Channing Lovely wrote:

Please review
http://cr.opensolaris.org/~lovely/microroot/

which addresses

835 DC microroot creation should be cleaner
http://defect.opensolaris.org/bz/show_bug.cgi?id=835

This modified webrev implements the following changes

 
1. The size of the /etc exclusions in build_dist.lib:populate_root_archive() 
are documented.
The size of /etc/gtk-2.0 is negligible, but I am excluding it to be consistent 
with the 
Indiana PV2 image.

2. Lines 63 and 64 in build_dist.bash are switched
BOOT_FILENAME=x86.microroot
BOOT_ARCHIVE=$DIST_PROTO/boot/$BOOT_FILENAME

3. The redirection of stderr to /dev/null on line 220 of build_dist.lib has 
been removed

4. livemedia_processing() in build_dist.lib has been cleaned up to remove 
the MICROROOT variable and the reference to SMF initialization

5. The icon theme cache cleanup has been manually re-merged

6. The Belenix artifact has been removed from usr_microroot_files

7. The gzip line in build_dist.lib has been changed to use the -9 option

In addition, per a discussion with Dave Miner, the gnome initialization has 
been moved before microroot creation

I also added a PATH variable definition to build_dist.bash, though I have not 
stripped the absolute paths off any of the commands

What is not implemented....

I have not changed build_dist.lib:populate_root_archive() to use a loop over 
the names of directories. My exclusions are hardcoded. As this is a throwaway 
prototype, I am not sure that level of elegance is going to add anything here. 
Such a construct will be in the design for the "real" distro constructor that 
will be written in python.

I have not encapsulated 148-169 of build_dist.bash into a function as that 
doesn't change how this works and doesn't fix a defect.

I have an iso built with the above modifications implemented
/net/pentagram.central/download/export/test1.iso

I have gotten this image to boot and install using VirtualBox on my laptop. 

I look forward to any and all comments. 

Channing


Reply via email to