On Thu, Aug 24, 2023 at 2:08 AM Rob Landley <r...@landley.net> wrote: > > On 8/23/23 11:35, enh via Toybox wrote: > > %t is the appropriate choice for pointer differences on both ILP32 and > > LP64, and shorter than casting to long: > > ``` > > external/toybox/toys/pending/expr.c:209:45: error: format specifies > > type 'long' but the argument has type 'int' [-Werror,-Wformat] > > Warning: long and int are literally identical in this context, and in the > context where they're not long is correct.
as unix people you and i (and posix) believe that, but the c standard doesn't require that. (maybe in another 40 years? we did eventually get 8-bit bytes and two's complement...) > Did we ever figure out what the standards committee was smoking? ptrdiff_t seems like an unfortunate natural consequence of uintptr_t? (the one i've never understood is intptr_t --- it's the more natural name, but i only ever see it when i'm rooting out obscure bugs and changing it to the uintptr_t that should have been used instead! i assume intptr_t is an unfortunate natural consequence of all other integral types having signed and unsigned variants?) > Rob > _______________________________________________ > Toybox mailing list > Toybox@lists.landley.net > http://lists.landley.net/listinfo.cgi/toybox-landley.net _______________________________________________ Toybox mailing list Toybox@lists.landley.net http://lists.landley.net/listinfo.cgi/toybox-landley.net
