db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Daniel John Debrunner <...@debrunners.com>
Subject Re: JSR169 support status?
Date Tue, 30 Nov 2004 01:27:16 GMT
Hash: SHA1

Jeremy Boynes wrote:

> Why the preference for a single jar(with lots of switching code inside)
> rather than different jars for different platform levels?

Ease of use for the software developer, assembler etc. That's the
approach Cloudscape took over the years and the customers and users
liked it. Never have to be concerned about which Derby jar file version
to use or ship.

> I would have thought for the ME platform the desire would be for as
> small a footprint as possible so eliminating all the non-169 code would
> be a benefit. At the other end, the 1.5 build could be optimized for
> that platform with features not in others (e.g. StringBuilder, atomics,
> ...) - plus no need to use reflection for toPlainString() :-)

Well, we do optimize the code for the platform within that single jar,
I've stated earlier that, for example, with J2SE 5.0 Derby should use
the MathContext class for arithmetic with BigDecimal.

As for footprint, the bulk of the code is in the SQL language
implementation, so while we could get a smaller jar for J2ME, currently
it wouldn't be much smaller. So not much benefit.

Anyway, the first thing is to get the code written to allow us to run on

Version: GnuPG v1.2.5 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org


View raw message