[postgis-devel] Adding size to LWGEOM

[Paul Ramsey]

Re: your note to self, look at how the _add functions work, and how
you would change them to not copy anymore and more efficient. I think
this is where having size around would be helpful.

P.

On Fri, Dec 26, 2008 at 2:13 AM, Mark Cave-Ayland
<mark.cave-ayland at siriusit.co.uk> wrote:
> Paul Ramsey wrote:
>
>> As I fart around with building multi-geometries, I am getting
>> increasingly annoyed that I can't just efficiently add new geometries
>> to an existing object.  If LWGEOM had a size_t indicating how big the
>> allocated dynamic segment was, I could build a bunch of simple
>> accessors that filled out the allocated area, and doubled it when it
>> get too small.  Basically like the dynptarray idea, but in the main
>> objects. Since the LWGEOMs aren't serialized directly, there's no real
>> size penalty to doing this, i don't think (time to read the
>> serialize/deserialize code). and the on-disk format wouldn't change,
>> so it would be backward compatible.
>>
>> This is an idle thought, it's certainly possible to code around this
>> (that's what I'm doing) but it just seems like I'm writing lots of
>> code that isn't strictly necessary.
>>
>> p.
>
> I'm not sure I quite follow what you mean here - the LWGEOM structure for a
> multi/collection type contains a pointer to the sub-geometry and not the
> geometry data itself so this shouldn't be relevant?
>
> If you want to add a sub-geometry in a certain position within a
> multi/collection type then look at the *_add functions in lwm*.c within the
> liblwgeom directory.
>
> NOTE TO SELF: These functions look as if they are slow and leak memory as
> they deep-copy each individual geometry into a new array. In fact, all you
> need to do is shuffle the pointer array to include the new entry.
>
>
> ATB,
>
> Mark.
>
> --
> Mark Cave-Ayland
> Sirius Corporation - The Open Source Experts
> http://www.siriusit.co.uk
> T: +44 870 608 0063
> _______________________________________________
> postgis-devel mailing list
> postgis-devel at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-devel
>