Shawn:

> I think that SQLite could provide real value to developers by being
> integrated as a shared component. This vague resistance to its
> integration seems rather odd to me, given that MySQL and PostgreSQL
> already have been integrated. SQLite compared to both of those is
> positively tiny and is becoming widely-adopted among software
> developers (thanks Mac OS X, etc.).

The issue here has to do with what Sun means when they say "support".
Sun has a policies that we try to only ship things that we know our
users can depend upon (won't break on upgrade, for example).  Database
interfaces tend to have lots of consumers, which make them trickier
to manage.

There are similar "tricky to manage" interfaces like libexpat and
libxml.  Just getting all the internal groups who depend on these
interfaces all talking and working together is time consuming.

In fact, these are part of the reasons why Sun has a policy
that all database interfaces should be managed by a single group, the
database group.  Our group isn't allowed to just add database interfaces
willy-nilly without engaging them.  In our past experience of getting
BerkeleyDB support for JDS (needed by evolution), this is a slow
process.  The JDS team is starting to engage the database team on the
ideas of bringing SQLite into Solaris, but it's too early in the
discussion to really set any expectations.

I'm not saying that it is impossible to get SQLite into Solaris, but
there are some extra hurdle's to manage.

Any help that people in the community want to do to assist with ARC
and other processes to get SQLite into Solaris is, of course,
very much appreciated.

Brian



Reply via email to