[postgis-devel] Vote on Merging postgis and raster installs andwhen

[Paul Ramsey]

On Tue, Jan 3, 2012 at 10:55 AM, Sandro Santilli <strk at keybit.net> wrote:
> On Tue, Jan 03, 2012 at 11:44:12AM -0800, Paul Ramsey wrote:
>> On Tue, Jan 3, 2012 at 10:42 AM, Sandro Santilli <strk at keybit.net> wrote:
>> >
>> > Speaking of which, I guess we want separate "script" versions
>> > in rtpostgis.sql and topology.sql, so you can find out if you
>> > have an old "core" with a new "raster" or "topology" or the
>> > other way around.
>>
>> Right, which seems an excellent argument for *not* being modular...
>> modularity means adding to our complexity stew with internal
>> versioning and dependencies.
>
> We don't really need different versions here, just the version
> from main release would do. Pretty much in the same way we already
> do for DLL vs. SQL -- the postgis_full_version() function compares
> the version in DLL with the version in SQL and tells you if they
> mismatch, we would be doing something similar with the versions
> in the different SQL files (3 of them).
>
> What's your take on configurability in the "not being modular" case ?
> Should it be possible at all to build PostGIS w/out having GDAL and
> latest and greatest GEOS ?

I think GDAL should be mandatory, but older GEOS should be OK (we've
gone to the trouble of stubbing those functions for a reason,
presumably. A configure WARNING makes sense, so at least ignorance
won't be an excuse.

P.

>
> --strk;
>
>  ,------o-.
>  |   __/  |    Thank you for PostGIS-2.0 Topology !
>  |  / 2.0 |    http://www.pledgebank.com/postgistopology
>  `-o------'
>
> _______________________________________________
> postgis-devel mailing list
> postgis-devel at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-devel