On 12/12/2015 08:03 PM, Jeff Epler wrote: > On IRC, seb mentioned considering whether to merge the rtai-5 branch > (suitably rebased, I assume) to 2.7. To complement my earlier review > notes on the rtai-5 branch, I offer these notes about my perspective on > merging to the stable branch. > > - should backport the realtime-math test if we do this
That's a little tricky. It would have to be a limited version of the test, since the current test fails on older versions of our RTAI packages, due to an advertised but missing realtime math function (I forget which one it was, but obviously one we never use). I guess the thing to do would be to change the header file to test the RTAI version, and only export the function if the underlying RTAI exports it. Then somehow in the test we'd try to detect if the function is exported, and only test it if it exists. Ok, that all sounds doable. > - udev is a new dependency, though probably not a large one. I think that part is in error and should be removed. I'm looking at reimplementing the CheckLoaded() function in the realtime script correctly, so that it actually checks for the condition that will allow the realtime environment to be used. The udev thing seemed like a good idea at the time, but i think you pointed out it's been deprecated for some time now. > - check whether the -msse block change can be dropped, since > rtai-5's rtai-config will get this right now I'll look into it. I owe you a response to some of your review comments from last time too. > Honestly I feel like it's a pretty minor change. If buildbot doesn't > turn up any problems, I'm not too worried about shipping it in 2.7. I agree. We'll see how it goes. -- Sebastian Kuzminsky ------------------------------------------------------------------------------ _______________________________________________ Emc-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-developers
