Johnny Billquist wrote:
As I recall (dimly, but I can find the bits if needed), it had to do with the hypot() call in one of the groff subroutines. I tracked it down once, if this is the problem I remember.I thought this had been sorted out a couple of years ago. Did that never get into the codebase, or do I remember wrong? pic: Unsupported relocation type 232 in non-PLT relocations/usr/src/tools/groff/obj/build/src/roff/groff/groff: grn: Segmentation fault (core dumped)grops: Unsupported relocation type 136 in non-PLT relocations
I think it ended up being that the vax floating point code didn't support NaN or something like that.
I substituted a different routine and the problem went away. I thought the fix was committed in the R3 branch,
but I might be mistaken. I can dig up the diff. -brad