On Mon, 25.11.13 19:12, Dave Reisner (d...@falconindy.com) wrote: > > On Tue, Nov 26, 2013 at 12:40:05AM +0100, Lennart Poettering wrote: > > On Mon, 25.11.13 15:20, Dave Reisner (dreis...@kemper.freedesktop.org) > > wrote: > > > > > uint64_t can be formatted correctly with %ju, rather than casting to > > > unsigned and potentially losing accuracy. > > > > Oh, shouldn't we be careful with that? %j is for intmax_t. Which might > > or might not be int64_t. Given that int128_t is already on the horizon > > (newer gcc already support __int128 on 64bit machines...), I wouldn't be > > surprised if intmax_t is growing to 128bit eventually. > > How do you change sizeof(uintmax_t) without breaking ABI?
Well, there's prior art for botching such type size changes up, they could just take inspiration from that and fail awesomely there too: off_t... > > Format strings don't really have a nice way to print fixed-size > > integers I fear... the only stuff that is correct is the "PRIu64" macro, > > but that's fricking ugly... > > > > I think PRIu64 is still a better, more future-proof option than %j though... > > Assuming uintmax_t really does increase in size, no type promotion, and > someone actually manages to conjure up a system with 4 billion > interfaces... Well, the latter only matters on LE machines. On BE you'd be immediately broken... Lennart -- Lennart Poettering, Red Hat _______________________________________________ systemd-devel mailing list systemd-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/systemd-devel
