Sorry Markku -- I didn't really finish answering your question....
The RGB color space used by all version of Radiance for the last 5 years at least should be the same, namely those given in src/common/color.h. The latest 3.4 release handles the transformation into other spaces differently, so using the -p option of pcond, ra_tiff, or ra_xyze will give you a different result, and this different will be significant for large white point changes. (If you stick with an equal-energy white, as you are encouraged to do by the outdated documentation, there will be no change in your results.)
The ra_tiff program does not normally report the Radiance primaries to the appropriate chromaticity tags in the TIFF output. It only does this if you use the ra_tiff -p option or a PRIMARIES= line appears in the Radiance header. The latter only happens when one of the other programs (pcond or ra_xyze) puts it there when it gets a -p option. All very confusing, but I'm sure I must have had a good reason for doing it this way. I just can't recall what it was.
P.S. to Mischler -- good find on the replmarks bug. Sure enough, it's in 3.4 as well. I've put in a fix for the next release.