https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91593
--- Comment #5 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> --- Looking back at the code I see we are translating from Front-end to run time regarding whether or not we actually are going to show the positive sign or not or suppress it the sign altogether in floating point writes. In this process we actually have three different enumumerators defined. I am going to analyze this a bit more since it "seems" that one ought to be able to do this all in one step at the right place. There is a comment about keeping things alligned so as to not break the ABI, so I would bet that as the standards evolved we had to evolve this code a bit. I will see what I can come up with that might make more sense and not break the ABI. (other than just doing a cast)