Yes it was actually me, who reimplement part of patchadd - dependancy check as 
pdo for Solaris10. Now it is not too much to speed up in patching - dependancy 
check if smallest fraction of patching and most time it spends in pkgadd.

I guess we may gain some performance this way but it will not really affect 
total patching time too much.

Thanks, Roland it is really nice to know that even in script old patching has 
enormous potential to improve 20-60 times. I remember that when db was 
introduced, major conclusion and argument for that introduction was "there is 
no way to improve performance of current technology, but db gives us 50% 
improvement in any case". I know that we have great potential to improve in 
packaging also - no need to revolutionary changes like going to different 
software delivery model, but if we just program it right - it will be totally 
different story, like this happened to patch dependancy check in Solaris 10. 
Contents file + compression + moving functionality from scripts may speed up 
current solution in many times, without changing customer experience, 
delivering procedure etc.

vassun
 
 
This message posted from opensolaris.org

Reply via email to