[postgis-devel] Re-organising the PostGIS codebase - to liblwgeom or not liblwgeom?

[Mark Cave-Ayland]

Paul Ramsey wrote:
> On Wed, May 14, 2008 at 8:41 AM, Mark Cave-Ayland
> <mark.cave-ayland at siriusit.co.uk> wrote:
>> - configure options have been renamed to better reflect their values
>>
>>        --with-pgsql -> --with-pgconfig
>>        --with-geos  -> --with-geosconfig
>>        --with-proj  -> --with-projdir
>
> Aesthetically, I think I prefer --with-pg-config, --with-proj-dir

That was one of my original thoughts, however the PostgreSQL config and
the GEOS config are subtly different - PostgreSQL uses pg_config whereas
GEOS uses geos-config. After getting annoyed at not being able to remember
which one used the underscore and which one used the hyphen, I removed
them all, so now the problem doesn't exist anymore :)

>> The other thing to point out is that my current SVN copy doesn't have any
>> JTS support included. Is this still being actively maintained, or has it
>> been deprecated due to reliability issues when compiling JTS with gcj?
>
> Not maintained but not deprecated... hard call. It got some love in
> recent memory when MarkL last tried to make it work (and got quite
> close...)

Okay, well the #defines are still in place, but the code is not referenced
in the Makefile. It won't be a problem to add it back in at a later date.

>> I haven't reworked this with the new autoconf code yet, since there seem
>> to be two systems in place: jade and xsltproc. Can we pick just one and
>> use that? I'm leaning towards xsltproc since that seems to work well on
>> Win32, so please let me know if that won't work for you.
>
> xsltproc. the jade stuff is historical and hasn't been used for many moons.

Great.

>> BTW if I commit with a broken documentation build Makefile, will this
>> break the auto-generated docs on postgis.refractions.net?
>
> Probably you'll break the autobuild in any event, so go ahead, we'll
> pick up the pieces after.

No problem.

>> - Remove non-ST_ prefix functions from the code base
>
> Not yet, grasshopper. I think our next move is to raise WARNINGs on
> all of them.

Hmmm. I think it may be possible to come up with a macro that will raise a
warning if the function name doesn't begin ST_. I'll look into this after
the initial commit.

>> - Re-work the index code based on newer versions of PostgreSQL (the newer
>> code seems a lot more compact and less error-prone)
>
> Good news. Perhaps we can do higher dimensionality at the same time :)

I'd have to revisit this and see how much work would be involved with the
new changes. Sponsorship from any interested parties always welcome ;)

> In the words of Prince, let's go crazy.

Excellent. I'll do the deed and see how we go...


ATB,

Mark.

-- 
Mark Cave-Ayland
Sirius Corporation - The Open Source Experts
http://www.siriusit.co.uk
T: +44 870 608 0063