[R] RE: msvrt.dll (was R-0.65.1 for WinNT/9X)

[Karl M. Syring]

> -----Original Message-----
> From: ripley at auk.stats [mailto:ripley at auk.stats]On Behalf Of Prof Brian
> D Ripley
> Sent: Sonntag, 10. Oktober 1999 13:03
> To: Guido Masarotto
> Cc: Karl M. Syring; r-devel at stat.math.ethz.ch
> Subject: Re: msvrt.dll (was R-0.65.1 for WinNT/9X)
>
>
> On Sun, 10 Oct 1999, Guido Masarotto wrote:
>
> > On Sun, Oct 10, 1999 at 10:33:19AM +0200, Karl M. Syring wrote:
> > > I think, we have now a good reason to make the binary
> distributions to load
> > > msvrt.dll. Otherwise, you would probably get both crtdll and
> msvcrt loaded
> > > if you use COM/ActiveX. As we already now, this is a safe way
> into difficult
> > > and hard to diagnose problems.
>
> Nothing in the (unreleased) (D)COM stuff makes any use of msvcrt.dll.
> I have just checked by removing from my system and re-running the tests.
> It is all ATL based.

Since I can not know what is not released, I will desits to comment
>
> >  Karl,
> >  my idea is to switch to msvrt.dll soon. However. some more tests
> >  are needed.
>
> In particular, we have to resolve what to do about machines without
> msvcrt.dll.  Not all Windows machines have it, and I have read the
> redistribution conditions on my copy of Visual Studio. I am not clear we
> have any right to redistribute it with an application not compiled under
> VC++6 (it might depend where we got it from) and if we do there seem to be
> complex conditions on checking version numbers which will demand a quite
> complex updater to be written.  It looks like having a local copy for the
> application is not approved of.  (All I know is that the redistibution
> conditions do depend on which of the many VS licences one has.)

As far as I know, there are no problems with different versions of
msvcrt.dll. To distribute it we could

a) recommend to install IE5 (deprecated...)

b) A perfect solution would be to make a diagnostic program that checks the
R dlls for the kind of C runtime it depends on. Of course, this diagnostic
program would be distributed with all potentially required dlls.


> Karl: it would be a great help if you could look into this for us, and
> if the way looks clear provide us with a suitable updater, or point us
> to a suitable ready-made one.
>
> Another side effect of requiring msvcrt.dll appears to be to stop the
> possibilities of using the cygwin and cross-compiling routes, since AFAIK
> only the binaries for the mingw32 compiler are available to use
> msvcrt.dll.
> Again, if you know more it would be very helpful if you could
> test this for
> us.

Yes it would make cross compilation impossible without modifying the cygwin
tools.
I think this is not too bad, as anyone who wants it could make still his
crtdll linked version at home.



>
> --
> Brian D. Ripley,                  ripley at stats.ox.ac.uk
> Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
> University of Oxford,             Tel:  +44 1865 272861 (self)
> 1 South Parks Road,                     +44 1865 272860 (secr)
> Oxford OX1 3TG, UK                Fax:  +44 1865 272595
>
>

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._