Boost Testing :
Subject: Re: [Boost-testing] Questions about "stale" (or not) tests
From: Steven Watanabe (watanabesj_at_[hidden])
Date: 2009-05-18 00:58:40
Robert Ramey wrote:
> Steven Watanabe wrote:
>> Robert Ramey wrote:
>>> Recently I made some changes in the trunk which provoked failure
>>> accross all platforms. Since I made more changes to correct the
>>> mistake, about half the platforms now pass the in failing test. I'm
>>> suspecting that the failing tests are some sort of artifact from the
>>> test system. The reason I suspect this is at some platforms pass
>>> and others fail - even with the same compiler/library configuration.
>>> How can this be?
>> Which test is this?
My best guess is that the overload of serialize is failing to match
on some compilers when size_t != unsigned. The declaration of bitset is
template <size_t N> class bitset;
according to the standard.