Almost succesfull build on AIX

[Thomas Vogels]

(If you don't care about R on AIX, happily ignore this email...)

First about Martin's question (concerning /usr/bin/make):
  I took a quick look at the Makefile.  The problem is with the
Makedeps target.  Replacing $< with $? solves the problem.  Shouldn't
but does.  Will this break on other platforms?  Shouldn't but may.

"Kurt" == Kurt Hornik <Kurt.Hornik@ci.tuwien.ac.at> writes:

>> Yes, there is still something wrong with the DLLFLAGS.  (Which also
>> shows up when you try to link local libraries (like a private copy of
>> the readline lib).)   I've played a little more with the way R.exp is
>> generated.  Let me try it on 64.2 and 65.

Kurt> Please, NOT with 0.64.2.  All my configure/make changes are in 0.65!
I know, but I also need a stable version.

Kurt> gcc (cannot find libgcc.a???)

If gcc cannot find libgcc.a for any other compilation, I'd get worried
in your place.  I might even get scared.
For R.binary linking it's simple and necessary to search for
libgcc.a and put it in the DLLFLAGS.  If you want to link with ld
instead of cc/f77, you have to add default libs, like -lxlf90 -lxlf
-lc -lm.  Which ones exactly, you can find out by calling ld with the
nobind option.  Most instructive since this will result in the binder
commands being printed.  

Kurt> The specs file is fine.  My question is, is that a known gcc problem on
Kurt> AIX or is that due to the rather outdated version of gcc we have?  If it
Kurt> is a general problem, I can easily add to DLLFLAGS.

See above.  But we may not be talking about the same.  When is
libgcc.a not found?

>> BTW, do you not get compiler errors about blkcnt not being defined
>> when you compile with gcc?  So far, I had to patch most programs
>> because there is a bug with the standard include files.

Hmm, I have gcc 2.8.1 (for AIX 4.1) which barfs at sys/stat.h and
friends. Somehow somebody forgot sometime about "typedef long
blk_cnt".  Or something close to this effect.

About DLLFLAGS, SHLIBLD etc:  Currently you'll probably need to know
the libraries (-lm -lxlf ...) when linking a shared lib since the
fortran library code is currently not exported via ldAIX4.  Is this
"bad"?  Is it worth spending time to get this right, i.e. R.binary
exports all symbols that are actually linked in?  Or can we live with
the baggage of extra libs in SHLIBFLAGS and a little bit bigger shared
libs?

Thanks,
  -tom

--
mailto:tov@ece.cmu.edu (Tom Vogels)   Tel: (412) 268-6638   FAX: -3204



-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._