Hi All.
Well its been a few months but I finally got around to working a bit no
a Matlab interface. Its nothing all that fancy and quite frankly the
C++ code is pretty bad, but it does implement most of the methods for
dates, calendars, daycounters, and the old pricers. I've also ported
the test cases for those namespaces from the python distribution. All
of it (included source and compiled cmex dll) is available at:
http://www.eng.uwaterloo.ca/~mrtreibe/QuantLib-Matlab-0.0.1.zip.
I'm basically going to scrap the architecture that I've used so far
because it was quick and dirty. One of the features I'd like to
maintain is the ability to save workspaces and unfortunately this method
doesn't save the cmex heap space (which would basically render any
restored workspace unusable). What I'm going to do is basically
replicate the Quantlib objects as matlab objects without a direct link.
Instead whenever methods are called they will in the cmex code create
the Quantlib objects and then copy all the applicable observers before
destroying the objects. While this'll make the internals function
based, it'll make the matlab objects appear very similar to the quantlib
objects from an outsiders point of view.
Anyways, hopefully I'll have another release soon that demonstrates this
architecture. Possibly documenting the architecture as well..
Mark.
Locked
