On Sun, Mar 04, 2007 at 01:23:58PM -0800, Brian Dessent wrote: > Er, but the entire reason for the existance of this module was so that > gdb can use %a to display floating point variables in hex, without > having to worry about whether the platform's snprintf natively supports > it or not.
No, that's why I started working on merging snprintfv, but this stuff that Bruno's done since then is different. It's useful independently, but I doubt it will be directly useful for GDB. What GDB wants to print in hex floating isn't a host double but a target double; if we cast to host double or long doube, then in some circumstances we lose precision. I suspect GDB will end up using gnulib to parse format strings, and then handling a few of the format specifiers on its own. There are lots of valid approaches to that. Anyway, I'm stalled for a while on snprintfv, so don't expect too much unless someone else wants to pick up the project of making GDB happier - which I'd be glad to assist with as time permits :-) -- Daniel Jacobowitz CodeSourcery
