Re R-beta: 0.61.3: Problems on DEC Unix 4.0

Peter Dalgaard BSA p.dalgaard at
Fri May 8 22:24:48 CEST 1998

"David Barber" <david at> writes:

> This then got me to another problem.  The compile failed because=20
> readline.h was to be included in unix/system.c   This problem apparently
> occurs because of the definition of HAVE_READLINE in Platform.h.
> That file appears to be generated by the ./configure script, so the probl=
> em
> must be the way that this script has been written.

As I mentioned in another post, the problem is that it doesn't want to
look all over the system disk for the location of the readline include
files, so what can happen is that it finds libreadline in a standard
location and decides that readline is there but can't find the include
files later on.

All you have to do is to supply a -I argument to gcc with the correct
location. Of course, there are some locations that are semi-standard
(/opt/gnu, e.g.) and probably configure can be persuaded to do a
better search. Or, make a symbolic link to the actual location from

> After I commented out this definition, everything compiles fine but now I
> get the error message=20
> =20
> Error in paste("echo $", i, sep = "") : negative length vectors are not
> allowed

This one is nasty! I bet something is not quite 64-bit clean. If you
know your way around gdb, it would be very nice to see a stack trace
at the point where the error is generated (line 307,

   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at             FAX: (+45) 35327907

r-help mailing list -- Read
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at

More information about the R-help mailing list