[postgis-devel] RE: [postgis-users] Problem with PostGIS 1.0.0 RC1
schabios at logi-track.com
Tue Jan 25 06:13:40 PST 2005
Mark Cave-Ayland schrieb:
>>We also have all of GeomFromText() which is currently taking
>>a 'geometry' instead of 'text'. This allows users to use
>>non-standard OGC functions, and I think should be changed.
>>Note that current behaviour makes this possible (as side-effects):
>> SELECT GeomFromText('00..FF');
>> SELECT GeomFromText('\\000\\001...'::bytea);
>>.. which I think should be forbidden.
>>What do you think ?
> Good catch. I would say definitely make it forbidden; I can't see a rational
> case for letting binary inputs into a *FromText() function at all(!).
> Markus, do you agree?
Yes, I agree. Strict OGC compliance for OGC functions is a good thing, I
I also remember that there was some discussion about some unexpecded
side-effects that were caused by the old "rely on auto-casts" approach.
And as we are just before releasing 1.0, I would not shed too much tears
about backwards compatibility for applications that abuse OGC functions,
as those apps are broken.
This may be different for _some_ of our own functions and interfaces,
they should be kept compatible (e. G. accepting old SRID;EWKT in
canonical text parser is needed for restoring dumps from hwgeom
databases. Another examile: The JDBC extension has no reason for
incompatible API changes).
As we are at client interfaces: Does someone know what happened to
Python Geotypes? http://people.initd.org/rjt/ seems down, and I think it
will need an update to work with PostGIS 1.0 as well...
markus schaber | dipl. informatiker
logi-track ag | rennweg 14-16 | ch 8001 zürich
phone +41-43-888 62 52 | fax +41-43-888 62 53
mailto:schabios at logi-track.com | www.logi-track.com
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 256 bytes
Desc: OpenPGP digital signature
More information about the postgis-devel