John Maddock wrote:
>>In Feb 2005 there was a discussion of boost::tuple lib being
>>replaced with a different more std::tr1 conforming version from
>>boost::fusion. What is the status of this change?
> No progress as such, as far as I know (at least as far as Boost.Tuple goes),
> however the TR1 lib is using the new tuple code behind the scenes.
That's not exactly true. There's progress! :)
Fusion is in the review queue. Some people say that we should just
go ahead and change the old with the new. However, fusion is too
big a departure from the original tuple library. There are lots
more sequences (e.g. vector/list/set/map), an extensive set of
views), there are algorithms and iterators, etc. I'd rather regard
Fusion as an entire new library of its own. The TR1 compatibility
layer is a mere fraction of its capabilities.
> Arkadiy Vertleyb wrote:
> > I can't see a column for one of these compilers even in the full
> > view... Does anyone still care to support them?
> > Does anyone care if typeof supports these compilers?
> And yes. IIUC typeof uses a bug-feature on those compilers that makes
> the implementation trivial, is that correct? In that case, IMO, there's
> little reason not to support them.
I didn't mean to drop this. The bugfeature is also applied to vc-7_1, which
makes it relevant even without tools mentioned. My question was actually
caused by some code in LVALUE_TYPEOF that needs to use ifdefs based on
whether or not partial template specialization is supported, and whether or
not a function can return a reference to array. I lately run into a problem
with gcc 4.0.2, that doesn't seem to like the workaround we used (returning
pointer to array), so I was looking to somehow reduce the number of ifdefs
to a manageble amount.
Also, if the regression tests no longer run for these tools, how can one
guarantee the support? Or is this temporary, and these tools are going to
re-appear in the regression tables?
Arkadiy Vertleyb writes:
> Also, if the regression tests no longer run for these tools, how can
> one guarantee the support? Or is this temporary, and these tools
> are going to re-appear in the regression tables?