Alexander Eremin wrote: > Hi caiman-team, > please review my initial changes in DC bootroot_archive.py after > investigation of 6361 bug (more details here - I found also that for > sparc 'cpio' runs twice: > http://www.opensolaris.org/jive/thread.jspa?threadID=114297&tstart=0) > > Test: > # distro_const build -r br-arch ai_sparc_image.xml > ... > Before changes: > > # df -h /mnt > Filesystem size used avail capacity Mounted on > /opensolaris/dc/build_data/pkg_image/boot/boot_archive > 116M 79M 36M 69% /mnt > > After: > > # df -h /mnt > Filesystem size used avail capacity Mounted on > /opensolaris/dc/build_data/pkg_image/boot/boot_archive > 87M 79M 7.6M 92% /mnt > > webrev: > http://cr.opensolaris.org/~alhazred/6361/ > > Also tested with bootroot_size > 150000 > > Hi Alexandar,
Thanks for making the changes. You are right about the fact that the first CPIO that's done in the existing bootroot_archive.py script is a bug. Here are my comments the changes: - line 103: shouldn't the check be done for dst + "/" + cpio_file? - Please add more detail to the comment on lines 327-329 explaining why different overhead multipliers are used for sparc. - I think it would make the code easier to read if you move the code between lines 370-415 around so there's only 1 if-then-else block for if (sparc)... else... - You didn't mention it in your message above. Did you also built and tested x86 LiveCD images to make sure the changes didn't affect those? - I noticed that the formatting of the file doesn't correspond to the new PEP8 formating requirement that Clay sent email about last week. Can you fix that? The link announcing the requirement is at: http://www.opensolaris.org/jive/thread.jspa?threadID=114047&tstart=15 Thank you for your work! --Karen
