[Rd] R Enhancements

Duncan Murdoch dmurdoch at pair.com
Wed May 28 17:33:36 MEST 2003


On Wed, 28 May 2003 13:10:19 -0700, you wrote:

>Hi, 
>
>We - Venkatesh Mysore and Salvatore Paxia of the New York University's Bioinformatics Group under Prof.Bud Mishra - have been working on enhancements to R for a while now. 
>
>We're happy to announce that we have some good stuff up and running now:
>
>1. A debuggable Microsoft Visual C version ( _MSC_VER) of R-1.7.0: R-1.7.0-msc (YES !!!!)

How many changes were necessary to do this?  Was it mostly a matter of
working out instructions to get MSVC to work, or did you need a lot of
changes to the R source code?  Does your version pass make check?

I don't use that compiler, so I can't test any of your changes.  I'm
not interested in maintaining multiple compiler targets, but if it's
easy, I'd consider it (or at least try to avoid doing things that will
make it hard for you to do it).

>2. COM/OLE-support that will allow invoking of functions defined elsewhere (similar to the .ole functionality in S+): R-1.7.0-msc-com

That should be put together as a package.  

>3. Minor Fixes/Extensions: R-1.7.0-msc-com-minor

In your minor extension I, could you submit a bug report showing how
to invoke the bug, and how to fix it?  There's still time to get this
into 1.7.1.

Could you write up a description of why the changes in II and III are
a good idea?  I'll put them into 1.8.0 if they make sense.

Re the suggestion for a global "dont_exit" flag:  My first answer is
that global flags are bad, we have way too many of them now.  What we
need are bigger changes that make all of our global flags unnecessary.

Duncan Murdoch



More information about the R-devel mailing list