have a number of RedHat-compiled RPMs in the > > distribution, which could be problematic because of the "update service > > license" for Red Hat Enterprise. I haven't checked whether all of these > > RPMs come directly from RHE3 (but at least some like xfig do, same timestamps): > > > > rpm -qp --nosignature --qf "%-30{NAME}\t%{VENDOR}\t%{BUILDHOST}\n" linux/scientific/30rolling/i386/SL/RPMS/*rpm | grep -i "redhat" | wc -l > > 300 > > These did NOT come from RHE3 they came from RedHat 9. Same exact rpm as They really came from Fermi Linux 90x. -Connie Sieh > shown by diff. If you look at the date of the rpm you will see that they > are from the RedHat 9 time frame. RedHat did NOT recompile them for > RHE3 they got them from RedHat 9. Since RedHat 9 is freely distributatble > then I can put them in and so I did. > > > > > I suggest at least to check whether all of these are "legal" to ship > > (coming e.f from Fedora), and perhaps replace them with e.g. the > > CERN-recompiled ones from http://linuxsoft/cern/cel3/i386/RedHat/RPMS/ > > I was hoping that you would provide us with the AMD-64 and maybe itanium > rpms as we are kind of short of development hardware in these > areas. > > > > > Best regards > > jan > > > > -Connie Sieh >