https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92472
--- Comment #10 from David Binderman <dcb314 at hotmail dot com> --- (In reply to Jonathan Wakely from comment #9) > Or in other words, of course whether a parameter can be const is separate > from whether a member function can be const. Agreed. > But that doesn't mean that changing a parameter from non-const to const > can't have any effect. Double negative, but I think I know what you mean. > It certainly has an effect on which member functions you can call on the > parameter. Agreed, but does it matter ? These are a bunch of comparison functions, they AFAIK shouldn't be changing the objects they are comparing. I am still not sure if the new code is ok or not, and by now, TBH not bothered either way. I gave up on C++ because of its ever growing complexity. This is another example. Suggest play safe and assume that JW is correct, so the proposed change to file multiway_merge.h should be removed. If this code is to die, should it be marked as unmaintained ?
