[postgis-devel] about serialization i PostGIS 3.0

[Nicklas Avén]

On 6 October 2018 06:00:11 GMT+02:00, "Darafei "Komяpa" Praliaskouski" <me at komzpa.net> wrote:
>>
>> My idea is about solving the second part, and I don't know how it
>will
>> affect the overall compression.
>> ...
>>
>> Any thoughts?
>>
>>
>I see a problem with this approach: what is the actual operation that
>is
>going to get benefit from it?
>
>If we're making the format coordinate-number-addressable, then we need
>to
>utilize that. It means writing two versions of functions that operate
>on
>it, on "compressed" and "uncompressed" representation. The number of
>functions that will not require a complete scan over geometry is also
>pretty small - just StartPoint / EndPoint / PointN?
>-- 
>Darafei Praliaskouski
>Support me: http://patreon.com/komzpa


Yes , I agree it is just a small number of functions that can benefit from that. But it has been an argument. Wouldn't it be enough to make the that iterates pointarray and returns points aware about dynamic length? 

Ah, you are right, we use a pointer directly into the point array sometimes. That will affect many functions. My thought was that we anyway put the value in a double on the stack. But that is not true.

/Nicklas





Sent from my Android device with K-9 Mail. Please excuse my brevity.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-devel/attachments/20181006/8bc76752/attachment.html>