Mailing-List: contact derby-user-help@db.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Derby Discussion" <derby-user@db.apache.org>
Received-SPF: pass (asf.osuosl.org: domain of tomdzk@gmail.com designates
 64.233.182.194 as permitted sender)
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
        s=beta; d=gmail.com;
        h=received:message-id:date:from:to:subject:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references;
        b=PC+87IOXzI6EGzTJVcge0556L253X+MOKX1xaxl9wil0OKu4+/8s4AxhOKC1QabuPKJ4bvRD4X2IO0TdfxerqAvJmg129GRAkNyJEcKCu6pa4UyawjbfIKtoOpHh1ERruY+BqZcCz5FyXp+ADloN+kk96scRRGtXHGsBCN15KCM=
Message-ID: <224f32340512080826v2ed2b780ie0904a0f9d0591e5@mail.gmail.com>
Date: Thu, 8 Dec 2005 17:26:47 +0100
From: Thomas Dudziak <tomdzk@gmail.com>
To: Derby Discussion <derby-user@db.apache.org>
Subject: Re: Why are classpath databases always read-only ?
In-Reply-To: <43984C40.3090507@debrunners.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
References: <224f32340512080309s75d9cb83xf8fbd1ca9d8ee172@mail.gmail.com>
	 <43984C40.3090507@debrunners.com>

On 12/8/05, Daniel John Debrunner <djd@debrunners.com> wrote:
> Thomas Dudziak wrote:
>
> > the manual says
> > (http://db.apache.org/derby/docs/10.1/devguide/rdevdvlp38881.html):
> >
> > "classpath   Databases are treated as read-only databases, and all
> > databaseNames must begin with at least a slash, because you specify
> > them "relative" to the classpath directory."
> >
> > Now I wonder why that is so ? Is there a particular reason to treat
> > these always as read-only ? If the classpath points to a local
> > directory (as opposed to, say, a jar), Derby could easily resolve this
> > to a real path and work with this.
>
> Only because no-one has asked for this feature and maybe at the time I
> didn't realise you could get to the real path from the resource. When
> this feature was added Cloudscape was running on JDK 1.1 where the
> getURL method was not available.

Would you consider adding it ? I think that is would be really useful
because it would decouple the app from where it runs (without having
to run Derby in server mode).

> > This would allow web applications to be created with an initial,
> > embedded Derby database without having to hard-code the absolute
> > database path (no assumption can be made as to where the current
> > working directory is when running in a servlet container, so the path
> > must be absolute).
>
> Also, you can package a read-only database in a jar file and archive
> that in the web-application archive, war or ear file. Then one can
> distribute read-only data with web-applications and access that data
> without ever unpacking the war/ear file and not knowing its location on
> disk.

Sure, but that is exactly the point: the database is read-only. If I
want to have an embedded database (e.g. co-located to the web app,
say, in WEB-INF/classes/derbydb) then I have to use a jdbc directory
url with an absoulte path which kind of defeats the concept of WARs.
Or I have to determine the JDBC url at runtime as a workaround, but
that IMHO is ugly.

regards,
Tom