![]() ![]() ![]() That single precision is default REAL *IS* specified by the fortran > A reasonable complaint might be that most Fortran compilers use single The wrong specific for generics, and you'd get argument kindĭisagreements for other cases. That gets messy to even define and I do agree that you'dĮnd up breaking standard conforming programs that way.Īlways treating 1.1 as a double precision literal, independent ofĬontext, would certainly break standard conforming programs. (including such things as possibly invoking different specific The RHS expression be interpreted differently depending on the LHS That would go down the path of having various parts of Then you'd get into more than just the question of conversion from Iregularity and bound to invite complaints about how the same thingĭidn't happen in very simillar cases. Where the RHS was a single literal constant? If so, that would be an Would it really be restricted to only the case ![]() Now a much bigger question in my mind would be how to define it and ![]() That assumes such a thing is making assumptions outside of the standard. Precision be exact (or even that there necesarily be a double valueĬorresponding exactly to every single value with no error). Standard does not specify that the conversion from single to double I don't see how it would break conforming programs if limitted to theĮxact case shown, where the RHS was just a literal constant. I'm quite sure that it would break conforming programs. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |