2007/7/6, Glynn Foster <[EMAIL PROTECTED]>:
Hey, Again, I've been dabbling in what we currently have. As a first step, I've just generated a whole bunch of package lists, with compressed sizes for Solaris Express build 66. I did this by basically going into each SUNW* directory under Solaris_11/Product, and generating a size vs package name vs package description triplet. http://www.gnome.org/~gman/solaris-66-cd-1-contents.txt http://www.gnome.org/~gman/solaris-66-cd-2-contents.txt http://www.gnome.org/~gman/solaris-66-cd-3-contents.txt http://www.gnome.org/~gman/solaris-66-cd-4-contents.txt http://www.gnome.org/~gman/solaris-66-cd-5-contents.txt http://www.gnome.org/~gman/solaris-66-cd-6-contents.txt and summarizing across the entire set - http://www.gnome.org/~gman/solaris-66-cd-total-contents.txt Here's the non-SUNW packages - http://www.gnome.org/~gman/solaris-66-cd-non-sun-contents.txt There's a whole heap of analysis that can be done on this - o Removing the non-distributable packages o Removing the localization packages o Removing the developer packages o Removing the potentially redundant multiple version packages
Well, I can give the first tip on redundancy. One of the APOC package SUNWapoc, has a duplicate version of the libdb.so.1, we have one in /usr/lib and another in /usr/lib/apoc/ (berkeley db) also, the java bindings libraries and the .jar file are duplicated (/usr/share/db.jar and /usr/share/lib/apoc/db.jar). Currently I'm working on splitting the apoc source code and using autotools to build it and reference to the different classpath elements and libraries installed. However, this will be available from the open source version, the "server side" of APOC won't be released as free software before of the end of the year (probably even later due to the lazyness of the process). It also includes the (former Netscape) Mozilla LDAP JDK (in SUNWapbas), which is not packaged so there is no duplicity, I'm planing to make a proper package of this library for Solaris so other packages can reuse it. Hope this information is useful. Btw, are we going to collect this on some kind of central/collaborative documentation thing? wiki? I think that as long as we collect this information, it would be useful to have it in an overview of what we know so we can get the whole picture. o Removing the stuff we actually don't care about
o Estimating the packages that would be useful for a single CD I plan to start looking at answering a few of these questions. If anyone else is keen, please feel free to investigate and do some write-ups. At the very least, there's *definitely* some low hanging fruit for Solaris Express alone. Glynn _______________________________________________ indiana-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/indiana-discuss
-- Un saludo, Alberto Ruiz
_______________________________________________ indiana-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/indiana-discuss
