Re: Building RTEMS on Solaris

[Eric Norum <norume@aps.anl.gov>]

wtw00992 wrote:
> Greetings from L.I.
>  
>  
> We currently run our Slow Controls for the STAR  experiment at BNL using 
> EPICS and VxWorks running on a Solaris.
> We  run solaris 8, gnumake 3.79.1, gcc 2.95.3, perl 5.005.
>  
> Last week we began our build of RTEMS and ran into several problems 
> right from the start.   We were trying to build for a target mvme167.
>  
> During the course of  this we looked at many list serv archives, and I 
> failed to see where anyone had had an issue with these 
> target/platforms.  I did see an email form Eric Norum regarding folks 
> building for a mvme167 using RTEMS on a Linux... and it got me thinking 
> that maybe no one is trying this.  (Either that or its easy and we are 
> just klutzes???)
>  
> Has any one tried building RTEMS for Solaris with this target?  If so  I 
> would be interested in hearing from you.
>  
> ------------------------------------------
> For the record we tried the following:
>  
> We followed the tutorial http://www.aps.anl.gov/epics/base/RTEMS/tutorial/
>  
> The first trouble we had was with the "Get and build the development
> tools".  We simply couldnt get the RTEMS patches for gcc3.2.3 from the 
> OAR site  to work when we applied them on the solaris.  We ignored this 
> finally but clearly the RTEMS build failed with out the right tools.  We 
> eventually tried building the gcc tools on a Mac OSX and a Linux machine 
> and they work ... so we took the patched files we created there and 
> tried to compile those. Still when we tried to build RTEMS itself it 
> stil failed. And despite some very powerfull wishful thinking our 
> attempts to get EPICS to compile with this failed as well  :-)
>  
> We also tried to ftp the latest version of the tools from the RTEMS 
> site. Still the RTEMS build problems persisted.
> ---------------------------------
>  

I'm checking to confirm this, but I suspect that the problem will go away if 
you use GNU make rather than the SUN make.  Just set the MAKE environment 
variable to gmake and try running the getAndBuild script again.

--
Eric Norum                                 norume@aps.anl.gov
Advanced Photon Source                     Phone: (630) 252-4793
Argonne National Laboratory