On 01/11/12 23:05, Andrew Miller wrote:
Hi all,
The CellML API release candidate version 1.11rc2 has been released with
Sorry, that should of course be 1.12rc2.
a minor change to correct uncertainty solvers as CellML API version 1.12.
The files making up this release candidate are available here:
* http://cellml-api.sourceforge.net/download.html#1.12
This release candidate has a number of significant improvements over the
CellML API 1.12:
and that should be 'over the CellML API 1.11'.
* Extra attributes are provided to make accessing a variable's
initial value more convenient.
* IVs on algebraic variables is supported with initial_value rather
than RDF.
* Documentation on using the CellML API from Java has been improved.
* KINSOL, rather than LEVMAR, is used as the non-linear system
solver. This avoids the global state used by LEVMAR, allowing multiple
simulations to be working concurrently to return results more efficiently.
* Error reporting for models underconstrained due to missing initial
values has been added.
* It is now possible to use the DOM to determine the order of
elements in the CellML model and map from the DOM element to the CellML
element.
* An ambiguity in the TeLICeMS serialiser around the nesting of the
equals operator has been fixed.
* You can now create a CellML model from an existing DOM representation.
* The C++ interface for performing query interfaces has been improved.
* The generated binaries are smaller due to several changes,
including not generating unnecessary Java bridges.
* pcmpy has been completely removed from the source tree now that
better Python bindings are provided by cgrspy.
* More comprehensive tests have been written for (and passed by) the
units validation part of the Validation Against CellML Specification
Service.
* The precedence rules in the MATLAB CeLEDS XML file have been
corrected.
* The error handling in TeLICeMS has been improved.
* A number of minor issues with the Doxygen generated documentation
has been cleaned up so it is now easier to use.
* A number of problems that made the PCM -> Java bindings used for
callbacks like the IntegrationProgressObserver unusable have been fixed.
* Threading issues in the Java bindings have been fixed, preventing
crashes when multiple threads are used.
* The CellML API no longer automatically retrieves DTDs over the
network, giving API users complete control over how models are loaded.
* More sample code on how to use the RDF container APIs has been
written.
* The CellML API libraries now use versioning to avoid accidentally
using the wrong version of the API; this allows multiple versions of the
API to co-exist in the paths on the same system.
* It is now possible to use a native libxml2 and SUNDIALS instead of
the one shipped with the API by selecting an option at CMake configure
time.
* The order RDF triples are returned in is now more predictable.
* Parallel builds now work correctly.
* An RDF triple object can retrieved by subject, predicate, and
object, allowing it to be unasserted straightforwardly.
* A bug in CeVAS that meant it didn't look at nested encapsulation
relationship_refs has been fixed.
* New API operations have been added for retrieving annotations that
don't throw exceptions (which are inefficient) when there is no
annotation found.
* A bug in the Java -> PCM bridge enumeration support (important for
public / private interface retrieval) has been corrected.
Best wishes,
Andrew
_______________________________________________
cellml-discussion mailing list
cellml-discussion@cellml.org
http://lists.cellml.org/mailman/listinfo/cellml-discussion
_______________________________________________
cellml-discussion mailing list
cellml-discussion@cellml.org
http://lists.cellml.org/mailman/listinfo/cellml-discussion
