[Rd] build errors in configure script (PR#697)

flip phillips flip@skidmore.edu
Mon, 16 Oct 2000 09:43:51 -0400

on 10/16/00 5:35 AM, Peter Dalgaard BSA at p.dalgaard@biostat.ku.dk wrote:

> Kurt Hornik <Kurt.Hornik@ci.tuwien.ac.at> writes:
>>> the lack of '-lm' in the Macintosh OS X / Dylan release seems to play
>>> a bit of havoc w/ the build. For example, the test for libf2c.a fails,
>>> not because the library is missing, but because -lm is missing. (the
>>> configure looks a little like it wants to check for it but it doesn't
>>> really need to)
>>> i fixed it temporarily by majing a null libm.a, but i suspect this
>>> should really be changed in the configure script
>> Do I understand this correctly that there is no libm.a on the above
>> platform?  What is used instead?
>> I think we could fix the problem, by first checking for libm.a (which we
>> do lateron anyway), and then do the Fortran checking.
> Maybe we should wait for MOSX to come out of beta before changing
> anything. I can imagine that more than one Unix program (to say the
> least -- -lm is probably the one thing that has *never* been broken
> before) would break because of this, so perhaps they want to fix it in
> the OS ...

This really isn't a MacOS X issue as much as a Darwin issue I believe.
Darwin is the open-source BDS based underpinnings of OS X. The TCL
implementation is unusual as well, and I'll take care of figuring that into
the equation as soon as I can get the rest of the thing to compile. Right
now I'm messing with shared library issues.

A look at the Darwin mailing lists might be in order to see why libm.a isn't
a part, etc. 

flip phillips, phd
skidmore college

r-devel 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-devel-request@stat.math.ethz.ch