On Fri, Aug 17, 2018 at 08:51:48PM +0100, Ramsay Jones wrote: > > > On 16/08/18 23:12, Luc Van Oostenryck wrote: > > So that the names match the ones used in the comments > > describing the simplification. > > > > Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@xxxxxxxxx> > > --- > > simplify.c | 7 ++++--- > > 1 file changed, 4 insertions(+), 3 deletions(-) > > > > diff --git a/simplify.c b/simplify.c > > index 5d5bd3f00..4c3036963 100644 > > --- a/simplify.c > > +++ b/simplify.c > > @@ -547,10 +547,11 @@ undef: > > } > > > > static int simplify_mask_or_and(struct instruction *insn, unsigned long long mask, > > - pseudo_t src, pseudo_t other) > > + pseudo_t ora, pseudo_t orb) > > I can't say that these names are any better myself! ;-) > > Is this supposed to be or_lhs and or_rhs? No, at the first call they're indeed LHS & RHS but at the second call they're RHS & LHS (see my reply for 02/15). I choose these names so that they are very symmetrical because their role can be very symetrical (even src1 & src2 could suggest an ordering that is not present) and it helped me to see them as such (better than 'src' & 'other'). But yes, I agree that these names are not very good. I'll be glad to change them to something better (but only after that part2 (and maybe part3) is merged). In truth, the one called now 'ora' is involved in the second masking operation (base case is AND but could be LSR, SHL or ZEXT) while the one called 'orb' is not. So, any new names should, I think, reflect this. Suggestions are very welcome. -- Luc
