From: Douglas Gregor (gregod_at_[hidden])
Date: 2003-02-07 21:07:08
On Friday 07 February 2003 05:58 pm, Alisdair Meredith wrote:
> I have a potential patch for the boost::array tests that will pass under
> Borland 0x561 and probably under MSVC as well [I think it is the same
> I am leary of resolving an issue by patching a test, but I think the
> issue is comparatively minor, especially in regard to most library use,
> and currently failing 40% of tests does not accurately reflect the case
> The problem is that borland (and likely MS) cannot use initialiser
> syntax for aggregates that are in turn made up of non-aggregate types.
> The failing tests (3 & 4) uses arrays of std::strings. Switching to
> const char * on the broken compiler 'resolves' the issue.
> Hence the proposed patch: [from CVS diff logs]
I'm okay with this patch, but I would like to also keep a testcase that checks
specifically for this failure so that we don't forget that this is indeed a
potential problem for users.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk