Hi Alok, Thank you for reviewing the changes. Please see my response inline.
On 03/24/10 12:09, Alok Aggarwal wrote: > Hi Karen, Tony, > > On Wed, 24 Mar 2010, Karen Tung wrote: > >> Please review my changes for the following bug: >> >> Bug 15148 - Required DC updates to work with Early Manifest Import >> project >> changes >> http://defect.opensolaris.org/bz/show_bug.cgi?id=15148 >> >> I would like to integrate my changes by COB of Wed 3/24, if possible. >> Therefore, please try to send your comments by 2pm, pacific time. >> >> webrev: >> http://cr.opensolaris.org/~ktung/15148/ >> >> Testing: >> >> The changes are provided by the SMF team, and they have built >> LiveCD and AI images and tested them successfully. >> >> I took their changes and did the following tests so far: >> - I built Slim CD, x86 AI, and sparc AI images using a build 135 repo. >> This verified >> the changes can be used to build pre-EMI images successfully. >> - I also built Slim CD, x86 AI, and sparc AI images using a repo >> containing >> all >> the EMI code. This verified the changes can be used to build post-EMI >> images successfully. >> - I verified that all the pre-EMI images can be booted correctly, and >> they >> install >> correctly. >> - I verified the post-EMI slim cd image is able to boot and install >> correctly. >> >> I will be performing the following tests Wed morning, and will not >> integrate >> until I confirm that they all pass. >> - Build x86 Text Installer and sparc Text installer images using a >> pre-EMI >> change repo, >> and make sure the images can boot and install correctly. >> - Build x86 Text Installer and sparc Text installer images using a >> post-EMI >> repo, >> and make sure the images can boot and install correctly. >> - Verify the post-EMI sparc and x86 images are booting and installing >> correctly. > > boot_archive_configure: line 260: A comment is needed > here that explains why this check exists. Yep. A comment will be added. > > boot_archive_configure: line 261: This is checking for > the case where you're doing a build with pre-EMI repo > but still have profiles in /etc/svc. Will this ever be > the case? With pre-EMI repo, it is true that profiles won't be in /etc/svc. The list of profiles we are applying comes from the list in the DC manifests. All the DC manifests are updated to include both paths of profiles in the pre-emi and post-emi "era" to help users with easy transition. So, for pre-emi images, we want to filter out the post-emi profiles. For post-emi images, we want to filter out the pre-emi profiles. That's basically what this block of code is doing. > > mkrepo: line 52, 126: X never seems to be set > > mkrepo: line 138: Should $basedir be appended to the log > location? Tony has responded to these comments. Thanks, --Karen > > Alok
