From: Preston A. Elder (prez_at_[hidden])
Date: 2007-06-01 20:32:45
Not entirely true ;)
Its consistent with iostream, but not DEFAULT iostream - only iostream
with increased precision. One can accept the same behavior as iostream
as not a bug only when they don't have to manipulate iostream (ie.
increase precision) to reveal said bug.
But another way, one expects ss << d; ss >> s; to work identically to s =
They DON'T expect ss.precision(digit10+1); ss << d; ss >> s; to be the
Just my $.02 :)
On Fri, 01 Jun 2007 22:47:57 +0400, Alexander Nasonov wrote:
> If it's broken in 1.33.1, then it's no my fault. I thought I introduced
> this bug when I increased precision to pass loopback test.
> I'm sure it's consistent with std::iostream behavior so it's not a bug
> Preston A. Elder wrote:
>> I tried on 2 platforms:
>> RedHat AS4 (2.6 kernel), x86
>> gcc 3.4.4 and boost 1.33.1
>> Gentoo (2.6 kernel), amd64
>> gcc 4.1.1 and boost 1.33.1 AND CVS Head
>> All cases the same result.
>> On Thu, 2007-05-31 at 15:52 +0400, Alexander Nasonov wrote:
>> > In HEAD or 1.34? which compiler?
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk