Matlab's practice of substituting an "8 byte integer" BLAS/LAPACK
library for the "4 byte integer" BLAS/LAPACK libraries that Cantera is
designed to use causes Cantera to crash. This can only be avoided by
statically linking Cantera to BLAS and LAPACK, e.g. by using the
BLAS/LAPACK implementations included with Cantera.
Since the Matlab toolbox itself needs to be a shared library, it is
not usually possible to link statically link it with vendor-provided
BLAS/LAPACK libraries, since they are not compiled with the
appropriate options, i.e. '-fPIC'.
Resolves Issue 34.
Change the variable named *species* to be named *component*. This
better matches the utility of this variable, since it can be used to
retrieve any of the state variables. Also, update the corresponding
documentation.
The name 'H' can mean either the species by that name or the entahlpy
of the reactor, in the case of ConstPressureReactor, and the previous
behavior always returned the index of the enthalpy.
This changes the behavior to preferentially return the species, and
adds alternative names for reactor state variables that are less
likely to generate namespace collisions: 'mass', 'volume',
'int_energy', 'enthalpy', 'temperature', 'distance', 'velocity'. The
single character names are still supported.
Resolves Issue 193.
The order of XML attributes within a node has no meaning, and is not
consistent when the attributes are represented as a Python dict, as in
ctml_writer.py. As a result, comparing the text representation of the
XML generates spurious errors when the order changes.
Adding the attribute allow_discontinuities="true" to a phase's
'thermo' node (CTML) or adding 'allow_discontinuous_thermo' to a
phase's 'options' argument (CTI) will disable the automatic adjustment
of the provided polynomials so that enthalpy and entropy are
continuous at the midpoint temperature.
This fixes the problem of some output (notably from the 1D solver)
going to std::cout when that is not the primary output location,
e.g. when using IDLE or the IPython QtConsole.
There are three problems with this function. It does not work, since
it calls ctmethods with the wrong number of arguments. Even if that
were fixed, it wouldn't do what it claims, as the method it calls
generates a CTI file, not an XML file. Finally, it is redundant, as
the 'ck2cti' function can be used for generating the CTI file, and the
CTI to XML conversion is handled automatically when importing the
phase.
Therefore, this function is being removed.
