This post was updated on .
Hi all,
I am considering migrating from QLNet to QuantLib-SWIG and I'm wondering if it would be a wise decision. You may remember that I published NQuantLib64 on NuGet a few months ago (https://www.nuget.org/packages/NQuantLib64), and before I make this move I would like to ask a few initial questions. - About what percentage of QuantLib is usable via QuantLib-SWIG? - If QuantLib is not 100% usable via QuantLib-SWIG, what are the fully, partially, and zero implemented features? - Is there active development for QuantLib-SWIG by the QuantLib team, or are new features only added by request and/or community support? - What is the lag between QuantLib releases and QuantLib-SWIG releases? I've written some simple test projects pricing fixed rate bonds using QuantLib-SWIG (NQuantLib64) and it worked out well! The thing is I may need access to more features in the future, and QLNet is quite limited in this regard. If QuantLib-SWIG is well supported and feature-full, I would surely migrate. Thanks, Grant |
Hello, On Thu, Oct 29, 2015 at 5:56 PM grantathon <[hidden email]> wrote: - About what percentage of QuantLib is usable via QuantLib-SWIG? Hard to say. It's not 100% for sure, and of course it may depend on what part of the library you're most interested in. The general rule is: a feature is available if it's declared in a .i file in the QuantLib-SWIG/SWIG folder. - Is there active development for QuantLib-SWIG by the QuantLib team, or are New features are usually added when someone needs them. - What is the lag between QuantLib releases and QuantLib-SWIG releases? Technically, a new QuantLib release doesn't need a new QuantLib-SWIG release; we keep backwards compatibility, so if I released, say, QuantLib 1.7, you could still use QuantLib-SWIG 1.6 with it. This said, I usually release QuantLib and QuantLib-SWIG in tandem (unless I make a bugfix release for QuantLib, such as 1.6.2). I've written some simple test projects pricing fixed rate bonds using Not feature-full and not extremely well supported, I'm afraid. However, it wouldn't be so hard for you to add the features you need. (Especially if they're missing from QLNet, too; it would surely be easier to export them through SWIG than to port them to C#). Hope this helps, Luigi -- <http://leanpub.com/implementingquantlib> ------------------------------------------------------------------------------ _______________________________________________ QuantLib-users mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/quantlib-users |
Thanks Luigi!
|
This post was updated on .
Hi Luigi,
Another question. If I were to add more interface files to QuantLib-SWIG for further support of C# would I need to add any unit tests? From looking at the git repository, it looks like this is currently not the case, but I'm not sure of how you validate and carry out pull requests regarding SWIG interface files on your end. Looking forward to contributing! Best, Grant |
Hi Grant, no, there's no requirement for tests for the SWIG interfaces. There's not a lot of logic there (creating the interfaces is usually just a matter of copying the corresponding declaration into the .i files) so if the wrapper compiles the odds are that the thing is correct. A little manual testing should be enough. Luigi On Mon, Nov 9, 2015 at 5:49 PM grantathon <[hidden email]> wrote: Hi Luigi, -- <http://leanpub.com/implementingquantlib> ------------------------------------------------------------------------------ _______________________________________________ QuantLib-users mailing list [hidden email] https://lists.sourceforge.net/lists/listinfo/quantlib-users |
Free forum by Nabble | Edit this page |