[EMAIL PROTECTED] writes: > > Thu, 2 Dec 1999 06:15:19 -0800, Simon Peyton-Jones > <[EMAIL PROTECTED]> pisze: > > > Indeed. But it is true to say that Select.select's timeout > > parameter could usefully be generalised to either an Integer or > > a TimeVal (i.e. (sec,usec) pair). Since this library isn't > > an H98 standard, it might be worth just doing this, no? > > I vote for Integer as opposed to (sec,usec). I guess that C uses two > longs because it did not have a longer type, but we do have a longer > type, so why not use it? I really can't think about a reason to use > (sec,usec)... > Uhm, because you have to interface to a C function that takes (sec,usec)? i.e., the extra range you get from using Integer is useless. --sigbjorn
- RE: GHC Select and Time modules - struct timeval Sigbjorn Finne
- Re: GHC Select and Time modules - struct ti... Hannah Schroeter
- RE: GHC Select and Time modules - struct timeval Sigbjorn Finne
- RE: GHC Select and Time modules - struct timeval Simon Marlow
- RE: GHC Select and Time modules - struct timeval Simon Peyton-Jones
- Re: GHC Select and Time modules - struct ti... Keith Wansbrough
- Re: GHC Select and Time modules - struct ti... George Russell
- RE: GHC Select and Time modules - struct timeval Simon Marlow
- Re: GHC Select and Time modules - struct timeval Marcin 'Qrczak' Kowalczyk
- Re: GHC Select and Time modules - struct timeval Sigbjorn Finne
- Re: GHC Select and Time modules - struct ti... Keith Wansbrough
- RE: GHC Select and Time modules - struct timeval Simon Marlow
- Re: GHC Select and Time modules - struct timeval Alex Ferguson
- Re: GHC Select and Time modules - struct ti... Hannah Schroeter