Changeset 3983 for TI01-discovery/branches/ws-Discovery2-upgrade/readme.txt

Timestamp:

03/06/08 15:28:32 (13 years ago)

Author:

cbyrom

Message:

Update system docs.

File:

• TI01-discovery/branches/ws-Discovery2-upgrade/readme.txt (modified) (5 diffs)

TI01-discovery/branches/ws-Discovery2-upgrade/readme.txt

@@ -2 +2 @@
 ------------
 
+Build Info/Set up
+--------------------
 NB, completing the build here assumes you have both apache axis2 and ant installed.  The AXIS2_HOME environment variable
 should be set to point to the top level directory of the axis2 install.  If you are deploying locally to tomcat, you need
@@ -14 +16 @@
 ant -f buildClient.xml deploy.client - build the service jar and deploy it to a local tomcat install
 ant -f buildClient.xml deploy.client - build the service jar and deploy it to a local tomcat install + restart tomcat
+ant -f buildClient.xml generatestubs - generate both client and server side stubs - and generate the XBeans jar using the client-side stubs
 ant -f buildClient.xml - defaults to building the service jar
 
@@ -23 +26 @@
 Eclipse set up
 -------------------
-To remove build errors in the code, ensure that the client build is ran, and eclipse is refreshed to pick the
+To remove build errors in the code, run the build with the 'generatestubs' target, and eclipse is refreshed to pick the
 generated code.  Add all the jars in the resources dir to the build classpath and also add all the jars in the AXIS2_HOME/lib
 dir.
@@ -29 +32 @@
 Running tests
 ----------------
-There are a few JUnit tests that can be ran to check the basic integrity of the web service.  From these to be ran from inside
-eclipse, you'll need to add the generated XBeans-packaged.jar (in build/lib) to the build path.
+There are JUnit tests that can be ran to check the basic integrity of the web services.  For these to be ran from inside
+eclipse, you'll need to add the generated XBeans-packaged.jar (in build/lib) to the build path.  The tests also assume that
+the files test_file.xml and test_file2.xml, in resources, have been ingested to the postgres DB as badc files (i.e. put them
+into the expected harvest directory (currently, /usr/local/jakarta-tomcat/webapps/oai/WEB-INF/harvested_records/badc-nerc-ac-uk-dif)
+and run oai_document_ingester.py badc
+
+Failing tests
+----------------
+There are currently 3 tests that fail:
+
+i) validAuthorTermSearch
+- this fails currently since unsure where the author data should be set and retrieved from
+
+ii) nullDateStartRangeSearch
+- Generated code is a little ropey atm; if a start date is not defined, this date range is automatically
+not set - however if it is it will be - i.e. the end date is not consulted when setting up this object
+- this makes it currently impossible to recognise the case where a start date is null and the end date is set
+- NB, would be better to tighten up WSDL definition so that you can/cannot set things up wrongly
+
+iii) basicPresentDIFFormatCompare
+- This naively assumes the original format returned will equal the DIF format returned (since the original was in 
+DIF format).  Clearly there are discrepancies introduced by the transform.
 
 Trouble Shooting
@@ -47 +70 @@
 
 when the webservice is deployed to tomcat.
+
+iii) Be very careful when playing around with the Setsrid sql calls - the following apostrophes must be ' and not ` - NB, eclipse
+may substitute these in which will cause problems which can be difficult to diagnose... 
+
+To do
+------------
+i) Currently the time searches use only the 'overlaps' function - i.e. the 'temporalOperator' element of the WS is unused.  This
+could be implemented to allow similar search types to the spatial operator - i.e. within, overlaps and no overlaps