| age |
author |
description |
| 2014-09-26 |
Stella Giannakopoulou |
in the case of select constructs there is no need to push the resulting geometry |
| 2014-09-26 |
Babis Nikolaou |
corrected by name |
| 2014-09-26 |
Babis Nikolaou |
added dependency of Strabon to testsuite and strabon-endpoint-client (in main pom.xml) |
| 2014-09-25 |
Babis Nikolaou |
fixed bugs #49 and #65; we do not do any decode on the query that is sent through the forms of the strabon endpoint beans; this is already handlded by the container |
| 2014-09-24 |
Babis Nikolaou |
merge with default branchtemporals |
| 2014-09-24 |
Stella Giannakopoulou |
fixed the tests of geof:distance and geof:buffer, that had wrong results due to the |
| 2014-09-24 |
Babis Nikolaou |
updated ChangeLog |
| 2014-09-23 |
Babis Nikolaou |
corrected tests that were expecting a WKT in 4326 to contain also the URI of the CRS; we do not include this anymore, if the CRS is the default for that datatype |
| 2014-09-23 |
Babis Nikolaou |
corrected another test that was expecting a WKT in 4326 to contain also the URI of the CRS; we do not include this anymore, if the CRS is the default for that datatype |
| 2014-09-23 |
Babis Nikolaou |
it turns out that there is no simple way of having the datatype of a constant geometry available in the query builder, thus it is not easy for geof:getSRID to determine the correct datatype and return the appropriate URI for the CRS, when that has SRID 4326. Therefore, we choose to always return CRS84, even for strdf:WKT arguments. Notice, that when geof:getSRID is called evaluated in Java, we have such information available and we can return the correct result. However, we chose in changeset 1420:37792dd7041b to always push the evaluation of all spatial functions in the database. The respective tests have been updated as well. |
| 2014-09-23 |
Babis Nikolaou |
added log4j.properties to endpoint-client; corrected test that using geof:getSRID() with a strdf:transform function with an input a geo:wktLiteral constant; in such a case we cannot return CRS84 simply because strdf:transform creates strdf:WKT datatypes for which the default CRS is EPSG:4326 |
| 2014-09-22 |
Babis Nikolaou |
CORRECT commit message: I de-activated evaluation of grounded expressions in SELECT queries in Java (see the respective #72 the reason I did so) |
| 2014-09-22 |
Babis Nikolaou |
fixed a test that was using strdf:distance in ORDER BY clause with two arguments (expecting three arguments always) and de-activated evaluation of grounded expressions in SELECT queries in Java (see the respective #72 the reason I did so) |
| 2014-09-22 |
Babis Nikolaou |
more type checking in extension functions in SELECT clause; added also some FIXME and TODO notes |
| 2014-09-22 |
Babis Nikolaou |
type checking in extension functions in SELECT clause |
| 2014-09-22 |
Babis Nikolaou |
log an info message that we do not support calculation in meters in java |
| 2014-09-22 |
Babis Nikolaou |
refixed bug #69 |
| 2014-09-22 |
Babis Nikolaou |
fixed bug that would cause a class cast exception for grounded expressions with function nesting in the SELECT clause |
| 2014-09-22 |
Babis Nikolaou |
buffer in meters in SELECT clauses containing only constants is not supported yet (TODO), but we log a corresponding message and we do the computation in degrees :) |
| 2014-09-22 |
Babis Nikolaou |
corrected some tests that were expecting the default SRID (4326) for WKT literals; from now, we do not include such URI, since this is the default and is assumed |
| 2014-09-22 |
Babis Nikolaou |
This is a merge commit that addresses the following issues: |
| 2014-09-21 |
Babis Nikolaou |
support also geof:convexHull in grounded expressions evaluated in the SELECT clause |
| 2014-09-21 |
Babis Nikolaou |
now StrabonPolyhedron carries information about the datatype of the underlying geometry, so as to be consistent with the presentation of results |
| 2014-09-21 |
Babis Nikolaou |
corrected output for test that was checking computation of boundary; now that grounded expressions are evaluated using JTS, the boundary of the polygon [Polygon((0 0, 1 0, 1 1, 0 1, 0 0))] is not a Linestring any more as PostGIS was computing (in the wrong way), but LinearRing! |
| 2014-09-21 |
Babis Nikolaou |
reverting back to dbpedia.org/sparql for testing endpoint-client on virtuoso |
| 2014-09-21 |
Babis Nikolaou |
revert the behaviour of considering EPSG:4326 as lat/long in order to be in line with PostGIS; this is because PostGIS defines several of their functions, like ST_Buffer(Geometry, meters) on SRID 4326. Therefore, if we change this semantics, we would get 4326 lat/long and that would be chaotic |
| 2014-09-21 |
Babis Nikolaou |
corrected a test output for getSRID that was set to returning integers (84000) instead of URIs |
| 2014-09-21 |
Babis Nikolaou |
a lot of refactoring and cleaning in GeneralDBEvaluation.evaluate(FunctionCall, BindingSet) + support of grounded ST_Centroid in SELECT (ST_MakeLine is still missing) |
| 2014-09-20 |
Babis Nikolaou |
Fixed bug #72. The same bug applies to other functions apart from these that return the SRID, such as all the property functions, postgis functions, etc. |
| 2014-09-20 |
Babis Nikolaou |
fixes bug #69: better error handling |
