[postgis-users] 0.7.4 Release
strk at freek.keybit.net
Fri Feb 14 02:31:40 PST 2003
There has never been a break... just a mis-interpretation of holes.
The PIP function (we talked about it already, do you remember?) loops
for one point more then the available once. The crossing check does not
do the right thing thus - sometimes - it makes shp2pgsql think that
an hole is not an hole (does not find any other ring containing it) and
consider it as a new outer ring. Visually it is clever that it IS an hole.
Anyway, instead of keep asking for data proves (send the shape...) why not
just check the PIP() function contained in shp2pgsql ? I compared it with
the PIP function found in the main postgis objects (_ops maybe?) and is
different, for just an i<(n-1) check in the for loop. We checked togheter
the shapefile format specification and we found out they are "closed" as
the pgsql once.. do you remember ?
It is not easy to discover the problem because many GIS viewer have their own
Polygon In Polygon functions to check for holes, and sometimes overlapping
layers will write something in the supposed hole thus making it invisible.
I've seen that latest modification of shp2pgsql are very few
(a couple of lines maybe?) from 0.7.3, so I don't see why should
it work now.
The question is: have you checked at all the PIP() function defined there ?
[ sorry for the tone, but I find a lot of resistance against my patches,
I find myself keeping patching all new releases with little response
from main developers -- END_OF_STRESS ]
BTW, Thanks a lot for your job.
> Do you have data which still breaks the current shp2pgsql?
> strk wrote:
> > There is no trace of fix for the PIP() bug in the new shp2pgsql.
> > Is there any problem with it ?
> > --strk;
> > pramsey wrote:
> >>Since more and more new users seem to be trying to use the release
> >>version with PostgreSQL 7.3.X, I have created a new release version
> >>0.7.4 which includes support for that PgSQL version. It also includes
> >>the several fixes to shp2pgsql which have occurred since the last
> >>official release. Spatial index selectivity is included as well, though
> >>that is pretty hard to figure out until we add the documentation. Dave
> >>explained it thoroughly on this list, for those who are interested
> >>(search for "selectivity" in the archives and read the Makefile carefully).
> >>The next release should be 0.8.0 which will include real spatial
> >>predicates, based on the GEOS library (geos.refractions.net). Before
> >>that major release, there will be a testing period where we will be
> >>asking folks to make use of the CVS version and exercise the predicates
> >>and the build system.
> | Paul Ramsey
> | Refractions Research
> | Email: pramsey at refractions.net
> | Phone: (250) 885-0632
> postgis-users mailing list
> postgis-users at postgis.refractions.net
More information about the postgis-users