Thread: jar naming consitency

Hi!

I'd like to see consistency in the jars naming, for instance, in the
change between 310 and 309:

postgresql-8.0-310.jdbc2.jar
              ^
postgresql-8.0.309.jdbc2.jar
              ^

Having consistency makes easier others process that depends on the
jars as RPM packing.

Regards,
Manuel.

On Fri, 4 Feb 2005, Manuel Sugawara wrote:

> I'd like to see consistency in the jars naming, for instance, in the
> change between 310 and 309:
>
> postgresql-8.0-310.jdbc2.jar
> postgresql-8.0.309.jdbc2.jar
>               ^

This change was intentional to try and indicate that it is build 310 of
the 8.0 driver series, not something to do with the postgresql minor
version 310.  Additionally for build 310 the source package name was
changed postgresql-8.0.309.src.tar.gz to
postgresql-jdbc-8.0-310.src.tar.gz to try and differentiate it from the
server as well.  We think we've finally hit upon a reasonable convention
and do intend to stick with it.

Kris Jurka

Kris Jurka <books@ejurka.com> writes:

> We think we've finally hit upon a reasonable convention and do
> intend to stick with it.

Excellent!

Regards,
Manuel.

Am Freitag, 4. Februar 2005 19:22 schrieb Kris Jurka:
> This change was intentional to try and indicate that it is build 310 of
> the 8.0 driver series, not something to do with the postgresql minor
> version 310.

But that would then imply that both releases are the same source, just
different "builds", i.e., compiled on different occasions from the same
source.  That is not true.  If you release an altered source code, then it
should get a new version.

> Additionally for build 310 the source package name was
> changed postgresql-8.0.309.src.tar.gz to
> postgresql-jdbc-8.0-310.src.tar.gz to try and differentiate it from the
> server as well.  We think we've finally hit upon a reasonable convention
> and do intend to stick with it.

I think it would be preferrable if "jdbc" appeared somewhere in the name.
There may be other contexts in which a jar appears in relation with
postgresql.

--
Peter Eisentraut
http://developer.postgresql.org/~petere/

Peter Eisentraut wrote:
> Am Freitag, 4. Februar 2005 19:22 schrieb Kris Jurka:
>
>>This change was intentional to try and indicate that it is build 310 of
>>the 8.0 driver series, not something to do with the postgresql minor
>>version 310.
>
> But that would then imply that both releases are the same source, just
> different "builds", i.e., compiled on different occasions from the same
> source.

I don't see why "8.0-310" implies anything different about the source
used than "8.0.310", beyond making it clear that it is version 310 of
the 8.0-based driver. "8.0-311" and "8.0.311" are both obviously
different versions to the corresponding 310 versions. Maybe "build" is
bad terminology, but it's the terminology that the JDBC driver has been
using for some time now.

If you rebuild from the same source with different options. I'd expect
the package to tack on a package version along the lines of
postgresql-jdbc3-8.0-310-4, etc.

Do you have a better suggestion for making it clear which source is
being used? We could call it "postgresql-jdbc-stable-310" -- which is
really what it is -- but then you have the problem that it's not clear
which "stable" release you should use for a particular server version.

Personally, I'd rather just use subversion and release with a tree
revision number -- then it's very obvious exactly what source you built
from -- but I don't see that happening any time soon :)

-O