Jeff Trawick wrote: > Amanda Waite wrote: >> Hi all, >> >> Attached is a draft Arc case for spawn-fcgi in SFW. As a reminder, >> spawn-fcgi is the component that was part of Lighttpd but is now >> distributed as a separate source bundle. > >> /usr/lighttpd/1.4 >> /bin >> /man >> >> Note: These directories are delivered by the Lighttpd >> packages. We continue to use the Lighttpd directory >> structure so as >> this is where spawn-fcgi is located in the Lighttpd ARC >> case >> (LSARC/2008/191). > > > What about delivering spawn-fcgi solely to /usr/spawn-fcgi and, for as > long as SUNWlighttpd14u depends on SUNWspawn-fcgi, create symlinks > within the SUNWlighttpd14u delivery from > /usr/lighttpd/1.4/bin/spawn-fcgi and /usr/lighttpd/1.4/man/whatever to > the appropriate SUNWspawn-fcgi deliverable.
I was just thinking along those lines, at the very least to show in the spawn-fcgi arc case that the /usr/lighttpd/1.4 paths aren't the exported interfaces. Your suggestion is better but it does mean that SUNWlighttpd will include links to files that it doesn't deliver which might be an issue. I'm also inclined to put the executable in /usr/bin, it will often be scripted but will likely just as often be run on the command line (not basing that on any evidence though, maybe Alvaro has a view on this). > > Then the spawn-fcgi deliverable is as expected for the long term, and > the package that must have the temporary dependency anyway > (SUNWlighttpd14u) is the only one handling any legacy issues. It is a tidy solution I have to admit, I can then go through the process of obsolescing spawn-fcgi in SUNWlighttpd*. Amanda > > _______________________________________________ > > > webstack-discuss mailing list > webstack-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/webstack-discuss
